Responsibility and the emotional structure of relationships

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Request accessible transcript

Transcript (if available)

Content Responsibility and the Emotional Structure of Relationships Stephen Bero A dissertation presented to the Graduate School of the University of Southern California in partial fulfillment of the requirements for the degree of Doctor of Philosophy (Philosophy) December 2017 Copyright 2017 | Stephen Bero for Beth iii Acknowledgements It is astonishing and overwhelming to step back and take account of all the debts I have incurred and all the help I have received in writing this dissertation. Having the opportunity to record some of them here is a great pleasure. It seems appropriate to start by recognizing the many exceptional teachers and mentors who supported and encouraged me on the path that eventually led me (back) to graduate school. These include: at Bard College, Daniel Berthold, William Griffith (my senior project advisor), Garry Hagberg, and William Mullen; at the Uni- versity of Texas at Austin, Brian Leiter (my master’s thesis advisor), Bob Solomon, and David Sosa; at Columbia Law School my law review note advisor Charles Sabel; and at University College London my LLM thesis advisor Stephen Guest. Two in- comparable professional mentors also belong on this list: Judge Jack Weinstein and Judge Reena Raggi, both beacons of legal brilliance, professional integrity, and per- sonal generosity. In connection with the writing of the dissertation I am, first, grateful to the USC philosophical community, which has provided an extraordinarily collegial, col- laborative, and challenging (in the best possible sense) setting to do philosophy. I have benefitted from incisive and constructive questions, feedback, and discussion in dissertation seminars, Speculative Society sessions, working group meetings, mock interviews, practice job talks, and over coffee or drinks in casual conversations too numerous to count. Just as important as this substantive input, I have received formal and infor- mal expressions of support, encouragement, and interest in my work. This has included substantial material support in the form of a five-year Provost’s Fellowship, a one-semester Flewelling Fellowship, and a Glovin Endowed Scholarship, all re- ceived through the assistance or sponsorship, in one form or another, of the philosophy department or its members. But it has also, crucially, included many peo- Responsibility and the Emotional Structure of Relationships iv ple whose judgment I deeply respect giving me the confidence to develop and pursue my interests in the direction that has resulted in this dissertation. I simply cannot overstate the value of this. No less indispensible has been the help, in countless matters large and small, provided with miraculous efficiency and unfailing patience and kindness by the phi- losophy department staff over the years that I have been here: J.N. Nikolai, Corey Resnick, Natalie Schaad, Barrington Smith-Seetachitt, and of course most of all Cynthia Lugo. I wish I had kept better track of all the occasions on which I have had help- ful conversations at USC, including conversations which did not at the time appear to bear on my dissertation, but ended up teaching me something that later turned out to be useful. A complete list of USC faculty—other than my advisors or committee members—from whom I learned something that I was eventually able to put to good use here would include more people than it would omit. For now I will limit myself to cases I specifically recall of questions, comments, or conversations that helped me to push this project forward; these came from Steve Finlay, John Hawthorne, Ed McCann, Scott Soames, Kadri Vihvelin, and Ralph Wedgwood. My debt to other USC graduate students is even greater, and is distributed over an even larger number of people. I won’t try to list all of the instances I can ac- tually remember—there would be literally dozens of names, and even then I would inevitably have forgotten a dozen others. Nonetheless, specific thanks are owed in a few cases. Alex Dietz for several semesters organized an ethics working group where I benefitted from reading others’ works-in-progress as well as from extremely helpful comments on my own (a large share of them from Alex). I received excellent, de- tailed written comments on a draft of chapter 2 from Alex Heape. Abelard Podgorski attended a number of presentations of my work in dissertation seminars and working group meetings and never failed on those occasions to offer an insight that expanded my understanding of some critical issue. I have been particularly lucky in sharing a wonderful grad student cohort with Rima Basu, Erik Encarnacion, and Caleb Perl. Each of them would merit spe- cific thanks just for their philosophical influence and assistance—I have learned Acknowledgements v more philosophy from each of them than I can properly account for. But much more important, though difficult to capture properly in a phrase, is that they have been exceptionally congenial, supportive, and companionable fellow travelers to share this experience with. I would happily do it over again with them. I have also been fortunate to be at USC during a high point for philosophy of law, together with a tight-knit group of grad students and post-docs, including: Kristen Bell, Renee Bolinger, Erik Encarnacion, Michael Pressman, Alex Sarch, and Aness Webster. I am particularly grateful to Kristen for helpful, sympathetic, and very encouraging conversations about the overall shape and approach of my project. This group also includes three of the people to whom my largest debts of gratitude are owed. First, since Alex Sarch arrived as a post-doc at USC, and continuing after he left, he has been a tireless source of energy, inspiration, ideas, and thoughtful, con- structive criticism. Alex is a personification of the best kind of collaborative spirit— always interested, engaged, and on point, and always ready with interesting ideas and projects of his own to share. I couldn’t be happier to be joining him as a colleague again soon. Second, Aness Webster, two years ahead of me in the program, had a key role in shaping my experience as a grad student. A model of charitable, productive dialogue, she has engaged with my project as deeply and sympathetically as anyone, and has always been able to see the best side of an idea or position—often better than I saw it myself. At the same time, she has often seen before I did the thing that I myself had most to be worried about. Having someone in Aness who always just got the wavelength that I was working on has been a tremendous boon. Third is Erik Encarnacion. In the case of many of these thanks, my gratitude pertaining to matters professional and academic is dwarfed by my unstated gratitude, admiration, and affection as a friend; but in Erik’s case the disproportion is especially vast. In terms of the dissertation, suffice to say that his influence has been immeas- urable; he has left a positive mark of one sort or another on nearly every page. But this is really the least of the things for which I am grateful to Erik. Responsibility and the Emotional Structure of Relationships vi This brings me to the advisors and committee members who have guided this project from uncertain beginnings through circuitous, inconsistent, and mean- dering progress to this eventual conclusion. Andrei Marmor was my faculty advisor from when I first arrived at USC through my qualifying exam and up until he left the department. I could not have asked for a more patient, generous, and supportive advisor, or for a more reliable and dependable source of counsel and advice, both philosophical and professional. Andrei, more than anyone, is responsible for bringing me to USC, for assembling the philosophy of law community that flourished while I have been here, and for setting me on the path of inquiry that led to this dissertation. The faith that Andrei had in my work from the very beginning, and the tireless support that he has offered throughout, have been an essential source of morale and confidence. I have learned as much from the work Pamela Hieronymi as I have from that of nearly any other philosopher, to the extent that a paper of hers—one of the very best on the subject, in my view—became a central point of reference in the disserta- tion. I was thrilled when she magnanimously agreed—in response to an email that I sent to her completely out of the blue—to be on my qualifying committee, and I have benefitted from her perceptive comments on early drafts of several chapters. Greg Keating has been an anchor of the USC legal theory community throughout my time here, a generous teacher and interlocutor in numerous seminars and reading groups, and a source of perspective and helpful feedback on my qualify- ing committee and now my dissertation committee. He has also provided a model of theoretically informed legal scholarship that I will always aspire to and be grateful for. Jon Quong has provided remarkably incisive and constructive criticism since he agreed to be on my qualifying committee. He has served as a kind of ideal skepti- cal reader—always going, with unfailing insight and accuracy, straight to the most critical or dubious points, and formulating his concerns with the kind of force and precision that makes them feel obvious in retrospect. I am sure that I have not ade- quately addressed all the issues that he has raised, but through his questioning I have come to understand the issues at stake much better, and I am at least confident that the final product is significantly improved from what it otherwise would have been. Acknowledgements vii My good luck when Mark Schroeder agreed to be a late addition to the dis- sertation committee, taking over Andrei’s spot, can hardly be exaggerated. But in a way this was only making official a role that Mark had already been filling without formal recognition (as he does for just about every grad student who passes through USC). If I were to accurately describe Mark’s qualities as an advisor—the amount of time he spends; the quality of his practical and professional advice; the insightfulness and above all the unique constructiveness and usefulness of his philosophical feed- back—no one who was not already familiar with these qualities would believe me. So I will leave it to those in the know to appreciate what I have to be grateful for. Last on this list but first among my teachers and advisors is Gary Watson, the chair of my dissertation committee. It is hard to do justice, by way of thanks, to having had the opportunity to write a philosophy dissertation under the direction of one of my favorite philosophers. It is quite something, at my rather advanced age as a student, to have the feeling of learning something genuinely new about what phi- losophy is actually capable of. In his written work and in person, Gary has served as a model of the kind of depth and discernment that the very best philosophical in- quiry can achieve. His guidance and example have gently corrected and redirected many (perhaps not quite all) of my worst intellectual habits, and he has shaped my philosophical technique and temperament—or at least, the philosophical technique and temperament that I aspire to—more than any other person. It has been a privi- lege that will always be an absolute highlight of my career. Finally, this dissertation has, unexpectedly and yet perhaps somehow inevita- bly, ended up being all about relationships. It would not be what it is—and I would not be who I am—without the people from whom I have learned the most about what good relationships are and what they are worth: my parents Cheryl and John Bero, my brother Alex, and my wife Beth Snyder. My debt to Beth, both personal and substantive, is the greatest of all. Indeed, it simply bears no comparison to the others; I will not try to reduce it to words. This dissertation is dedicated to her. Contents Introduction 1 1 Taking Responsibility and Relationships 7 2 Relationships and the Participant Stance 36 3 The Audience in Shame 76 4 Blame, Shame, and the Emotional Economy of Responsibility 110 Conclusion 155 References 157 1 Introduction This is an inquiry into responsibility and the emotional structure of relation- ships. It proceeds in two directions: on the one hand, I examine the emotional structure of our interpersonal relationships in the hope of gaining some insight into the nature of responsibility; on the other hand, I use responsibility as a point of entry into questions about interpersonal relationships that I believe have a broader signifi- cance. So the order of terms in the title should not be taken to mean that responsibility is the primary target; I think of the second item as in a sense the larger prize. Our theoretical understanding—as opposed, for instance, to empirical or lit- erary understandings—of the richness, complexity, and significance of our interpersonal relationships is, I claim, surprisingly thin. At the same time, such un- derstanding is of vital importance, in at least two respects. First, our relationships are much of the stuff that makes life worth living; as such they serve as focal points around which our lives are organized. Second, a great number of other things that we desire to understand better are, in sometimes unappreciated ways, fully intelligible only in terms of their roles in our lived relationships. In my view these things in- clude, among other things—and in addition to responsibility—attitudes like respect, trust, and love; character traits like sensitivity, loyalty, and integrity; and moral statuses like personhood, dignity, and autonomy. These claims are not theses that I aim to establish here; they are rather gen- eral insights that I hope to build towards by exploring a number of more manageable questions about responsibility and the emotional structure of relationships. The fol- lowing brief programmatic remarks are intended to provide some orientation, as well as a sense of the way in which pursuit of these larger insights guides the efforts of the following chapters. They will also help to situate this inquiry in relation to other work, and in particular the work of Peter Strawson and those who have been influ- Responsibility and the Emotional Structure of Relationships 2 enced by him. While I do not intend my central claims or arguments to turn on in- terpretive claims about Strawson, he has served as a starting place and a touchstone, and I take myself here to be trying to carry forward an important legacy of his work. Strawson started what can, without too much exaggeration, be described as a revolution in thinking about freedom and responsibility when he suggested, in his seminal 1962 article “Freedom and Resentment,” that responsibility should be un- derstood in terms of our sociality—that is, in terms of the characteristic ways in which we, as human beings, interact with and relate to one another. An important part of Strawson’s idea was that responsibility is, at its root, a matter not only of the connection between an agent and her actions, but also of the connections between the agent and the various other parties who, in the ordinary course of social life, are in a position to take an interest of one kind or another in those actions. And in par- ticular, Strawson thought that in order fully to understand the nature and conditions of responsibility we must consider the character and significance of the ways in which these interested parties are prone to respond to actions of different kinds. These responses prominently include the interpersonal emotional responses that Strawson dubbed the “reactive attitudes”—gratitude, resentment, indignation, and the like. In contemporary work on responsibility—as well as in moral theory more generally—Strawson’s influence is pervasive. Discussion of the reactive attitudes and their significance for understanding responsibility—as well as other normative no- tions like obligation and authority—is widespread. “Strawsonian compatibilism” has become a prominent school of thought in the free will and responsibility literature. In these ways, Strawson’s revolution has been a notable success. And yet in another respect, the revolution that Strawson began remains, in my view, incomplete. Strawson sought, among other things, to chart a path out of a familiar, nar- row circle of concepts that have typically been the stock-in-trade of philosophers who theorize about responsibility: reason, rationality, freedom, control, agency, de- sert, punishment, blame, and so on. Strawson urged us to broaden our view: We should think of the many different kinds of relationship which we can have with other people—as sharers of a 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. Introduction 3 Then we should think, in each of these connections in turn, and in others, of the kind of importance we attach to the attitudes and in- tentions towards us of those who stand in these relationships to us, and of the kinds of reactive attitudes and feelings to which we our- selves are prone (1962: 76). When we hold someone responsible, Strawson observed, we are prone in different situations to a range of different attitudes and feelings towards them—as well as as- sociated dispositions, intentions, and expectations—and these responses derive their meaning and importance from the kinds of relationships that we stand in to each other, and from the kinds of concern that we have for each other and for those rela- tionships. This is the way in which responsibility is embedded in our flesh-and-blood interpersonal interactions. By stepping back and situating responsibility within this broader horizon, Strawson thought, we could avoid a certain myopia that is charac- teristic of theoretical inquiry about responsibility, and put ourselves within reach of a richer and more satisfying understanding. While a number of the particular theoretical tools that Strawson intro- duced—particularly the reactive attitudes—have been enthusiastically adopted by subsequent writers, the full extent of his effort to broaden the horizon of inquiry has been met with more resistance. Rather than situating responsibility in the context of our social lives quite generally, for instance, there is a tendency to carve out a special dimension of our interpersonal relations as particularly relevant to responsibility, and to theorize about this special dimension in isolation. The designated domain is, con- sistent with a tendency that is natural in theorizing, often conceived in terms that are highly moralized, rationalized, intellectualized, and de-personalized; connections with other areas of human concern—attachments, motivations, and cares that are in vari- ous ways personal, intimate, affectionate, visceral, and so on—are frequently downplayed or ignored. I will say more to substantiate these claims in the following chapters; none- theless, the simple picture I have sketched in the previous paragraph is obviously a bit of a caricature. The point of its exaggerations is to highlight tendencies that are often subtle and implicit, but that nonetheless play an important role in setting the terms of discussion. Part of the impetus for this dissertation is a desire to counter- balance these tendencies, and push back towards the broader horizon indicated by Responsibility and the Emotional Structure of Relationships 4 Strawson. My sense is that something important has been lost, or at least diluted, as Strawson’s ideas have been disseminated and exerted their influence. It is at least a missed opportunity—and perhaps a substantive mistake—to see Strawson merely as developing new theoretical resources for neatly resolving existing questions about determinism and responsibility; the relationship dynamics that he identified have a significance that is both deeper and broader than this. And in particular, the original idea that animated Strawson’s introduction of the reactive attitudes—that our rela- tionships, understood quite generally, are structured and regulated by a broad but unified set of emotional responses—represents, in my view, a fundamental insight into the nature of our sociality, one that reaches well beyond questions about re- sponsibility. There is then an opportunity to return to what I take to be the original spirit of Strawson’s contribution, in pursuit of some more general insights. But at the same time, exploring broader questions about the structure of our relationships can only further expand our resources for understanding responsibility. The more connec- tions we can make between responsibility and a wider range of human concerns, the better we will be able to understand the nature, function, and significance of respon- sibility as it appears in our lives. One of the threads that runs through the following chapters is thus an effort to develop an account of a distinct class of attitudes that play a constitutive, structur- ing role in our relationships, in order to use this account to shed light on our responsibility practices. So my starting point will be questions about responsibility, but rather than dipping into our relationships only in order to extract resources that we can use to answer those questions—and then getting out as quickly and efficiently as possible—I will treat this as an opportunity for an extended exploration of the nature and structure of our interpersonal relations. Then I will return at the end to questions of responsibility, to see what further progress can be made with what has been learned. Chapter 1 begins with an inquiry into what it means to take responsibility. Drawing on recent work by Joseph Raz, I argue that taking responsibility is an im- portant but neglected aspect of our responsibility practices, complementary to but distinct from holding responsible (including holding oneself responsible). I embrace Introduction 5 Raz’s insight that taking responsibility serves as a way to respond to threats to our self-conceptions as agents, but argue that there is a significant gap in Raz’s account. To close the gap, I suggest that we attend to the way in which our self-conceptions as agents make reference to and incorporate the attitudes towards us of others with whom we stand in various relations. And this means that in order fully to understand responsibility, we need a better model of the emotional structure of those interper- sonal relations. Chapter 2 offers an account of this emotional structure in terms of a class of attitudes that I characterize in broadly Strawsonian terms—though I rechristen them the “participant attitudes,” to distinguish them from the narrower set of blaming re- sponses (typically resentment, indignation, and guilt) now commonly called the “reactive attitudes.” I go on to defend this classification against the objection that it is not sufficiently unified to be theoretically useful. The participant attitudes are, I propose, a discrete class of emotional responses that reflect the distinctive, personal way in which we care about the attitudes of others towards us. Once fully articulated, this proposal sheds light on a dynamic of reciprocal emotional involvement that is central to our interpersonal relations, yet easily overlooked or misunderstood. Chapter 3 demonstrates the explanatory power of this account by using it to shed light on a fundamental social emotion: shame. There is a longstanding theoreti- cal disagreement concerning the role in shame of an audience. An elegant solution to this problem, I propose, lies in conceiving of shame as a participant attitude—and thus as an emotional response to attitudes towards the subject—and attending care- fully to the distinction between the contents of the attitudes that shame is a reaction to, and the attitudes themselves. Notably, thinking of shame in these terms makes it easier to understand how it is that in shame we can serve, in a sense, as our own audience. Just as shame before others responds to an unwelcome gaze and reflects a sense that relations with others are not as the subject would wish, shame before one- self responds to an unwelcome self-assessment and reflects a sense that the subject’s relations with herself are not as she would wish. This approach brings diverse experi- ences of shame under a common account, and affords insight into the way in which we often take the participant stance towards ourselves. Responsibility and the Emotional Structure of Relationships 6 Finally, chapter 4 returns to the theme of responsibility, and uses the model of the emotional structure of relationships developed in chapters 2 and 3 to offer a deeper and more satisfying account of the fundamental emotional economy of re- sponsibility. There is wide agreement in the literature that to blame someone or to hold her responsible is something more than simply making an adverse assessment of her, but also something less than actually treating her adversely. The challenge is to locate something that lies between these two extremes yet is connected in the right way to both; this is often described as the problem of accounting for the special force of blame. Despite their resourcefulness, I argue that even the most sophisticated existing accounts do not fully account for the force that is present in many paradig- matic instances of blame. I propose that the missing element of force can be captured by understanding blame as a participant attitude, and by coming to terms with the distinctive emotional and personal character of blame that this implies. I also suggest that there is a natural fit between blame and shame that exhibits both atti- tudes’ roles in an interpersonal emotional economy of responsibility. This fit corresponds to the complementary fit between holding responsible and taking re- sponsibility, the topic with which we began. I do not think of the conclusions of chapter 4 as end points. The hope is that they might serve rather as illustrations of the kind of inquiry that it is possible to un- dertake once we have situated a notion like responsibility within the broader horizon of our interpersonal relationships and the emotional dynamics that structure them. 7 Chapter 1 Taking Responsibility and Relationships When the notion of responsibility has application in our lives, there are often (at least) two sides to the transaction: a wronged or injured party who holds the agent responsible, and the agent herself who (ideally) takes responsibility for her conduct. The first side of this transaction has been the object of intense theoretical scrutiny, due in large part to the influential conjecture that an agent is responsible if and only if it is (in some sense) appropriate to hold her responsible, and moreover that what it is for an agent to be responsible perhaps just is for it to be appropriate to hold her responsible. 1 According to this way of thinking, understanding what it is to hold someone responsible, and when it is appropriate to do so, takes us to the very heart of responsibility. The preoccupation with holding responsible—with its focus on the perspec- tive of the wronged party, and on responses like blame, reproach, and sanction— contrasts with a relative neglect of the other side of the transaction, including the perspective of the agent and the role that she has to play by taking (or refusing to take) responsibility. 2 The aim of this chapter is to establish that taking responsibility 1 This conjecture, at least in its contemporary versions, can be traced to P.F. Straw- son’s “Freedom and Resentment” (1962). 2 This neglect has not been universal; for instance, reconciliation theorists like Mar- garet Urban Walker (2006) and Linda Radzik (2009) and punishment theorists like Antony Duff (2001, 2003), John Tasioulas (2006, 2007), Christopher Bennett (2008), and Ambrose Lee (2017) have made important contributions to our understanding of taking responsibility and the agent’s role in the aftermath of wrongdoing. In my view, however, reconciliation theorists’ primary focus on the end state of relational repair, and punishment theorists’ primary focus on punishment and its justification, lead both groups largely to neglect (and in some cases, I would argue, to misrepre- sent) the underlying emotional economy of responsibility in which the phenomenon of taking responsibility is embedded. One main goal of this dissertation is to better understand this emotional economy. Responsibility and the Emotional Structure of Relationships 8 is theoretically interesting in its own right—that it is an important and independent phenomenon, complementary to but distinct from holding responsible. To do this I will borrow a key idea from Joseph Raz’s (2011) account of responsibility for inad- vertent conduct, though I will first need to extract the idea that I want from Raz’s larger account, which I think faces significant problems. If successful, the result of this effort will be to place taking responsibility in a wider emotional economy that plays a central role both in our self-understandings (as Raz’s account suggests) and in our interpersonal relations. Later chapters will fill out the picture of this interper- sonal emotional economy, and eventually return to apply this picture to advance our understanding of the complementary phenomenon of holding responsible. 1 Holding responsible and taking responsibility First we need a clearer sense of how I am understanding the difference be- tween holding responsible and taking responsibility. One initially appealing thought is that taking responsibility is not really distinct from holding responsible, but rather just boils down to holding oneself responsible. This thought is perhaps encouraged by the Strawsonian idea that holding someone responsible is fundamentally a matter of hav- ing a type of emotional reaction towards them that comes in the personal, vicarious, and reflexive versions that Strawson labeled resentment, indignation, and guilt, re- spectively (1962: 83-85). According to this idea, my holding you responsible for an action that showed ill-will towards me is fundamentally a matter of (or involves, or is somehow connected to) my resenting you; my holding you responsible for an action that showed ill-will towards a third party is a matter of my feeling indignant towards you—which is another way of saying that that I feel resentment, on behalf of the third party, towards you; and my holding myself responsible for an action that showed ill-will towards another is a matter of my feeling guilty—which is another way of saying that I feel resentment, on behalf of that other, towards myself. This way of carving up the phenomena suggests the hypothesis that what it is to take re- sponsibility just is to hold oneself responsible, in particular by undertaking the emotional burden of guilt and being disposed to respond and to act in the ways that are associated with a sense of guilt. Chapter 1 | Taking Responsibility and Relationships 9 But I wish to suggest that this turns out, on closer inspection, to be incorrect. My case for this conclusion will ultimately rest on the strength of the positive ac- count of taking responsibility that will emerge over the course of this chapter. But to set the stage and provide some initial motivation for this enterprise, I will offer a few suggestive observations in favor of distinguishing between holding oneself responsi- ble and taking responsibility. One reason for thinking that we have here two distinct phenomena is that we can identify two different sets of emotional reactions that are distinguished by their characteristic objects. On the one hand are the emotional reactions that you, the agent, have towards yourself as wrongdoer; these include guilt, in the standard Straw- sonian sense, and more generally self-directed anger, blame, frustration, and the like (perhaps also disappointment, contempt, and a range of other responses). These are self-directed versions of the attitudes that we take when holding another person re- sponsible for their wrongdoing, and they are associated with self-directed versions of the dispositions and responses that are associated with holding another responsible, including reproach, instruction, sanction, and so on. On the other hand are the emotional reactions that you, as agent, have to- wards another person as someone affected by your conduct and as a party to the relationship that you stand in to them. Here anger, blame, reproach, and the like are clearly out of place. What is called for instead are contrition—or as we say, feeling sorry—towards the person you have wronged and remorse for the effect that you have had on them. 3 As- sociated with these reactions are dispositions to apologize, to “undo” the wrong, to make amends, and so on. This distinction between guilt and contrition (or feeling sorry) is observable not only in principle but also in practice. It is possible, and indeed fairly common, to feel sorry for something you don’t feel guilty for, and vice versa. For instance, sup- pose you make plans to see a friend who is very eager to spend time with you, after having cancelled several previous meetings (all with excellent reason). As luck would have it, when the time comes for the meeting another legitimate crisis arises and you 3 Another response that is likely to be called for, particularly if the person you have affected is resentful or disapproving of your conduct, is shame. For further discussion of the role of shame in this dynamic, see chapter 4, § 5. Responsibility and the Emotional Structure of Relationships 10 have no choice but reluctantly to cancel again. In such cases it is common and, I think, quite reasonable not to feel guilty (after all, you had excellent reason to can- cel—there’s nothing else you could reasonably have done), but at the same time to feel sorry towards your friend for letting her down again. And this feeling will tend to manifest itself in a disposition to apologize to her, a desire to “make it up to her,” and so on, rather than in a disposition to reproach or sanction yourself. When you think about what happened, the focus will primarily be on your friend, and on the disappointment you have occasioned, rather than on yourself, your choices, or your conduct (though it is of course an important premise of feeling sorry that it was you who occasioned this disappointment—in this way contrition is distinct from imper- sonal regret). Now consider a case in which you recklessly risk the safety of an indetermi- nate number of people, but through sheer good luck the risk of harm is never realized. For instance, on a night out you have several more drinks than you planned, and foolishly decide to drive home when you should know better. On waking the next morning, you might feel guilty about what you’ve done (“What was I thinking? I could have killed someone!”), but without feeling sorry towards anyone—after all, no one was hurt, and the people whose safety you risked are unidentifiable. This feeling will naturally manifest itself in a disposition, for instance, to reproach yourself, rather than in a disposition to apologize or make amends (to whom?). When you think about what happened, the focus will primarily be on yourself, your choices, and what you did, rather than on any other person. For another illustrative example of feeling guilty without feeling sorry we can repurpose a case imagined by R. Jay Wallace (1994: 43), in which a lapsed Catholic feels irrationally guilty for having recreational sex, despite no longer accepting the rule that this guilt is based on; while Wallace does not remark on this aspect of the case, presumably this person need not also feel sorry. He may, for instance, be angry at himself and disposed to reproach himself for his lasciviousness, but without being disposed to apologize to, or feel sorry towards, anyone at all. The previous two examples are useful for isolating feelings of guilt from feel- ings of contrition because they involve no other person to whom it would make sense to feel sorry. But it is quite possible to have only feelings of guilt even in cases Chapter 1 | Taking Responsibility and Relationships 11 where there is a readily identifiable victim. Suppose you allow yourself an intemper- ate remark addressed to a person for whom you have unmitigated contempt. This person, you are convinced, deserved no better—indeed, much worse than what you said. And yet . . . you prefer to hold yourself to a higher standard. Far from feeling sorry or disposed to apologize, you might continue to take satisfaction in having given this person a bit of what they had coming, even as you blame yourself for your lack of self-control. In this way, you might feel guilty, be disposed to reproach your- self, and so on, despite not being at all sorry towards the person who bore the brunt of your remark. There are then, I submit, two sets of emotional reactions that are clearly dis- tinguished from one another by their characteristic objects, as well as by their associated dispositions and responses. And these different sets of emotional reac- tions correspond to a further distinction between what I am calling holding oneself responsible and taking responsibility. Whereas the characteristic emotional dimension of holding oneself responsible is feeling guilty, the characteristic emotional dimension of taking responsibility is feeling sorry. Guilt, and holding oneself responsible, are self- directed; contrition or feeling sorry, and taking responsibility (in the sense that I am concerned with), are directed towards the person whom one’s conduct has negatively affected. Thus, the same cases that illustrate how someone can feel sorry but not guilty and vice versa, also illustrate how it is possible both to hold oneself responsi- ble without taking responsibility, and to take responsibility without holding oneself responsible (though of course it is common and often appropriate to do both). The person who misses an appointment with a friend for excellent reasons may take re- sponsibility (with respect to her friend) without holding herself responsible; similarly, the lucky drunk driver may hold herself responsible without taking responsibility (with respect to anyone). As these remarks make clear, I am giving special, stipulative senses to the terms holding responsible and taking responsibility, in order to bring out a distinction that everyday usage does not always clearly demarcate. In ordinary parlance, it can make perfect sense to say that in apologizing to your friend for the missed appointment and making it up to her, you are holding yourself responsible, in the broad sense that you accept that the missed appointment is something that it is your responsibility to Responsibility and the Emotional Structure of Relationships 12 respond to. Similarly, it can make perfect sense to say that by entrusting her keys to a friend on future nights out, the lucky drunk driver is taking responsibility for her conduct, in the sense that she is stepping up and assuming the responsibility to do something about it. We do not typically have occasion to distinguish these various senses of responsibility, and the taking versus the holding of it, in ordinary speech. 4 Nonetheless, the examples we have considered highlight a real and, I will argue, theoretically significant distinction that I will use the terms holding responsible and tak- ing responsibility to identify. We now have at least some reason to think that the phenomena that I am re- ferring to as taking responsibility and holding oneself responsible are distinct, but we lack as yet an analysis or account of taking responsibility—of what it is, what it means, and why we do it—that makes good on this thought. The goal of the remain- der of this chapter is to develop such an account. But we will proceed indirectly, via an extended detour through Joseph Raz’s novel and resourceful treatment of respon- sibility for inadvertent conduct. Understanding that treatment, and the ways in which I think it goes awry, will put us within reach of a significant advance in our under- standing of taking responsibility for conduct quite generally. 2 Raz on responsibility for inadvertence Raz is concerned with what we might call the puzzle of responsibility for in- advertent conduct. He does not observe the distinction drawn in the previous section between holding responsible and taking responsibility; he tends in some re- spects to lump them together, and he uses language like “holding oneself responsible,” “taking responsibility,” and “acknowledging one’s own responsibility” interchangeably. But as we will see, Raz is ultimately concerned with (what I am call- ing) our practice of holding ourselves responsible, rather than taking responsibility; indeed, one of his innovations is to offer an explanation of our responsibility for in- advertent conduct in terms of the special reasons that we have to (as I would say) hold ourselves responsible for such conduct. In discussing Raz’s account, I will ac- cordingly use the language of holding responsible. 4 On the different senses of “responsibility,” see also note 7 below. Chapter 1 | Taking Responsibility and Relationships 13 The puzzle of responsibility for inadvertent conduct arises in the following way. On the one hand, we often hold ourselves and others responsible for various inadvertent failures: forgetting an appointment, unintentionally bumping into some- one, botching an operation, and so on (231 5 ). On the other hand, this can seem mysterious: It is (relatively) easy to see why we are “on the hook” for successful con- duct that turns out just the way we intended, but harder to see why we should be on the hook for conduct that results inadvertently from failures of our powers (232). In particular, responsibility for inadvertent conduct appears to violate a plausible condi- tion that would limit responsibility to conduct over which we successfully exercise some sort of control, and thereby subjects us to a potentially troubling kind of moral luck (243). 6 There seems then to be at least some reason to think that we are on firmer ground in holding agents (including ourselves) responsible for conduct that is successfully guided by their powers than we are in holding them responsible for con- duct that isn’t. Raz aims to show that this is wrong. According to the principle of responsi- bility that he defends, we are responsible for “conduct that is the result of the functioning, successful or failed, of our powers of rational agency, provided those pow- ers were not suspended in a way affecting the action” (231, emphasis added). 7 By “powers of rational agency” Raz means quite broadly all of the powers that we can call upon in the service of our activity as rational agents who respond to reasons, form intentions, and execute those intentions in normal ways. Our powers of ra- 5 Unless otherwise indicated, parenthetical page references in this chapter are to Raz 2011. Most references are to chapter 12, “Being in the World” (227-54), though there are occasional references to other chapters, especially chapter 13, “Responsibil- ity and the Negligence Standard” (255-69). 6 For discussion of the conflict between moral luck and a plausible control condition on responsibility, see generally Enoch & Marmor 2007. 7 As we will see, for Raz an agent’s “being responsible” is a matter of the reasons that the agent has to hold herself responsible. Among other things, this means that Raz’s principle concerns what we might call “backward-looking” responsibility for past conduct, rather than forward-looking responsibility (as in, “It is her responsibility to secure the building.”) or responsibility as a status of competent rational agents (“She is not of sound mind; she is not responsible.”). Raz uses subscripts to distinguishes between these senses of “responsibility,” and refers to backward-looking responsibil- ity as “responsibility 2 ” (227-28). My focus throughout will be on backward-looking responsibility, and I will accordingly suppress the subscript in quotations from Raz. Responsibility and the Emotional Structure of Relationships 14 tional agency thus do not include powers that are merely physiological or otherwise entirely unresponsive to reasons, such as our power of digestion (2). But they do in- clude our sub-personal capacities of memory, attention, perception, bodily control, will power, and all of the other abilities that we regularly call upon to translate our intentions into conduct (227). Raz’s principle, then, makes us responsible not only for conduct that results from the successful functioning of our powers of rational agency, but also for con- duct that results from failures of those powers 8 ; as a result, it makes us responsible for a wide range of conduct that results from lapses of memory, coordination, will power, and all the rest (including some conduct traditionally classified as “negligent” (244)). Thus, when you fail to pick up your child from school because you simply forgot (267); when you hit the car in front of you because your foot simply slipped off the brake (244); and so on, you are responsible for those inadvertent lapses, ac- cording to Raz. 9 But these sorts of examples alone do not quite convey the breadth of this view of responsibility for inadvertent conduct. Consider that one familiar position in the literature on responsibility is that we can be responsible for an inadvertent lapse of memory, coordination, or the like, provided that the lapse is a manifestation of an objectionable character trait, attitude, or judgment. The idea is that, for instance, someone can be responsible for forgetting a friend’s birthday if the forgetfulness is a manifestation of a general lack of sufficient concern or care for her friend. 10 Another common thought in the literature on responsibility is that an agent can be responsi- ble for an inadvertent lapse of memory, coordination, or the like, provided that the 8 With the important qualification, to be explained below, that we are only responsi- ble for failures of our powers that fall within what Raz calls our “domain of secure competence” (244-45). 9 The lapses at issue are all simple in a sense defined and discussed by Julie Tannen- baum; namely, “there is no further causal explanation that will make a difference to our moral evaluation of that lapse” (2015: 62). By contrast, when a lapse is not sim- ple, whether the agent is responsible for the lapse will depend on the causal explanation for the lapse; for instance, the kind of character-based or tracing expla- nations discussed below may be inculpatory, while explanations that involve excusing conditions like duress or temporary incapacity may be exculpatory (cf. Tannenbaum 2015: 63). 10 This now-standard example is from Smith 2005. Chapter 1 | Taking Responsibility and Relationships 15 lapse is in some way causally traceable to a previous error on her part that was not inadvertent. For instance, an agent can be responsible for forgetting a birthday if she had previously deliberately decided not to bother with marking the birthday in her calendar, consciously disregarding the risk that she might later forget. Raz’s view goes further than either of these ideas: He says that someone who simply forgets a friend’s birthday—her memory just fails her on this occasion, as memories sometimes do—is (barring the presence of a condition, like illness or seizure, that suspends the agent’s powers of rational agency in a way affecting her conduct) responsible for the lapse even if her concern and care for her friend and her prior conduct have both been beyond reproach, or even exemplary. The breadth of Raz’s view, it should be noted, is somewhat mitigated by the fact that he considers responsibility to encompass more than just blameworthiness or liability to sanctions. 11 He deplores the “mesmerizing fascination with responsibility as a condition of nothing but praise and blame,” which while “natural in our blame soci- ety, misses the central role of responsibility” (265). Instead, he stresses that “[t]here is a whole variety of responses and attitudes mandated by responsibility, other than blaming or praising,” including for instance, at “the practical end of the range . . . the duty to make amends” (251, 265). 12 Responsibility, in Raz’s sense, may thus open an agent up to blame in some cases, but in others it may license only other, less harsh attitudes and responses. 13 Nonetheless, Raz’s view makes us responsible, in one way or another, for a very broad range of conduct. 11 For discussion of this feature of Raz’s view, see Gary Watson 2014: 3-4. Contrast, e.g., Gideon Rosen’s stipulation: “Just as legal responsibility amounts to liability to legal sanctions (monetary damages, criminal punishment), so moral responsibility, as I understand it, amounts to liability to ‘moral sanctions’” (2004: 296). 12 Cf. Julie Tannenbaum’s discussion of “mere moral failure” (2015). Tannenbaum argues that in addition to blameworthiness and innocence, there is an intermediate condition of mere moral failure that licenses an emotional reaction that lies, in some sense, between full blame and mere regret (including agent-regret). Raz’s understand- ing of responsibility, on my reading (though see the next footnote), leaves room for this kind of non-blaming response to a responsible agent. 13 In one place, however, Raz seems to draw a closer connection between his notion of responsibility and blame. “Still,” he remarks, “responsibility is also a condition of blameworthiness. Agents are to blame for conduct for which they are responsible unless they are excused. It follows that . . . people are to blame for conduct that they do not control, i.e. that they are subject to ‘moral luck’” (251). These comments are Responsibility and the Emotional Structure of Relationships 16 We should here pause to observe that Raz’s enumeration of the attitudes and responses that may be licensed by responsibility includes responses associated with holding responsible (blame) as well as responses associated with taking responsibility (making amends). But in actually developing his account, as we will see, Raz focuses exclusively on attitudes and responses that the agent takes towards herself as agent, rather than attitudes and responses that the agent takes towards the person affected by her conduct. This is why I read Raz as offering an account of our practice of holding our- selves responsible, rather than of taking responsibility. I will accordingly understand Raz’s comments to the effect that responsibility encompasses more than blamewor- thiness as meaning that in some cases responsibility may involve liability only to responses that are less harsh than blame. Raz’s case for his broad view of our responsibility for inadvertence draws on the notion of a “domain of secure competence” that each of us possesses (244-45). The powers of rational agency that we “securely command” define the boundaries of our domains of secure competence (268). Within our domains of secure competence we can be confident that (barring some competence-defeating condition, like a sei- zure) if we undertake to act we will succeed; we are entitled to act without reflecting on the prospects for success (244-45). For instance, if I sit down to play chess with a grand master, I can try to win but I cannot be confident of success; beating such a player is not within my domain of secure competence. I can, however, be confident that I will succeed in moving my pieces in accordance with the rules of chess and, puzzling and apparently inconsistent with other features of the account. Elsewhere, for instance, Raz says that it is “a conceptual truth that negligent conduct is a prima facie wrong for which one is responsible” (260) and that liability for negligent con- duct, “not being subject to excuses, applies also to cases which do not warrant blame” (265). Negligent conduct, being a form of responsible conduct that cannot be excused and yet may not warrant blame, thus constitutes a counterexample to the claim that “agents are to blame for conduct for which they are responsible unless they are excused.” I think the most charitable way to resolve this inconsistency, given the overall shape of Raz’s account, is simply to ignore the latter claim as mistaken; I thus take it that for all Raz says it is at least an open question whether an agent can be responsible but not blameworthy for unexcused conduct. This also means that Raz is too quick to conclude that his account allows for blame in cases of moral luck; his account explains how agents can be responsible in moral luck cases, but because the account leaves the connection between responsibility and blame quite loose, it does not entail that agents can be blameworthy in moral luck cases. Whether they can is a further question that Raz’s account (as I read it) does not address. Chapter 1 | Taking Responsibility and Relationships 17 more fundamentally, that I will be physically able to grasp and move the chess pieces, because these more modest skills are within my domain of secure competence. Agency in general, Raz argues, presupposes the stable existence of a domain of secure competence. It is our confidence in our domains of secure competence, for instance, that allows us to deliberate, decide, and act without constantly second- guessing ourselves or going back to verify the soundness of our deliberations (250); in the same way, we rely on our domains of secure competence to navigate our physical world without constantly questioning and verifying the nature and extent of our basic abilities to control our bodies and manipulate our surroundings. Without taking a substantial domain of secure competence for granted, we would be unable to rely unreflectively on our various lower-level capacities in order to perform any complex action, and would be effectively paralyzed. 14 Because possession of a domain of secure competence is an essential condi- tion of effective human agency, Raz contends, it is also central to our self- conceptions and identities as agents. Our self-esteem, our self-respect, our pride (or shame) in who we are and what we can do, our sense of our own potential—and thus our projects and ambitions (245)—“all these and various other self-directed atti- tudes and emotions depend in part on competence in using our faculties of rational agency” (268). To have a shorthand for the complex of self-directed attitudes (in- cluding relevant emotions, beliefs, evaluative judgments, etc.) that Raz has in mind, I will refer to them collectively as the agent’s “self-conception as an agent.” For Raz, our self-conceptions as agents comprise many (though not all 15 ) of those self- directed attitudes that are most critical to our senses of who we are, of our place in the world, and of our own worth. And according to Raz’s account, our self- 14 Compare George Sher’s related observation about memory: “No agent could func- tion at all if he did not have confidence that his mind will, just of its own accord, dip into his memory bank to deliver up just the information he needs at just the time he needs it” (2009: 127). 15 Other important elements of our self-understandings include, for instance, “gender or ethnicity and their social meanings” (239). In future chapters I will shift to talk of our “self-understandings” in order to distinguish a larger set of self-directed attitudes from the narrower set that Raz focuses on, which I am referring to as our “self- conceptions as agents.” See particularly chapter 2, § 5. Responsibility and the Emotional Structure of Relationships 18 conceptions as agents depend on establishing and maintaining the domains of secure competence that we claim for ourselves. Because our domains of secure competence play a central role in our self- conceptions as agents, Raz concludes that any failure to perform an act that the agent considers to be within her domain of secure competence has a special signifi- cance—it threatens to undermine her very self-conception as an agent: Failure to control conduct within our domain of secure competence threatens to undermine our self-esteem and our sense of who we are, what we are capable of, etc. We must react to it. We may conclude that we are no longer able securely to perform that kind of action. We have grown frail, our competence is diminishing. We come to recognize our limitations. Commonly this is not the case, and we do not allow it to be. We assert our competence by holding ourselves re- sponsible for it. To disavow responsibility is to be false to who we are (245). 16 When this kind of failure occurs, Raz says, “We must react to it,” and there are two options: the agent can either revise her self-conception by conceding that her domain of secure competence is not as wide as she thought, or she can reassert her compe- tence by holding herself responsible for the failure. (By this point it is clear that Raz is offering an account of our practice of holding ourselves responsible, not of taking responsibility. The agent’s focus in his account is squarely on herself, her perform- ance, and her self-conception as an agent, rather than on the affected party, any harm he may have suffered, or his relationship to the agent.) This need to uphold our self-conceptions as agents, according to Raz’s ac- count, explains our practice of holding ourselves responsible for conduct that results from failures of our powers of rational agency. That practice upholds our self- conceptions as agents by defending and maintaining our sense of our domains of secure competence: “In acknowledging responsibility for actions due to our rational powers we are simply affirming that they are our secure rational powers” (245). In other words, by holding ourselves responsible we remove the threat that failures of our powers of rational agency pose to our domains of secure competence, and 16 Raz’s canonical formulation of his principle of responsibility for inadvertence (231) does not explicitly include the restriction that we are only responsible for failures of our powers of rational agency that fall “within our domain of secure competence.” This appears, however, to be an oversight, as the quoted passage indicates. Chapter 1 | Taking Responsibility and Relationships 19 thereby reaffirm our self-conceptions as agents. 17 By the same token, “To disavow responsibility is to be false to who we are,” because it is equivalent to renouncing our own self-conceptions as agents (245). 18 In the next section I will adduce what I think are two significant problems for this account of responsibility for inadvertence, but first it is worth pausing to ap- preciate how much there is to admire in Raz’s proposal. By introducing the idea of a domain of secure competence and elaborating its role in our agency, Raz connects responsibility to what may have seemed to be a mostly unrelated area of human con- cern: our vital interest in maintaining our own self-conceptions as agents. This furnishes us with some genuinely new materials for understanding our responsibility practices. Indeed, the idea that responsibility is connected to our self-conceptions as agents strikes me as an important insight, and I will try to develop it in a different way in § 4, in connection with what I am calling taking responsibility. 3 Two doubts about Raz’s account But first I have two doubts about Raz’s own approach. The first doubt con- cerns whether the kinds of threats to our self-conceptions as agents that Raz has in mind are anywhere near as widespread as he suggests. This doubt is significant be- cause it suggests that in order to make Raz’s underlying idea theoretically useful we 17 What about one person holding another responsible? This appears to be an after- thought for Raz, but he does at least briefly indicate that when others hold us responsible for inadvertent conduct within our domains of secure competence, they “acknowledge our mastery of those abilities,” and thereby express their respect for (and perhaps also appeal to or invoke) our self-conceptions as agents and our rea- sons to defend and maintain those self-conceptions (268). 18 To be clear, for Raz holding myself responsible for my conduct is something more than simply acknowledging that the conduct in question is my conduct. For instance, I have forgotten what I had for dinner ten years ago today. This instance of forget- ting is mine; it is a normal product of my very average memory, and thus not something that I can disclaim (“It isn’t really me who forgot.”). In this broader sense, we might say that I am “responsible” for forgetting—in the sense that I am forced to admit that my memory is not perfect and that it is in fact I who have forgotten. At the same time, however, in another sense there is no pressure on me to hold myself responsible for forgetting (e.g., blame myself, reproach myself, or whatever), because remembering such a minor detail so many years later is well beyond the scope of the domain of secure competence that I claim for myself. In this sense—the sense that interests Raz—I am not responsible for forgetting, even though it is none other than I who forgot. Responsibility and the Emotional Structure of Relationships 20 may need an alternative understanding of what constitutes a threat to our self- conceptions as agents. The second, more serious and fundamental doubt concerns whether Raz’s account succeeds in identifying genuine reasons to hold ourselves re- sponsible. Raz’s difficulties on this score will motivate a shift from thinking in terms of holding oneself responsible to thinking in terms of taking responsibility. 3.1 What is the threat? In Raz’s account, failures of our powers of rational agency are significant be- cause they threaten our domains of secure competence, which are critical to our self- conceptions as agents. The first doubt concerns whether the kinds of threats that Raz has in mind are as widespread as he suggests. We are all, of course, familiar with failures of our powers of rational agency that reveal something surprising and unwelcome that casts doubt on our domains of secure competence: I set out confidently for my destination, and quickly realize I am lost; I confidently launch into an explanation of a philosophical argument, and soon trip myself up in confusion; I confidently take aim and miss the target by a mile, not once but repeatedly and consistently, even under favorable conditions. These types of failures may indeed reveal that I am not as good as I thought I was, and this often can be jarring to my sense of self. 19 But the difficulty for Raz’s account is that not all failures are like this. Return again to Raz’s own examples: you fail to pick up your child from school because you simply forgot (267); you hit the car in front of you because your foot simply slipped off the brake (244), and so on. Of course it may well be appropriate in such cases to hold ourselves and others responsible (or, for that matter, to take responsibility); the doubt I wish to raise concerns not the appropriateness of such responses, but rather the explanation that Raz offers of this appropriateness. For in these cases there may be no reason to revise our view that your memory is generally reliable, or that you are generally a competent driver. Even the best memories, even the best drivers, sometimes fail. We know this. Raz’s account claims that we must hold ourselves re- 19 Though even in these cases, we need not accept Raz’s suggestion that holding one- self responsible for the failure functions to remove the threat that it poses to my self- conception. See § 3.2 below for further discussion of the relationship between hold- ing oneself responsible and threats to our self-conceptions as agents. Chapter 1 | Taking Responsibility and Relationships 21 sponsible even for these failures, on pain of being “false to who we are” (245). But why should truly one-off failures threaten our self-conceptions as agents in the first place? It cannot be that in order for our domains of secure competence to play their assigned role in our self-conceptions as agents, we must believe that we can never fail within our domains of secure competence. As Raz acknowledges: “We are always liable to fail to control actions within our sphere of secure competence, even when no competence-defeating condition obtains” (246). So in placing a glass on the table, to take another of Raz’s examples (245), I am ordinarily entitled to presume for prac- tical purposes that my efforts to judge spatial relationships, to grip firmly but not too firmly, and so on, will succeed as normal, even while I continue accurately to believe that these powers are liable to a small but real chance of failure. Our question, then, is: If our domains of secure competence can play their assigned role in our self- conceptions as agents consistent with our accepting that we are always liable to some small possibility of failure, why should the occasional realization of that possibility— this time, the glass tumbles off the table when I try to set it down—threaten our self- conceptions as agents? Gary Watson (2014b) has raised a concern related to this one, focusing on an aspect of Raz’s view that I have so far ignored: his proviso that we are responsible for conduct that is the result of failures of our powers of rational agency only “pro- vided those powers were not suspended in a way affecting the action.” This proviso reflects Raz’s recognition that we are not responsible if certain familiar competence- defeating conditions obtain: seizure, sleep, hypnosis, temporary paralysis, etc. (231, 244). Such conditions, Raz explains, “negate the proper connection between the ac- tion and the powers of rational agency” (231). But Watson suggests that this proviso threatens to expand until it erases the distinctive breadth of Raz’s view (2014b: 407- 08). After all, every malfunction of our powers of rational agency will be explained by some factor or other; how then can we distinguish the competence-defeating fac- tors from the factors that merely explain a failure without being competence- defeating? When our powers of rational agency simply fail in some case, why should we suppose that “nothing relevantly similar to what goes wrong in competence- defeating circumstances obtains?” (2014b: 408). Why, that is, should we not assume Responsibility and the Emotional Structure of Relationships 22 that behind every inadvertent failure there is a condition that “negate[s] the proper connection between the action and the powers of rational agency”? If Raz has no criterion for distinguishing competence-defeating conditions from other factors that explain our failures, Watson suggests, then it would seem that he has not explained why we should ever accept responsibility for conduct that results from failures of our powers of rational agency. Watson’s concern and the question that I have raised are, I believe, related and complementary. Watson points out that Raz provides no principled criterion to distinguish between competence-defeating conditions and other factors that explain our failures. I am further suggesting that the absence of such a criterion from Raz’s account is symptomatic of an underlying failure to explain why inadvertent errors that do not give us some affirmative reason to doubt or reevaluate our domains of secure competence should be experienced as threats to our self-conceptions as agents in the first place. This further aspect of the problem is important, because it means that Raz’s account cannot be saved simply by formulating a more-or-less workable criterion for distinguishing between competence-defeating conditions and the other conditions that explain our failures; also needed is an explanation of exactly how it is that the latter conditions constitute threats to our domains of secure com- petence. Various criteria for distinguishing the two types of conditions might suggest themselves. For instance, we could try to distinguish conditions that tend to render an agent incompetent across a wide range of activities, or for extended periods of time, from conditions that typically affect only one area of competence, or only for limited periods. Alternatively, we might distinguish conditions that typically result in more serious, out-of-the-ordinary failures, from conditions that typically result in more minor, run-of-the mill failures. Or, we might distinguish conditions whose competence-defeating character is broadly recognized—that is, the way in which they “negate the proper connection between the action and the powers of rational agency” is more or less well understood, and they are subject to relatively stable, public, and transparent standards of assessment—from conditions that are idiosyn- cratic, heterogeneous, poorly understood, and subject to no stable, public standards. Chapter 1 | Taking Responsibility and Relationships 23 Any of these candidate criteria would face serious questions concerning whether they succeed in picking out all and only the kinds of familiar competence- defeating conditions—seizure, sleepwalking, etc.—that Raz has in mind. 20 But the deeper and more serious problem is that none of these proposed criteria, taken by itself, helps us to understand why it is that one-off failures that do not give us reason to doubt or reevaluate our domains of secure competence should be experienced as threats to our self-conceptions as agents in the first place. That is, these distinctions are all consistent with the view that, whenever our powers of rational agency simply fail in some case, something “relevantly similar to what goes wrong in competence- defeating circumstances obtains” (Watson 2014b: 408), such that underlying the fail- ure is a condition that “negate[s] the proper connection between the action and the powers of rational agency” (231). A more-or-less workable criterion to distinguish between two types of condition is of little use if we still lack any explanation of why either type of condition should threaten our self-conceptions as agents. Perhaps, however, we could pursue, on Raz’s behalf, a more direct approach. Instead of distinguishing between competence-defeating conditions and other condi- tions that explain our failures, perhaps we could simply distinguish directly between failures that do, and those that do not, give us reason to doubt or reevaluate our do- mains of secure competence. Failures of the former type—for instance, suspicious or repeated failures that give an agent reason to suspect that her domain of secure com- petence is not as wide as she thought—would then be regarded as threats to our domains of secure competence, while failures of the latter type—for instance, fail- ures due to temporary, extraordinary competence-defeating conditions, as well as one-off, isolated failures—would not. Unlike the previous proposals, this approach does not face the further chal- lenge of showing what the distinction has to do with threats to our self-conceptions as agents: the proposal is simply to distinguish directly between threats and non- threats. But this proposal effectively concedes rather than resists the point that 20 For instance, regarding the distinction between longer-lasting conditions and more temporary ones, Watson observes: “The answer cannot simply be that what goes wrong [in the case of non-competence-defeating conditions] is episodic, because competence-defeating conditions might in principle and practice be one-off (a ‘mini- stroke’ causing confusion or motor disruption)” (2014: 14). Responsibility and the Emotional Structure of Relationships 24 threats to our domains of secure competence are much less widespread than origi- nally suggested; in so doing it sacrifices the breadth that was distinctive of Raz’s account and largely abandons his ambition of resolving the puzzle of responsibility for inadvertent conduct. Accepting this proposed fix would mean, for instance, that a driver whose foot slipped off the brake, causing a serious accident, would only have reason to hold herself responsible for the accident if the slip gave her reason to doubt her general competence as a driver; if the slip was just a one-off failure that gave her no reason to doubt her general competence, she could disclaim responsibil- ity. Moreover, this would seem to have the further strange implication that a driver whose foot slipped off the brake could avoid responsibility for the resulting accident if she could successfully demonstrate afterwards that she remained generally compe- tent to operate a car safely. If this first doubt is sound, then Raz’s account of responsibility for inadver- tence is significantly narrower than originally claimed, and it fails to track our existing practices in important ways. In § 4 I will suggest that we can better leverage the theo- retical power of Raz’s underlying idea by shifting to intentional conduct and exploring a quite different idea about what constitutes a threat to our self- conceptions as agents. As we will see, this will also involve a shift in focus from our reasons for holding ourselves responsible to our reasons for taking responsibility. But first, there is a second doubt about Raz’s account. 3.2 Why hold oneself responsible? Consistent with the doubt raised in the previous section, Raz’s account could still justify our practice of holding ourselves (and, by extension, others 21 ) responsible for certain instances of inadvertent conduct (namely, those instances in which fail- ures of our powers of rational agency provide an affirmative reason to reevaluate our domains of secure competence)—were it not for a second source of doubt. This more serious and fundamental doubt about Raz’s account concerns whether the ac- count really identifies reasons to hold ourselves responsible. 21 See note 17 above. Chapter 1 | Taking Responsibility and Relationships 25 Raz claims that holding ourselves responsible functions to remove the threat to our self-conceptions as agents that certain failures of our powers can pose. But his account offers no explanation of how holding ourselves responsible could perform this function. Raz contends that because a failure to perform an act within our do- mains of secure competence threatens our self-conceptions as agents, “We must react to it” (245). To answer this challenge, he says, “We assert our competence by hold- ing ourselves responsible for it”; or again, “In acknowledging responsibility for actions due to our rational powers we are simply affirming that they are our secure ra- tional powers” (245, emphases added). But how exactly does holding ourselves responsible for a failure serve to reassert our competence? The mechanism is deceptively difficult to identify. Following Raz’s own re- marks, we have been thinking of responsibility broadly as a condition that opens up the agent to “a whole variety of responses and attitudes” (251; cf. 265). We know at least, then, that an agent who holds herself responsible thereby adopts some subset of these attitudes or responses towards herself (perhaps this could also involve ac- cepting that others would be warranted in adopting some subset of these attitudes or responses towards her). But it is not obvious what the connection is between the at- titudes and responses that are associated with holding responsible and the agent’s understanding of her own domain of secure competence. It is one thing to “assert” our competence or “affirm” the scope of our powers of rational agency, and a quite different thing to take up (or accept that others would be warranted in taking up) the responses and attitudes that are associated with holding oneself responsible. What Raz’s account needs is an explanation of how it is that responding in these ways to failures of our powers of rational agency could serve to address or negate the threat to our self-conceptions as agents that such failures pose. Return to the case of trying and failing to place a glass on the table. When the glass tumbles to the floor, let us suppose (setting aside the doubts raised in the pre- vious section) that we experience this as undermining our domain of secure competence. In response, Raz says, we “tend to feel annoyance, and to blame our- selves” (245). This response is one way of holding ourselves responsible; that is, for Raz it is one way of addressing the threat by reasserting our competence. It is meant to be both explained by the felt need to address this threat and justified in light of Responsibility and the Emotional Structure of Relationships 26 the nature of the threat. But these claims are not self-explanatory. We need, to re- peat, some account of how it is that holding ourselves responsible in this way could be effective as a means of reaffirming the scope of our domain of secure compe- tence. To make the problem as stark as possible, it is helpful to contrast some other possible reactions. For instance, an agent who drops a glass might buck herself up, reminding herself that “nobody’s perfect,” that “you’re only human,” or that “every- one makes mistakes,” until she regains her equilibrium and feels assured that the accident is not inconsistent with her understanding of her own competence. Another agent might look around for a competence-defeating explanation for the accident (“It’s dark in here, I couldn’t see what I was doing”). In this way, the agent could come to terms with the accident in a way that did not imply that he was not generally competent to perform this sort of action under more favorable conditions. A third agent might flex her hand to ensure that nothing is wrong with her grip, and then pick up the glass, refill it with water, and practice lifting and setting it down several times, until she is satisfied that nothing is amiss with her relevant powers of rational agency. Yet a fourth agent might decide to undertake a routine of exercises to strengthen his hands and improve his motor skills, until he is satisfied that he has attained the level of competence that he originally claimed for himself. Finally, a fifth agent might simply accept, with a sigh, that she is not as capable as she thought, and revise her sense of her own competence accordingly. Of course, these strategies could also be combined with each other in various ways, or with a range of other possible measures. Each of these responses is a natural and potentially effective way for the agent to address any threat that her failure might pose to her sense of her own com- petence. But note that none of these responses necessarily involves the agent’s holding herself responsible for dropping the glass; at best, each of these responses is neutral with respect to responsibility (a few of them might more naturally accompany efforts to evade responsibility). Moreover, when we set these approaches alongside the response that Raz himself cites—feeling annoyance, and blaming ourselves (245)—the latter seems conspicuously ill-suited to the task at hand. It is easy to see how the five strategies just described could defuse a threat to the agent’s understand- Chapter 1 | Taking Responsibility and Relationships 27 ing of her own competence, but it is hard to see how blaming herself could achieve this result. The same goes for other responses and attitudes that are associated with holding responsible. It is difficult to see, for instance, how reproaching or sanction- ing herself, or accepting that others would be warranted in blaming, reproaching, or sanctioning her, could address a threat to the agent’s sense of her own competence. Again, the contrast with the five strategies above is suggestive. There is a difficulty, then, in explaining how holding oneself responsible could be efficacious as a way of addressing threats to our self-conceptions as agents. And the problem is deeper than merely insufficient efficacy—it consists, rather, in what appears to be a basic mismatch between the threat and the proposed response. The threat is some kind of evidence or suggestion of incompetence; the proposed response is one or another combination of emotion, expression, and sanction: self- blame, guilt, reproach, and so on. Unless the strategy is one of distraction or redirec- tion (which Raz does not suggest and which seems foreign to his approach), this kind of response just seems like the wrong sort of thing to meet the threat. And con- versely, the need to respond to this kind of threat seems like the wrong sort of thing to justify the kinds of response that are involved in holding oneself responsible. 22 Consistent with the general emphasis of Raz’s discussion, we have been fo- cusing on ways in which the agent might hold herself responsible, rather than taking responsibility. But a shift in focus to taking responsibility cannot help Raz’s account with the present difficulty; if anything, it makes the problem worse. After inadver- tently spilling red wine on my neighbor’s upholstery, for instance, it is ordinarily very much incumbent upon me to take responsibility, apologize, and try to repair the damage. But here again, taking responsibility does not seem to be a promising way to preserve my self-conception as an agent. How, for instance, could feeling sorry, or making an apology, shore up my confidence in my ability to wield a wine glass? Once again, the difficulty is deeper than merely insufficient efficacy—there is the same basic mismatch between threat and response. Taking responsibility— contrition, apology, amends, etc.—just is not the right sort of thing to meet this kind of threat to our sense of our own competence. Moreover, and conversely, the need 22 For further discussion of the distinctive force and significance of these responses, see chapter 4. Responsibility and the Emotional Structure of Relationships 28 to preserve my self-conception as an agent is again not the right sort of thing to jus- tify my taking responsibility. For one thing, taking responsibility is not so self- regarding. My reasons to take responsibility have to do rather with the fact that I have done something that has injured someone else without justification, and per- haps also with the fact that I stand in a relationship to that person that I have good reason to want to maintain. These are the circumstances that seem to make contri- tion, apology, amends, and so forth intelligible and appropriate (as we will explore further in the next section). There is then a significant gap in Raz’s account. It lacks an explanation of how it is that holding oneself responsible (or, for that matter, taking responsibility) could be efficacious as a way of preserving our self-conceptions as agents, and in- deed of how holding oneself responsible (or taking responsibility) could even be the right kind of response to meet this need. 4 What is it to take responsibility? We have just completed a long detour through Raz’s account of responsibil- ity for inadvertence. It is finally time to return to our main theme: taking responsibility. Despite the problems that I have raised for Raz’s account, I think that he is on to something essential with his ideas that our own conduct can threaten our self-conceptions as agents and that our responsibility practices can be understood and justified, in part, as a way of addressing such threats. But I propose to develop this idea in a different direction; the result will be an account of our reasons to take responsibility for ordinary intentional conduct, rather than an account of our reasons to hold ourselves responsible for inadvertent conduct. A central feature of Raz’s approach is his narrow focus on the individual who is concerned with her own conduct and its implications for her own self-conception as an agent. For Raz, the case in which “the glass we put on the table tumbles off it . . . [and] we tend to feel annoyance, and to blame ourselves” (245) exhibits the full structure of responsibility: a threat to the agent’s self-conception that is addressed by the agent’s holding herself responsible by blaming herself. This narrow focus on the individual agent, considered in isolation, is both fruitful and problematic. On the positive side of the ledger, this first-personal approach marks an interesting contrast Chapter 1 | Taking Responsibility and Relationships 29 with accounts that understand responsibility in the context of interpersonal relations, particularly accounts that focus on the second-personal perspective of the other party to the transaction: the offended, blaming party who holds the agent responsible. 23 Raz’s emphasis on the distinctive perspective and concerns of the responsible agent brings welcome and useful attention to the relatively overlooked half of transaction, and this is, in my view, the root of his main insight. But on the other side of the ledger, Raz’s narrow focus on the individual agent causes him, perhaps without even noticing it, to focus on holding oneself re- sponsible to the neglect of taking responsibility, and more generally to ignore the social or interpersonal dimension of our responsibility practices. The failure to ap- preciate this social dimension is, I think, the root of the difficulties that I have raised for his account. Recapturing the missing social dimension will require expanding and elabo- rating upon Raz’s understanding of our self-conceptions as agents, in order to include the vital concern that we have for others’ attitudes about our agency. Pro- ceeding in this way will naturally suggest a shift in focus from the agent’s reasons to hold herself responsible to her reasons to take responsibility with respect to other persons—in the way in which you might take responsibility for missing a meeting by feeling sorry for letting your friend down, apologizing to her, and offering to make it up to her. The picture of taking responsibility that will emerge will support the claim with which we started, that taking responsibility is a distinct and independently sig- nificant phenomenon. We can start in the same place as Raz, with his notion of a domain of secure competence. Let’s accept Raz’s suggestion that complex agency of any kind requires a domain of competence within which we are entitled to act without reflecting on the prospects of success. And now let’s go a step further by observing that, in the case of human agents living in communities, the domain of secure competence also has a social dimension and plays a critical social role. To operate as a rational agent in the world, I need to be confident of my ability to place a glass on the table or to step on 23 I have in mind here views like Strawson’s (1962) and views that bear Strawson’s influence, including those of Watson (1987), Wallace (1994), Hieronymi (2004), Darwall (2006), Scanlon (2008). Responsibility and the Emotional Structure of Relationships 30 the brake at the appropriate moment; but to function as a full, participating member of my community, I also need to be able to hold myself out to others as someone who is able to place a glass on the table or step on the brake. When I accept a glass of red wine from my neighbor, I implicitly represent to her that I am someone who can be trusted to wield a glass securely in the normal manner, without undue risk to her up- holstery. These sorts of representations have benefits for both parties. They assure my neighbor that she can serve me a glass of wine without the need to reflect on the possible risk to her furniture, and they thereby also put me in a position to enjoy her hospitality. Just as our domains of secure competence are essential to our self- conceptions as agents, how others perceive our domains of secure competence plays a critical role in our self-conceptions as agents. We are creatures who relate to one an- other, expect things of each other, hold each other to standards, and expect—indeed, demand—that others hold us to standards in return. We need not only to see our- selves as competent within certain domains, but for others to see us and acknowledge us as competent, and to treat us accordingly as potential partners who are suited to participate in a wide variety of social activities. To be seen as incompe- tent by others, and thereby as unfit to participate in the various activities that make up the life of the community, is not only to our serious material disadvantage, but also threatens our very sense of who we are, of what we can do, and of our own worth. Our self-esteem, our self-respect, our pride (or shame) in who we are, our sense of what social and cooperative activities are open and available to us, and thus our projects and ambitions and our sense of our own potential—all of these critical aspects of our self-conceptions as agents depend in important ways on how we un- derstand others to assess our domains of secure competence. But there is still more to our self-conceptions as agents. We can extend the notion of a domain of secure competence one step further, by recognizing that the social dimension of our self-conceptions as agents is not limited to others’ assess- ments of our sub-personal powers of rational agency—like memory or physical coordination—but also involves what Strawson called the quality of our wills (1962: 83). Normal, well-socialized agents care not only about the quality of other agents’ wills towards them (as Strawson observed), but also about their own qualities of will Chapter 1 | Taking Responsibility and Relationships 31 and, crucially, about how other agents perceive their qualities of will. That is, we care a great deal about whether we are well-meaning, kind, reliable, trustworthy, and so on, and about whether others see us as such. This is because we need to be able to participate in various trusting, caring, and otherwise meaningful relationships with others. Someone who was not perceived as good-willed in various ways could never be trusted, befriended, or loved in the way that adults mutually love one another, and to realize that we were in this situation would once again threaten our very sense of who we are, of what we can do, and of our own worth. 24 The various elements of our self-conceptions as agents—self-esteem, self-respect, pride (or shame) in who we are, our sense of our own potential, and so on—thus depend in significant ways on how we understand others to assess the qualities of our wills. We construct our prac- tical identities in light of how well suited we understand ourselves to be, and how well suited others perceive us to be, to participate on mutually acceptable terms in meaningful relationships and activities with them. 25 With these pieces in place, we can identify an important class of potential threats to our self-conceptions as agents. Assuming we are well-socialized, our self- conceptions as agents make reference to and incorporate others’ assessments of our competences and qualities of will. It is critical to our self-conceptions as agents that 24 Cf. Pamela Hieronymi’s observation: “It . . . seems quite plausible that standing in relations in which the quality of one’s will is recognized, both by oneself and by oth- ers, is of considerable importance. A change in what you or another person thinks about the quality of your will, in itself, changes your relations with them. Insofar as it is important to stand in relations in which goodwill is recognized, the judgment that you have shown ill will itself carries a certain force” (2004: 124). I will have occasion to return to this observation in the course of discussing Hieronymi’s account of blame in chapter 4. 25 Where perceptions are important there is also incentive for deception. While I cannot do this important topic justice here, it is worth observing that it is not ordi- narily enough merely to be seen by others as, for instance, trustworthy if we know privately that we are not—that they would not so see us if they knew everything that we know. Maintaining such deceptions is cognitively and (at least in non-pathological cases) emotionally exhausting, and because we know that in these matters it is nearly impossible to fool everyone all of the time, the cognitive effort tends to be accom- panied by persistent fear of tripping up and being unmasked. Ordinarily, the most effective way to feel secure in others’ trust is to convince oneself that one is in fact worthy of it. The realization that one is less trustworthy than others believe is thus often enough to subvert one’s self-conception as an agent, in the extended sense be- ing developed here, even if others have not (yet) discovered this fact. Responsibility and the Emotional Structure of Relationships 32 we are, and are seen to be, eligible in various ways to participate in meaningful rela- tionships and cooperative activities with others. And when we perform a wrong or vicious action that belies the view of ourselves that we want others to accept, this naturally threatens our self-conceptions as persons who are esteemed, respected, trusted, admired, loved, and so on. 26 On pain of revising our self-conception to re- flect a diminution in social eligibility of the relevant kind, we must react to this threat by reestablishing our eligibility in others’ eyes as well as our own. All of this sets us up to understand how our self-conceptions as agents— taking into account their social dimension—can ground reasons to take responsibility for our ordinary intentional conduct. 27 Recall the questions that caused difficulties for Raz’s account: How does our conduct threaten our self-conceptions as agents? and, How does holding ourselves responsible (or indeed taking responsibility) ad- dress the threat? We are now in a position to give quite different answers to these questions. Regarding the threat, when our intentional conduct 28 reflects a quality of will that is inconsistent with the expectations of trustworthiness, kindness, and so on that we want others to have of us, we thereby disappoint the expectations of those with whom we participate (or wish to participate) in meaningful relationships and activi- ties, and give them reason to doubt that we are suitable objects of cooperative or trusting attitudes. Because it is an important aspect of our self-conceptions as agents that we are, and are recognized by others to be, fit to participate in various meaning- ful relationships and shared activities, these doubts are, quite literally, threats to our self-conceptions as agents—they are inconsistent with how we see ourselves, and how we want to see ourselves, in relation to others. When our conduct disgraces us 26 It may do this either immediately, by changing the way others actually think of us, or premonitorily, by creating a reason which, if it were discovered, we know would be likely to change the way they think of us. On this point, see the previous footnote. 27 Responsibility for inadvertence raises special questions; I have raised doubts about Raz’s answers to those questions, and do not have answers of my own to offer here. For this reason I limit myself to discussion of intentional conduct. 28 This line of thought could be extended in much the same way to inadvertent con- duct that expresses or otherwise manifests the agent’s character traits, judgments, or evaluative attitudes. Chapter 1 | Taking Responsibility and Relationships 33 in the eyes of others, we can no longer (without self-deception) think of ourselves as respected, trusted, or loved in the same way as before. Regarding the second question, a direct and effective way to address this kind of threat to our self-conceptions as agents is to commit ourselves to vindicating our good will and general social eligibility, by doing whatever is warranted to make amends and put ourselves back in the good graces of those whom our misconduct has affected—that is, to take responsibility for our conduct. Taking responsibility thus serves as a way to reassure others (and ourselves) of our fitness to participate in meaningful relationships and activities by demonstrat- ing to them that we value our relationships in the right way, that we take our infractions seriously, that we are unwilling to allow others to be disadvantaged through their reliance on our previous representations of our fitness to participate in meaningful relationships and activities with them, and that we are resolved to do bet- ter in the future. These functions unite the complex matrix of attitudes, dispositions, and responses that are characteristic of taking responsibility, including feeling sorry or contrite towards those one has affected, feeling regretful or ashamed about one’s conduct, being disposed to apologize and make various gestures of penitence and conciliation, and being willing to repair or compensate for the harm done, even at significant cost to ourselves. 29 In particular, feeling contrite or sorry towards those affected by our conduct plays a crucial role, because it demonstrates that our concern for the relationship and for the effect we have had is not merely instrumental, but intrinsic; we feel badly because we value the other’s recognition of our good will, and the relationship of which this recognition is a constituent part, for their own sake. 30 If sound, this proposal vindicates Raz’s ideas that our own conduct can threaten our self-conceptions as agents and that our responsibility practices can be 29 While all of these elements can play a part in taking responsibility, I do not aim to specify a set of necessary and sufficient conditions for taking responsibility, and I am not sure that it would be useful to try. Taking responsibility is something that seems to come in different kinds and degrees, and moreover different elements can be in- volved in different ways on different occasions. In any event, I will leave detailed questions about what we might call the internal mechanics of taking responsibility for another occasion (though chapter 4 further discusses the role of shame in taking responsibility). 30 Subsequent chapters will develop this thought in detail. Responsibility and the Emotional Structure of Relationships 34 understood and justified, in part, as a way of addressing such threats. But in structure it is quite different from Raz’s account. For Raz, holding oneself responsible was conceived as an autopoietic act that served to reestablish the agent’s self-conception in a direct, unmediated way. In this revised proposal taking responsibility’s connec- tion to the agent’s self-conception is mediated by its interpersonal significance—that is, by the part that taking responsibility can play in rehabilitating the agent in others’ eyes (as well as her own) through a demonstration of her continued eligibility for the relevant range of shared activities and relationships. The advantage of this deviation from Raz’s approach is that the social dimension of our self-conceptions as agents helps to explain the efficacy of taking responsibility as a way to address threats to our self-conceptions. If the threats consist in diminution of our eligibility in others’ eyes to participate in meaningful relationships and activities with them, then an intrinsic concern for what has been done to those others and the relationships, together with a readiness to conciliate them and restore ourselves in their good graces, is just the kind of response that is likely to be effective. Moreover, the reasons that this proposal identifies are also, at least plausibly, reasons of the right kind to justify taking responsibility. Rather than the purely self- focused reasons that featured in Raz’s account of holding oneself responsible, the reasons that feature in this alternative proposal are inextricably bound up with our concern for others and for the quality of the relationships and activities that we par- ticipate in with them. If I hurt or offend someone, my reasons for taking responsibility plausibly include the fact that I do not wish to think of myself, or to be thought of, as someone who holds the other and our relationship in so little regard— that’s not who I am. Conclusion I have tried to identify a connection between our sociality and our self- conceptions as agents, and to sketch the role that this connection can play in explain- ing our practice of taking responsibility. That explanation is inspired by Joseph Raz’s core insight that our own conduct can threaten our self-conceptions as agents and that our responsibility practices can be understood and justified as a way of address- ing such threats, but I have argued that Raz’s own development of this insight runs Chapter 1 | Taking Responsibility and Relationships 35 into difficulties. Instead, I have developed the idea in a different way, focusing on the way in which our self-conceptions as agents make reference to and incorporate oth- ers’ attitudes towards us. This serves to vindicate and flesh out the picture offered at the outset, ac- cording to which taking responsibility has a quite different character and function from holding responsible, including holding oneself responsible. To recap: Holding oneself responsible is similar to holding another person responsible; it consists in attitudes and responses that you take towards yourself as wrongdoer: blame, reproach, instruc- tion, sanction, and so on. Taking responsibility, on the other hand, according to the account that I have developed, is a response to threats of a certain kind to your social self-conception, and consistent with this social orientation it consists of attitudes and responses you take towards the other as a victim of your conduct and as a party to the rela- tionship. Taking responsibility is thus complementary to but distinct from holding responsible—it is, in this sense, the converse of blame. If our reasons to take responsibility are grounded in the way in which our self-understandings incorporate the attitudes of others towards us, then fully to un- derstand taking responsibility (as well as its place in the wider emotional economy of responsibility) will require a better and more general grasp of this aspect of our self- understandings. To this end, chapter 2 will further explore the connection between our self-understandings and our interpersonal relations, and in particular the ways in which our relationships are constituted, structured, and regulated by our emotional sensitivities to the attitudes that others take towards us. 36 Chapter 2 Relationships and the Participant Stance According to the argument of chapter 1, our reasons to take responsibility are grounded in the way in which our self-understandings incorporate the attitudes of others towards us. A full understanding of our responsibility practices accordingly requires a better and more general grasp of this aspect of our self-understandings. That is the goal of the present chapter; while we will later return to responsibility (in chapter 4), the aim here will be to advance our understanding of the social dimension of our self-understandings by building on Peter Strawson’s (1962) conception of the reactive attitudes and his related distinction between the objective and participant stances. This will involve developing these theoretical tools in a novel way. In recent normative theory there has been much discussion of the relational character of our moral practices and, in particular, of the central role in those prac- tices of the Strawsonian reactive attitudes. The reactive attitudes make significant appearances, for instance, in the work of Stephen Darwall (2006), T.M. Scanlon (2008, 2013, 2015), and R. Jay Wallace (1994, 2011, 2014). But whereas Strawson saw the reactive attitudes as encompassing the very wide range of responses that we are susceptible to in virtue of participating in interpersonal relationships—resentment, hurt feelings, love, shame, etc.—this has often been rejected as too broad. The reac- tive attitudes are now typically conceived as a narrow class of moralized responses closely associated with responsibility and blame. In fact, it has become a common practice simply to stipulate that “reactive attitudes” refers specifically to the blaming attitudes of resentment, indignation, and guilt. 1 1 This practice is largely due to the influence of Wallace (1994); for one recent exam- ple, see Shabo 2012: 132. A notable exception to this trend is Susan Wolf (2015), who has recently proposed a conception of the reactive attitudes that is, I think, much broader even than Strawson’s own. The reactive attitudes are, for Wolf, all of Chapter 2 | Relationships and the Participant Stance 37 It seems to me that something important has been lost in this shift. I want to suggest that Strawson’s original idea—that our relationships are structured and regu- lated by a broad but unified set of affective responses—represents a fundamental insight into the nature of our sociality. One the one hand, we depend heavily—and during certain periods of our lives, entirely—on the assistance or cooperation of oth- ers for the satisfaction of our needs and desires. This is what we might call the instrumental dimension of our sociality, and there is a long tradition in social phi- losophy—Hobbes is one standard bearer—of treating our instrumental reliance on one another as a theoretically fundamental fact. But, on the other hand, our sociality also has an intrinsic dimension: We care directly about others’ attitudes towards us, and many of our fundamental interests and ultimate ends concern such attitudes. We strive avidly for the approval, trust, admiration, and love of those who matter to us, and just as zealously seek to avoid disapproval, disrespect, disdain, or disgrace. De- spite extensive empirical work in this area, this dimension of our sociality has not always received the theoretical attention it deserves. Strawson’s reactive attitudes are a useful tool for remedying this philosophical neglect. But Strawson never tried to say exactly what the diverse attitudes that he categorized as “reactive” have in common. My aim is to develop an account of this broad class of attitudes and to defend the classification against the objection that it is not sufficiently unified to be theoretically useful. The attitudes in question are, I pro- pose, a discrete class of emotional responses that reflect the distinctive, personal way in which we care about the attitudes of others towards us. As we will see, it is this personal way of caring that connects our concern for others’ attitudes to our self- understandings. I contend that this proposal sheds light on a dynamic of reciprocal emotional involvement that is central and pervasive in our interpersonal relations, yet easily overlooked or misunderstood. the attitudes that we are susceptible to in virtue of seeing another person as “one of us” in some very wide sense; this extends, for instance, to the way in which we “like or love some artists on the basis of their artwork” (2015: 132-33). I will urge a more circumscribed conception of the reactive attitudes (or, as I will prefer to call them, the participant attitudes) than Wolf’s. Responsibility and the Emotional Structure of Relationships 38 1 Strawson on relationships and reactive attitudes Strawson’s own examples of the reactive attitudes include: gratitude, resent- ment, forgiveness, and hurt feelings (75 2 ); “the sort of love which two adults can sometimes be said to feel reciprocally, for each other” (79); “all the essentially personal antagonisms” (80); the “sympathetic or vicarious or impersonal or disinterested or generalized analogues” of these, like indignation on another’s behalf (83-84); and “self-reactive attitudes associated with demands on oneself for others,” including guilt, remorse, and shame (84-85). It is natural to ask what these diverse attitudes have in common that justifies grouping them together into a single class. Strawson does not directly answer this question, but instead characterizes the class in looser, more evocative ways. For instance: We should think of the many different kinds of relationship which we can have with other people—as sharers of a 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. Then we should think, in each of these connections in turn, and in others, of the kind of importance we attach to the attitudes and in- tentions towards us of those who stand in these relationships to us, and of the kinds of reactive attitudes and feelings to which we our- selves are prone (76). In this and several other passages, Strawson sounds two themes in characterizing the reactive attitudes: first, they are the “non-detached” (75) attitudes that are character- istic of participation in relationships; second, they are reactions to the attitudes and intentions that other people have towards us. (Strawson also several times sounds a third theme, concerning a connection between the reactive attitudes and demands, that we will return to below.) The first two themes are connected; Strawson thinks that being involved in relationships is, at least in significant part, a matter of according a certain importance to others’ attitudes and intentions towards ourselves. He thus characterizes the atti- tudes that reflect the importance we accord to others’ attitudes and intentions as “participant” attitudes—that is, as the attitudes that are characteristic of participation in interpersonal relationships (79-81). Strawson then contrasts the participant atti- tudes with the “objective” attitude: 2 Unless otherwise indicated, parenthetical page references in this chapter are to Strawson 1962. Chapter 2 | Relationships and the Participant Stance 39 To adopt the objective attitude to another human being is to see him, perhaps, as an object of social policy; as a subject for what, in a wide range of sense, might be called treatment; as something certainly to be taken account, perhaps precautionary account, of; to be managed or handled or cured or trained; perhaps simply to be avoided . . . (79). The objective attitude thus represents a “detached” perspective that preserves a cer- tain emotional distance. It is not, however, to be confused with being entirely cold and unfeeling: The objective attitude may be emotionally toned in many ways, but not in all ways: it may include repulsion or fear, it may include pity or even love, though not all kinds of love. But it cannot include the range of reactive feelings and attitudes which belong to involvement or participation with others in inter-personal human relationships (79). The objective attitude, then, is strictly distinct from the participant attitudes, but nevertheless includes a significant range of emotional responses and attachments, including even a “detached” kind of love. We are inclined to adopt the objective attitude, Strawson suggests, “just in so far as the agent is seen as excluded from ordinary adult human relationships by deep- rooted psychological abnormality—or simply by being a child” (81). But the objec- tive attitude is also available to us more generally, as “a resource” that we can use “as a refuge, say, from the strains of involvement; or as an aid to policy; or simply out of intellectual curiosity” (79-80). In this way, the objective attitude appears in contexts in which one party has chosen to view the other as an object of study, treatment, control, or the like; for instance, in the attitude of a therapist towards her patients 3 or a social scientist towards her subjects. In sum, then, we take the objective attitude, at least in some respect and to some degree, towards those with whom we cannot have, or choose not to have, ordinary interpersonal relationships. 3 A therapist’s attitude towards her patient is one of the standard examples of the objective attitude, but it is worth pointing out that it can also be important in therapy that the patient take an objective attitude towards her therapist, at least to some ex- tent. Taking the objective attitude can free the patient up to focus on the content of the therapist’s judgments and reactions (“you seem like a very selfish person”; “you seem to have difficulty being honest”), without taking them personally in the way that would be normal in an interpersonal relationship. The nature of this dynamic will become clearer as we proceed. Responsibility and the Emotional Structure of Relationships 40 Strawson thus strictly distinguishes the objective attitude from the participant attitudes, and treats the participant attitudes as equivalent to the reactive attitudes. From this it follows that “being involved in inter-personal relationships as we nor- mally understand them precisely is being exposed to the range of reactive attitudes and feelings” (81). I will refer to this as Strawson’s Relationship Thesis, and under- stand it as follows: Strawson’s Relationship Thesis: To participate in interpersonal re- lationships just is to be susceptible to the reactive attitudes; conversely, the reactive attitudes just are those attitudes that we are susceptible to in virtue of participating in interpersonal relationships. 4 (This version of the thesis is labeled “Strawson’s” in order to distinguish it from a modified version that I will ultimately defend.) Strawson’s Relationship Thesis is significant in at least two ways, one widely recognized and the other neglected. First, it plays a celebrated role in Strawson’s case for compatibilism about responsibility. He argues (very roughly 5 ): (1) that our prac- tice of holding one another responsible is to be understood in terms of our susceptibility to the reactive attitudes; (2) that Strawson’s Relationship Thesis is true, meaning that to participate in interpersonal relationships just is to be susceptible to the reactive attitudes, and vice versa; and (3) that we neither could nor should aban- don our participation in interpersonal relationships. He concludes (4) that we neither could nor should abandon our susceptibility to the reactive attitudes, and therefore (5) that we neither could nor should abandon our practice of holding one another responsible—regardless of whether the thesis of determinism is true. Conclusions (4) and (5) do not follow without some version of premise (2), Strawson’s Relationship Thesis. 6 4 To streamline discussion, my formulation of Strawson’s Relationship Thesis leaves out “vicarious” reactive attitudes like indignation. These attitudes could be captured by including a reference to the attitudes that we are susceptible to in virtue of sympa- thetically inhabiting the perspective of another’s involvement in an interpersonal relationship. 5 And with apologies to those who have worked to reveal and clarify the structure of Strawson’s subtle argument; see for instance Bennett 1980, Hieronymi MS, Russell 2013, and Watson 1987 and 2014a. 6 Strawson’s Relationship Thesis is a strong, constitutive claim about the connection between reactive attitudes and interpersonal relationships; strictly speaking, Straw- son’s compatibilist argument could make do with the weaker claim that we cannot Chapter 2 | Relationships and the Participant Stance 41 But it is the second, neglected significance of Strawson’s thesis that primarily interests me. Tightly connected in Strawson’s original conception of the reactive atti- tudes were two ideas: that the reactive attitudes play an important role in our responsibility practices, and that the reactive attitudes play a broad and basic role in our social lives more generally. According to the latter idea, Strawson’s Relationship Thesis identifies a distinctive dynamic that is pervasive in interpersonal relations. For continuity’s sake I will follow Strawson in describing relations in which this dynamic appears as “relationships,” though the label is not crucial (and has been, I will suggest, in some ways misleading). Whatever we choose to call it, I think this dynamic is im- plicated in a very wide range of theoretically interesting phenomena, including (as Strawson emphasized) our responsibility practices, but also interpersonal emotional responses like shame, hurt feelings, and love, as well as diverse configurations of so- cial involvement like friendship, loyalty, respect, and subordination. But before these claims about the broader theoretical significance of Strawson’s Relationship Thesis can be vindicated, it is necessary to address serious doubts that have been raised about its truth, and indeed its determinateness. 2 Wallace on relationships and reactive attitudes R. Jay Wallace (1994, 2014) has raised an important and influential challenge to Strawson’s Relationship Thesis, in the form of two significant objections. First, he argues that the thesis embodies a conception of the reactive attitudes that is too inde- terminate, and seemingly too broad, to be theoretically useful. Second, he suggests that the thesis is false for the straightforward reason that we can be susceptible to the reactive attitudes outside of the context of an interpersonal relationship. Underlying the first objection is a worry about the theoretical usefulness of the category of reactive attitudes for understanding responsibility (see premise (1) of Strawson’s compatibilist argument as outlined above). It is, admittedly, not easy to see what participant attitudes like hurt feelings, for example, or “the sort of love abandon susceptibility to the reactive attitudes without also abandoning participation in interpersonal relationships. But Strawson’s Relationship Thesis, assuming some version of it could be substantiated, has the important advantage of explaining why we cannot abandon susceptibility to the reactive attitudes without abandoning par- ticipation in interpersonal relationships. Responsibility and the Emotional Structure of Relationships 42 which two adults can sometimes be said to feel reciprocally, for each other,” have to do with holding someone responsible. If our conception of the reactive attitudes is too broad or indeterminate (the worry goes) then there will be no illuminating way to connect the reactive attitudes as a class with responsibility. In Wallace’s view, Strawson is led to an overly diffuse conception of the reac- tive attitudes through his tendency to conflate what are, properly speaking, two different distinctions: (1) the distinction between reactive attitudes and other attitudes one might take towards persons; and (2) the distinction between (in Strawson’s words) “the attitude (or range of attitudes) of involvement or participation in a human relationship, on the one hand, and what might be called the objective attitude (or range of attitudes) to an- other human being, on the other” (1994: 27-28, quoting Strawson 1962: 79). In other words, Wallace thinks that it is important not to conflate the reactive atti- tudes with what we might refer to as the participant attitudes properly so-called—that is, those attitudes that we are susceptible to just in virtue of participating in interper- sonal relationships. Wallace’s reason for thinking this is that the latter category is not well defined. As Wallace says: It will not do, for instance, simply to define the reactive attitudes as the emotions one has insofar as one is involved in interpersonal rela- tionships. This puts all of the weight of the account of reactive attitudes (and of their incompatibility with objectivity of attitude) on the notion of interpersonal relationships; but we do not have an in- dependent concept of an interpersonal relationship suitable to play the required role in the account (1994: 29). For Wallace, then, collapsing the reactive attitudes into the participant attitudes leads to uncertainty about which attitudes count as reactive, because we do not have a suf- ficiently clear notion of an interpersonal relationship. 7 Strawson clearly intends for the emotional responses that are commonly found in the relations between, for ex- ample, parents and infants or therapists and patients not to count as reactive attitudes. But why not? By what criterion are we to determine whether these relations (and other relations that fall in the uncertain zone between fully interpersonal and 7 As Wallace observes (2014: 122), this concern about the role that the notion of an interpersonal relationship plays in Strawson’s account traces back to Jonathan Bennett (1980: 34-36). Chapter 2 | Relationships and the Participant Stance 43 impersonal) qualify as interpersonal relationships of the right kind? As Wallace asks: “What is it, for instance, about the relation between therapists and their subjects that is not ‘interpersonal,’ or that is ‘interpersonal’ in the wrong way?” (1994: 29; cf. 2014: 122). The upshot of this is that we have no firm grasp of which attitudes count as participant attitudes, and thus if we treat the participant attitudes and the reactive attitudes as equivalent we are left with no firm grasp of which attitudes count as re- active. Moreover, identifying the reactive attitudes with the participant attitudes threatens to make the category of reactive attitudes very wide (as illustrated by Straw- son’s own diverse examples). This gives Wallace further grounds to worry that the category of reactivity will become “so capacious that there is no illuminating way to say what reactive emotions are reactions to, and this exacerbates the difficulty of characterizing these emotions as a class” (1994: 29). Wallace’s second objection to Strawson’s Relationship Thesis is that partici- pation in an interpersonal relationship is simply not necessary for susceptibility even to core reactive attitudes like resentment and indignation. In articulating this objec- tion, Wallace relies on what he calls the “familiar conception of an interpersonal relationship, understood as a socially salient pattern of historical interaction and at- tachment, as when we speak of our relationships to our friends, family members, colleagues, and so on” (2014: 119). He is thus thinking of a relationship as something that develops over time and that is constituted, at least in part, by a history of inter- action and emotional involvement between the parties. This conception of relationships will admit of gray areas; it is not clear how much interaction and at- tachment, and in what kinds and proportions, counts as enough to constitute a relationship. But the commonsensical idea, I take it, is that friends of long acquain- tance, happily and unhappily married couples, professional colleagues with long experience working together, and the like all have relationships; Romeo and Juliet, on first seeing each other across the room, do not—at that moment a relationship is precisely what they both ardently want, but do not yet have. The underlying thought here is familiar, and there is something natural about this way of putting it: that what Romeo and Juliet want is a relationship. But at the same time, there is something distinctive that Romeo and Juliet already have once Responsibility and the Emotional Structure of Relationships 44 their eyes meet—something that Benvolio and Juliet, for instance, may not, even if their eyes also recently happened to meet across the room. 8 What Romeo and Juliet have is a distinctive way of relating to one another, and we might wonder why the title of “relationship” should be withheld from this kind of relating. We will return to this question. Wallace contends that a relationship, in his historical sense, is not necessary for susceptibility to the reactive attitudes. In support, he points out that Strawson’s “vicarious” reactive attitudes, like indignation and moral disapprobation, “can in principle be experienced towards individuals with whom one has never interacted at all—war criminals, for instance, whose misdeeds one has only read about in the newspaper” (2014: 123). In such cases there has been “no antecedent pattern of his- torical interactions between the subject of the reactive emotion and its proper object, nor does the emotion itself seem to constitute an interpersonal nexus that could properly be described as a relationship” (2014: 123). In these kinds of cases, Wallace concludes, there is no interpersonal relationship—and yet we experience no impedi- ment whatsoever in feeling indignation towards war criminals. Now, in these passages Wallace seems to overlook what might seem like an obvious Strawsonian response; namely that the attitudes Wallace focuses on are “vi- carious” precisely in the sense that they are felt by imaginatively adopting the point of view of someone else—such as an actual victim—who is related to the target of the reactive attitude through a concrete historical pattern of interactions. Perhaps this neat response is not fully satisfying; but in any event, the problem can be stated in a more general way that renders it ineffective. Consider what we can call the “problem of the stranger” 9 : Imagine that your foot is trampled, or your pocket is 8 Of course, Benvolio and Juliet may well have a “relationship” in whatever weaker and wider sense it is that seems to be required by Strawson’s Relationship Thesis. The question implicit in Wallace’s objection is whether this sense of “relationship” is both specifiable independently of Strawson’s Relationship Thesis and has enough content to be informative about the nature and scope of the reactive attitudes. Wallace suggests not; in §§ 3-5 I will try to develop a conception of relationships that satisfies these desiderata. 9 Cf. George Sher’s “Wrongdoing and Relationships: The Problem of the Stranger” (2013). The problem that concerns Sher is actually slightly different from the one I discuss here. Sher questions whether wrongdoing by a stranger can be understood as impairing a (previously existing) relationship (as Scanlon’s account of blame seems to Chapter 2 | Relationships and the Participant Stance 45 picked, or your good name is impugned, by some malicious person whom you have never previously met and who makes such a clean escape that your paths never cross again. In a case like this it is at least highly doubtful whether the pattern of historical interactions between you and your wrongdoer are extensive enough to constitute a relationship in the familiar sense that Wallace has in mind; the one-off instance of mistreatment seems an insufficient “interpersonal nexus” to qualify as a relationship. And yet, there is no doubt whatsoever that you may, quite justifiably, resent a stranger who has mistreated you in any of these ways. If this is right, then Wallace would seem to be on firm ground in asserting that a relationship, in the sense of “a socially salient pattern of historical interaction and attachment,” is not a necessary condition for susceptibility to the reactive attitudes. Wallace’s own response to these two problems, following a path originally indicated by Gary Watson (1987: 229-31), is to re-characterize the reactive attitudes in terms of normative demands, expectations, or requirements. In this he is pursuing what I previously called the third theme that Strawson sometimes sounds when dis- cussing the reactive attitudes. Strawson says, for instance, that “these attitudes of disapprobation and indignation are precisely the correlates of the moral demand in the case where the demand is felt to be disregarded. The making of the demand is the proneness to such attitudes” (90; quoted at Wallace 1994: 26 and 2014: 122). Wallace accordingly defines reactivity in terms of a common propositional content; namely, that the target of the attitude has violated a demand or expectation to which the reactive subject holds that person (2014: 122). This results in a much narrower, and apparently more definite, class of reactive attitudes than Strawson thought he had envisioned. Wallace limits the reactive attitudes to resentment and its vicarious and reflexive analogues, indignation and guilt. All of Strawson’s other reactive atti- tudes—gratitude, shame, love, hurt feelings, and so on—are excluded. And because the class of attitudes that we are susceptible to in virtue of participating in interper- suggest). (A version of Sher’s question has been discussed by Scanlon (2008: 140-41, 2013: 87-88) and also by Linda Radzik (2009: 79-80).) But I am concerned rather with the question of whether reactive attitudes towards strangers can be understood in terms of some sort of relationship between the stranger and the reactive subject. It is thus the reactive attitude itself, rather than the wrongdoing, whose connection to a relationship is in question. Responsibility and the Emotional Structure of Relationships 46 sonal relationships (however we conceive those) is clearly much broader than just resentment, indignation, and guilt, Wallace’s modified conception of the reactive atti- tudes automatically renders Strawson’s Relationship Thesis untenable. In the wake of Watson’s and Wallace’s work, it has become common to characterize the reactive attitudes in terms of demands. 10 A few writers have argued that this conception of the reactive attitudes is too narrow. For instance, Coleen Macnamara has argued both that Wallace is mistaken to exclude positive responses like gratitude and praise from the reactive attitudes (2011: 89), and that he is mis- taken to limit the negative reactive attitudes to the deontic domain (that is, to responses to moral wrongness) to the exclusion of the evaluative domain (responses to moral badness) (2013: 142-48). And T.M. Scanlon has urged that the reactive atti- tudes be broadened beyond just emotional responses to include changes in attitudes like intentions (2013: 89). But these writers nonetheless share with Wallace the defi- nitional presumption that the reactive attitudes are essentially those attitudes that are involved in holding responsible, where these represent a narrow subset of the much broader class of attitudes that we are susceptible to in virtue of participating in inter- personal relationships. 11 That is, together with Wallace, they reject Strawson’s Relationship Thesis. We can think of this rejection of Strawson’s Relationship Thesis as part of a strategy to preserve what I earlier described as premise (1) of Strawson’s compati- bilist argument—that our practice of holding one another responsible is to be understood in terms of our susceptibility to the reactive attitudes. The idea is to re- treat from Strawson’s Relationship Thesis in order to ensure that the reactive attitudes encompass just those attitudes that it is plausible to understand holding re- sponsible in terms of. Of course, this strategy amounts to rejecting premise (2), 10 Examples include Darwall (2006: 17), Margaret Urban Walker (2006: 25-27), Lucy Allais (2008: 54), and Victoria McGeer (2012: 299, 303); Coleen Macnamara cata- logues several others (2013: 141-42 & n.1). 11 This pattern is also reflected in the work of Kate Abramson and Adam Leite, who raise objections similar to Macnamara’s to characterizing the reactive attitudes nar- rowly in terms of demands (Abramson & Leite 2011: 690-92). Like Macnamara, they assume that “Underlying the notion of the reactive attitudes, as introduced by Straw- son, is the thought that these attitudes are ways of holding one another responsible” (2011: 692). Chapter 2 | Relationships and the Participant Stance 47 leaving the compatibilist argument very much in doubt. (Wallace himself accepts premises (1) and (3) of the argument, but acknowledges that because he denies (2) he must give up on the conclusions (4) and (5).) Overlooked in discussions of the trade-offs that this strategy involves has been the potential independent significance of Strawson’s Relationship Thesis as a general thesis about the nature of our sociality. But in order to assess that signifi- cance, Wallace’s charges of indeterminacy and falsity must be addressed. And for this, we need a better understanding of the range of possible ways of understanding the notion that is at the center of the thesis: the notion of a relationship. 3 What is a relationship? Defending Strawson’s Relationship Thesis against Wallace’s objections will require a conception of relationships that is independently viable and that can give us an “illuminating way of saying what reactive emotions are reactions to”—one that will unite Strawson’s diverse examples of reactive attitudes into a well-defined class. First, in order to avoid confusion, as well as the host of associations that have accumulated around the term “reactive attitudes,” it will be useful for our pur- poses to rechristen the class of attitudes that Strawson labeled reactive. Fortunately, we have already at hand the other label that Strawson used for the same set of atti- tudes: the “participant attitudes.” We can reserve the term “reactive” for the class of attitudes that are involved in holding responsible, however that class is to be fully characterized and whatever it may include. We will be interested in the broader class of participant attitudes, and we will reformulate Strawson’s Relationship Thesis ac- cordingly. Call the reformulated version simply the “Relationship Thesis”: Relationship Thesis: To participate in interpersonal relationships just is to be susceptible to the participant attitudes; conversely, the participant attitudes just are those attitudes that we are susceptible to in virtue of participating in interpersonal relationships. This reformulation does nothing to affect the force of Wallace’s objections. To meet them, we will need to address his concern that we cannot derive the content of (what we are now calling) the “participant” attitudes from the notion of a relationship, be- cause we have no independent, well-defined notion of a relationship to work with. Responsibility and the Emotional Structure of Relationships 48 3.1 The varieties of relationships Wallace argues that thinking of a relationship as a “socially salient pattern of historical interaction and attachment” will not do the trick, because this kind of rela- tionship could include paradigmatically objective relations, like the relations between therapists and their patients. But this is not the only way to think of relationships; the general notion of a relationship admits of many different possible conceptions, and different theoretical purposes may call for different conceptions. To get a sense of the variety of possible ways to conceive of relationships, consider two alternatives derived from Niko Kolodny (2003) and Scanlon (2008, 2013). Kolodny develops a conception of relationships that is related to, but much broader than, Wallace’s. According to Kolodny, interpersonal relationships are distin- guished from other interpersonal relations (being exactly twice the age of, sitting to the left of, etc.) by the conjunction of three features: relationships are ongoing, that is, they persist through time; they obtain between particular people (unlike, for in- stance, the relation that consists in someone’s “having, at any given time, some dentist”); and they are historical, in the sense that they depend on facts about the pasts of the particular people who participate in them (2003: 148). To illustrate the last feature, Kolodny notes that the relation between an emergency room doctor and the next patient on the triage list is not a relationship, in this sense, because it does not depend on facts about the pasts of either participant; being siblings or friends or lovers, by contrast, does constitute a relationship, because we must look to the pasts of the participants to see if the relationship holds (2003: 148). Wallace’s conception of a relationship as a “socially salient pattern of histori- cal interaction and attachment” satisfies Kolodny’s definition—such patterns persist through time, obtain between particular individuals, and are (trivially) historical—but Kolodny’s definition also encompasses a much wider range of interpersonal rela- tions. Twins separated at birth who have never met and are ignorant of each other’s existence do not have a relationship in Wallace’s sense, because there is no relevant pattern of attachment or interaction between them; but they do have a relationship in Kolodny’s sense, because they are related in a manner that is historical, particular, Chapter 2 | Relationships and the Participant Stance 49 and ongoing. 12 Another type of relationship that falls within Kolodny’s definition but outside of Wallace’s involves common personal histories or situations. For instance, two people who share and participate in the same culture have a kind of relationship that is ongoing, particular, and historical, even if they have never met or even heard about one another (Kolodny 2010: 51-53). 13 (As Kolodny argues, this sort of rela- tionship is significant because it may provide participants with reasons to be partial towards one another as compared with outsiders.) Kolodny also, however, observes a distinction between types of relationship (in his sense) that brings us closer to Wallace’s conception. Sibling and other family relationships (considered merely as such) are examples of what Kolodny calls “atti- tude-independent” relationships, which exist “independently of any historical pattern of concern” (like the twins separated at birth); friendship and romantic love, on the other hand, are examples of “attitude-dependent” relationships, which depend for their existence on “the pattern of concern that participants have for one another, for the relationship, and for the pattern itself” (2003: 149). Without such an historical pattern of concern, two people cannot count as friends or lovers: “A friendship or romantic relationship just is an ongoing pattern of concern” (2003: 149). Kolodny’s definition of an attitude-dependent relationship closely resembles Wallace’s conception of a relationship, especially when it is characterized in terms of patterns of mutual concern. But it is also worth noting that attitude-dependent rela- tionships could be understood in a less restricted way that would be significantly broader than Wallace’s conception. Consider the relation of having as a role model. If 12 The example of sibling and other biological family relationships is Kolodny’s (2003: 148-49). But Kolodny’s conception of relationships would seem to encompass much more than this, though Kolodny does not explicitly acknowledge this breadth. For instance, the relation of being older than holds between me and billions of other people; as between me and each of these people, the relation is ongoing, particular, and historical. 13 Samuel Scheffler discusses an overlapping and perhaps similar sense of relation- ship which he describes in terms of a “socially salient connection” or “social tie” between participants (1997: 198). For example: “two members of a socially recog- nized group do have a relationship in the relevant sense, even if they have never met, and if they value their membership in that group they may also value their relations to the other members. Thus, the fact that you are a member of the John Travolta Fan Club means that you have a relation to each of the other club members, and if you value your membership you may also value those relations” (1997: 198). Responsibility and the Emotional Structure of Relationships 50 someone is my role model, then there are qualities about them that I esteem and ad- mire and that I use, on at least more than one occasion, as ideals to guide my conduct or personal development. That is just what it is to have someone as a role model. Such a relation is thus ongoing, particular, and historical, and it is attitude- dependent in the sense that it would not hold in the absence of my historical and ongoing admiration and esteem. Nonetheless, it would not satisfy Wallace’s concep- tion of a relationship because it need not involve any actual interaction (my role model may, for instance, be dead). Kolodny’s treatment of relationships, then, places Wallace’s conception in a larger context. While Kolodny describes (one subset of) attitude-dependent relation- ships in terms similar to Wallace’s conception, his account also provides two successively broader categories of which Wallace’s relationships are a proper subset: attitude-dependent relationships in the widest sense (including, e.g., having as a role model ), and relationships in general (comprising both attitude-dependent and atti- tude-independent relationships). Scanlon, like Kolodny, has a conception of relationships that is, at its fullest scope, much broader than Wallace’s, but Scanlon’s conception is also different in character than either Kolodny’s or Wallace’s. Scanlon conceives of relationships in terms of normative ideals; according to this conception, a relationship is “a set of intentions and expectations about our actions and attitudes towards one another that are justified by certain facts about us” (2013: 86). For instance, two people count as friends when they have a certain pattern of common interests and experiences and enjoy each other’s company; under these circumstances, they are subject to a norma- tive ideal of friendship, which prescribes that they have certain intentions and expectations towards each other, including the intention to help each other when needed, to share and keep confidences, and so on (2013: 86-87). Two people count as siblings when a certain biological condition (sharing biological parents) or histori- cal condition (being raised by the same parents) holds between them; under these circumstances, they are subject to a normative ideal that prescribes that they care for and support each other in certain ways. 14 14 The sibling example is mine, not Scanlon’s, but conforms to the template laid down by Scanlon’s account. Chapter 2 | Relationships and the Participant Stance 51 The relevant facts about us can thus include our attitudes towards each other, but they can also include a wide variety of other facts about our histories, features, or situations. This means that Scanlon’s conception of relationships can accommodate a distinction between attitude-dependent and attitude-independent relationships, in much the way that Kolodny’s does (but Wallace’s does not). But the variety of facts that Scanlon recognizes as potentially relationship-grounding also includes some non-historical facts. And this means that Scanlon’s conception encompasses relations that even Kolodny’s widest conception would exclude. Scanlon refers, for instance, to the relationship that people stand in to one another in virtue of sharing “a com- mitment to some group, cause or ideal” (or even, in appropriate cases, in virtue of sharing “some properties in virtue of which they should have such a commitment”) (2013: 87). And at its broadest, Scanlon’s conception embraces what he calls the “moral relationship,” which holds between “all rational creatures” (2013: 87); thus all rational creatures are, in relation to each other, subject to a normative ideal that pre- scribes that they conduct themselves in accordance with general moral duties, and that they exhibit a general attitude of good will towards each other (2013: 87). On this view, the relationship that two people stand in to each other may not be exclusively a matter of what intentions or expectations they actually happen to have towards each other, or even the ones they are disposed to have or would have under appropriate conditions; it can also be a matter of what intentions and attitudes are justified or called for by the relevant facts about them. These relationships are, we might say, relationships in the normative sense, even if they are not relationships in Kolodny’s or Wallace’s descriptive senses. All of this is offered to illustrate the flexibility and malleability of the notion of a relationship; depending on our purposes, we can conceive of relationships in normative or descriptive terms, as attitude-dependent or attitude-independent, as historical or ahistorical, and so on. We thus need not simply accept as a default Wallace’s conception of a relationship as a “socially salient pattern of historical inter- action and attachment.” For present purposes we are interested in whether we can identify a distinct and theoretically useful conception of relationships according to which participation in a relationship (of the relevant kind) just consists in susceptibility to the particular Responsibility and the Emotional Structure of Relationships 52 class of attitudes that we have designated the “participant” attitudes. As we will see, this will require a conception of relationships that diverges from all of the concep- tions we have encountered so far. But the diversity of conceptions that we have already seen suggests that this alone should not deter us. 3.2 Relationships in the participant sense Our goal, it is important to emphasize, is to find a conception of relation- ships that will help us to arrive at an account of what the participant attitudes are; so one approach that will not work, on pain of circularity, is simply to say that for two people to be in a relationship (in the relevant sense) just is for them to be susceptible to the participant attitudes with respect to each other. In order to do the work we need it to do, our conception of relationships must be such that it can be identified without invoking the participant attitudes. What we need is an alternative to Wallace’s conception of an interpersonal relationship as a “socially salient pattern of historical interaction and attachment.” The latter is doubtless a familiar way to think about relationships; but there is also, I think, another familiar way of conceptualizing relationships, and one which I would venture is closer to Strawson’s original idea. According to the way of thinking about relationships that I have in mind, a relationship consists in the parties’ taking up what we might call a certain stance or orientation towards one another. (Conceiving of relationships in terms of a stance or orientation is different from Wallace’s and Kolodny’s approach because the occupation of a stance is not a historical matter; it is different from Scanlon’s approach because it is descriptive rather than normative.) In subsequent sections I will try to say what the relevant stance or orientation consists in. The goal of this section is to get an intuitive grasp of the alternative sense of rela- tionship that I have in mind, and to explore some structural features of an account that understands relationships in terms of the parties’ adoption of a stance or orien- tation. We earlier noted that, on the one hand, when Romeo and Juliet first meet there is something that they want to have between each other, but do not have yet. But on the other hand, Romeo and Juliet already do have a distinctive way of relating to one another. This way of relating involves Romeo’s orienting himself in a certain Chapter 2 | Relationships and the Participant Stance 53 way towards Juliet and vice versa, and this mutual orientation consists in various forms of reciprocal emotional receptivity and sensitivity. To take just one illustration, Juliet cares in an exquisitely sensitive way about how Romeo feels towards her—as she urgently asks herself, “Does he feel this too?” And Romeo, of course, is asking himself just the same question. According to the logic of the play, they are in love; we might say more cautiously that they are mutually infatuated, or at least that they are keenly romantically interested in each other. For most of us, this is a very signifi- cant way in which we can mutually relate to other persons. This is obviously not yet anything close to a fully satisfying characterization of what is happening between Romeo and Juliet, and I will say much more about this later. The point for now is just that there is some distinctive dynamic in place between them that is worth taking notice of, one that distinguishes their way of relating to each other from the way Ju- liet relates to Benvolio, or the way she relates to Romeo’s father (whom she never encounters in the play), or indeed the way she relates to historical figures like Julius Caesar or Cleopatra. She is related in a minimal way to all of these persons (she’s at least heard of all of them), and she may relate to each of them in various ways (per- haps she admires Caesar, or would have liked to meet Cleopatra); but she and Romeo relate to each other in a special way that is distinguishable from the others. Does this dynamic constitute a relationship? Of course, there are relationships and there are relationships; but to my ear at least it does not sound strange to say that Romeo and Juliet have a relationship already at this stage, albeit a very nascent and uncertain one. But there is, ultimately, no need to fight over the term “relation- ship.” Those who wish to reserve the term for cases that feature a more substantial historical element could introduce another term to describe the dynamic that is dis- tinctive in Romeo and Juliet’s way of relating to each other. We will have occasion to revisit this terminological issue, but for now suffice it to say that rather than intro- duce a new term for the general type of relating of which Romeo and Juliet’s is an instance, I will speak of two different senses of “relationship.” There is Wallace’s historical interaction and attachment sense—or simply historical sense, for short—in which Romeo and Juliet do not have a relationship (yet), but there is also another sense, in which Romeo and Juliet have a relationship in virtue of the orientation or stance that they take towards one another. Call this the participant sense of relationship. Responsibility and the Emotional Structure of Relationships 54 In a moment we will turn to the task of saying with more precision what ex- actly a relationship in the participant sense consists in. But we can already note an important possibility that the participant sense opens up. If the notion of a relation- ship in the participant sense proves to be a coherent one, then the example of Romeo and Juliet suggests that a relationship in the historical sense is not a necessary precondition for a relationship in the participant sense. At the fateful moment when their eyes meet, Romeo and Juliet are strangers to each other, in the historical sense that motivated the “problem of the stranger.” This means that if we can articulate a viable sense in which Romeo and Juliet have a relationship at this moment, we can start to see how an interpretation of the Relationship Thesis that incorporates this way of conceptualizing relationships might have resources to avoid the problem of the stranger. (I will say more about this once the details of the account have been filled in.) There is also an interesting converse possibility. I have just suggested that a relationship in the historical sense is not a necessary precondition for a relationship in the participant sense. But neither is an historical relationship a sufficient condition for what I am calling a participant relationship, at least in principle. Consider the case of a therapist and his patient of many years. We can suppose that the therapist and patient have an extensive history of interaction, involving the communication and discussion of the most intimate details of the patient’s life and thoughts. Moreover, we can suppose that they are quite attached to each other, in the following ways. The patient has profound respect for the therapist’s insights and therapeutic prescrip- tions; she takes the view that she never could have progressed as well or achieved what she has through therapy with another, less skillful and attuned therapist. At the same time, she keeps their relationship firmly in perspective: the therapist is a pro- vider of professional services, and she is a consumer. This explains why she is comfortable allowing the therapist to access even the darkest, most protected corners of her psyche, and why she is willing to tolerate—even insists upon—the therapist’s most brutally honest, unvarnished judgments. As she says, “I would be much too ashamed to reveal some of the things I say to him even to my closest friends, but he is my therapist, not my friend—we don’t have a personal relationship, and I don’t Chapter 2 | Relationships and the Participant Stance 55 take his reactions to my revelations personally.” 15 The therapist, meanwhile, regards his work with the patient as some of the most rewarding of his career; hers is an un- usually fascinating case, and he derives much of his sense of professional accomplishment from what he has achieved with her. As he says, “She is a patient, not a friend, and of course we do not have a personal relationship. But purely in pro- fessional terms, she is my favorite patient.” In the historical sense of relationship—in the sense of a “socially salient pat- tern of historical interaction and attachment,” this patient and her therapist certainly have a relationship. It is, we might say, a very successful instance of a certain kind of professional relationship. But at the same time, as their comments suggest, there is another sense in which they do not have a relationship, despite their deep historical ties. A significant dimension of emotional receptivity and sensitivity that is present between Romeo and Juliet is absent from the therapist and patient’s way of relating to each other. What is missing in the case of the therapist and the patient can, I think, usefully be characterized as a certain kind of participant relationship. Of course, before the end of the play Romeo and Juliet will have both a par- ticipant relationship and a (tragically brief) historical relationship. The two kinds of relationship can, and very often do, coexist. And moreover, the most paradigmatic relationships—between friends of long acquaintance, married couples and divorced ones, professional colleagues with long experience working together, etc.—are gen- erally relationships in both senses. The cases of Romeo and Juliet (when they first meet) and of the therapist and his patient are both unusual, in different ways. One way to think about the participant sense of relationship is that it encompasses the dimension of normal or paradigmatic relationships that seems to be missing from the case of the therapist and the patient, even though they have a well developed histori- cal relationship. But let me stress that the cases we have discussed so far are meant to be illustrative, and to provide an intuitive grasp of the distinction between two dif- ferent senses of “relationship.” We will not arrive at a more complete understanding 15 This, by the way, is an illustration of the way in which successful therapy may re- quire both parties to adopt the objective attitude towards one another, at least to some extent. See also note 3 above. Responsibility and the Emotional Structure of Relationships 56 of the distinction until I have given a fuller account of what a relationship in the par- ticipant sense actually consists in. From now on, unless noted otherwise all talk of relationships should be un- derstood as referring to participant relationships. I have already suggested that a relationship (in this sense) consists in two people taking a certain orientation or stance towards each other. Let’s call this stance or orientation—whatever it is—the “participant stance.” Introducing the idea of a stance or orientation provides a no- menclature with which to distinguish between particular, episodic other-directed attitudes—love, hurt feelings, resentment, and so on—and the general openness or receptivity towards others that is a precondition for such episodic attitudes. This means that we can use the idea of the participant stance as a bridge to an informative account of what the participant attitudes are, in two steps: First, we need an account of exactly what the participant stance consists in—this will tell us what relationships (in the participant sense) are. Second, we can then apply the Relationship Thesis, and conclude that the participant attitudes are just those attitudes that we are susceptible to in virtue of taking the participant stance towards another person. So far, of course, this is just a sketch of the structure of a certain way of con- ceptualizing relationships and the participant attitudes. I will turn to filling in the substance shortly, but first it is worth pausing briefly to consider some structural fea- tures. If we think of a relationship as consisting in the parties’ taking a certain stance towards each other, then there can be bilateral cases in which each party takes the participant stance towards the other (think of normal adult friendships or romantic relationships), as well as unilateral cases in which only one party takes the participant stance while the other takes an objective stance (think of a patient who becomes at- tached to her therapist while the therapist remains strictly detached). We can also distinguish between bilateral relationships in which the parties’ reciprocal adoption of the participant stance is common knowledge between them, and relationships in which it isn’t. As an illustration of the various possible configurations, consider the case of Fran the chef and Gabrielle the food critic. 16 First, we might suppose that Fran 16 Thanks to Erik Encarnacion for discussion on these points, and for suggesting the example. Chapter 2 | Relationships and the Participant Stance 57 deeply respects and is (in the way that, I will soon propose, is characteristic of the participant stance) personally invested in Gabrielle’s judgments about her food, while Gabrielle has no particular emotional receptivity or sensitivity towards Fran, but re- gards her as merely material for another review. This would suggest that we have a sort of unilateral relationship, in which one party (Fran) but not the other (Gabrielle) takes up the participant stance. Alternatively, we could suppose that Gabrielle also respects Fran, and is personally sensitive to Fran’s judgment concerning whether she is fair and balanced in her reviews. This could be a case in which each party takes the participant stance, at least in some respect, towards the other. But if the two have never met and neither knows or suspects that the other takes the participant stance towards her, then we have something a bit thinner than many full-blown relation- ships; Fran and Gabrielle’s relationship decomposes into something like two disconnected, unilateral relationships, constituted separately by Fran’s participant stance towards Gabrielle, on the one hand, and Gabrielle’s participant stance towards Fran, on the other. Finally, if Fran and Gabrielle share common knowledge of their mutual concern for each other’s good opinion, then we begin to approach the sort of more full-blown, reciprocal interpersonal involvement that we think of as a paradig- matic relationship (though the relationship may still be quite limited in scope). It is also worth noting that this could be a description of the way in which the relation- ship develops over time, such that the historical relationship involves a deepening of participant relations. This suggests an account of how thicker, more intimate and entangled interpersonal relations can be built up out of thinner layers of relating. It is, I claim, a strength of the idea of the participant stance that it picks out a very basic structure that is found in a wide and diverse range of different kinds of relations. This will provide some useful flexibility that we can use to specify various ways in which different types of more elaborate relationships are built onto the basic structure. But for the moment all we have is a structure; in order to put the idea of the participant stance to work, we need to fill in what this stance actually consists in, and explain how it can help us to understand what the reactive attitudes are. Responsibility and the Emotional Structure of Relationships 58 4 The participant stance: two initial passes Let’s turn, then, to what the participant stance is. To repeat the earlier caveat, if our goal is better to understand what interpersonal relationships are, then on pain of circularity we cannot be content to characterize the participant stance in terms of susceptibility to the participant attitudes, since we understand the participant atti- tudes as just those attitudes that we are susceptible to in virtue of participating in interpersonal relationships. As we will see, my view is that the participant stance con- sists in a particular way of caring about or being invested in someone’s attitudes towards you; once we understand what this kind of caring consists in we will have a substantive new answer to the question of what the participant attitudes are reactions to, which will give us a criterion that defines and unifies the class of participant atti- tudes. I will work progressively through two unsatisfactory accounts of the participant stance, as a way of building up to my own proposal. 4.1 First Pass As a first pass, we might look back to Strawson’s remark about the “the kind of importance we attach to the attitudes and intentions towards us of those who stand in . . . relationships to us.” One perhaps natural and straightforward way to draw the distinction between the participant and objective attitudes is to say that the participant attitudes just are responses to the attitudes and intentions of other people towards us. Working backwards from this might yield something like the following understanding of the participant stance: Participant Stance Thesis [First Pass]: To take the participant stance towards someone is to care intrinsically about their attitudes towards oneself (at least in some respect and to some degree). The participant stance, recall, was introduced as a way to characterize what it is to be in a relationship, in the participant sense. And according to this way of thinking about relationships, when we combine First Pass with the Relationship Thesis (ac- cording to which the participant attitudes just are those attitudes that we are susceptible to in virtue of participating in an interpersonal relationship), the partici- pant attitudes will be just those attitudes that we are susceptible to in virtue of caring intrinsically about someone’s attitudes towards us. So First Pass has the virtue of Chapter 2 | Relationships and the Participant Stance 59 suggesting a determinate and informative way to characterize what the participant attitudes are. We will turn in a moment to weighing other advantages and disadvantages of First Pass. But first, four preliminary interpretive observations are in order. First: According to First Pass two people are in a (normal, bilateral) relation- ship when each of them cares intrinsically about the other’s attitudes towards herself. But First Pass does not say that each needs to care about all of the other’s attitudes, nor that she needs to care to any great extent. First Pass thus imposes no threshold that must be met to count as adopting the participant stance; rather, the participant stance comes in different scopes and degrees. One person takes the participant atti- tude towards another precisely to the extent that she cares intrinsically about the other’s attitudes towards herself. In principle, the scope and extent of the caring could be indefinitely small. We will return to this point later in this section. Second: First Pass is limited to intrinsic caring. This serves to exclude merely instrumental or strategic reactions to the attitudes of others—fear, concern, relief— that are not distinctive of interpersonal relationships. We can see intuitively that such attitudes are not participant attitudes because there would be nothing the least bit unusual about our having them even towards wild animals. That is, just as we can be concerned about a tiger’s intention to eat us, we can be concerned about a thief’s intention to rob us or a bully’s intention to harass us—but in virtue of these instru- mental responses we are no more engaged in an interpersonal relationship with the thief or the bully than we are with the tiger. (Contrast the different level of engage- ment involved in resenting the thief’s or the bully’s intentions, as opposed to merely fearing them for what they portend.) Third: First Pass invokes a basic notion of “caring” for which I will not here offer an analysis or philosophical account. 17 I have in mind a broad class of attitudes, including valuing and deeming important but also simply desiring or being averse to. For my purposes, it is not important exactly how we think of caring, because ulti- mately my favored proposal will replace the broad language of “caring” with a 17 There will, however, be occasion to say more on this topic in Chapter 4, particu- larly §§ 3.4 and 4. Responsibility and the Emotional Structure of Relationships 60 narrower kind of attitude that is described in different terms—so we can think of “caring” here as a kind of placeholder. Fourth and finally: I do not have a precise account on offer of what it means for an attitude to be towards or about the subject in the relevant sense. The rough idea is that if I learn that someone thinks that I, SB, am (for instance) ridiculous or con- temptible, that is an attitude that I could intelligibly have a participant response to (for example, shame); but if I learn instead that the person thinks that someone who is a total stranger to me is ridiculous or contemptible, then a participant response (like shame) is not available to me (barring some empathetic or vicarious or other mechanism of identifying with or responding on behalf of the targeted person). There are, however, cases that fall between these extremes. It seems clear, for in- stance, that the attitude need not make direct reference to me in order for it to be about me in the relevant way. I could have a participant response (like shame) when I learn that someone has contempt for my ethnicity, nationality, or religion, even when that person has no idea that I exist. A more precise delineation of the class of attitudes that are possible objects of the participant stance (and thus of the partici- pant attitudes) would require a broader inquiry into how we construct our identities. This is a task that I will not undertake here; I will make do instead with a rough-and- ready sense of what counts as an attitude that is (or is not) about someone, and I will stick to relatively clear cases. With these preliminary matters out of the way, we can confirm that First Pass is on a potentially promising track by considering again Strawson’s examples of reac- tive (or in our terms, participant) attitudes: gratitude, resentment, forgiveness, love, hurt feelings, personal antagonism, and so on. It does seem at least plausible, I sub- mit, that all of these involve attaching intrinsic importance of one kind or another to the attitudes of others towards us, and reacting on the basis of that importance. Moreover, First Pass sounds very similar to something Strawson himself says at one point about the reactive attitudes, namely that they “are essentially reactions to the quality of others’ wills towards us, as manifested in their behaviour: to their good or ill will or indifference or lack of concern” (83). Subsequent commentators have ech- oed this remark. For instance, in the course of a reconstruction and interpretation of Strawson’s argument Pamela Hieronymi offers the following definition of a reactive Chapter 2 | Relationships and the Participant Stance 61 attitude: “In general, a reactive attitude is x’s response to x’s perception of the quality of y’s will towards z” (MS: 5, fn. 6). Similarly, in explaining his Strawson-influenced account of responsibility and blame Scanlon suggests that the reactive attitudes are “attitudes towards a person that are reactions to the attitudes towards others that are reflected in his or her actions” (2013: 89). 18 For our purposes, however, First Pass and these similar sounding formula- tions from Strawson, Hieronymi, and Scanlon are not going to be satisfactory. (This is not to suggest that Hieronymi’s and Scanlon’s formulations are not satisfactory for the quite different purposes to which they put them—neither is concerned, in the discussions from which I have quoted, with the set of problems surrounding the Re- lationship Thesis that are my focus.) The problem, for our purposes, is that these formulations are all in one important respect too inclusive. Nonetheless, because First Pass provides a useful stepping stone to more promising proposals, before identifying its basic flaw it is worth briefly clarifying two respects in which First Pass is even more inclusive than Hieronymi’s, Scanlon’s, and Strawson’s own remarks would suggest, and indicating why we should see this as a potential virtue rather than a defect. (This virtue is one that subsequent proposals will inherit.) First Pass is, in the first place, not limited to the attitudes of others that are manifested or expressed in their conduct—it characterizes the participant stance in terms of our caring about attitudes as such. In this respect, the language of First Pass is broader than both Strawson’s suggestion that the reactive attitudes are “reactions to the quality of others’ wills towards us, as manifested in their behaviour” (83, emphasis added) and Scanlon’s similar suggestion that the reactive attitudes are “attitudes to- wards a person that are reactions to the attitudes towards others that are reflected in his or her actions” (2013: 89, emphasis added). 19 Of course, when theorizing about responsi- 18 Hieronymi and Scanlon’s formulations are both worded so as to encompass the standard interpersonal case (e.g., x resents y’s treatment of x) as well as the vicarious case (x is indignant about y’s treatment of z) and the reflexive case (x feels guilty about x’s treatment of y). As previously noted, for simplicity I focus here on the standard interpersonal case. 19 In fairness to Scanlon, this remark may be intended as a loose characterization of the reactive attitudes rather than a precise definition. (I suspect that something simi- lar is true of Strawson’s remark.) In other work Scanlon argues that attitudes alone can be proper objects of blame, in the absence of actions (2008: 157-58). Responsibility and the Emotional Structure of Relationships 62 bility there are important questions about the conditions under which it is appropriate for one person to hold another responsible for her attitudes alone, in the absence of any conduct that manifests those attitudes. But we are inquiring into the nature of relationships and the participant stance more generally; we should beware lest look- ing ahead to questions about responsibility may mislead us and distort our understanding of relationships. In relationships it seems that we often do care as much about others’ attitudes, as such, as we do about their conduct. Our feelings can be hurt, for instance, by another’s disesteem or disapproval, even if is revealed only by an involuntary blush or grimace and never acted upon (indeed, even if it is never revealed at all, but only confidently inferred by us). It thus seems premature at this stage, in developing an account of relationships, to limit the participant stance to concerns about others’ attitudes as manifested in overt behavior. In the second place, First Pass’s reference to “attitudes towards oneself” is meant to be construed quite broadly, to include cognitive attitudes (beliefs, judg- ments, assessments, evaluations, etc.), conative attitudes (desires, intentions, motivations, wishes, hopes, etc.), and emotional attitudes (positive and negative af- fective responses, affinities, aversions, etc.); as well as the lack or absence of any of these. 20 In this respect, again, First Pass looks more inclusive than Strawson’s remark, which seems to focus on a particular subset of attitudes towards us, namely others’ “good or ill will or indifference or lack of concern,” which he refers to as the “quality of others’ wills” (83). Hieronymi, in the course of elucidating Strawson’s argument, echoes this, defining reactive attitudes as “x’s response to x’s perception of the quality of y’s will towards z.” While Strawson never says exactly what he means by “quality of will,” it seems that he has in mind things like the intention to harm or benefit an- other, concern or indifference for another’s well-being, and the like—he seems not to have in mind other attitudes that are covered by First Pass, such as another’s assess- 20 Moreover, First Pass is meant to include both occurrent and dispositional versions of the relevant attitudes, and (as previously noted) attitudes “towards oneself” are meant to include more than attitudes that have the reactive subject (or propositions making reference to the subject) as intentional objects—also included will be atti- tudes towards classes of which the subject is a member, and perhaps also general dispositions or inclinations (e.g., generalized hostility, insensitivity, generosity, etc.) that the subject has reason to expect may be directed towards her under a variety of possible circumstances. Chapter 2 | Relationships and the Participant Stance 63 ment that you are terrible at checkers or pretty good looking, that you have an infec- tious laugh or a way with words, and so on. Here again, it seems advisable to proceed cautiously and to err on the side of inclusivity. In questions of responsibility, we are often particularly concerned with motives and attitudes of a certain kind that might be well captured by Strawson’s talk of “quality of will.” But in our relationships in general we are concerned in a familiar, distinctive way with a much broader class of attitudes. We can see this by considering common grounds for feelings of shame, which Strawson lists as a reactive attitude. We often feel shame when we are discovered to have wronged another, and thereby to have had a poor quality of will towards them; but we also often feel shame when we perceive or fear that others take a dim view of our physical appearance, our table manners, our ability to control our own bodily functions, or innumerable other fac- ets of our capacities, attributes, and general situation. This suggests that it would once again be premature to limit the participant stance to concern about “quality of will.” It is at least worth leaving open, at this stage, the possibility that the participant stance might encompass a wider concern about attitudes towards us. This naturally raises questions about how widely the relevant type of concern for others’ attitudes can, or should, range. Clearly we can, and arguably should, care (in the way that grounds susceptibility to shame, for example) about whether others see us as kind, generous, honest and the like; at the same time, we sometimes care more than we should about whether others see us as wealthy, successful, or popular; and we simply don’t (and shouldn’t) care in this way about many other things, like whether others know our favorite beer or brand of toothpaste. These details are im- portant for understanding the scope and character of our actual relationships, and for determining what distinguishes a typical or a healthy relationship from a deviant or dysfunctional one. But by design, First Pass abstracts from all of this; it represents an attempt to say something general about what the participant stance is, without settling any ques- tions about its natural or appropriate objects, scope, or intensity. Instead, First Pass is satisfied whenever someone cares about an attitude in the relevant way, whatever that attitude happens to be, without restriction. So if an idiosyncratic person happens to care, in the relevant way, only about whether others believe that she was born on a Responsibility and the Emotional Structure of Relationships 64 Wednesday, according to the thesis she thereby adopts the participant stance (only) with respect to that belief. This approach is consistent with the general descriptive thrust of the present inquiry, and has the benefit of ensuring that the theoretical tools we are developing are maximally flexible. It may, however, seem to be a cost of this way of proceeding that it means that the person who cares in the relevant way only about whether someone believes she was born on a Wednesday will, simply on that basis, count as standing in a “rela- tionship,” in the participant sense, with that other person. For my part, I do not mind incurring this cost and am happy to call this a relationship, albeit a very strange, limited, and perhaps unilateral one. This has the advantage of putting this way of re- lating on the same spectrum with other, richer ways of relating with which it shares what I regard as an important distinguishing feature. Then we can go on to describe narrower types of relationships by placing restrictions on the objects, scope, or in- tensity of the participant stance. For instance, we could say that two parties stand in a Scanlonian moral relationship (understood as a subspecies of relationships in the participant sense) when each party cares in the relevant way about whether the other reasonably sees her conduct as consistent with principles that no one could reasona- bly reject. 21 This is just one example of a limitless number of different ways that we could build on the basic conception of relationships in the participant sense. But others may object that this way of putting things dignifies idiosyncratic and seemingly fetishistic patterns of concern (like the concern with whether others believe you were born on a Wednesday) with the titles of “participant stance” and “relationship.” They may prefer to reserve these terms for more involved—or more normatively attractive—relations. In principle I have no deep objection to this, and feel no need to insist on the terminology that I have chosen, as long as we could also agree on an alternative terminology for describing the wider range of what I am call- ing “participant” phenomena. To recap, First Pass is broad in two significant respects: first, it includes among the participant attitudes our reactions to others’ attitudes towards us, whether or not those attitudes are manifested in conduct; second, it includes our reactions to 21 See Scanlon 1998, especially chapter 2, § 3 (“A Contractualist Account of Motiva- tion”). Chapter 2 | Relationships and the Participant Stance 65 others’ attitudes towards us quite generally, not just our reactions to their “qualities of will.” But it is, I submit, a point in favor of First Pass that it seems broad enough in these ways to capture the full range of the participant attitudes that Strawson had originally cited (as well as the full range of possible cases in which those attitudes might be manifested). When our feelings are hurt by well-meaning but exacting criti- cism, when we are grateful for someone’s generosity, when we resent a slight, when we are ashamed of our shabby apparel, and when we partake in “reciprocated adult love” (80)—in each of these cases in various ways we are, at least in part, responding to the attitudes of another person towards us in a way that reflects the intrinsic im- portance that we attach to those attitudes. First Pass, I submit, gets this much right. The problem with First Pass, however, is that in a different respect it is too broad: It encompasses the full range of Strawson’s participant attitudes, but it also includes a wide range of attitudes that fall intuitively into the objective category. Re- call that as Strawson sees it, “the objective attitude may be emotionally toned in many ways, but not in all ways: it may include repulsion or fear, it may include pity or even love, though not all kinds of love” (79). Social scientists or therapists, for in- stance, might respond to the attitudes that their subjects or patients take towards them with feelings of admiration, disdain, approval, disapproval, and even a sort of detached love—reactions that reflect their recognition of the intrinsic worth of the target attitudes. The therapist might privately disapprove of or find disagreeable her patient’s violent or abusive impulses (including impulses directed towards the thera- pist herself). Not only is this consistent with the therapist’s general objectivity of attitude towards the patient, it seems that at least a limited readiness to assess the patient’s attitudes in this emotionally engaged way may be useful or even necessary in order for the therapist to evaluate her patient’s state of health and treatment needs— the therapist should be guided by a thick assessment of whether the patient’s atti- tudes (including attitudes towards her) are healthy or impaired. Similarly, a social scientist studying altruistic motivations might feel admiration for a subject’s altruistic impulses (including impulses directed towards the social scientist herself), consistent with maintaining her objectivity of attitude towards the subject. 22 22 Other examples of this dynamic might include someone conducting a job inter- view or a college admissions interview—the interviewer might regard the applicant’s Responsibility and the Emotional Structure of Relationships 66 This suggests that with First Pass we have not yet hit on a fully satisfactory characterization of the participant stance. 4.2 Second Pass The problem with First Pass is that it classifies as participant attitudes a range of assessments and responses that still seem in some sense too “detached,” too arm’s-length, compared to the sorts of examples that Strawson used as illustrations of the participant attitudes. Still missing is an element of personal investment in the attitudes of the other person that seems essential to the participant stance. Seth Shabo, who has written an important series of recent papers on the Strawsonian conception of relationships, seems to be getting at something like this point when he remarks: “To be involved in a personal relationship with someone, we must care in a particular way about that person’s attitudes towards us” (2012: 133, emphasis added). He goes on to characterize this as caring “in an essentially personal way,” and suggests that to be involved in a relationship with someone is to “take [their] treatment of us personally” (2012: 139). 23 Following Shabo’s lead, we could modify First Pass in or- der to capture the missing personal element in the following way: Participant Stance Thesis [Second Pass]: To take the participant stance towards someone is to care in an essentially personal way about their attitudes towards oneself (at least in some respect and to some degree). If we then combine Second Pass with the Relationship Thesis, the participant atti- tudes will be just those attitudes that we are susceptible to in virtue of caring in an essentially personal way about others’ attitudes towards ourselves. Second Pass is narrower than First Pass, but does not seem excessively nar- row. It is still broad enough to capture the range of canonical reactive attitudes— responses like hurt feelings, gratitude, resentment, shame, and reciprocal love all do seem to involve caring in some essentially personal way about someone’s attitudes generous or considerate attitudes towards her as intrinsically admirable, and treat them as positive qualifications in the context of the selection process, consistent with maintaining an objective stance. 23 Cf. Wallace’s suggestion that when the social scientist or therapist deviates from the objective attitude, “[a]s we might colloquially put the point, they have started to ‘take it personally’ that their subjects act as they do” (2014: 123). Chapter 2 | Relationships and the Participant Stance 67 towards us. 24 Moreover, unlike First Pass, it does not seem, at least at first blush, that Second Pass is necessarily too broad as a characterization of the participant stance, provided at least that we can appropriately fill out what it is to care in an “essentially personal way” about others’ attitudes towards us. Second Pass also inherits the dis- tinctive breadth and flexibility (for instance, with regard to the variety of possible objects of concern) that I suggested were advantages of First Pass. Second Pass thus seems at least to be narrowing in on our goal, and indeed the account that I will propose in a moment will essentially be a version of Second Pass. But the remaining issue with Second Pass as it stands is that it merely labels what is missing from First Pass—namely, an “essentially personal way” of caring— without offering much insight into what the missing form of caring actually is. Shabo himself, pursuing a similar line of thought, ultimately does not attempt to say what it is to care in this special way about others’ attitudes towards us—or as he also puts it, to take others’ attitudes towards us personally. He admits: “While it is not easy to say what it is to take something personally, indications that someone has taken something per- sonally (or has ‘taken offence’) are often easy to recognize” (2012: 139, emphasis added). Shabo thus settles for a set of earmarks—characteristic emotional and behav- ioral dispositions and susceptibilities—that we can use to identify cases in which someone has taken something personally, without offering an account of what it ac- tually is to take something personally in the relevant sense. 5 Self-understanding and the participant stance Second Pass, then, is not obviously overbroad like First Pass, but as an ac- count of what the participant stance is, it is incomplete. We can at least aspire to a more articulate account. To this end, I offer the following: Participant Stance Thesis: To take the participant stance towards someone is to treat their attitudes towards oneself as (non- derivatively) central to one’s self-understanding (at least in some re- spect and to some degree). 24 Remember that for the sake of simplicity we are focusing on the “non-vicarious” reactive attitudes; to capture vicarious attitudes like indignation we will ultimately have to widen the account to include attitudes that we are susceptible to when we imaginatively inhabit the perspective of another person’s involvement in an interper- sonal relationship. Responsibility and the Emotional Structure of Relationships 68 And combining the Participant Stance Thesis with the Relationship Thesis yields that the participant attitudes are just those attitudes that we are susceptible to in virtue of treating someone’s attitudes towards ourselves as (non-derivatively) central to our self-understandings. This final Participant Stance Thesis represents an attempt to give more specific content to the idea of caring in an “essentially personal way” about someone’s attitudes; at the same time, as a modified version of Second Pass, it inherits the advantages and structural features of the earlier passes. But this proposal stands in need of some explication. It may be useful to start by understanding the Participant Stance Thesis in the context of a very general ob- servation about our human condition. The general—and, I hope, uncontroversial— observation is that we are self-aware creatures, each of whom conceives of herself under a variety of more or less stable descriptions, ranging from those that are quite marginal to our self-understandings (e.g., born on a Wednesday) to those that are quite central (e.g., wife or husband, mother or father, expert or novice in various fields, trombonist or philosopher, and so on). The self-descriptions that are more central to our self-understandings are often, for instance, those that we rely on fre- quently and extensively in our deliberations, those that are important to our sense of self worth, those that play a prominent role in the narrative through which we under- stand our personal history and development, or those that we take as important points of comparison and contrast between ourselves and others (this list is meant to be illustrative, not exhaustive). Together, the descriptions that are most central to our self-understandings in this multifaceted sense constitute our sense of identity and relative position; they are, in short, the descriptions through which we understand who we are and where we stand in the world. These core self-descriptions thus include the self-conceptions as agents that we examined in chapter 1 (borrowing from Raz), as well as a wide range of other self-descriptions—concerning, for instance: our personal attributes; personal histo- ries; institutional and other affiliations; tastes, preferences, and goals; hopes and fears; ultimate values; social identities (gender, race, ethnicity, religion, nationality, etc. 25 ); social connections; and general social standing and reputation, that is, how we 25 Raz specifically notes that “our sense of who we are” is “in part determined inde- pendently of our activities (say by gender or ethnicity and their social meanings)” Chapter 2 | Relationships and the Participant Stance 69 are related to others in our community and to other persons more generally. The last few social items in this open-ended list bring us to the particular facet of this general observation about the human condition that is especially pertinent for present pur- poses, namely that among the descriptions that are quite central to our self- understandings we invariably find—at least among typical, well-socialized adults— descriptions of the contents of other people’s attitudes towards us; for instance, that we are loved, hated, respected, relied upon, trusted, suspected, feared, misunder- stood, and the like. It is perhaps easy to forget, considering the matter from a theoretical remove, just how deeply integrated into our self-understandings these relational or social self- descriptions are. 26 To appreciate this, it may help to retrace a few steps from chap- ter 1. Consider first the constituents of our basic self-conceptions as agents: our senses of our various basic physical and mental capacities and competences, such as our abilities to walk, talk, drive, dance, eat, write, remember, recognize objects and people, perform simple addition, and so on. Raz pointed out that order to exercise complex agency of any kind, we need to be able to rely implicitly on our understand- ing that we reliably command these and innumerable other basic capacities. 27 But chapter 1 further argued that to function as normal, participating members of our communities, we also need to be able to hold ourselves out to others as competent to do these things. In order to coordinate our activities with others, engage in reciprocal schemes of cooperation and mutual aid, and feel confident that we are valued and respected in these roles, we need to be assured that others recognize us as competent in various ways. To be seen as incompetent by others, and thereby as unfit to partici- pate in the various social activities that make up the life of the community, is not only to our serious material disadvantage, but also threatens our very senses of who we are, of how we stand in relation to others, and of our own worth. (2011: 239); that is, by self-descriptions that fall outside of our narrower self- conceptions as agents. 26 As Strawson observed, “it is easy to forget when we are engaged in philosophy, especially in our cool, contemporary style, . . . what it is actually like to be involved in ordinary interpersonal relationships, ranging from the most intimate to the most cas- ual” (77). 27 Another illuminating discussion of the nature of these competences and their role in complex agency can be found in Railton 2009. Responsibility and the Emotional Structure of Relationships 70 In a similar way the relational or social dimension of our self-understandings incorporates our own attitudes towards others, including (but not limited to) what Strawson called the quality of our wills. That is, we care a great deal about whether we are well-meaning, kind, reliable, trustworthy, and so on, and about whether others see us as such. Someone who was not perceived as good-willed in the relevant ways could not expect to be cooperated with, trusted, befriended, or loved in the way that adults mutually love one another, and to realize that we were in this situation would once again threaten our very senses of who we are, of where we stand in relation to others, and of our own worth. These points are familiar from chapter 1. But they serve here as illustrations of a much more general phenomenon, which is the intensely personal way in which we are often concerned about the attitudes that others take towards us. We care about those attitudes not just strongly and intrinsically, but in this special manner that involves incorporating those attitudes into our self-understandings—our very senses of who we are and where we stand in the world. And a moment’s introspec- tion reveals that this kind of concern extends to a broader range of the attitudes that others take towards us, beyond their perceptions of our competences and qualities of will. We are often equally, if not more, emotionally invested in how others react to, among other things: our personal appearances; past accomplishments; institutional affiliations and positions; and conformity with, or deviation from, the norms and expectations of various aspects of our social identities (gender, class, ethnicity, etc.). It is, moreover, a mark of the special importance that we attach to others’ at- titudes towards us that we appear often to care about them in a way that is distinct even from the way that we care about other, more impersonal aspects of our self- understandings. A change in our fortunes, attributes, or abilities that results from impersonal causes can have a devastating or exalting effect on someone’s self- understanding—a concert pianist’s loss of the use of her hands through a random accident, say, or a young scholar’s stumbling on a career-making discovery in a dusty archive. But no matter how momentous such events may be for us, taken on their own they are not occasions for hurt feelings, shame, offense, resentment, or grati- tude. What sets these reactions apart, even from other reactions that reflect the importance that we accord to facts that we incorporate into our self-understandings, Chapter 2 | Relationships and the Participant Stance 71 is their essential connection to others’ attitudes towards us. The pianist is devastated by—rather than offended by or resentful of—the loss of the use of her hands; how- ever, she resents and her feelings are hurt by the way her musician friends lose interest in her company after the accident. The young scholar is elated by—rather than grateful for—her discovery; but she is grateful to the mentors and supporters who believed in her enough to encourage her research. So I claim that we care in an especially distinctive way about others’ attitudes towards us. But this point deserves to be treated with some care. It is possible that rather than caring about others’ attitudes in a way that is somehow qualitatively dif- ferent from the way that we care about impersonal facts that are central to our self- understandings, we instead have a special vocabulary for describing the way that we care about others’ attitudes. According to this alternative view, it is not that the pian- ist’s hurt feelings when her friends abandon her are a qualitatively different emotion—a different kind of emotional reaction—from her feelings of devastation at the loss of the use of her hands; but rather that we have different terms for describ- ing the same kind of emotional reaction, depending on what it is a reaction to. This could reflect the special interest that we have, as social creatures, in being able to identify and describe significant aspects of one another’s interpersonal attitudes and relationships. A third possibility, of course, is the mixed one: that we have a different emotional repertoire for responding to others’ attitudes towards us, but that we also sometimes use different terms for describing the same kind of emotional reaction depending on whether it is a reaction to another’s attitudes or to some impersonal event. I will not try to settle here which of these three alternatives is the correct one; I intend the Participant Stance Thesis to be consistent with all of them. These observations put us in better a position, I hope, to understand the substance of the Participant Stance Thesis. Consider again: Participant Stance Thesis: To take the participant stance towards someone is to treat their attitudes towards oneself as (non- derivatively) central to one’s self-understanding (at least in some re- spect and to some degree). Again, this represents an effort to capture what it is to take the participant stance towards someone in a more articulate way than simply saying that it is to care in an essentially personal way about the person’s attitudes or that it is to take their attitudes per- Responsibility and the Emotional Structure of Relationships 72 sonally. The thought is that taking someone’s attitudes personally consists in caring about them in a distinctive way that involves treating them as central to our self- understandings. Being a (more specific) version of Second Pass, this proposal excludes from the participant attitudes the sorts of “detached” attitudes that got First Pass into trouble. For instance, the therapist might disapprove of her patient’s violent or abu- sive impulses towards herself—or admire her patient’s altruistic impulses—while maintaining her objectivity of attitude, provided that she disapproves or admires in a “detached” way, without being personally invested in the sense that she treats the patient’s impulses as significant for her own self-understanding. Moreover, the social scientist might be elated (or crushed) to discover that one of her subjects has a par- ticular attitude towards her, but only due to the fact that the presence of this attitude confirms (or falsifies) a hypothesis that she has dedicated her career to establishing. In this case, the social scientist might well view the subject’s attitude towards her as central to her self-understanding, but only derivatively so; that is, only because she views the truth of the hypothesis that the attitude confirms (or falsifies) as central to her self-understanding as a successful (or failed) researcher. Treating the subject’s attitude in this way would thus also be excluded from the participant stance. At the same time, the Participant Stance Thesis provides what I hope is a more determinate and satisfying account of what unites the participant attitudes, in- cluding attitudes—like hurt feelings, shame, and pride—that Wallace had sought to exclude from the “reactive” class. When our feelings are hurt by well-meaning but exacting criticism, when we are ashamed of our shabby apparel, and when we display an accomplishment with pride, according to the Participant Stance Thesis the distinc- tive feature that makes each of these a participant attitude is that we are treating others’ attitudes towards us as intrinsically significant for our self-understandings, either for good or for ill. That is, we are registering that our poor performance has caused us to lose valued ground in another’s estimation of us, that our threadbare clothes make us less qualified in others’ eyes for the esteem we crave, or that our ac- complishments are recognized and valued by others in a way that gratifies or fulfills us. In each case, we see others’ attitudes towards us as an important determinant of who we are or of where we stand in the world. Chapter 2 | Relationships and the Participant Stance 73 Now we are also in a position to reconsider Wallace’s objections to the Rela- tionship Thesis. The Relationship Thesis, as we reframed it, states that to participate in interpersonal relationships just is to be susceptible to the participant attitudes, and conversely that the participant attitudes just are those attitudes that we are suscepti- ble to in virtue of participating in interpersonal relationships. Wallace had objected, first, that the Relationship Thesis embodies a conception of the reactive attitudes that is too indeterminate, and seemingly too broad, to be theoretically useful; second, he objected that the Relationship Thesis is false because participation in an interper- sonal relationship is unnecessary for susceptibility to the reactive attitudes. In an effort to clarify and defend the Relationship Thesis, I have proposed that we think of relationships in terms of the parties’ taking a certain orientation or stance—the “participant stance”—towards each other, and that the participant stance consists in treating the other’s attitudes towards oneself as central to one’s self-understanding. Given this account of what a relationship is, the Relationship Thesis entails that the participant attitudes are just those attitudes that we are suscep- tible to in virtue of being invested in another’s attitudes towards us in this distinctive way. This account delimits the class of participant attitudes using two criteria: the objects that we are reacting to (someone’s attitudes towards us), and the manner in which we react to them (in a way that reflects their significance for our self- understandings). Regarding Wallace’s first objection, I submit that the Relationship Thesis in conjunction with the Participant Stance Thesis yields a conception of the participant attitudes that is both sufficiently determinate to be theoretically useful and that tracks—while also elucidating—Strawson’s examples of reactive (or participant) and non-reactive (or non-participant) attitudes, as well as his way of distinguishing be- tween the participant and objective stances. Regarding Wallace’s second objection, the proposed account straightfor- wardly guarantees that participation in an interpersonal relationship is necessary for susceptibility to the participant attitudes, by characterizing both relationships and participant attitudes in terms of the participant stance—that is, in terms of caring in a particular, personal way about others’ attitudes towards us. Unlike an approach that characterizes relationships in a way that leaves a potential gap between participation Responsibility and the Emotional Structure of Relationships 74 in a relationship and susceptibility to the participant attitudes—for instance, Wallace’s suggestion that we think of relationships in terms of patterns of historical interaction and attachment—the proposed account leaves no such gap. This means that the “problem of the stranger” never arises; we are susceptible to participant atti- tudes towards strangers, and we stand in relationships (in the participant sense) with strangers, precisely to the extent that we take the participant stance towards strang- ers, which is just to say to the extent that we treat strangers’ attitudes towards us as central to our self-understandings. This is actually something that we do quite commonly. Imagine realizing in a flash that two complete strangers across the subway car who have been in poorly concealed fits of laughter for several stops have been amused because they find your outfit, hairstyle, hygiene, or demeanor to be ridiculous. Even if they get off at the next stop, never to be seen again, it is easy to imagine finding this experience fairly deeply discomfiting. This is not to say that you should find it discomfiting, that most people would, or that a mature person who has achieved a reasonable level of self- confidence and emotional independence would not be able to shrug off this kind of experience. Rather, I mean only that it is easy to imagine a person with certain nor- mal and familiar kinds of insecurities and vulnerabilities finding this experience fairly deeply discomfiting. The characteristic depth of this reaction, when it occurs, is in my view a function of the fact that in those instances the subject is taking the par- ticipant stance towards strangers, at least in some respect and to some degree. On this view the participant stance, and by extension relationships, come in a wide range of different scopes and degrees. In the case just described, for instance, there is only a very shallow, narrow, and perhaps one-sided relationship, but still one that shares a distinctive and important structural feature with broader, deeper, and more devel- oped mutual relationships like friendships and romances, namely that one party treats the other party’s attitudes towards her as central (in some respect and to some degree) to her self-understanding. The sense in which there is a relationship in this kind of case is precisely what I have labeled the participant sense of relationship. Chapter 2 | Relationships and the Participant Stance 75 Conclusion This completes my elucidation and defense of the Relationship Thesis. If the defense is successful, it shows that the wide range of attitudes that Strawson cited as reactive or participant attitudes form a unified and potentially useful class, one that is centrally involved in a structure of emotional involvement that appears in many of our interpersonal relations. I have tried, moreover, at least to suggest that this repre- sents an important insight into the nature of our sociality, and that by identifying and characterizing this distinctive, pervasive dynamic in interpersonal relations we gain a useful point of entry for theorizing about a wide range of independently interesting emotions, attitudes, and relationship types. But I have not defended or properly de- veloped these more ambitious claims about the significance of the Relationship Thesis. At most I hope to have provided grounds for optimism, and opened up a path for future investigations in this under-explored area. In the next chapter I will take a first step towards vindicating this optimism, through an examination of one of Strawson’s participant attitudes: shame. Shame will serve as a useful case study both because it is, in my view, a basic and important member of the participant class, and because it often features prominently when agents take responsibility for their conduct (in the sense developed in chapter 1). In particular, chapter 3 will demonstrate the usefulness of this conception of the par- ticipant attitudes by proposing that conceiving of shame as a participant attitude helps to resolve a longstanding disagreement about the role of an audience in the structure of shame. Chapter 4 will then take up again the theme of responsibility, and explore how this conception of the participant attitudes can help to advance our un- derstanding of the interpersonal emotional economy of responsibility, including our practice of holding responsible. 76 Chapter 3 The Audience in Shame The previous chapter identified and described a class of attitudes that we are susceptible to in virtue of treating others’ attitudes towards us as central to our self- understandings; this class is meant to include a wide range of interpersonal re- sponses, from hurt feelings and resentment to gratitude and certain kinds of love. Borrowing from Strawson, I have labeled these the “participant attitudes” and pro- posed that they form a distinct and theoretically useful category that helps us to understand the basic structure of many of our interpersonal relations. This is, in my view, a rich and important but widely overlooked legacy of Strawson’s work. As we have seen, discussion of Strawson has tended rather to focus more narrowly on blaming attitudes, and on the way in which Strawson deploys these attitudes in his case for compatibilism about responsibility. 1 As a result, theoretical attention has centered on a small subset of (what I am calling) the participant attitudes, particularly resentment, indignation, and guilt (and in deference to current convention, I will continue to reserve the label “reactive” for these latter attitudes). 2 1 For discussions of the role of the reactive attitudes in our responsibility practices see, e.g., Gary Watson (1987), R. Jay Wallace (1994), Pamela Hieronymi (2004), Ste- phen Darwall (2006), and T.M. Scanlon (2008); for discussion of the role of the reactive attitudes in Strawson’s case for compatibilism about responsibility see, e.g., Pamela Hieronymi (MS) and Gary Watson (2014). 2 As we saw in chapter 2, § 2, Wallace explicitly limits the reactive attitudes to these blaming attitudes (1994: 29-30). Scanlon broadens the class of reactive attitudes to include changes of intention or goodwill (e.g., changes in one’s disposition to take pleasure in another’s successes), but only insofar as these are involved in our prac- tices of holding responsible (2013: 86-89). Watson (1987, 2014) and Darwall (2006) follow Strawson in classifying a wider range of attitudes (e.g., gratitude and love) as reactive attitudes, but nonetheless focus their discussions on blaming attitudes like resentment that are directly involved in holding responsible. Chapter 3 | The Audience in Shame 77 I have raised the concern that this narrow focus on the blaming or “reactive” attitudes risks losing track of the relationship between two ideas that were tightly connected in Strawson’s original conception: the idea that the reactive attitudes play an important role in our responsibility practices, and the idea that the reactive atti- tudes play a broad and basic role in our interpersonal relations more generally. This connection was critical for Strawson’s understanding and defense of our responsibil- ity practices. In order to regain the insight afforded by the original conception, we would do well, I think, to revert to the more broadly conceived class of participant attitudes and to examine how our responsibility practices are grounded in, and intel- ligible in terms of, the overall emotional structure of our interpersonal relationships. This chapter and the next represent first steps in this larger undertaking. This chapter focuses on a core participant attitude—shame. Understanding shame as a participant attitude, I argue, enables us to resolve a longstanding problem about the role of an audience in shame. The next chapter will return to the questions about responsibility with which we began, and in particular to an examination of our prac- tices of holding responsible. There I will apply the account of relationships and the participant attitudes—including shame—that we have been developing to address a gap in current understandings of blame. 1 Shame as a participant attitude According to the account of the participant attitudes offered in chapter 2, to say that shame is a participant attitude is to say that it is an attitude that we are sus- ceptible to in virtue of treating someone’s attitudes towards ourselves as (non- derivatively) central to our self-understandings. This puts shame into the same cate- gory with other participant attitudes like resentment, gratitude, hurt feelings, and (certain kinds of) love, and distinguishes it from non-participant attitudes like fear, disgust, disappointment, and relief. 3 3 Or at least, the familiar impersonal, non-participant versions of fear, disgust, disap- pointment, and the like—as when you are afraid that another person will block your view at a concert, or disgusted by their slovenly appearance. It is consistent with the account of the participant attitudes offered in chapter 2 that there may also be par- ticipant versions of at least some of these attitudes—as when you are afraid that your friend will stop trusting you when he discovers what you’ve done, or disgusted by Responsibility and the Emotional Structure of Relationships 78 This way of categorizing shame diverges from what is sometimes found in the reactive attitude literature. Strawson himself listed shame as one of the reactive attitudes, alongside resentment, hurt feelings, and the rest (1962: 75, 85). 4 But as dis- cussed in chapter 2, subsequent treatments have often narrowed the class of reactive attitudes to exclude shame; indeed shame, particularly when contrasted with guilt, is sometimes taken as a paradigmatic example of a non-reactive attitude. For instance, R. Jay Wallace contends that guilt, unlike shame, involves “the quasi-evaluative stance of holding oneself to an expectation, so that in feeling guilty about an action one thinks of the action as violating some expectation”; and because the reactive atti- tudes are, for Wallace, just those attitudes that involve holding someone to an expectation, guilt but not shame qualifies as a reactive attitude (1994: 240-41). Ste- phen Darwall reaches the same conclusion for different reasons. Guilt, he argues, unlike shame, involves recognition of another’s second-personal authority to make claims on the guilty party; because the reactive attitudes are, for Darwall, just those attitudes that invoke or recognize this sort of second-personal authority, guilt but not shame is a reactive attitude (2006: 67, 71-72). 5 Gary Watson is a noteworthy dissenter from this view. Unlike Wallace or Darwall, Watson follows Strawson in listing both guilt and shame as reactive atti- tudes (1987: 220). Moreover, Watson has recently argued that shame, like guilt, (i) can be a natural expression of our concern for the quality of will that we display towards others, (ii) can demonstrate our commitment to moral standards and expec- someone’s cynical attempt to manipulate you (here disgust shades into taking of- fense, another participant response). 4 My view parallels Strawson’s in that I consider shame to be a participant attitude. But there is an important difference between Strawson’s view of shame and mine (or at least, there appears to be; in fairness to Strawson, his treatment of shame is ex- tremely brief). Strawson suggests that shame, together with things like compunction, guilt, and remorse, are “self-reactive attitudes associated with demands on oneself for others” (1962: 84). But in my view, as we will see, shame is in most cases an ordinary participant response to others’ attitudes, just like resentment or gratitude. Shame is thus not a “self-reactive” response that one has towards oneself on behalf of others. This is one important distinction between shame and guilt that we will return to in chapter 4, § 5. 5 I have been unable to find any reference to shame in Scanlon’s writings on respon- sibility and the reactive attitudes (see, e.g., his 2008 and 2013), though he classifies guilt as a reactive attitude (2008: 143). Chapter 3 | The Audience in Shame 79 tations, and (iii) “plays a vital role in securing conformity with our ethical norms” (2014a: 28-31). In all of these ways, Watson observes, shame and guilt are similarly bound up with our responsibility practices. Thus, even if we conceive of the reactive attitudes fairly narrowly as just those attitudes that are in some sense closely con- nected to responsibility, the considerations Watson identifies would seem to weigh in favor of classifying at least some instances of shame as reactive. Watson’s case strikes me as persuasive, and I intend for my claims about shame to be compatible with his. But for present purposes I do not need to take a side in this disagreement about whether shame counts as a reactive attitude. I will have more to say about the blaming attitudes that preoccupy Wallace and Darwall, and about the connection between those attitudes and shame, in the next chapter. For now it is sufficient to point out that whether shame is, as I contend, a participant attitude (in my sense of that term) is independent of whether shame is a reactive atti- tude, however the latter category may be defined (e.g., in terms of holding to expectations, exercising second-personal authority, or holding responsible). There is thus no inconsistency in placing shame and guilt in different categories as regards “reactivity” and yet also in the same category of the participant attitudes. It is also worth noting that, in contrast to Watson’s claims about shame as a reactive attitude, my proposal that we conceive of shame as a participant attitude is meant to apply to all instances of shame. Watson’s arguments tend to show that at least some instances of shame are appropriately considered reactive, in virtue of both the way they demonstrate our attachment to moral standards and the role they play in securing conformity with those standards. But at the same time, there are clearly instances of shame that are not connected to moral standards in this way (e.g., shame relating to one’s physical appearance). My aim here is to clarify the structure of shame quite generally, including both what we might call the reactive and non- reactive instances of shame. Part of my proposal will thus be that all instances of shame are participant attitudes in the sense developed in chapter 2, even if not all in- stances of shame are reactive (or otherwise responsibility-relevant). 6 6 Even if I am wrong and shame is not always a participant attitude (in my sense), my claims about shame would still apply to a wider range of instances of shame than Watson’s claims about the reactivity of shame. Responsibility and the Emotional Structure of Relationships 80 As I will argue, characterizing shame as a participant attitude is useful for un- derstanding the widely shared but difficult-to-pin-down sense that shame is, like gratitude and hurt feelings but unlike impersonal fear or relief, a social emotion. And at the same time, thinking of shame as a participant attitude will help to advance our understanding of the participant attitudes and the participant stance, because in order to make sense of shame as a participant attitude we will also need to account for the ways in which we can feel shame in response to our own attitudes. This will serve as a useful illustration of the ways in which we can, and I think often do, adopt the par- ticipant stance towards ourselves. Let us turn then to an examination of the theoretical difficulties that have arisen around the idea that shame is a social emotion. 2 The problem of the audience in shame Adam and Eve were naked without shame before they ate of the tree of knowledge; after eating they became ashamed of their nakedness and covered them- selves with fig leaves. Nakedness, of course, is a matter of being exposed to the view of others. Other familiar experiences of shame also centrally involve exposure; think of the shame associated with some bodily functions, or the shame of being “caught in the act.” Reflection on these and similar cases has produced a long theoretical tra- dition of understanding the gaze or, more generally, the appraisal of an audience to be an essential element in experiences of shame—call this the Audience Thesis. The Audience Thesis is a common way of trying to capture the sense in which shame appears to be a social emotion. John Deigh, for instance, contends that “a satisfactory characterization must include in a central role one’s concern for the opinion of others,” and goes on to suggest that this is “really a lesson in recall. From Aristotle onward, discussions of shame have focused attention on the subject’s con- cern for the opinions others have of him” (1983: 240). 7 Similarly, Bernard Williams asserts that “shame and its motivations always involve in some way or other an idea of the gaze of another” (1993: 82); Cheshire Calhoun contends that shame “seems 7 In addition to Aristotle, Deigh cites Aquinas, Descartes, Spinoza, Darwin, and Sar- tre as incorporating a concern for others’ opinions into their definitions of shame (1983: 240). Chapter 3 | The Audience in Shame 81 intrinsically tied to the thought of social others’ actual or imagined contempt” (2004: 131); Heidi Maibom says that “Central to shame . . . is the notion of a seeing other, of an audience” (2010: 569); and recently Phillip Galligan has said that in shame the subject is “intrinsically concerned about the negative perceptions of others” (2014: 64). These are all versions of the Audience Thesis—that shame essentially involves being exposed to the appraisal of others. 8 But the Audience Thesis, despite its pedigree and intuitive appeal, appears to be subject to decisive counterexamples in the form of cases of private shame. Con- sider, for instance, the case of Sam, who has been working productively on an article all afternoon and initially feels quite pleased with the result (Deonna & Teroni 2011: 202). After a coffee break, however, he rereads his draft and notices an unsupported premise. Suddenly it dawns on him that the conclusion he had meant to be arguing for is essentially equivalent to the unsupported premise. What he had thought to be a compelling argument turns out to be a circular muddle, “contrived and badly written, the assertive tone unwarranted, the boldness pure defensiveness”; as a result, “Sam feels so ashamed” (Deonna & Teroni 2011: 202). In this case there appears to be no audience that can be invoked in connection with Sam’s shame, as he has not shared his draft with anyone. While his experience of shame involves a change of view on Sam’s part, in which he perceives his writing in a new, less favorable light, this new appraisal is still very much his own—there is no other evaluative perspective, distinct from Sam’s, from which the appraisal issues. Sam is, in short, alone in his shame. 8 The claim that an audience’s perspective is an essential element in experiences of shame also has proponents in the psychological literature. Mark Leary, for instance, contends that shame—together with other emotions like guilt, embarrassment, and pride—belongs to a family of “self-conscious” emotions that arise “not from self- reflection per se . . . but rather from imagining oneself through the eyes of other people” and in particular how one is “perceived and evaluated by other people” (2007: 45). For similar views see, e.g., Baldwin & Baccus 2004, Kemeny et al. 2004, and Elison 2005. The Audience Thesis has also played a role in anthropological discussions of the distinction between “shame cultures” and “guilt cultures.” Ruth Benedict, for exam- ple, asserts: “Shame is a reaction to other people’s criticism. A man is shamed either by being openly ridiculed and rejected or by fantasying to himself that he has been made ridiculous. . . . But it requires an audience or at least a man’s fantasy of an audi- ence. Guilt does not” (1946: 223). Responsibility and the Emotional Structure of Relationships 82 On the strength of examples like Sam’s, a number of theorists have rejected the Audience Thesis outright. Anthony O’Hear, for instance, contends that “it is quite possible to think of people, such as writers or craftsmen, with high standards of their own, feeling shame just because they have let themselves down (not produced a masterpiece), without thinking of them imagining other craftsmen inspecting and condemning their work”; he concludes, “shame does not then require the back- ground of a real or imaginary audience” (1976: 77). Like O’Hear, John Kekes argues simply that “we often feel shame when no one is present to observe us. So audience cannot be necessary” (1988: 284). Recently Julien Deonna and Fabrice Teroni (the authors of the Sam example) have run the same argument, noting that cases in which “shame is felt as a result of a failure to uphold a privately held ideal” are “common at least for some of us,” and concluding from this that a “concern for others’ opinions . . . is not always the relevant determinant of shame” (2011: 200). 9 We have then a neat stalemate pitting critics of the Audience Thesis against its advocates, in which each side can point to apparent commonplaces and platitudes in support of its view. Broadly speaking, theories that describe shame as essentially “social” 10 or “group-centered” 11 or “heteronomous” 12 accept the Audience Thesis; theories that describe shame as essentially “asocial” or “subject-centered” or 9 Deonna and Teroni, together with a third co-author, Raffaele Rodogno, further elaborate this argument in In Defense of Shame (2012). The same position is also repre- sented in the psychological literature. Jessica Tracy and Richard Robins, for instance, contend that “self-conscious emotions [like shame] are experienced when a person’s identity is threatened or elevated—which can occur in public or private, and in inter- personal or task contexts, as long as the eliciting event is relevant to the aspirations and ideals (as well as the fears) of the self” (2007: 11). What is important, in this view, is that the subject’s own aspirations and ideals are engaged; exposure to an audience plays no essential role. 10 For this terminology see, e.g., Deonna and Teroni’s “Is Shame a Social Emotion?” (2011). 11 Maibom (2010: 570-71) and Galligan (2014: 57-58) contrast “group-centered” and “subject-centered” (or in Maibom’s case, “agent-centered”) understandings of shame. 12 Calhoun (2004: 129), Maibom (2010: 567, 570, 572), Michelle Mason (2010: 405, 408), Deonna and Teroni (2011: 199), and Chloë FitzGerald (2015: 228-33) contrast “heteronomous” and “autonomous” understandings of shame. Chapter 3 | The Audience in Shame 83 “autonomous” reject it. 13 Advocates insist that the Audience Thesis captures some- thing that is distinctive and essential about shame, while critics shrug and rejoin that this cannot be so, as there are familiar cases of shame in which the feature that the Audience Thesis identifies as distinctive and essential is simply absent. Both sides in the debate over the Audience Thesis face serious difficulties. My aim is first to develop these difficulties and then to outline a way of thinking about shame that transcends them. I will suggest that the Audience Thesis reflects a fundamental truth about shame, but does so imperfectly. The truth that it reflects is that shame is not a first-order self-appraisal, like disappointment in or disapproval of oneself; it is, rather, a participant attitude, and as such it is essentially a response to appraisals of oneself—an appraisal of appraisals. The flaw in the Audience Thesis as it has traditionally been understood is that it assumes that shame must be a response to appraisals that issue from an evaluative perspective that is distinct from the sub- ject’s own. But if we remove this restriction and allow that shame may also be a response to one’s own self-appraisals—a response that arises when we adopt the par- ticipant stance towards ourselves—then the difficulties surrounding the Audience Thesis disappear, opening up a promising path for further theorizing about shame (and also, by extension, about other reflexive participant attitudes). Moreover, this proposal also sheds light on other facets of shame. For instance, it makes sense of some of the rhetoric and imagery that are associated with shame; it clarifies what is happening when we feel shame in the face of negative appraisals with which we do not agree; and it explains how it may be possible to feel shame in the face of apprais- als that are neutral or even positive in character. In order for my reinterpretation of the Audience Thesis to deliver these re- sults, it will be necessary to accept an important substantive claim that I will not fully defend here, which is that we can be emotionally invested in our own appraisals of ourselves in a way that is similar to the way in which we are emotionally invested in 13 The Audience Thesis is of particular interest to moral psychologists, because whether shame is governed by the subject’s own values and standards (i.e., is “autonomous”) or rather by the (perhaps alien) values and standards of an audience (i.e., is “heteronymous”) may determine the role that shame can play in moral moti- vation, including the circumstances under which shame can be treated as a reliable guide, rather than a possible threat, to proper conduct. For one nuanced discussion of these issues, see Calhoun’s “An Apology for Moral Shame” (2004). Responsibility and the Emotional Structure of Relationships 84 others’ appraisals of us—which is just to say, that we can take the participant stance towards ourselves. For instance, we are often emotionally invested in the deeply per- sonal way that is distinctive of the participant stance in whether or not, and to what degree, certain others respect us; and I claim that most of us are emotionally invested in a similar way in whether or not, and to what degree, we respect ourselves. While I will not argue in detail for this claim, I think it is independently plausible and I will try to indicate why. But looking at things in another light, if my proposal about the nature of shame offers a promising resolution to the problems surrounding the Audience Thesis, and is otherwise attractive, this will provide at least some corrobo- ration for the underlying claim that we can, and often do, adopt the participant stance towards ourselves. I will next, in §§ 3 and 4, trace a few more steps in the dialectic between ad- vocates and critics of the Audience Thesis, to show why neither camp’s attempts to seize an advantage over the other are satisfying. I will then turn in § 5 to the proposal that we conceive of shame as a participant attitude. 3 Internalizing the audience Advocates of the Audience Thesis often resist the critics’ interpretation of experiences like Sam’s by insisting, despite appearances, that there is an audience— an evaluative perspective distinct from Sam’s own—before which Sam feels shame. This requires the advocates to locate the missing audience somewhere within the psychic mechanisms that produce Sam’s experience of “private” shame. This internal audience can be located shallowly or more deeply within the subject’s mind. On a shallower level, experiences of private shame can be explained in terms of an imagined audience. Williams, for instance, calls it a “silly mistake” to suppose that shame always requires an actual audience, and suggests that “for many of [shame’s] operations the imagined gaze of an imagined other will do” (1993: 81-82). Here he invokes Sartre’s example of a voyeur who is looking through a keyhole and suddenly realizes that he himself is being watched from behind; Williams contends that “an imagined watcher could be enough to trigger the reactions of shame” (1993: 82). Somewhat more subtly, Williams further proposes that the imagined audience might consist not in any specific, identifiable person, but could rather be an “inter- Chapter 3 | The Audience in Shame 85 nalized other” who is “abstracted and generalized and idealized,” representing simply someone, other than the subject, who “can provide the focus of real social expecta- tions, of how I shall live if I act in one way rather than another, of how my actions and reactions will alter my relations to the world about me” (1993: 84). And in the same vein but perhaps more subtly still, Phillip Galligan proposes that in shame we “sometimes adopt others’ perspectives without marking the shift as it happens, and the social perspective need not even be that of any particular person or group”; for instance, a cheater’s shame “might result from imagining what it would be like to be known as a cheater, full-stop,” again without imagining any particular audience as the subject of such knowledge (2014: 67). According to proposals of this kind, cases of “private” shame are to be un- derstood as shame before an imagined audience of a more or less rarified nature that supplies the foreign evaluative perspective that is necessary for shame. So we should suppose that Sam feels ashamed because he explicitly imagines what his discerning colleague would think if she were to read his sloppy draft, or explicitly imagines what some abstracted and idealized discerning reader would think, or perhaps without re- alizing it shifts in evaluating his draft to the imagined perspective of some unspecified, generalized discerning other. On a deeper level, experiences of private shame might be explained in terms of an audience that is not merely imagined, but rather internalized and incorporated into the subject’s psyche in some more complex way. This is the route taken by ad- vocates of the Audience Thesis like Richard Wollheim (1999) who draw on the resources of psychoanalytic theory. 14 Wollheim proposes that in developmental terms shame arises when what was simply fear of losing the love of an external figure is transformed, via “introjection” of the external figure, into shame. Here the audience, rather than a figment of the imagination of a unified subject, is conceived as a dis- tinct psychic agency within a divided subject, an internal critic that operates in some sense independently and beneath the level of the subject’s conscious awareness. In this picture, the audience is buried so deeply that it falls out of the subjective experi- ence of shame, to play a behind-the-scenes explanatory role. So we should suppose 14 Influential psychoanalytic accounts of shame include Piers and Singer 1953 and Lewis 1971. Responsibility and the Emotional Structure of Relationships 86 that Sam, for instance, feels shame because he is, beneath the level of his conscious awareness, subject to criticism from an introjected figure. But Sam need not be aware of any of this; he simply feels shame. In presenting these strategies I have drawn a contrast between shallower and deeper versions in order to suggest that, in attempting to explain cases of private shame in terms of an internalized audience, defenders of the Audience Thesis face a dilemma. On the one hand, against the proposal that private shame is to be under- stood in terms of an imagined audience of one sort or another, critics of the Audience Thesis can simply hold firm and deny that this explanation fits the evi- dence. We have already seen a version of this response in O’Hear’s description of craftsmen who feel shame about their shoddy product “without imagining other craftsmen inspecting and condemning their work” (1976: 77). Kekes similarly rea- sons: “It is no use to postulate an imagined audience in whose hypothetical eyes our unobserved selves must feel shame. For we often feel shame for things that outsiders would not regard as shameful, such as not achieving our personal best when we want to or falling short of some higher than accepted standard we have set for ourselves” (1988: 284). Moreover, as Deonna and Teroni observe, imagination is ordinarily “subject to the will and open to introspection” (2011: 196). This is not to say that imagination is a mental operation—like mentally reciting the Gettysburg address or calculating the square root of ten—that we ordinarily undertake only with full awareness of what we are doing; rather, though imagination can function spontaneously and with- out our attending directly to it, it nevertheless seems ordinarily to be at least accessible to introspection and some degree of control. And this means that if cases of private shame like Sam’s are to be understood in terms of an imagined audience of one sort or another, then when we put ourselves in Sam’s shoes and imaginatively reconstruct his experience as one of shame, we must thereby imagine some audience that his (and our empathetic) shame is felt in the eyes of—an audience that we ought to be able to locate introspectively, even if we had not originally attended to its presence. But it is at least far from obvious that merely in understanding Sam’s experience as one of shame we are imagining such an audience. It seems open to critics of the Audience Thesis, like O’Hear and Kekes, to deny this. Chapter 3 | The Audience in Shame 87 On the other hand, against the proposal that private shame is to be under- stood in terms of a more elaborately and unconsciously internalized audience of the sort described by Wollheim, critics of the Audience Thesis can object, first, that this sort of explanation involves extravagant theoretical presuppositions that they can plausibly reject. The difficulty is that the sort of internal critic that Wollheim appeals to needs to be both sufficiently internalized by the subject to be firmly incorporated into her own psyche (and thus to operate at a deeper level than a merely imaginary audience) and sufficiently independent of and distinct from the subject to serve as the kind of foreign evaluative perspective that, according to the Audience Thesis, is re- quired for shame. Critics of the Audience Thesis who are not sympathetic to the basic tenets of psychoanalytic theory may reasonably be skeptical about the possibil- ity of a psychic agency within the subject that straddles the divide between internal and external in this way. 15 Second, this sort of view begins to look quite difficult to distinguish, in terms of the available evidence, from alternative views that reject the Audience Thesis and understand shame as simply a matter of the subject’s own perception of a discrep- ancy between the ideals to which she holds herself and her actual situation. The evidence from intuition and introspection that we can glean by considering a case like Sam’s does not clearly distinguish between these two sorts of account, precisely because the mechanism by which the perspective of an audience is supposedly en- gaged is, by hypothesis, operating beneath the level of the subject’s conscious awareness. And in the absence of decisive evidence, critics might argue, the simpler view that does not appeal to unconscious processes or dubious psychic agencies should be preferred. 16 15 Many thanks to Alex Sarch for clarification of this problem. For a different criti- cism of Wollheim’s account of shame, see Deonna & Teroni 2011: 199 and Deonna, Rodogno & Teroni 2012: 129. 16 The dilemma I have just described (minus the concern about the internal/external status of Wollheim’s internal critic) is nicely summed up by the psychologists Jessica Tracy and Richard Robbins: “One could argue . . . that an imagined audience is pre- sent in [cases of private shame], but it would be quite a broad claim to assume that every time someone is ashamed of his or her score on an exam, the way he or she looks, or of a cruel thought he or she had, the shame is due to the presence of an imagined audience. It seems more likely that an imagined other exists at an uncon- scious level, but when we begin to think of the observing ‘other’ as unconscious, the Responsibility and the Emotional Structure of Relationships 88 Advocates of the Audience Thesis are faced, then, with two unpromising op- tions in cases of private shame. Either they can locate the audience that they claim to be essential to shame on a relatively shallow level that is accessible to the subject’s introspection, in which case critics of the Audience Thesis can simply deny that this is borne out by the evidence. Or they can locate the audience on a deeper, uncon- scious level, in which case critics can respond both that the proposed explanation is inherently dubious and that alternative explanations, not involving an audience, han- dle the cases just as well with significantly fewer theoretical commitments. 4 Dispensing with the audience I have just suggested that critics of the Audience Thesis might explain cases of private shame directly, without reference to an audience’s distinct evaluative per- spective, in terms of the subject’s perception of a discrepancy between the ideals to which she holds herself and her actual situation. This represents an advantage that critics have over advocates of the Audience Thesis who, we have seen, struggle to account for cases of private shame. Here it may seem natural for the advocates to respond that the Audience Thesis has its own advantages, in that it better accounts for experiences of shame in which the presence of an actual audience is central, like the case of Adam and Eve. But in fact it is not so difficult to accommodate such cases without the Audience Thesis, because those who deny the Audience Thesis deny only that that exposure to an audience is necessary for shame, not that such ex- posure can be an element in some experiences of shame. It thus turns out that the challenge for those who reject the Audience Thesis lies elsewhere. In cases of private shame like Sam’s it does appear that the subject perceives a significant discrepancy between his ideals and his actual situation—in particular, between his ideals of good writing and the quality of the draft he has produced. Ac- counts of shame that reject the Audience Thesis often focus on this as the defining feature of shame experiences. John Kekes, for instance, holds that “What is essential to shame” is that we view “what we are, have, or do” as “falling short of some stan- concept becomes difficult to disentangle from activation of an ideal self- representation,” i.e., a set of ideals that could be used to explain shame directly in terms of the agent’s own self-assessment (2004: 173). Chapter 3 | The Audience in Shame 89 dard” (1988: 284); R. Jay Wallace defines shame as the feeling one has about “the lack of an excellence that one values and aspires to possess” (1994: 240); Martha Nussbaum describes shame as “a painful emotion responding to a sense of failure to attain some ideal state” (2004: 184); and Michelle Mason contends that shame “pre- sents one’s character as meriting some degree of withdrawal of esteem in virtue of one’s failure to approximate a legitimate character ideal” (2010: 418-19). 17 For con- venience I will refer to these and similar accounts as “personal-ideals” accounts of shame. 18 Personal-ideals accounts can deal smoothly with Sam’s case. But how then, a defender of the Audience Thesis might ask, are we to account for the shame of Adam and Eve? What ideals have they fallen short of? There is actually a fairly straightforward answer to this question available to personal-ideals accounts. Any appearance of difficulty is likely due to thinking that the only relevant ideals are ide- als of personal conduct, achievement, or excellence; but many of our most important personal ideals—ideals on which our identities and senses of self-worth are partially premised—are social or relational ideals that concern how we stand in relation to others. In this broader sense, a person’s personal ideals include not only the non- relational standards of behavior and performance that she holds herself to, but also her aspirations concerning how she prefers to be seen by others and the place in the world that she expects to occupy. Most of us, for example, treat it as a personal ideal 17 The psychologists Jessica Tracy and Richard Robins similarly characterize shame as occurring “when a person’s identity is threatened or elevated . . . [by an] eliciting event [that] is relevant to the aspirations and ideals . . . of the self” (2007: 11). 18 Another family of accounts in the literature are what might be called self-esteem ac- counts of shame. John Rawls, for instance, defines shame as “the feeling that one has when he experiences an injury to his self-respect or suffers a blow to his self-esteem” (1971: 442). (For a rich but critical discussion of the interaction between Rawls’s conceptions of shame and self-esteem, see Deigh 1983.) While I will not directly ad- dress self-esteem accounts here, I believe that self-esteem accounts that do not incorporate some version of the Audience Thesis confront problems that are similar to the issues discussed below that plague personal-ideals accounts. For a self-esteem account that incorporates the Audience Thesis, see Galligan 2014, in which shame is defined as “a feeling of low self-esteem, experienced in response to the subject’s per- ception that he is not thought of in the way that he intrinsically values himself for being though of by some relevant other” (2014: 65). Galligan, of course, must then account for cases private shame; he explains these cases in terms of an imagined audience (2014: 67). Responsibility and the Emotional Structure of Relationships 90 not only to be worthy of respect or love but actually to be respected or loved by those who are closest to us. With the benefit of this broad conception of personal ideals, we need only suppose that Adam and Eve (after eating of the tree) treat it as an important ideal that each not be seen naked by the other (this might be an instance of some more general ideal). Then, when each of them perceives that she is seen in the way that she desires not to be, this represents a discrepancy between her personal ideals and her actual situation, which is experienced as shame. 19 So there is little apparent difficulty in accounting for the shame experienced in cases of exposure to an unwelcome gaze. 20 And this personal-ideals approach yields an account of shame that dispenses with the Audience Thesis while covering both private cases and cases of unwelcome exposure in a unified way. The same personal-ideals strategy can be used to account for a wide range of other cases in which an audience appears to play an essential role in experiences of shame, like the following: In India, it is common for men who are close friends to hold hands in public. Indian men who immigrate to the U.S. generally give up the practice though, presumably in response to the discovery that it is looked down upon here. Chintan and Nilay refuse to assimilate in this respect, however, as they continue to see nothing wrong with the custom they grew up with. Holding hands as they walk together around their neighborhood is nonetheless an uncomfortable experi- ence—it makes them feel as though everyone is staring at them, they have trouble meeting their American neighbors’ eyes, and they find themselves looking for excuses to break the contact. Though they do not think there is anything genuinely wrong with holding hands, they feel ashamed to be seen doing it (Galligan 2014: 66 21 ). 19 For an example of this strategy for accommodating exposure shame in the context of a personal-ideals account, see Deonna, Rodogno & Teroni 2012: 129-31. 20 Ultimately, even setting aside the other difficulties confronting personal-ideals ac- counts that I develop below, in the final analysis I consider this account of Adam and Eve’s shame to be unsatisfactory for reasons that do not concern us here. The point here is just that for purposes of the present dialectic, critics of the Audience Thesis can maintain that this account of cases of exposure shame is at least arguably no more problematic than the advocates’ account of cases of private shame in terms of an internal audience. 21 Galligan borrows the template for this example from Maibom (2010: 573), who uses it to make a similar point. Chapter 3 | The Audience in Shame 91 Chintan and Nilay see nothing wrong with holding hands, and if there were no one around they would feel no shame in doing so; exposure to a disapproving audience, however, causes them to experience shame. A similar dynamic, in which someone finds herself susceptible to shame for failing to measure up to the standards of an audience that she does not endorse, tends to arise in other situations where subjects switch social contexts, as when a person from a working-class background enrolls in a prestigious university and finds herself ashamed to use working-class language around her new peers. 22 But changed social contexts are not the only circumstances in which this occurs, as the following example illustrates: Consider the case of Crito and his great concern for what the good citizens of Athens will think of him for failing to deter Socrates from meeting his demise. “I am ashamed,” he says in vainly trying to argue Socrates out of accepting his fate, “both on your account and on ours, your friends’; it will look as though we had played something like a coward’s part all through this affair of yours.” And though Crito is in the end convinced that Socrates’ course is the right one and knows all along that he has done everything one can expect of a friend, we still have, I think, no trouble picturing this good-hearted but thoroughly conventional man feeling ashamed when before some respectable Athenian, who reproaches him for what he believes was cowardice on Crito’s part (Deigh 1983: 234-35). Another type of case that is relevantly similar is the situation of abused or subordi- nated persons who suffer shame in the face of demeaning or oppressive treatment at the hands of the more powerful, even though they reject the false value judgments on which such treatment may be based. 23 In all of these cases, because the subject of shame does not share or endorse the audience’s critical perspective, exposure before the audience plays an indispensi- ble role in explaining the experience of shame, in the sense that in the absence of the exposure there would be no occasion for shame. For this reason, these sorts of cases are a standard part of the arsenal deployed by advocates of the Audience Thesis. But these cases are not difficult for personal-ideals accounts to accommodate. 24 We need 22 This example is Maibom’s (2010: 573); John Deigh provides two further examples of this kind (1983: 233, 238). 23 For discussion and specific examples see Calhoun 2004: 128, 135-38 and Maibom 2010: 572-73. 24 Again, I ultimately have doubts about this sort of explanation that are not directly relevant to the present dialectic. See note 20, above. Responsibility and the Emotional Structure of Relationships 92 only suppose that Chintan and Nilay treat standing in the favorable regard of their new neighbors as a personal ideal; that the working-class student treats fitting in with her posh new peers as a personal ideal; that Crito treats the approbation of other re- spectable Athenians as a personal ideal; and that (at least some) victims of abuse and oppression treat being respected (or, at a minimum, not being profoundly disre- spected) by others as a personal ideal. The difficulty for personal-ideals accounts lies elsewhere. Defenders of the Audience Thesis have raised one pointed worry, but there is also a more general and, I think, potentially more concerning one. The pointed worry is that without the Audience Thesis, personal-ideals accounts will need some other way to account for the difference between shame and negative self-appraisals like disappointment in oneself. 25 Contrast Sam’s case with a nearly identical one: Amy’s high opinion of herself is derived in part from her belief that she writes excellent poetry, but she never shares any of her work with anyone else because she worries that she would be unable to express herself as authentically if she were thinking about the way her poems would be received by someone else. Then Amy reads some critical essays on poetry, and she becomes convinced that her own work lacks the characteristics for which she had previously valued it—her poems are stylistically ordinary and not especially insightful (Galligan 2014: 62). Other than the fact that Amy writes poetry while Sam writes prose, Amy’s case is virtually indistinguishable from Sam’s. But unlike the authors of Sam’s case, the author of Amy’s case presents it as a situation in which the subject might not experi- ence shame: “it does not seem to follow from this description that Amy must experience shame; certainly disappointment with herself, and maybe even depression, but not necessarily shame” (Galligan 2014: 62). The problem here is not that we cannot see how Sam could experience shame in his circumstances; perhaps he might. The problem is rather that it would be just as consistent with the facts that we have been given for Sam not to experience shame, but only disappointment in himself, disapproval of himself, frustration with himself, or the like. This means that the factor that is responsible for explaining the shame that Sam experiences (but Amy does not) has been left out of the description 25 For discussion of this point, see e.g. Deigh 1983: 240-41 and Deonna & Teroni 2011: 211. Chapter 3 | The Audience in Shame 93 of Sam’s case. Here an advocate of the Audience Thesis might step in and helpfully suggest that we could explain the difference between the two cases by supposing that Sam imagines an audience that disapproves of his prose, while Amy simply assesses her poetry directly, without in addition imagining any audience. But without some resource of this kind, personal-ideals accounts face a general difficulty in explaining why private shame occurs sometimes and not others, and thus what makes for the difference between Sam and Amy. Falling short of one’s ideals, by itself, seems to leave out something crucial that is distinctive in experiences of shame. It is also worth noting that personal-ideals accounts face the converse diffi- culty in exposure cases. Just as they do not explain why someone might experience private shame rather than merely disappointment, they also do not explain why we would expect someone in a case of unwelcome exposure to feel not merely disap- pointment but shame. Personal-ideals accounts seem to suggest that disappointment would be just as natural a thing for Adam and Eve to feel as shame when they are exposed to unwelcome view, but this is wrong: to feel only disappointment in such circumstances would be quite unusual; shame is not just one possible response to unwelcome exposure, but rather the characteristic response. Even if personal-ideals accounts could be somehow amended or supple- mented so as to correctly predict when we should expect to find shame versus negative self-appraisals, 26 there would still be a more general underlying worry. The more general issue is that in accounting for cases of exposure shame in the way that they do, personal-ideals accounts treat those cases as mere instances of a much more general phenomenon; as a result, such accounts lose hold of the intuition that cases of exposure shame are somehow paradigmatic, central cases of shame. While it would be contentious to deny that someone like Sam can experience shame privately, it nonetheless seems clear that cases like Sam’s are not the most vivid or exemplary illustrations of the nature of shame. If we were trying to teach an English learner the 26 For one attempt to respond to this problem from the perspective of a personal- ideals account, see Deonna, Rodogno & Teroni 2012: 114-18. Very roughly, Deonna et al. suggest that negative self-appraisals like disappointment are generally “less se- vere” than shame (2012: 115). In my experience, at least, this does not seem to be true, and in any case it does not appear fully to address the problem; it does not, for instance, advance our understanding of why Adam and Eve feel shame rather than merely disappointment. Responsibility and the Emotional Structure of Relationships 94 meaning of “shame,” for instance, it would be confusing to offer Sam’s case as a paradigmatic example; much better would be a case of unwelcome exposure, even a variation of Sam’s case in which he realizes that his shoddy draft has been discovered and read by a discerning and disapproving colleague. Cases of this sort, in which an audience is prominent, would be much more likely to put the learner on the right track. Accounts that understand shame in terms of a perceived discrepancy between personal ideals and reality leave this seemingly significant fact unexplained. This worry is more general, and potentially deeper, than the first worry in two ways: it might continue to apply even if we could supply personal-ideals accounts with some accurate non-audience-based criterion for predicting the occurrence of shame versus negative self-appraisals like disappointment; and it would seem at least potentially to present a serious obstacle, not just to personal-ideals accounts, but to any alternative account of shame that does not incorporate something like the Audience Thesis. 27 5 The participant stance and the audience in shame The previous two sections have elaborated on the stalemate between critics and advocates of the Audience Thesis. This sets the stage for my suggestion that by reinterpreting the Audience Thesis we can transcend this impasse, retaining what is intuitive and correct in the Audience Thesis while also accounting for experiences of private shame like Sam’s. The way out of the impasse, I propose, is to conceive of shame as essentially a participant attitude; that is, as a reaction that we are susceptible to in virtue of treat- ing someone’s attitudes towards ourselves as (non-derivatively) central to our self- understandings. This way of thinking makes shame a response to appraisals of, or more generally attitudes towards, the subject of shame. In order to appreciate the significance of this proposal, it is important to be clear about what it entails for the structure of shame. Consider the case of Samantha, who like Sam has written a draft that she initially feels quite proud of, but who then immediately shows the draft to a discerning colleague for feedback. She feels shame when the colleague judges the draft to be a shoddy piece of work and disapprovingly tells Samantha “You can do 27 For instance, self-esteem accounts of shame that do not incorporate the Audience Thesis (see note 18) would face a version of this problem. Chapter 3 | The Audience in Shame 95 better.” If we are to understand Samantha’s shame as a participant reaction to her colleague’s appraisal, we need to be careful to distinguish the content of the col- league’s appraisal—Samantha can do better—from the appraisal itself that has that content—the colleague’s thinking that Samantha can do better. If Samantha were to come to agree with her colleague that she could do better, she might have various non- participant attitudes about this—disappointment, say, that she has fallen short of her ideals of good writing. But on the assumption that shame is a participant reaction to appraisals (rather than to their contents), none of these attitudes would be shame; rather, shame is an attitude that Samantha can only have about the distinct fact that (as it would appear to her) my colleague thinks that I, Samantha, can do better. Of course, Samantha’s emotional reaction to her colleague’s appraisal and her reaction to that appraisal’s content (assuming she endorses the content) often will overlap and may be difficult to distinguish in introspection; nonetheless, they are distinct. We can draw the same distinction in all of the cases of shame that we have considered. When Crito is ashamed before some respectable Athenian, the basis for that shame is that the Athenian thinks that Crito is a coward. But Crito’s shame is not, I am proposing, a direct reaction to that content, but rather a participant reaction to the appraisal itself, to the fact that (as he would see it) this Athenian thinks that I, Crito, am a coward. Of course, if Crito were to agree with the Athenian, then he would also be liable to have some overlapping attitude about the fact that I, Crito, am a coward, but again this non-participant attitude would not be shame on the assumption that shame is a response to appraisals or attitudes. Similarly, when Adam is ashamed of his nakedness before Eve, his shame is a reaction not to the content of Eve’s percep- tion—that Adam is naked, or if you prefer, simply Adam’s nakedness—but rather a reaction to Eve’s perception that I am naked (or Eve’s perception of my nakedness). In each case, the experience of shame consists in the subject’s response to the fact that he is thought of or perceived in some way—rather than to the way in which he is thought of or perceived. In a slogan, then: Shame is a reaction to attitudes, not (just) to their contents. In a moment, I will draw out some of the features that recommend this proposal and distinguish it from the other views that we have been considering. But first I want to identify two important questions that the proposal raises but that I will not discuss Responsibility and the Emotional Structure of Relationships 96 here. The first question is: If what distinguishes shame from first-order self- appraisals like disappointment is that shame is a participant reaction to attitudes to- wards the subject, then what is it that distinguishes shame from other participant reactions that a subject might have to attitudes towards her—reactions like anger, offense, or gratification? For instance, I might respond to another’s criticism with shame, but I might instead feel angry or offended, or even gratified that they have even paid attention to my performance. These other reactions, like shame, are par- ticipant reactions to the other’s appraisal, not (just) to its contents. What differentiates these reactions and explains why one, and not the others, is felt on a particular occasion? This is an important issue that a complete account of shame would have to address. 28 In my view, the answer will have to do with the wide variety of ways in which we can be invested in the different kinds of attitudes that are taken towards ourselves, and the similarly wide variety of positions that we can occupy with respect to those attitudes and the people who hold them. But for now I set this to one side; my present goal is only to develop and defend a view about the place of the audience in the structure of shame. 29 The second question concerns the contents of the attitudes that shame is a reaction to. In saying simply that shame is a participant reaction to attitudes, I place no restriction on what the contents of those attitudes might be. I regard this as a vir- tue of the proposal, and I am inclined towards the view that, at least in principle, there may be virtually no limit on the types of attitudes that it is possible to respond to with shame, provided only that the attitudes are appropriately about or concerning the subject of shame. 30 The following discussion will touch on some examples that I 28 One possible route to take here would be to combine my proposal that shame is a participant attitude with the idea that shame is connected to personal ideals (or alter- natively to self-esteem—see note 18). Then we could say that shame is a response to perceived discrepancies between one actual situation and one’s personal ideals regard- ing attitudes towards oneself. For reasons I will not go into here, this is not the approach that I favor (see also note 20), but it is available as a fix for advocates of personal- ideals (or self-esteem) accounts. 29 In another paper currently in progress I address one aspect of this question by in- vestigating the relationship between shame and anger. On this relationship, also see note 37, below. 30 This raises yet a third question that I previously set aside in chapter 2, § 4.1, and also will not address here, namely what it means for an appraisal or attitude to be Chapter 3 | The Audience in Shame 97 think tend to support this. But it is important to be clear that the main proposal (that shame is a response to attitudes rather than (just) their contents) is perfectly com- patible with views according to which shame can only be felt in response to attitudes that have some restricted sort of content—for instance, only in response to negative appraisals of the subject (not neutral or positive ones), or only appraisals that con- strue the subject as ridiculous or contemptible (not negative appraisals of other kinds), and so on. Whether there are such restrictions on the contents of the atti- tudes to which shame may be a response, and if so what the relevant restrictions might be, are important questions that I will not directly address here. And to repeat, my main proposal is neutral on these questions and compatible with a wide range of possibilities. Return now to the main line of inquiry, and to the advantages of accepting that shame is a response to attitudes, not (just) their contents. At first blush, this may seem to represent no progress with respect to the Audience Thesis, since we have been considering cases in which shame consists in a reaction to the attitudes of other people towards the subject. But there is nothing about the structure just described that dictates that the appraisal or attitude that is the object of shame must be an atti- tude of some other person. In Sam’s case we can just as well draw a distinction between the content of his self-appraisal (that Sam’s draft is a shambles) and the ap- praisal itself (Sam’s judgment that his draft is a shambles). And once we have distinguished these two things, we can distinguish Sam’s attitudes about the fact that (as he would see it) my draft is a shambles—disappointment in himself, say—from his attitudes about the fact that I, Sam, judge that my draft is a shambles. Private shame, I propose, is an atti- tude of the latter kind. (And these observations open up the possibility of identifying “private” versions of other participant attitudes, understood as participant responses to one’s own attitudes concerning oneself.) This proposal fits with the rhetoric and imagery that is often associated with shame before oneself. We say things like “I couldn’t live with myself if I did that,” or about the subject, in the way that is relevant for shame (as well as other participant attitudes). It seems clear, for instance, that the attitude need not make direct refer- ence to me in order for it to be about me in the relevant way. I could feel shame when I learn, for instance, that someone has contempt for my ethnicity or religion, even when that person has no idea that I exist. Responsibility and the Emotional Structure of Relationships 98 “I’m so ashamed I can’t look myself in the mirror.” Imagine that Sam (being a bit of a dramatic personality) says “I’m so ashamed of this draft I can’t look myself in the mirror!” Why might he say this, rather than, or perhaps in addition to, “I’m so ashamed of this draft I can’t look at it ”? The answer suggested by the current pro- posal is that these two acts—looking at the draft, and looking in the mirror— correspond to two different emotional objects. When he looks at the draft, Sam can- not avoid the thought that my draft is a shambles, which is painful and disappointing because it means that Sam has fallen short of his ideals of good writing. By contrast, when he looks in the mirror, the painful thought that Sam cannot avoid is quite dif- ferent; it is the thought that I, Sam, judge that my draft is a shambles. And it is this thought, which describes an attitude that someone (namely, Sam) has towards Sam, that is a possible basis not just for disappointment or other direct negative self- appraisal, but for the participant, second-order appraisal of shame. This interpretation of Sam’s experience thus does not require two different evaluative perspectives; what does the work is rather the difference between two dif- ferent objects of appraisal—Sam’s draft, and Sam’s own appraisal of his draft—both of which can be appraised from the same, unified perspective; namely, Sam’s own. This marks an important difference between my proposal and another treatment of shame to which mine bears a close resemblance. Sarah Buss has suggested both that “to experience shame is to experience oneself as an object of appraisal,” and that this can be extended to cases of private shame by recognizing that “insofar as one is the source of one’s ideals, one is critical of one’s behavior; and insofar as one is the object evaluated in light of these ideals, one is ashamed of this behavior” (1999: 531). For Buss, then, private shame is a matter of responding to one’s own self-directed atti- tudes; on this fundamental point our accounts agree. 31 31 Gabriele Taylor’s view of shame is similar in this respect. She says: “to feel shame about his inferior work a craftsman need not think, i.e., either believe or imagine, that there is another craftsman looking at his work. . . . All that seems necessary is that he shift his viewpoint from that of the creator of the work to that of the critical asses- sor, and he himself can fulfil both functions” (1985: 58). It would require an extended discussion to do full justice to Taylor’s sophisticated account of shame and the differences between it and my proposal. But one key distinction is that Taylor takes it to be essential to shame that the subject make a negative self-appraisal: “A person feeling shame judges herself adversely. . . . Her final judgment concerns her- Chapter 3 | The Audience in Shame 99 But in another respect my proposal marks a further departure from the tradi- tional understanding of the Audience Thesis than Buss’s view. Buss believes that the structure of private shame involves a “divided consciousness” in which the subject occupies two different “evaluative perspectives,” corresponding to the position of “the critic” who appraises and the distinct position of the object of that appraisal (1999: 531-32). So the Audience Thesis and Buss’s view both invoke two different evaluative perspectives to explain cases of private shame; the Audience Thesis posits a foreign evaluative perspective that is distinct from the subject’s own, whereas Buss suggests instead that a divided subject herself occupies both evaluative perspectives. Buss’s approach here is motivated in part by larger theoretical commitments concerning the role and significance of shame that are beyond the scope of this dis- cussion. 32 But taken solely on its own terms, Buss’s resort to a “divided consciousness” appears unnecessarily to conflate two distinctions: the distinction between two different positions—the positions of subject and object of appraisal— and the distinction between two different “evaluative perspectives.” But these two distinctions can come apart; someone can occupy the positions of both subject and object of appraisal without occupying different evaluative perspectives. A vain per- self only: she is degraded not relatively to this audience, she degraded absolutely” (1985: 68). In my view, no such negative self-appraisal is essentially involved in shame; the emotional object is not the self, but an attitude directed at the self. Michelle Mason may also be gesturing towards an understanding of private shame that is similar to Buss’s when she suggests that subjects who experience such shame “shrink from their own views,” rather than from the view of an external audi- ence. She concludes: “That someone actually distinct from himself or herself occupies the perspective in question thus need not enter into the content of the emotion” (2010: 416-17). But ultimately this observation plays no substantive role in Mason’s account of shame, which is, in my terminology, a straightforward “personal- ideals” account. According to Mason, shame is “an emotion of self-assessment” that “presents one’s character as meriting some degree of withdrawal of esteem in virtue of one’s failure to approximate a legitimate character ideal” (2010: 418-19). This leaves out a key structural feature that is present in both my account and Buss’s and that, in my view, distinguishes shame from structurally simpler forms of negative self-appraisal: shame before oneself is precisely not a first-order self-assessment (like self-disapproval or self-disappointment), but rather a reaction to one’s own self- assessments. 32 Buss assigns shame a key role in her account of how it is that we ordinarily develop a sense of respect for other persons generally; shame is, according to her view, one of the chief ways in which we come to recognize that others’ evaluative perspectives are independent sources of reasons for us (1999: 525-27). Responsibility and the Emotional Structure of Relationships 100 son, for example, can both admire his own appearance (just as he would admire the appearance of a similarly dapper stranger) and take pleasure in the fact that he is so admired—admired, that is, by himself. (Thus I take feelings of vanity, like shame, to be participant attitudes that can manifest “privately” or reflexively.) Here the vain person simultaneously occupies the position of subject and object of appraisal, but without any divided consciousness. His two relevant attitudes—roughly: I’m so dapper and It’s wonderful that someone thinks that I’m so dapper—reflect these two different posi- tions, but they are mutually compatible and issue from a single evaluative perspective. 33 According to both Buss’s view and my own, someone who feels private shame, like Sam, is similar to the vain person in both appraising himself in a certain way (thus occupying the position of subject of appraisal), and simultaneously re- sponding to that appraisal (thus occupying the position of object). But as in the case of the vain person, Sam’s relevant attitudes can reflect the two different positions that he occupies while being mutually compatible and issuing from within a single evaluative perspective. We thus need not accept Buss’s view that this structure must involve a consciousness divided across different evaluative perspectives. Now, my proposal can only make sense of cases of private shame like Sam’s by invoking the further substantive claim that I mentioned at the end of § 2, namely that we are often emotionally invested in our own appraisals of ourselves in a way that is similar to the way in which we are emotionally invested in others’ appraisals of us—that is, that we can, and often do, take the participant stance towards our- selves. 34 Sam’s assessment of his own writing cannot serve as a basis for shame unless he cares about his own assessment in something like the way in which he 33 In this, the vain person’s relevant attitudes differ from incompatible sets of atti- tudes that genuinely require a divided consciousness. For instance, the working-class student mentioned earlier, who enrolls at a private university and develops mixed feelings about her working-class manners, perhaps may come to feel both that those manners are to her credit and that they are to her discredit. This might involve mutu- ally incompatible appraisals of the same object that can only be made from different evaluative perspectives. 34 Buss’s account would seem to require an analogous claim, regarding our emotional investment in the different evaluative perspectives that we ourselves occupy in cases of private shame. Chapter 3 | The Audience in Shame 101 cares about the assessments of other discerning readers before whom he is liable to shame. As indicated in § 2, I will not argue in detail for this general claim here. But it strikes me as independently plausible. Shame before others—like all participant atti- tudes—is grounded in the kinds of relations that we have with those others and the ways in which we are emotionally invested in those relations. We are not, after all, equally liable to feel shame (or other participant attitudes) before everyone on the same occasions: a student might feel ashamed to be caught goofing off by her teacher but not by her peers; she might be ashamed to be caught studying diligently by her peers but not by her teacher. And she may be liable to shame before one teacher, or one peer, but not another. What determines these differences is the ex- pectations and other attitudes of the various parties and the different ways in which the subject is invested in those attitudes—including whether, and to what degree and in what respects, the subject takes the participant stance towards the relevant others. What makes a student liable to shame before one teacher but not another is that she is emotionally invested (in the way that is characteristic of the participant stance) in the former’s attitudes towards her but not the latter’s. Before the latter’s disapproval she might feel (impersonal) annoyance or distress, but only the former’s disapproval touches on her self-understanding in the special, penetrating way that triggers the participant response of shame. Just as we have relationships of many different kinds with other people, we have relationships of many different kinds with ourselves, relationships that in vari- ous respects resemble (and developmentally, perhaps may be modeled upon) the relationships we have with others. We love or hate ourselves, admire or ridicule our- selves, respect or disrespect ourselves, enjoy or dislike our own company, find ourselves funny or clever or tedious, and so on. We are, of course, deeply emotion- ally invested in many of our own basic attributes—in whether we are attractive or kind or admirable or ridiculous, and the like; and moreover, chapter 2 advanced the view that we are often emotionally invested in a distinctly personal way in others’ attitudes towards ourselves. Now I am further suggesting that we can be deeply emo- tionally invested in much the same way in our own attitudes towards ourselves. To repeat an example from earlier: just as we are often emotionally invested (in the way Responsibility and the Emotional Structure of Relationships 102 characteristic of the participant stance) in whether or not, and to what degree, certain others respect us, we are often emotionally invested in a similar way in whether or not, and to what degree, we respect ourselves. This is the kind of emotional invest- ment that makes us susceptible to the participant attitudes. Thus, just as shame before others is a response to an unwelcome gaze and reflects a sense that relations with others are not as the subject would wish them to be (in a way that touches on the subject’s self-understanding), shame before oneself is a response to an unwelcome self-assessment and reflects a sense that the subject’s relations with herself are not as she would wish them to be (in a way that touches on the subject’s self-understanding). If Sam, for instance, cares about his own attitudes towards his writing in this participant way, then we can make sense of the idea that Sam might react to his own disappointment in his writing in much the same way that he would react to the disappointment of a discerning other—with shame. A complete understanding of shame would include, in addition to chapter 2’s account of the distinctive kind of emotional investment that renders us liable to the participant attitudes generally, much more thorough accounts of, among other things: the various natural or typical objects of such emotional investment; the proper or appropriate or healthy scope of such investment; and the specific condi- tions in which such emotional investments render us susceptible to the various participant attitudes—shame as opposed to hurt feelings, resentment, and so on. These topics are beyond the scope of this discussion (though we will touch on the different significances of shame and resentment in the next chapter); nevertheless, I hope that my proposal opens up a promising path for making progress on them by identifying the type of thing that can be an object of shame—namely, attitudes to- wards the subject (as distinct from the contents of those attitudes)—and by clarifying the distinctive structure of shame experiences that differentiates them from first- order negative self-appraisals. Moreover, even without wading into those deeper waters, my proposal clari- fies a number of puzzles about the nature and structure of shame. We have just seen how it provides a model with which to understand private shame. If my proposal is correct, then at least some cases of private shame turn out to be simpler in structure than many defenders of the Audience Thesis have thought. While it is consistent Chapter 3 | The Audience in Shame 103 with my account that there are subjects (like Williams’s version of Sartre’s voyeur) whose shame is grounded in the gaze of an imagined audience, the account makes it unnecessary to posit any audience at all—in the sense of an evaluative perspective that is distinct from the subject’s own, whether imagined, “introjected,” or other- wise—in order to explain the shame in a case like Sam’s. Rather, the explanation of Sam’s shame can lie in the fact that Sam serves as his own audience, simply in virtue of his capacity to have self-directed attitudes and to be emotionally invested in the right deep and distinctively personal way in what those attitudes are. The key to the puzzle is that Sam’s shame is not just a self-directed attitude (like disappointment in himself, or disapproval of himself), but rather a self-directed attitude of a very par- ticular kind; namely, a participant attitude about a self-directed attitude. At the same time, in light of this proposal it is not difficult to see why the theoretical significance of the distinction between the contents of one’s own self- directed attitudes and the attitudes themselves might be overlooked, particularly in thinking about shame. Cases like Crito’s show that the subject of shame need not agree with the assessment that forms the basis for his shame. If Crito does not agree that he is a coward, then it is clear that his shame is not based on the thought that I am a coward but rather on some version of the thought that I am thought to be a coward. His shame is a reaction to the attitude of the audience, not that attitude’s content. But this distinction is more difficult to see in cases of private shame, because Sam’s shame, for example, is based on the thought that I, Sam, judge that my draft is a sham- bles, and in order to have this thought Sam ordinarily must also have the thought that my draft is a shambles. So when speaking of private shame we cannot point to any fa- miliar sort of case in which the “audience” and the subject disagree to illustrate that it is the attitude, rather than its content, that is the focus of shame. But even though the subject of private shame ordinarily accepts the content of the attitude that is the focus of shame (simply because it is her own attitude), with the benefit of the parallel to cases of public shame like Crito’s we can see that the subject’s attitude about the content, and her attitude about the attitude that has that content, are distinct. Careful attention to this distinction can also help to resolve another poten- tially perplexing issue that is illustrated by, though not unique to, Buss’s view of shame. Buss holds that feeling shame in response to some appraisal—whether one’s Responsibility and the Emotional Structure of Relationships 104 own or that of another—must involve in one way or another agreeing with or credit- ing the content of that appraisal. She says, for instance, that to feel “shame is to discover that one’s actions have a meaning which one must affirm even though it is not conditional upon one’s own point of view” (1999: 526), that “shame is an ac- knowledgment that one is, indeed, this very object—that one is as seen” (1999: 527), and that “shame is, intrinsically and essentially, the experience of an external authority— the experience of the authority of an external point of view” (1999: 533). 35 These characterizations may fit some shame experiences, but are at least doubtful when applied to cases of shame in which the subject disagrees, on the level of explicit judgment, with the audience whose attitude is the focus of shame (for in- stance, when Chintan and Nilay are ashamed to be seen holding hands, or when Crito is ashamed to be thought of as a coward). And they founder completely when applied to a case of unwelcome exposure like that of Adam and Eve. The reason for this is that Eve’s attitude, after eating of the tree, towards Adam’s nakedness (like Adam’s attitude towards Eve’s) need not be critical—it might indeed be admiring— but Adam could still very well respond to Eve’s perception of his nakedness with shame. If we are to suppose that Adam’s experience of shame involves an experience of Eve’s “external authority,” or an experience that he “is as seen” by Eve, then in this scenario his shame would involve in some way experiencing his naked body as, or accepting on Eve’s authority that his nakedness is, pleasing to the eye. But it is not obvious how an experience of this sort would be a basis for shame. Another example that illustrates this problem concerns an artist and his nude model. The model understands her relationship with the artist to be purely profes- sional and is perfectly comfortable with her role, until one day in the middle of a sitting she realizes that the artist no longer sees her solely as a subject to be painted, but has come to see her in a sexual light. With this realization, the model suddenly and for the first time has the sense that she is somehow compromised in her nudity, and feels a rush of shame. 36 The setup of this example, and the values that inform it, 35 These claims are connected to Buss’s view that shame has an important role in al- lowing us to recognize that others’ evaluative perspectives are independent sources of reasons for us. On this point, see also note 32 above. 36 This case comes originally from Max Scheler’s “Über Scham und Schamgefühle,” in his Schriften aus dem Nachlass (1957). It was, as far as I know, introduced into the Chapter 3 | The Audience in Shame 105 perhaps strike us as a bit Victorian; nonetheless, if we assume that the model herself has a somewhat Victorian sensibility and make allowances for this, it is easy enough to understand that she might experience shame. 37 As with Adam and Eve, the model’s situation is one in which her shame con- sists in feeling exposed to an unwelcome gaze. But this example illustrates more clearly than the example of Adam and Eve that the basis for shame is not simply nu- dity, or the subject’s discomfort with her own naked body, but rather more specifically the attitude of the audience—after all, the only difference between the original situation, in which the model is at ease, and the later situation, in which she is ashamed, is the artist’s attitude towards her. And here again, we can easily imagine that the model understands the artist’s attitude to be entirely approving. If the model’s experience of shame is supposed to involve an experience of the artist’s “ex- ternal authority,” or an experience that she “is as seen” by the artist, then the model’s shame will involve in some way seeing herself as, or accepting on the artist’s author- ity that she is, sexually attractive. But, again, it is not obvious how an experience of this sort could account for the model’s feeling of shame. There is a tendency in theoretical discussions of shame to treat as paradig- matic those cases in which the audience makes an adverse assessment of the subject; seen in this light, the model’s experience of shame can seem atypical. But in fact shame in the face of neutral or approving attitudes seems to be possible, and perhaps even common—particularly, as we have already seen, in cases of unwelcome expo- sure, both visual and otherwise. Nakedness, sexual behavior, and bodily functions of various kinds provide particularly vivid examples of this dynamic, but it can also be found in a wider variety of cases in which the zone of privacy that we consider im- portant to maintain is intruded upon, exposing even entirely admirable and widely admired activities (artistic, sentimental, intimate, etc.) to unwelcome publicity. To Anglophone literature by Gabriele Taylor (1985: 61), and is also discussed by Wil- liams (1993: 220-21) and Wollheim (1999: 159-60). I am relying on Taylor’s recounting of the case. 37 Contemporary readers may be inclined to say that if the artist’s leering makes the model uncomfortable, then she has grounds to be annoyed or offended rather than ashamed—the fault being the artist’s rather than hers. This difference in sensibility is, in my view, intelligible in terms of the relationship between shame and anger, which I explore in another paper currently in progress. Responsibility and the Emotional Structure of Relationships 106 borrow from Thomas Nagel: “The boundary between what we reveal and what we do not, and some control over that boundary, are among the most important attrib- utes of our humanity”; for this reason, it is often the case that we “don’t want to expose ourselves completely to strangers”—or even, we might add, to friends and intimates—“even if we don’t fear their disapproval, hostility, or disgust” (1998: 4). Another class of cases in which shame may be felt in the face of approving attitudes involves what we might call the wrong kind of approval. As Williams re- marks, “the view taken by the observer need not itself be critical: people can be ashamed of being admired by the wrong audience in the wrong way” (1993: 82). 38 This can occur, for instance, for subjects who have experienced significant changes of heart. Consider someone who had once been an open and prominent white su- premacist or homophobe, but then came to see the error of her ways and to feel deep remorse for her past conduct. When she hears admirers who share her former hateful views (and who may not realize that she has had a change of mind) praising her contributions to the cause, she might well be ashamed before their admiration. 39 38 Taylor agrees: “[I]n feeling shame the actor thinks of himself as having become an object of detached observation, and at the core to feel shame is to feel distress at be- ing seen at all. How he is seen, whether he thinks of the audience as critical, approving, indifferent, cynical or naïve is a distinguishable step and accounts for the different cases of shame” (1985: 60). Compare also David Velleman’s observation: “We keep some things private not because we fear disapproval of them but rather because we fear approval of a sort that we would experience as vulgar or cheap. Even if we think that others would admire our poetry, for example, we may not like the idea of exposing it to their un- discerning admiration” (2001: 43). This remark combines the two circumstances that I am discussing separately: unwelcome exposure of admirable qualities or activities, and admiration of the wrong kind. Consistent with his general theory of shame, Velleman also thinks that we can experience shame when another’s positive attitudes towards us interfere with our efforts to control how we present ourselves to them, as when another person is overwhelmed by our beauty (2001: 45) or lavishes praise on us (2001: 49) and thereby foregrounds an aspect of our person or situation other than the ones that we wish them to focus on. For reasons I will not go into here, I am uncertain whether these should be interpreted as cases of shame; but if they are then they would be fur- ther examples of shame felt in the face of positive or approving attitudes. 39 Note that this class of cases involves a confounding factor that is absent from the unwelcome exposure cases (including the model case), namely that in these cases the subject of shame sees the quality or circumstance that is the ground for others’ admi- ration as a reason for negative self-evaluation (independently of others’ admiration). Chapter 3 | The Audience in Shame 107 All of these cases, in which the attitude that is the basis for shame is not negative or critical of the subject, illustrate the error when thinking about shame of conflating two different emotional objects: the content of an attitude, and the atti- tude itself. The model’s shame, for instance, cannot be made sense of as a reaction to the content of the artist’s attitude—to the thought that (as the model would see it) I am sexually attractive; after all the same thought, had in connection with her lover or simply as a direct self-assessment would tend to arouse no shame at all. It is the fact that she is sexually attractive to this observer on this occasion that is experienced as unwelcome and that engenders her shame. To make sense of her shame we must then understand it as a reaction to the perception that I am seen by the artist as sexually attractive; it is this that bothers the model and that she is reacting negatively to, when she experiences shame. Contrary to Buss’s characterization, it has nothing to do with the “authority” of the artist with respect to her sexual attractiveness, or with whether the model somehow affirms or acknowledges the content of his attitude; it has to do, rather, with her discomfort at being seen in this way on this occasion. 40 To be so seen does not sit well with her, in a way that reflects the kind of emotional invest- ment that is characteristic of the participant stance. And something similar could be said about the other cases that we have described in which shame is felt in the face of neutral or approving attitudes. Of course, when we feel shame the attitudes that we are reacting to often are negative or critical in content, and sometimes we agree with the contents of those attitudes. This agreement can form the basis for other negative attitudes— disappointment in ourselves, disapproval of ourselves, anger with ourselves—that This means that the former bigot might also feel shame before her own disapproval, in addition to the shame she feels before her bigoted admirers’ admiration. 40 Another way to put the point is that we need to locate the artist’s authority in the right place. In order to be liable to shame before the artist, the model need not treat the artist as an authority with respect to whether or not she is sexually attractive. But she does need to be emotionally invested in the right way in the artist’s attitudes to- wards her—for instance, in whether he sees her in a professional or a sexual light. It is precisely this kind of emotional investment that characterizes the participant stance; and in this sense, she has invested him with a kind of authority, in that his attitudes have the power to make her feel shame. But this “authority” concerns the effect of his attitudes on her, not the validity of his attitudes as regards their con- tents. Responsibility and the Emotional Structure of Relationships 108 are painful and unpleasant, and this cocktail of concomitant emotions very often ex- acerbates the unpleasantness of shame experiences. But if the foregoing analysis is correct, then it is important not to conflate these common concomitants with shame itself; they are not the essence of shame. In particular, and unlike shame, these first- order self-assessments are not participant attitudes. Conclusion The problem of the audience in shame is that the Audience Thesis—that shame essentially involves exposure to an audience—seems both to be inconsistent with certain familiar experiences of private shame, yet also to capture something im- portant and essential about shame. Rejecting the Audience Thesis entirely and understanding shame in terms of perceived discrepancies between our personal ide- als and our actual situation provides a way to accommodate cases of private shame; but this approach both leaves out of account the special character of shame that dis- tinguishes it from (non-participant) self-directed attitudes like disappointment, and more generally leaves the close connection between shame and exposure unex- plained. An elegant solution to this problem, I have suggested, lies in conceiving of shame as a participant attitude—and thus as a response to attitudes towards the sub- ject—and attending carefully to the distinction between the contents of the attitudes that shame is a reaction to, and the attitudes themselves. Keeping this distinction in mind allows us both to capture the defining feature of shame that distinguishes it from first-order self assessments like disappointment, and to accommodate cases of private shame by recognizing that we sometimes take the participant attitude towards ourselves, and can thus serve as our own audience in shame. This also explains why exposure cases strike us as particularly paradigmatic instances of shame. It is because in exposure cases the shame and the attitude that is the object of shame are separated across different subjects, thus clearly exhibiting the distinctive participant structure of shame. Shame has also served here as a case study for purposes of better under- standing the significance and explanatory power of chapter 2’s account of the participant attitudes and of the way in which they structure our relationships— Chapter 3 | The Audience in Shame 109 including our relationships with ourselves. And we are now in a position to return to the theme with which we started in chapter 1, and to make some further progress in understanding the interpersonal emotional economy of responsibility. 110 Chapter 4 Blame, Shame, and the Emotional Economy of Responsibility Blame is an important part of our social and moral lives, and it is connected in theoretically significant ways to fundamental normative phenomena like responsi- bility and obligation. Yet, despite a surfeit of recent work on blame, it has proved resistant to analysis. 1 The difficulty is often described as the challenge of accounting for the force or sting of blame—the distinctive manner in which blame seems to be something more than mere disapproval yet something less than overt reproach or sanction. A variety of theories have been advanced that account for the force of blame in quite different ways. I think that blame and its force are complex, and that we should not be sur- prised if no simple analysis can encompass all of the relevant data. But in addition to this, and despite the cumulative richness of existing accounts, there is a dimension of blame’s force that has not yet been adequately reckoned with, largely due to a ten- dency (natural among theorists) to intellectualize and de-personalize blame. The special force of many paradigmatic instances of blame is, in my view, precisely a function of blame’s participant character, and in particular of the emotional and per- sonal dimensions that this entails. My aim in this chapter is to describe and better understand these dimensions of blame. I will also propose that doing so puts us in a better position to understand the role that blame plays in the interpersonal emotional 1 Cf. R. Jay Wallace: “Blame is one of those phenomena that is both familiar and phi- losophically puzzling. . . . A satisfactory analysis is elusive” (2011: 348); Michael McKenna: “Despite the pervasiveness of the phenomenon in ordinary life, blame is an elusive notion. It is maddeningly hard to nail down a theory that gets the exten- sion even close to right” (2013: 119); T.M. Scanlon: “Although experiences of blame and being blamed are familiar parts of everyday life, it is not entirely clear what blame involves” (2013: 84). Chapter 4 | The Emotional Economy of Responsibility 111 economy of responsibility, including both the interplay between holding responsible and taking responsibility, and the little-noticed way in which the participant reaction of blame can call for a matched participant response—namely, shame. 1 Taking responsibility and the participant attitudes Before delving into blame, it will be helpful to review how we have arrived at this point. We began, in chapter 1, with an investigation into what it means to take responsibility. According to the argument of that chapter, when we engage in con- duct that tends to give others grounds to doubt our eligibility to participate in meaningful, reciprocal interpersonal relations, this can threaten our self- understandings as persons who are trusted, admired, loved, and so on. A direct and potentially effective way to address this kind of threat to our self-conceptions as agents is to show that we take our misconduct seriously, where this includes commit- ting ourselves to vindicating our good will and general social eligibility by doing whatever is warranted to make amends and put ourselves back in the good graces of those whom our misconduct has affected. This, I suggested, is just what it is to take responsibility for our conduct in relation to another person. Taking responsibility thus serves as a way to reassure others (and ourselves) of our fitness to participate in meaningful relationships and shared activities, by demonstrating to them that we value our relationships in the right way, that we take our infractions seriously, that we are unwilling to allow others to be disadvantaged through their reliance on our fitness to participate in relationships and activities with them, and that we are resolved to do better in the future. This significance unites the complex matrix of attitudes, dispositions, and responses that are characteristic of tak- ing responsibility, including feeling sorry or contrite towards those one has affected, feeling remorseful or (as will be emphasized below) ashamed about one’s conduct, being disposed to apologize and make various gestures of penitence and conciliation, and being willing to repair the harm done. 2 2 While all of these elements can play a part in taking responsibility, as noted in chap- ter 1 (note 29) I do not aim to specify a set of necessary and sufficient conditions for taking responsibility, and I am not sure that it would be useful to try. Taking respon- sibility is something that seems to come in different kinds and degrees, and moreover different elements can be involved in different ways on different occa- Responsibility and the Emotional Structure of Relationships 112 This picture shows taking responsibility to have a quite different character and function from holding responsible, including holding oneself responsible. Hold- ing oneself responsible is similar to holding another person responsible; it consists in attitudes and responses that you take towards yourself as wrongdoer: blame, reproach, instruction, sanction, and so on. Taking responsibility, on the other hand, is a re- sponse to threats of a certain kind to your social self-conception, and consistent with this social orientation it consists of attitudes and responses you take towards the other as a victim of your conduct and as a party to the relationship. Taking responsibility is thus complementary to but distinct from holding responsible—it is, in this sense, the flip side of blame—and it forms an important but theoretically neglected dimension of our responsibility practices. According to the argument of chapter 1, then, our reasons to take responsi- bility are grounded in the way in which our self-understandings incorporate and depend upon the attitudes of others towards us. A full understanding of taking re- sponsibility (and thus responsibility in general) accordingly requires a better and more general grasp of this aspect of our self-understandings. To this end, chapter 2 turned to the connection between our self-understandings and our interpersonal rela- tions, and in particular to the ways in which our relationships are constituted, structured, and regulated by our emotional sensitivities and reactions to the attitudes that others take towards us. In our relations with other people, I suggested, we often adopt what I called (borrowing and developing ideas from Strawson) the participant stance, which entails susceptibility to the participant attitudes, a class of affective re- sponses that reflect the distinctive, personal way in which we care about others’ attitudes. Members of this class of attitudes include shame, pride, love, hurt feelings, anger, gratitude, and so on. Shame in particular both is a core participant attitude, in my sense, and often features prominently when agents take responsibility for their conduct. Chapter 3 sions. Feeling ashamed, for instance, is not (at least ordinarily) alone sufficient for taking responsibility; but in the right circumstances it can serve as part of a jointly sufficient set of responses by manifesting the personal way in which the agent takes seriously the significance of what she has done. There may also be cases in which shame is necessary for fully taking responsibility, because other responses without shame would seem lacking in some important respect—in a word, shameless. I will try to say more to clarify the role of shame in taking responsibility in § 5 below. Chapter 4 | The Emotional Economy of Responsibility 113 developed a conception of shame as a participant attitude, and used this conception to develop a resolution to a longstanding disagreement about the role of an audience in the structure of shame. According to the resulting account, we should think of shame as a certain kind of negative response to attitudes towards the subject of shame. The foregoing chapters, then, to the extent that they have been concerned with responsibility, have focused primarily on the position and perspective of the responsible agent, rather than the perspective of the party who holds the agent re- sponsible. But with these resources in hand, we are now in a position to advance our understanding of the other side of the transaction—the blaming side. Blame is, I will propose, itself a participant attitude that is part of an emotional economy of respon- sibility—and of relationships more broadly—in which shame, among other attitudes, also plays an important role. By understanding how blame functions in this emo- tional economy, we can better understand the character, and special force, of (at least many paradigmatic instances of) blame. 2 The problem of the force of blame We ought first to get clearer about what is meant by the “force” or “sting” of blame, and what it is about the force of blame that has made a satisfactory theoretical account of blame elusive. When we blame someone (at least in one familiar sense of “blame”) we do more than simply make an adverse assessment of her conduct 3 ; but we can also blame her without punishing her or otherwise treating her adversely, and without reproaching her or even letting her know that we blame her—we can blame without there being any overt expression or manifestation at all. Blame is thus something more than adverse assessment yet something less than adverse treatment. This ob- servation has become a commonplace in theoretical discussions of blame. 4 Something similar, by the way, goes for gratitude: When we are grateful to someone we do more than simply make a favorable assessment of her conduct, but we can be 3 To take just one example, we can judge a young child’s conduct to have been selfish or even malicious without blaming her—as we might say, “She is only a child.” 4 See, e.g., Wallace 2011: 348; Coates & Tognazzini 2012: 202; Scanlon 2013: 85-86; Smith 2013: 27-28. Responsibility and the Emotional Structure of Relationships 114 grateful without actually rewarding her or otherwise treating her favorably, and with- out praising her, thanking her, or even letting her know that we are grateful—we can be grateful without there being any overt expression or manifestation at all. Grati- tude is thus something more than favorable assessment yet something less than favorable treatment. But blame is also somehow connected or closely related to adverse assess- ment, and it finds natural expression in reproach and punishment; gratitude is somehow connected or closely related to favorable assessment, and it finds natural expression in thanks, praise, and reward. The challenge in understanding blame or gratitude, then, is to locate and elucidate something that lies between these two ex- tremes and is distinct from each, yet nonetheless is connected in the right way to both. As already indicated, in the case of blame (which has been the subject of more theoretical attention than gratitude), this is often framed as the challenge of account- ing for blame’s distinctive “force” or “sting.” 5 Different theoretical accounts of blame have different resources for answer- ing this challenge. 6 One family of theories, which we can label cognitive, hold that blame consists in adverse assessments of a particular kind, suggesting that the force of blame is to be explained in terms of some special feature of the relevant assess- ments (for instance, the object or dimension of assessment). 7 Other theories approach the problem from the opposite direction. Conative theories hold that blame consists, at least in part, in certain motivational states—desires, intentions, disposi- tions, etc.—suggesting that the force of blame is to be explained in terms of its 5 For the “force” characterization see, e.g., Hieronymi 2004; for the “sting” charac- terization see Pickard 2013. Another way to describe this is to say that it is the challenge of accounting for the way in which blame “go[es] beyond mere descrip- tion” (Hurley & Macnamara 2010: 373) or “goes beyond simply judging [the agent] to be blameworthy” (Smith 2013: 33). As we will see, the “goes beyond” locution may be misleading, insofar as it seems possible irrationally to blame even those who are not, in our settled judgment, blameworthy. 6 The following informal taxonomy of theories of blame is not meant to prejudge any substantive issues, but merely to suggest the lay of the land. Compare the various slightly different taxonomies in, e.g., Coates & Tognazzini 2012, Pickard 2013, Smith 2013, and Tognazzini & Coates 2016. 7 See, e.g., Feinberg 1970; Glover 1970; Scanlon 1998; Hieronymi 2004. Chapter 4 | The Emotional Economy of Responsibility 115 connection to action. 8 Communicative theories hold that blame consists in some kind of either overt or (more commonly) potential or incipient complaint, protest, or other form of address, suggesting that the force of blame is to be explained in terms of its connection to expression. 9 Each of the foregoing approaches faces its own characteristic difficulties. Cognitive theories can seem too thin; they need to explain how merely descriptive assessments could carry the distinctive force of blame. In addition to this, cognitive theories face a challenge in accounting for the close connection between blame and its overt expressions or manifestations, as well as the way in which we sometimes seem to blame as it were irrationally, contrary to our considered assessments. Cona- tive and communicative theories, on the other hand, can seem too thick; they need to explain the force that even purely interior, private episodes of blame seem to carry. In addition, conative and communicative theories lie open to the charge that the force that they identify is not really the distinctive force of blame, but rather the force of the forms of action or expression that they associate with blame. It is open to doubt whether this identification of one sort of force with another is a licit substi- tution. Occupying a potentially favorable position with respect to these problems are emotional theories of blame, which hold that blame’s distinctive force is a matter of the hot or agitated affect that is characteristic of blaming reactions. 10 Emotions, at least arguably, have a different character than (and thus in a sense “go beyond”) de- scriptive assessments, while also remaining fundamentally internal and thus distinct from action or expression. At the same time, emotions are associated in a familiar way with characteristic dispositions towards action and expression. But emotional theories face a difficulty of their own: They owe an explanation of what it is that makes this affective element—dismissed by Pamela Hieronymi as “a certain unpleas- ant emotional disturbance” (2004: 121)—the right kind of thing to carry the special depth and seriousness of blame. Blame seems to involve more than just a sort of 8 See, e.g., Sher 2006; Scanlon 2008, 2013, 2015. 9 See, e.g., McKenna 2012, 2013; Talbert 2012; Smith 2013. The idea that the reactive attitudes involved in blame are “incipient forms of communication” was originally suggested by Gary Watson (1987: 230). 10 See, e.g., Hurley & Macnamara 2010; Wallace 1994, 2011. Responsibility and the Emotional Structure of Relationships 116 mental indigestion that tends to be associated with judgments of disapproval or blameworthiness. Hieronymi thus objects that the “affect, itself, seems insufficiently robust” to carry the force of blame (2004: 121). Many things have been said by advocates of the various theories to address these difficulties, and this discussion has been productive. The emerging result is a multifaceted understanding of blame and its significance, in light of which it seems that talk of “the” force of blame can be misleading. The force of blame is, plausibly, a complex phenomenon, and it is not clear that we should expect any single, mo- nadic explanation to be complete. 11 Moreover, many of the points that have been offered on behalf of the various theories of blame are complementary rather than rivalrous; after all, everyone can accept that each of the basic elements that the com- peting theories traffic in—assessments, intentions, demands, emotions, etc.—are involved in significant ways in at least some instances of blame, even if there is dis- agreement about which elements are essential to blame as such. Still, despite this progress, it seems to me that there remains a significant gap in our understanding of the complex force of blame. In my view, existing accounts all in one way or another fail to do justice to the emotional, personal character of many paradigmatic instances of blame. The next section will attempt to locate and explore the nature of this gap; the following section will then suggest how we might close the gap by conceiving of blame as a participant attitude. (The same general ap- proach would serve, I think, to close a similar gap in our understanding of the “force,” or what we might call the special warmth, of gratitude, as well as a range of related phenomena, though I will focus here on blame.) 11 This possibility has arguably not been taken seriously enough in theoretical writ- ings on blame. It is mentioned only to be set to one side by R. Jay Wallace (2011: 370 n.1) and Angela Smith (2013: 28 n.3). Wallace and Smith frame the issue in terms of whether there is a “single” or “well-defined” phenomenon of blame that is “suscep- tible to philosophical analysis.” Perhaps there is no such phenomenon; but Wallace and Smith may overlook another possibility, which is that there are phenomena in this vicinity which are susceptible to fruitful philosophical inquiry despite being nei- ther single nor, perhaps, well-defined. Chapter 4 | The Emotional Economy of Responsibility 117 3 Three accounts of the force of blame Here I aim to illustrate the nature of the gap in existing accounts of blame by examining and comparing what I regard as three of the most compelling and well worked-out theories: those of T.M. Scanlon (2008, 2013, 2015), Pamela Hieronymi (2004), and R. Jay Wallace (1994, 2011). I’ll proceed by pointing out a gap in Scan- lon’s view (§ 3.2), and then exploring how Hieronymi (§ 3.3) and Wallace (§ 3.4) each, in a different way, makes some progress in closing the gap. I will ultimately ar- gue that each succeeds only partially, because they each address themselves to only one aspect of what is at least a twofold problem. But before turning to the three ac- counts, it will be useful to have some background on earlier attempts to understand blame in cognitive terms—that is, in terms of purely descriptive judgments or as- sessments (§ 3.1). 3.1 Prelude: blame as assessment We observed in § 2 that blame is somehow connected or closely related to adverse assessment; ordinarily at least, when we blame someone we also make an adverse assessment. This observation, as well as the fact that we can blame without overtly expressing or manifesting our blame (together with considerations of theo- retical parsimony), naturally suggest that blame might be understood as nothing more than an adverse assessment or collection of such assessments. Taking this ap- proach results in what I am calling a cognitive theory of blame. But we also made reference in § 2 to the seeming platitude that blame is more than an adverse assessment of someone’s conduct; it is, it seems, one thing to assess or to judge that someone has acted it a way that would warrant blame, and a differ- ent thing actually to blame her—or as we sometimes say, actually to hold it against her. This puts pressure on cognitive accounts to explain how it is that, contrary to initial appearances, blame turns out just to be an adverse assessment (or collection of such assessments). One way to try to meet this burden is to argue that the seeming plausibility of the idea that blame is something more than an adverse assessment is due to a ten- dency to focus too narrowly on assessments of the quality of conduct as such, ignoring other possible objects of assessment. The thought here is that blame goes Responsibility and the Emotional Structure of Relationships 118 beyond an assessment of conduct as bad, wrongful, or otherwise objectionable, but not because it involves anything other than an assessment; it does this, rather, by comprising an assessment of a different object—for instance, the agent herself, or perhaps her motive, her character, the quality of her deliberations, the exercise of her rational agency, or her moral worth. Why should assessments of the agent (or her motive, character, etc.) be thought to carry a different sort of force than assessments of her conduct? Here a metaphor is often invoked of something like debits and credits in a ledger. Thus Jonathan Glover suggests that “in our present practice of blame is a kind of moral accounting, where a person’s actions are recorded in an informal balance sheet, with the object of assessing his moral worth” (1970: 64); and Michael Zimmerman ob- serves that blame can be understood as “judging that there is a ‘discredit’ or ‘debit’ in [the actor’s] ‘ledger,’ a ‘negative mark’ in his ‘report-card,’ or a ‘blemish’ or ‘stain’ on his ‘record’” (1988: 38). 12 The metaphor of a ledger or account in which character assessments are re- corded takes us beyond isolated assessments of conduct on particular occasions. But we still need some further grasp of what significance this ledger has or of what func- tion it serves in practice in order for the metaphor to yield an advance in our understanding of the force of blame. On this point, many subsequent writers have agreed with Gary Watson’s criticism that descriptive assessments or judgments alone “leave out something integral to the practice of holding responsible and to the con- cept of moral responsibility . . . . It is as though in blaming we were mainly moral clerks, recording moral faults, for whatever purposes (the Last Assizes?)” (1987: 226- 27). The difficulty for ledger views is that blame seems to function as more than just a way of keeping score (even in a particularly hallowed or official scorebook). Indi- vidual episodes of blame, moreover, have an interpersonal significance that goes beyond their contribution to a notional dossier on the agent’s conduct, and this in- terpersonal significance seems connected to the force of blame. For this reason, the idea of an account or ledger—that is, of merely assembling a record of an agent’s 12 See also Joel Feinberg’s discussion at 1970: 124-26. He says, for instance: “To de- feat the charge of being to blame by presenting a relevant strong excuse is to demonstrate that an action’s faultiness is not properly ‘registrable’ on one of the agent’s records, not chargeable to ‘his account’” (1970: 125). Chapter 4 | The Emotional Economy of Responsibility 119 character-revealing conduct—seems inadequate to capture the significance of blame as it actually functions in our interpersonal interactions. The limitations of the “ledger” metaphor are further illuminated by Watson’s (1996) distinction between responsibility in the attributability sense and responsibility in the accountability sense. It is, it seems, one thing to attribute conduct to someone, even in the distinctively deep way that connects it to her character (or, if you prefer, to the exercise of her rational agency, etc.), and a quite different thing to hold her ac- countable for that conduct. To blame someone in the manner described by the metaphor of recording a debit in a ledger does not seem to go beyond attribution; whatever force such an attribution may have (and this may, as Watson suggests, be considerable), it still lacks the depth and seriousness that is characteristic of holding her responsible in the accountability sense. And the fullest force of blame appears to reside on the accountability side of this distinction. The influence of Watson’s criticisms can be heard, for instance, in Scanlon’s treatment of the view that blame consists in assessments that take an agent’s action “as showing something negative about that person’s character” (2008: 126). This view, Scanlon concludes, “does not explain the distinctive weight that moral blame seems to have. Unless we say more about why we are interested in this kind of char- acter assessment, it may seem to be a pointless assignment of moral ‘grades’” (2008: 127). This dissatisfaction with the ledger metaphor is now quite widely shared. To cite two further illustrative examples: Angela Smith quotes Scanlon’s remark with approval, and observes: “The notion that in blaming we may be dispassionately evaluating a person’s moral record simply does not do justice to the emotional sig- nificance of blame and to the important role that this attitude plays in structuring our moral relations with one another” (2013: 31). And Christopher Bennett similarly comments: “Blame involves us in a more intimate and charged relationship with a wrongdoer than does ordinary grading or moral criticism” (2013: 66). In these remarks we can separate out two connected but distinct grounds for dissatisfaction with the idea that blame is fundamentally a form of grading or record- keeping. The first is that this leaves out an essential affective or emotional dimension of blame (hence Smith’s “dispassionately evaluating” and “emotional significance,” and Bennett’s “charged relationship”). In response to this point, defenders of cogni- Responsibility and the Emotional Structure of Relationships 120 tive accounts can only hold firm and deny that blame has an essential affective or emotional element; they cannot satisfy the demand for an emotional element without abandoning the purely cognitive character that is distinctive of their theories. We will return to this issue later and explore the feasibility of holding firm on this point. The second ground for dissatisfaction with the ledger metaphor is that, taken on its own, it leaves out an essential personal dimension of blame. Grading or record- keeping seems like a more detached and disinterested activity than blaming, which involves some kind of personal involvement; this seems to be the concern that lies behind Bennett’s observation that blame “involves us in a more intimate . . . relation- ship,” as well as Smith’s suggestion that blame plays a role in “structuring our moral relations with one another.” On this second point, defenders of cognitive accounts have had more to say. Scanlon, in What We Owe to Each Other (1998), in effect develops a response to the concern that cognitive accounts fail to do justice to the personal dimension of blame; he does this by going beyond the metaphor of an account or ledger to say more about the nature of our interest in descriptive character assessments. According to Scanlon’s proposal in What We Owe, 13 the “special force” (1998: 269) of moral blame (or, as he there interchangeably calls it, “moral appraisal” or “moral criticism” (1998: 267-68)) derives from the special significance of the kind of agential failure with which such blame is concerned. According to Scanlon’s account, moral blame is justified when “the agent’s mode of self-governance has ignored or flouted requirements flowing from another person’s standing as someone to whom justification is owed” (1998: 271). This kind of failure is important because it changes the ways in which it is possible for the agent to relate to others: “the reasons one has failed to respond to are grounded not just in some value that others also recognize but in their own value as rational crea- tures. These violations therefore have particular importance for one’s relations with them” (1998: 271-72). Moral blame, in other words, is essentially a response to facts about the agent (in particular, the agent’s regard for the blaming party) that affect the 13 I emphasize that this account appears in What We Owe to Each Other because in more recent work Scanlon has developed an account of blame that supplements his earlier view in important ways. See the next section (§ 3.2) for discussion of these developments. Chapter 4 | The Emotional Economy of Responsibility 121 kinds of relationship in which it is possible for the blaming party and the agent to stand. A significant strength of this approach, for our purposes, is that it places the kind of character assessment that is (by hypothesis) at the heart of blame in a wider context, and describes a vital function that blame plays in that context. Blame, on this account, is not merely keeping score or moralistic grading for its own sake; blame consists, rather, in assessments that determine the quality and type of relation- ship that it is possible for us to stand in to others. When we blame someone, we are not making a mark in their ledgers whose relevance is limited to bragging rights or the Last Assizes; we are taking note of the way in which another person’s conduct, and what that conduct reflects about their character, affects the relationships in which they stand to us—that they are, for instance, not the friends that we thought, or more generally that they do not have the sort of regard for us that is a prerequisite for the kinds of relationships that we thought we had (or desired to have) with them. This way of thinking brings the metaphor of the ledger to life, and identifies an important sense in which blame is more personal than detached or disinterested assessment. Instead, we stand in a wide variety of human relationships whose terms and existence conditions make reference to the attitudes of the participants— friendship, for instance, ceases to exist without a minimum level of mutual regard between the friends. The ledgers in question are thus an indispensible tool for under- standing, navigating, and sustaining these relationships. In a sense, there would be no intimate interpersonal relationships, at least as we ordinarily think of them, without such ledgers. By locating blame within relationships and connecting it to facts that are significant for those relationships, this proposal recovers a personal dimension of blame that was lacking in less developed ledger accounts. But there nevertheless remains something unsatisfying about this picture (which, as we will see in the next section, seems to have led Scanlon himself in later work to supplement it). According to the account in What We Owe, the significance of blame is derived from the significance of the facts that blame is a response to. Scanlon is quite explicit on this point: The significance of [the agent’s] moral faults is shown in part in the ways in which others have reason to respond to them: with expres- sions of criticism and hurt feelings, withdrawal of friendship, and so Responsibility and the Emotional Structure of Relationships 122 on. But the primary significance of moral criticism lies not in what others may do in the future as a result of believing it, but rather in what is, if the criticism is cor- rect, already the case. If I have injured someone by failing to take their interests into account in the way one should, then my relation with them is already altered by that fact, whatever they do (1998: 272, em- phasis added). 14 What this means is that though Scanlon’s proposal gives some needed content to the ledger metaphor, and identifies an important personal dimension of blame, it also appears to drain blame of independent force. Blame registers or acknowledges facts that are themselves significant for the agent and her relationships—but blame adds little if any further force to the fact of fault itself. In this way, the proposal actually leaves the force of blame out of the picture. It explains an aspect of the personal nature of our interest in the kind of record- keeping that (according to cognitive accounts) blame consists in, but it fails to cap- ture a critical fact about the force of blame, which is that whether we are blamed is often something that has a significance that is over and above the significance of the misconduct that the blame is a reaction to. Or as we might also put the point: the distinctive force that we seek to account for is the force of blame—a force that blame itself possesses. So far, this remains unexplained. Cognitive theories are not without further options for addressing this diffi- culty, as we will see when we turn to Pamela Hieronymi’s account in § 3.3. But first it will be useful to consider Scanlon’s own eventual response to the problem, which has been to supplement the What We Owe account with additional, non-cognitive elements. 3.2 Blame as relationship modification In more recent work, Scanlon has come to view the kind of descriptive as- sessment that in What We Owe was interchangeably called “moral blame,” “moral 14 He returns later to reemphasize this: “moral criticism differs from other criticism of a person’s judgment-sensitive attitudes, such as criticism that points out our errors in mathematical reasoning, spelling, or chess, because the failings to which it calls attention have a particular kind of significance for the agent’s relations with other people” (1998: 276, emphasis added). Here again it is the kind of failing to which (moral) blame is a reaction that explains the special force of blame; the blaming reaction itself carries no independent force. Chapter 4 | The Emotional Economy of Responsibility 123 appraisal,” or “moral criticism,” as not itself actually constituting blame. Instead, Scanlon distinguishes these descriptive assessments of an agent’s moral faults both from judgments of blameworthiness and from blame itself (2008, 2013, 2015). The original descriptive assessment consists in the judgment that the agent has commit- ted a moral fault of a particular kind. A judgment of blameworthiness goes further; it is, Scanlon says, “a judgment that an action shows something about the agent that impairs, in various ways, his or her relations with others” (2013: 88; cf. 2008: 128). That is, as we might say, the judgment of blameworthiness consists in a determina- tion that the original descriptive assessment counts as a debit to be entered into the “ledger” that we use to keep track of the status of our interpersonal relationships. But actually to blame someone, according to Scanlon’s revised account, is something further still; it is, in addition, to “modify one’s understanding of one’s relationship with that person (that is, to alter or withhold intentions and expectations that that relationship would normally involve) in the particular ways that that judgment of blameworthiness makes appropriate” (2013: 89). These alterations may include things like decreased “readiness to enter into special relations such as friendship” and de- creased “willingness to help the person with his projects” (2015: 92). This abandons the purely cognitive approach. Blame, according to Scanlon’s revised account, consists in part in attitudes—like intentions and behavioral disposi- tions—that go beyond descriptive assessments, and includes (perhaps among other things) motivational elements. 15 The addition of these motivational elements gives Scanlon more resources with which to explain the force of blame. In particular, this addition suggests a straightforward answer to the problem raised at the end of the previous section. Ac- cording to Scanlon’s earlier account, the significance of blame was derived from the significance of the relationship-relevant facts about the agent’s conduct (and the character or regard that it revealed) that blame was a response to; this left blame with no independent force. According to the revised account, however, blame consists in part in revisions to the blaming party’s intentions, dispositions, and expectations to- 15 Because the additional attitudes include motivational elements, I count Scanlon’s revised account as a conative theory according to the taxonomy in § 2. This follows the taxonomy in Tognazzini & Coates 2016 (§ 1.3). Another, somewhat different conative account is George Sher’s (2006). Responsibility and the Emotional Structure of Relationships 124 wards the agent. And it is easy enough to see how the addition of motivational states like intentions adds a certain force that is lacking from the earlier account. As a mat- ter of definition such motivational elements are connected to various forms of action—confrontation, reproach, giving a cold shoulder, withdrawal, and so on. These forms of treatment have an undeniable, and potentially burdensome and un- pleasant, significance for anyone who is subject to them. On this basis, Scanlon offers his account as an answer to the question of blame’s force: This way of understanding [blame] . . . enables us to see how blame in particular is neither simply evaluative on the one hand nor punitive on the other. Blaming someone goes beyond mere evaluation because it involves modifications of our relationship with him or her, such as the withdrawal of trust and friendship. These are changes that the other person has reason not to want. Unlike punishments, however, they are not made because they are bad things from that person’s point of view, but for reasons having to do with our own concern with our relationships with them. (2015: 93) The force of blame, then, on Scanlon’s revised account, consists in the fact that the agent has reason not to want the blaming party to adjust her intentions, dispositions, and other attitudes towards the agent in adverse ways, ways that are warranted by the impairment to their relationship that the agent’s conduct has brought about. This is the significance that blame has over and above the significance of the misconduct that blame is a reaction to. It seems clear that this revised account describes something that often does occur in our reactions to wrongdoing; indeed, in substance the account strikes me as largely true, as well as insightful and important. What I think is less clear, however, is whether the additional force that the motivational elements contribute to Scanlon’s revised account is really the force of blame. Or perhaps it is better to be more cau- tious, in order to avoid verbal dispute about what counts as “blame,” and to say that it is unclear whether Scanlon’s account fully captures the distinctive force that is pre- sent in many paradigmatic instances of blame. We can grant that what Scanlon is describing is an important range of blaming responses, and that his account accords these responses a genuine force. But at the same time, there seems to be a further dimension of force that is present in at least certain paradigmatic instances of blame and that remains unaccounted for, even in the revised account. Chapter 4 | The Emotional Economy of Responsibility 125 Support for this suspicion can be found in critical responses to Scanlon’s re- cent work by Susan Wolf (2011) and R. Jay Wallace (2011), who argue separately that what Scanlon’s account describes is in an important sense both more (Wolf) and less (Wallace) than blame. For present purposes we can—and in fairness to Scanlon, quite probably should—read their responses more cautiously, as arguing that what Scanlon describes is both more and less than a certain kind of blame. I will thus be content to say that the issue is not that Scanlon misses out on blame entirely, but rather that his account of blame is at best only partial. For her part, Wolf notes that the kinds of adjustments that Scanlon pre- scribes—including decreased readiness to enter into relations with the agent and decreased willingness to help the agent—can be characterized as different forms of withdrawal from a relationship. But she argues that blame does not always, or even generally, have the character of withdrawal. Quite the contrary—in blaming some- one, “[r]ather than get some distance between you and the person you’re angry with, you might as likely want to ‘get in his face’” (2011: 338). Blame is as much associated with approach and engagement as withdrawal. Moreover, “[g]etting angry and express- ing it, and demanding a response, may bring people together and make them closer, rather than pushing them away”; this is due in part to the fact that “[g]etting angry, as opposed to withdrawing one’s trust, shows that one does not regard the person ex- clusively with the objective attitude” (2011: 339). These observations suggest that what Scanlon has identified as the force of blame quite generally is actually the force only of some drastic and perhaps unusual subset of blaming responses to relation- ship-impairing conduct. If this is right, then Scanlon’s account gives us insight into one dimension of the force that our responses to wrongdoing sometimes have, while still failing to capture the force of blame in at least an important range of cases. Wolf’s observations identify cases of blame that do not fit within Scanlon’s account, suggesting that the force that Scanlon has identified is not really the force of blame (or at least, of blame quite generally). Further support for this suggestion from a different direction is provided by Wallace, who contends that the responses that Scanlon’s account focuses on do not actually amount to blaming responses at all. Or as we can more cautiously put the point, the responses that Scanlon focuses on lack a familiar dimension of force that is present in at least many paradigmatic instances of Responsibility and the Emotional Structure of Relationships 126 blame. The problem, as Wallace argues, is that we can judge that someone has acted in a way that impairs their relationship with us, and alter our intentions and expecta- tions in accordance with that judgment, but do all of this in a “perfectly dispassionate” way and in a “perfectly cool and calculating frame of mind,” one that “lacks the emotional tone that seems essential to blame” (2011: 353-54). In such a case, we would satisfy the conditions of Scanlon’s account without, Wallace says, really blaming the other person; we would instead simply be “manag[ing] our rela- tions with them” (2011: 354). The kinds of attitude adjustments that Scanlon identifies with blame lack, Wallace suggests, “the affective quality that intuitively seems essential to blame, con- sidered as a characteristic and familiar way of responding to moral lapses” (2011: 354); he repeatedly refers to this affective quality as a “quality of opprobrium” that is distinctive in blame (2011: 355, 368). Thus Scanlon, as Wallace summarizes in a quip, “leaves the blame out of blame” (2011: 349, 355). Again, perhaps this way of putting the objection is too glib, and for present purposes we do not need anything so sweeping. It seems better to think of what Scanlon has to offer as an account of one particular kind of blame, or as a partial account of blame. But even if this is right and he has not left all of the blame out of blame, it still seems that he has left at least some of it out; there is something more to the complex phenomenon of blame and its force that the addition of the motivational elements that Scanlon identifies does not suffice to capture. Let’s pause and take stock. We started with the metaphor of blame as a kind of record-keeping, or entering debits into a ledger, and we distinguished two dimen- sions of dissatisfaction with that metaphor: on the one hand, it seems to leave out an essential affective or emotional dimension of blame; on the other hand, it seems to leave out an essentially personal dimension of blame. We set the first point to one side, and have been focusing on the second. Scanlon’s proposal in What We Owe de- scribed one sense in which blame can be said to have personal import, but failed to capture the significance that blame has over and above the significance of the rela- tionship-impairing circumstances to which blame is (according to Scanlon) a response. Scanlon’s more recent account addresses this problem, but does so by in- voking a set of attitude adjustments that appear, on closer examination, to be in one Chapter 4 | The Emotional Economy of Responsibility 127 respect harsher than many familiar blaming responses, and in a different respect more detached and dispassionate than many familiar blaming responses. Wolf’s and Wallace’s objections both make reference to the emotional quality of blame—the anger or opprobrium that is closely associated with, and arguably essential to, certain paradigmatic instances of blame. This suggests that the emotional dimension of blame and its personal dimension may be connected—we may not be able to ac- count for blame’s personal quality without reference to emotion or affect. While I think this is correct, there is still more that can be said about the personal element of blame, without yet resorting to emotion. 3.3 Blame as assessment, revisited We should consider one more attempt to explain the force of blame within a cognitive framework—that of Pamela Hieronymi (2004). Rather than seek some other kind of element—motivational, emotional, or otherwise—that could supply the force that appears to be missing from mere adverse assessments or (as she calls them) judgments, Hieronymi argues that descriptive judgments themselves can, given the right content, have a force that has not been fully appreciated, a force that is in- dependent of the facts that they describe. She thus urges a renewed focus on the judgments themselves. It should not, Hieronymi suggests, be mysterious that descriptive judgments can have significant force in interpersonal relations, because it often matters to us both whether something is true and whether others think that it is true: “Quite stan- dardly . . . a judgment gains force from the importance of its content and the importance of the opinions of others on that topic” (2004: 122, emphasis added). In particular, something that matters to us a great deal in interpersonal relations is the attitudes towards others that are manifested in our actions, and the attitudes towards us that are manifested in the actions of others. When someone acts in a way that manifests disregard or ill will for another, we don’t just judge them as having failed or per- formed poorly relative to some standard—that is, we don’t simply grade them— rather, we judge them as potentially ineligible to stand in “relationships of mutual regard” with others. Responsibility and the Emotional Structure of Relationships 128 Thus far, as Hieronymi notes (2004: 138-39 n.8), this hews closely to Scan- lon’s earlier account in What We Owe, and reflects the influence of that account. But Hieronymi then takes an important further step, by observing that it is important to us not only to stand in relationships of mutual regard with others, but also to stand in relationships in which mutual regard is recognized on both sides; a relationship in which we have due regard for the other, but the other erroneously believes us to be indifferent or ill-willed towards them, falls well short of the kind of relationship that we typically aspire to. This means that another’s judgment that we are not eligible to stand in relationships of mutual regard has an automatic force, because the judgment itself constitutes a significant change in our relations (beyond whatever change the fact that is the object of the judgment may constitute). It therefore cannot be the case (as Scanlon seems to have supposed in What We Owe) that the force of such judgments is exhausted by the force or significance of the facts that they describe. As Hieronymi says: A change in what you or another person thinks about the quality of your will, in itself, changes your relations with them. Insofar as it is important to stand in relations in which goodwill is recognized, the judgment that you have shown ill will itself carries a certain force. That judg- ment—even if incorrect—makes it the case that you no longer stand in relations in which your good will is recognized on all sides (2004: 124, emphasis added). Hieronymi concludes that the force of blame is, in large part, a matter of the force of judgments like those just described—judgments of disregard or ill will. Their force “derives from the importance of standing in relations of recognition of mutual regard” (2004: 124, emphasis added). There is, however, Hieronymi says, more to blame than just these judgments of disregard or ill will. Blame itself consists in the Strawsonian “reactive attitudes,” which are premised upon judgments of ill will, but which are also sensitive to other judgments that concern the particular significance that a given display of ill will has for our interpersonal relations. These collateral judgments concern, for instance, the standing of the blaming subject, the standing of the wrongdoer, the seriousness of the disregard shown, and so forth (2004: 132-33, 135). 16 Altogether, then, blame con- 16 Perhaps these additional judgments would include, or somehow amount to, what Scanlon in his revised account refers to as a judgment of blameworthiness—that is, a Chapter 4 | The Emotional Economy of Responsibility 129 sists in a group of attitudes that should be taken “to involve (to be sensitive to, or to entail, or to be the affective face of, or simply a label for) a complex set of judgments or evaluations,” including “the judgment that disregard was shown” as well as “judgments along a number of other dimensions” (2004: 133). Fully specifying all of the individual judgments that constitute a particular instance of blame, or that tend to constitute blame in general, would be “a difficult task”; but regardless, the “force of the reactive attitudes can be understood as the force of a set of judgments” (2004: 133). Hieronymi’s discussion shows that in a passage quoted earlier Scanlon, per- haps inadvertently, sets up a false choice. In What We Owe Scanlon had said: The significance of [the agent’s] moral faults is shown in part in the ways in which others have reason to respond to them: with expres- sions of criticism and hurt feelings, withdrawal of friendship, and so on. But the primary significance of moral criticism lies not in what others may do in the future as a result of believing it, but rather in what is, if the criticism is correct, already the case (1998: 272). This is too restrictive. The significance of moral faults need not be exhausted by, on the one hand, the changes that those faults themselves effect in the agent’s relation- ships and, on the other, the non-descriptive responses of others (“expressions of criticism and hurt feelings, withdrawal of friendship, and so on”). There is an ele- ment of significance or force that Scanlon has left out: the distinctive, special significance of others’ descriptive assessments of the agent—what those assessments themselves mean for the agent’s relationships with others. This is the missing piece that Hieronymi brings to light. Hieronymi thus seeks to capture the characteristic force of blame by expand- ing our conception of the kind of force that descriptive judgments can have in interpersonal relationships. In pursuing this strategy she is, I think, strikingly success- ful. The key to this success is the way in which she moves beyond looking at descriptive judgments merely in isolation, in terms of their intrinsic features (for in- judgment not merely that ill will was shown, but further that this display of ill will “shows something about the agent that impairs, in various ways, his or her relations with others” (2013: 88). But either way, Hieronymi holds that the force of blame consists in judgments alone—it does not require the further step of adjusting inten- tions, dispositions, and the like in the way that the judgments would render appropriate. Responsibility and the Emotional Structure of Relationships 130 stance, their contents, the object of appraisal, the dimension of appraisal (moral or otherwise), the valence of appraisal, etc.). Hieronymi looks instead to the important, independent role that judgments play in the interpersonal relations in which they are ordinarily made. In this picture, merely descriptive judgments of ill will (together with the concomitant judgments that for Hieronymi constitute the “reactive atti- tudes”) have a distinctive force because they play a distinctive, essential role in constituting and regulating relations of mutual regard that matter deeply to us. This substantially advances our understanding of the force of blame and, in a sense, completes and makes good on the ledger metaphor. Hieronymi’s account elaborates on the significance of the “ledger” and the role that it plays in regulating our interpersonal relations, and it locates the force of blame in the right place: in the attitudes themselves that (according to her view) constitute blame. This represents significant progress with respect to the concern raised earlier about the personal di- mension of blame; it is both consistent with and fleshes out Bennett’s suggestion that blame “involves us in a more intimate . . . relationship,” as well as Smith’s sug- gestion that blame plays a role in “structuring our moral relations with one another.” Nevertheless, I wish to suggest that there remains an important gap in Hi- eronymi’s understanding of blame and its force. We earlier set aside the problem of the emotional dimension of blame in order to focus on the problem of the personal dimension. But we have seen reference to the emotional dimension in Bennett and Smith’s comments, as well as in Wolf and Wallace’s objections to Scanlon. We should now consider this issue. Hieronymi explicitly denies that emotion or affect can carry any of the force of blame. The force of blame just is, for her, the force of a set of descriptive judgments. We will consider Hieronymi’s reasons for discounting the role of affect in the next section. But the difficulty this raises for Hieronymi’s account is similar to the one that faces Scanlon’s more recent account of blame: On the one hand, it seems possible to make all of the judgments that Hieronymi thinks are constitutive of blame, and yet to fail actually to blame in a way that carries a familiar force; on the other hand, it seems possible to blame in a way that carries a familiar force without making the judgments. Chapter 4 | The Emotional Economy of Responsibility 131 The first side of the difficulty is analogous to Wallace’s concern that Scanlon “leaves the blame out of blame.” The problem is that it seems at least in principle (and likely also in practice) possible to make all of the judgments that compose the complex set of judgments that for Hieronymi constitute blame, but to do so in a “perfectly dispassionate” way and in a “perfectly cool and calculating frame of mind” that “lacks the emotional tone that seems essential to blame” (Wallace 2011: 353)— or as we might more cautiously put it, the emotional tone that seems to lend a dis- tinctive force to at least certain paradigmatic instances of blame. In such a case, there is no doubt an important sense in which we do blame the other person; the elements that satisfy Hieronymi’s account thus amount to a genuine kind of blame. But at the same time, there is also an important sense in which we do not blame the other per- son, or at least, do not blame them as fully or as forcefully or in quite the same way that we might have done. Imagine, for instance, that someone we have wronged makes all of the judg- ments that are supposed to constitute blame (that we have shown disregard, that the failure was significant, that it constitutes an impairment in the relationship, that they have the requisite standing, etc.), but due to reasons of excessive affection, or mag- nanimity, or distraction, or depression, or general meekness, or the like, simply fails to muster any emotional reaction to accompany the judgments. The judgments are all in place, and yet this person finds their feelings towards us unchanged; there are, as we sometimes say, “no hard feelings.” In a case like this, we might reasonably take the view that even though we have not fully “gotten away with it,” in a sense we have gotten off easier than we might otherwise have done. In the right circum- stances, it would be perfectly intelligible, and indeed reasonable, for the agent to greet the absence of hard feelings in a case like this with relief. The reasonableness of this relief indicates that Hieronymi, like Scanlon, has left at least something out (even if it is also the case that she has not entirely left the blame out of blame). This illustrates that, at least in certain paradigmatic instances, blame can have a force—and a significance for the agent—that is at best incompletely captured by Hieronymi’s account. Her view would seem to provide only a partial account of blame, or perhaps an account of only a limited range of blaming reactions. Moreo- ver, the kind of case we have been considering also provides another intimation of Responsibility and the Emotional Structure of Relationships 132 the way in which the affective or emotional dimension and the personal dimension of blame are connected. When we fail to muster any blaming emotions towards those we judge to be blameworthy (or otherwise to satisfy the conditions that war- rant blame), whether from reasons of distraction or depression or magnanimity or the like, there is a sense in which we fail to care about the other person’s conduct—or as we sometimes say, to hold it against them or to take it personally. There is, we might also say, an element of personal engagement in blame that does not consist in any de- scriptive assessment, or even the full suite of descriptive assessments (whatever they may be) that Hieronymi designates as the constitutive judgments of blame. The first side of the problem for Hieronymi’s account, then, concerns this possibility of judgments without blame—or, more cautiously, without the full force of blame, or without the distinctive kind of force that we find in at least certain para- digmatic instances of blame. The other side of the problem concerns the possibility of blame (and its force) without judgments. Hieronymi’s account appears to leave no room for episodes of blame that are independent of our descriptive judgments. As Hannah Pickard (2013) has argued, blame can be, and indeed sometimes is, irra- tional—directed at those we know not to be blameworthy. In this respect, blame behaves similarly to paradigmatic emotions like fear, which can be stubbornly recalci- trant even in the presence of considered judgments that the object of fear is not dangerous and that the feeling of fear is misplaced. This is a kind of irrationality that is distinctive and characteristic of the emotions, rather than descriptive judgments (Pickard 2013: 613). Moreover, even episodes of irrational blame can have an important kind of force, which appears vividly in dramatic cases. Suppose an irrationally overprotective parent (he may even be aware that his overprotectiveness is irrational) is reluctant to let his child go on an overnight trip with a school group, but his more relaxed (and reasonable) spouse persuades him to go along with it. During the trip, some chance mishap occurs—a car accident say, in which the child is seriously injured. The over- protective parent might not be able to prevent himself, despite his considered judgment and his best efforts, to avoid on some level blaming his spouse. Here, there are recalcitrant and (we can imagine, given the stakes) intense hard feelings, despite the absence of supporting judgments. And it is, moreover, not difficult to imagine Chapter 4 | The Emotional Economy of Responsibility 133 that the spouse might feel the force or sting of this blame quite acutely, even without entertaining any doubts at all about the ex ante reasonableness of allowing the child to go on the trip. This possibility of (one kind of) blame (and its force) without judgment, to- gether with the possibility of the relevant judgments without (one kind of) blame (and its force), suggest that, despite the insight and explanatory power of Hiero- nymi’s proposal, there remains a dimension to the force of blame that her picture leaves out of account. The nature of the problem once again points naturally towards a solution in terms of emotion or affect, and so we will turn in the next section to consideration of an account of blame’s force in terms of emotion. But first, I want to consider a possible reply that might be offered on behalf of Hieronymi’s account. Perhaps the foregoing concerns are based on an insufficient appreciation of the resources available to cognitive views, and in particular on an in- adequate and overly narrow notion of what a descriptive assessment—or judgment, or evaluation, etc.—is and of what such assessments can do. I have suggested that descriptive assessments are not sufficient fully to ac- count for the force of blame because they cannot account for elusive dimensions of emotional and personal engagement that are distinctive and important in at least cer- tain cases of blame. And I have suggested that descriptive assessments are not necessary for the distinctive force of certain cases of blame, because we can blame irrationally, even despite failing to make the descriptive assessments (whatever they are) that are supposed to comprise blame. But on both fronts an advocate of the as- sessment view might conceivably push back. Descriptive assessments, they might say, are in one way or another capable of encompassing the elements of emotional and personal engagement that I am concerned to capture, because descriptive as- sessments can be “loaded” in some way that answers this concern. Moreover, they might add, the descriptive assessments in question need not be “partly or wholly identified with consciously accessible, personal-level judgments or beliefs” (Pickard 2013: 617). Rather, the kind of assessments we are after might be wholly or partially unconscious or sub-personal, in a way that would explain the possibility of conflict- ing assessments, and thus of cases of “irrational” blame, in which the assessments Responsibility and the Emotional Structure of Relationships 134 that comprise blame are in tension with the subject’s considered and settled per- sonal-level judgments. To be clear, there is nothing in Hieronymi’s discussion that suggests an inten- tion to embrace this kind of broadening of the notion of a descriptive assessment. She is quite clear that she thinks that blaming attitudes are “simply reactions to—that is, they simply mark or acknowledge—the importance of a display of ill will or disre- gard between persons standing in relation” (2004: 135). The response that I am considering would involve significantly loosening this constraint. I take no position here on whether this sort of strategy is either viable or at- tractive. If it is, then we could recast the present inquiry. If blaming assessments can be “loaded” in a way that covers the emotional and personal dimensions of blame, and if blaming assessments are of a different character than consciously-accessible, personal-level judgments, then it would be worthwhile to say something more about what this “loaded” dimension consists in, and about the particular ways in which loaded assessments differ from settled personal-level judgments. (Hieronymi, as I have noted, does not attempt this.) So there would at least be room to supplement the account, and in particular to consider the possibility that the “loaded” blaming assessments at issue could usefully be conceived as participant attitudes. 3.4 Blame as emotion Wallace, we have seen, criticizes Scanlon’s account on the ground that it “leaves the blame out of blame” (2011: 349) because the kinds of attitude modifica- tions that, according to Scanlon, constitute blame can be made “in a way that is perfectly dispassionate” (2011: 353). While Wallace’s way of putting the objection is perhaps too sweeping and dismissive, the objection does seem to have some force. And in the previous section I suggested that Hieronymi’s view may be subject to a version of the same objection. Even if Scanlon and Hieronymi each captures an im- portant dimension of blame (or kind of blame), there also seems to be a dimension (or kind) of blame that each of them leaves out. What is missing from accounts like these that do not accord an essential role to emotion, Wallace contends, is “the affective quality that intuitively seems essential to blame, considered as a characteristic and familiar way of responding to moral Chapter 4 | The Emotional Economy of Responsibility 135 lapses” (2011: 354). Wallace repeatedly characterizes this missing affective element as a “quality of opprobrium” (2011: 349, 355), and he argues that this quality of oppro- brium can only be made sense of in terms of the “reactive sentiments” of resentment, indignation, and guilt (2011: 349, 357-58, 366). 17 The force of these sentiments, he suggests, can be seen in the way in which they transform the meanings of the things that we do in response to others’ conduct, including the attitudinal adjustments that are central to Scanlon’s account. Wallace remarks: It is one thing to modify your standing intention to confide in a per- son who has betrayed you out of a sorrowful recognition that this attitude is no longer appropriate to the nature of the relationship be- tween you. It is quite another to modify the attitude when you also feel warranted resentment of the person for letting you down in this way (2011: 357). The resentment “gives you additional reasons for adjusting your other attitudes and activities” and lends your other responses an “expressive significance” that they would not otherwise have (2011: 357). Thus, “[r]esentment and the other reactive emotions do not leave everything else as it otherwise would be; they transform your activities, giving them an expressive character that would be completely missing in their absence” (2011: 358). This transformative, expressive power of the reactive emotions is a manifestation of the “special quality of blame” (2011: 358). Wallace thus proposes that to blame someone “is to be subject to a reactive emotion towards them, where this emotional response lends your reactions to wrongdoing an expressive dimension that would otherwise be missing” (2011: 366). 18 In order to complete this account, something more needs to be said about what ex- 17 As noted in chapter 2, § 2, Wallace inherits the term “reactive” from Strawson, but uses it in a special, restricted way. In the following discussion, I will follow Wallace’s convention of using the term “reactive attitude” (or “emotion,” “sentiment,” etc.) to refer exclusively to resentment and its vicarious and reflexive analogues (indignation and guilt, respectively). 18 The quoted formulation appears to presume that in order to blame someone you must both have a reactive emotion and react in other ways that acquire an expressive dimension from the reactive emotion. I think what Wallace means to say is rather something along the lines of: to blame someone is to have a reactive emotion to- wards them, where this emotional response is of the kind that would tend to give certain further reactions (if you had them) an expressive dimension that would otherwise be missing. Responsibility and the Emotional Structure of Relationships 136 actly the reactive emotions are, and thus about what exactly it is that is being ex- pressed when the reactive emotions lend an “expressive dimension” to other reactions to wrongdoing. As we will see in a moment, Wallace has more to say about this. But first it is worth pausing to take note of the potential advantages of this gen- eral kind of approach to accounting for the force of blame. First, as Wallace intends and as Hieronymi concedes, appealing to the reac- tive emotions “is attractive, in part, because it seems to allow blame to be more than description—to carry a certain force—but still to be less than action” (Hieronymi 2004: 119). This feature answers to the dilemma with which we started, that blame appears to be something more than adverse assessment, yet something less than ad- verse treatment. An emotional response like resentment seems favorably positioned relative to this way of framing the challenge; it is at any rate something different from—and thus at least potentially “more than”—adverse assessment, while also something less than adverse treatment (an emotional response need not be overtly manifested or expressed at all). At the same time, emotions are associated in a famil- iar way with characteristic dispositions towards action and expression, suggesting a possible explanation of the connection between blame, reproach, and sanction. So if we can make good on the idea that a resentful emotional response carries a force that is “more than” that of adverse assessment, and show that it is “more than” in the right respects, then we will have located something that sits in the sweet spot be- tween assessment and treatment. Second, conceiving of blame as an emotional response addresses the concern that we raised for Hieronymi’s account about irrational blame, including both judg- ments that fail to produce (one kind of) blame and blame that fails to be supported by judgments. This is a kind of irrationality that is distinctive and characteristic of the emotions, and by conceiving of blame as itself an emotional response, Wallace avoids this difficulty. 19 These are at least signs, then, that Wallace is on a promising path in conceiv- ing of blame as an emotional response. But we still need to understand how it is that 19 Wallace explicitly embraces this implication of his view in earlier work; see espe- cially his discussion of “irrational guilt” in Responsibility and the Moral Sentiments (1994: 40-50, § 2.4). Chapter 4 | The Emotional Economy of Responsibility 137 an emotional response can be the right kind of vehicle to carry the distinctive force of blame. Hieronymi states this point forcefully, as an objection: [I]t is unclear how the affective accompaniment of a judgment could, itself, carry the characteristic force of blame. An affective accompa- niment of a judgment would be a certain unpleasant emotional disturbance, occasioned by the judgment. But, the force of blame seems deeper, more serious or weighty than simply being the object of certain [sic] unpleasant emotional disturbance. The affect, itself, seems insufficiently robust (2004: 121). Hieronymi’s immediate target is a view according to which blame consists in an ad- verse judgment plus some “affective accompaniment”; this does not fit Wallace’s view, according to which blame just is an emotional response. Nonetheless, Hiero- nymi here raises a significant question for either type of view. We have seen that Scanlon’s and Hieronymi’s theories account for the force of blame by reference to the way in which the attitudes that are (by hypothesis) con- stitutive of blame themselves determine the status of the relationship between the blaming party and the agent. The question that Hieronymi raises for emotional theo- ries of blame is whether they can provide an explanation of the force of blame that has the same depth. It is not enough merely to point to the agitated or exercised quality of the affect or emotion that, according to a theory like Wallace’s, constitutes blame. We also need some account of what that agitated or exercised quality means, and why it matters. Affect alone, conceived as “a certain unpleasant emotional dis- turbance,” does not obviously have the same significance for relationships as adverse assessments, intentions, dispositions, or expectations. The latter attitudes derive their special importance from the constitutive role that they appear to play in relationships. To appreciate the force of this point, consider that I might, through some mysterious mechanism, develop indigestion every time I take you to have regarded or treated me in a certain way, and just in virtue of its unpleasantness this might af- fect the kind of relationship that it is possible for us to have. But this unpleasant reaction—no matter how unpleasant it might be—would lack the characteristic seri- ousness or force of blame. This is because it would be a merely practical impediment to closer relations; the unpleasant reaction would not itself constitute a change, im- pairment, or other disturbance in the relationship. The question, then, for emotional accounts of blame is: Why should an unpleasant emotional or affective distur- Responsibility and the Emotional Structure of Relationships 138 bance—the equivalent of a kind of mental indigestion—have a greater depth or seri- ousness? Like gastric indigestion, the affective reaction alone seems “insufficiently robust” to account for the characteristic force of blame. Wallace is acutely aware of this difficulty, 20 and he repeatedly emphasizes that the emotional element in his account “is not a matter of merely private or phenome- nological biography” (2011: 357) and “is not merely a matter of the private or phenomenological coloring of our experienced interactions” (2011: 366). It is not, in other words, merely a form of mental indigestion that is triggered by certain kinds of conduct or regard. To understand how emotions carry a significance that indigestion does not, Wallace starts with the distinction between “acknowledging something to be valu- able, and actually valuing it in the way that is called for by its value” (2011: 366-67). In Wallace’s example, two people might agree that opera and philosophy are both valuable pursuits—that there is reason to support them, to learn about them and ap- preciate them, to engage in them, and so on—and yet one might actually value opera deeply and have no personal interest in philosophy, while the other values philoso- phy deeply and has no personal interest in opera. As Wallace says: There is an additional quality of emotional engagement that charac- terizes the attitudes of people who genuinely value these pursuits; they take a real interest in them, care about whether they are in a good or a bad way, become excited when there are opportunities to engage in activities related to those pursuits, and are subject to dis- tress when opportunities of this kind are lacking (2011: 367). The kind of emotional engagement that distinguishes actually valuing some- thing from merely acknowledging or judging it to be valuable constitutes a particular way of caring about things. It is, Wallace suggests, a “deep fact about human beings that they are capable of caring about things, in the way that involves these forms of emotional engagement and vulnerability” (2011: 367). This capacity is a significant part of our “evolutionary inheritance”; it is crucial to the survival of our offspring and to our ability to achieve various objectives; and it contributes to our sense that our lives and activities have meaning (2011: 367). “Caring about things gives them a 20 Particularly in his 2011 paper. In earlier work, like Responsibility and the Moral Senti- ments (1994), it is not as clear that Wallace fully appreciated the importance of this problem. Chapter 4 | The Emotional Economy of Responsibility 139 kind of significance for our lives that they would otherwise entirely lack” (2011: 367). We might further say that to care about things in this fundamental way is just what it is for things to matter to or to be significant to us (as distinct from being judged by us to matter or to be significant in a general sense). And such caring is constituted, at least in significant part, by patterns of emotional engagement and vulnerability. To care about opera—in the sense of actually valuing opera, rather than merely judging it to be valuable—just is, at least in part, to be happy and excited when a new opera production is scheduled, to be disappointed when it is cancelled, and so on. These emotional reactions are instances of caring; they are exercises our of capacity to value. Blame, Wallace argues, should also be understood as a kind of caring or valu- ing. Judging that someone has acted in a way that warrants blame is a kind of judgment that something has a value or significance; blaming someone is a manifesta- tion of actually valuing it. Blame reflects that we are emotionally invested in the agent’s misconduct or lack of regard in a way that is distinctive of caring or valuing: “The immoral attitudes of the person who is blamed matter to you, in the way that is characteristic of the more general phenomenon of valuing” (2011: 367). More spe- cifically, Wallace contends, blame is a manifestation of our caring about moral values: “blame reflects our internalization of the values at the heart of morality.” That is, according to Wallace’s view of morality, blame is a manifestation of the way in which we care about relating to others in the distinctive way that compliance with moral requirements makes possible—“a distinctive form of mutual recognition or regard” (2011: 368). To internalize a concern for the values around which morality is orga- nized, Wallace says, “is to care about relating to people on these terms, valuing this form of relationship in a way analogous to the way we value and care about many of the other important goods that fundamentally shape our lives” (2011: 368). This, then, is the expressive significance that blame adds: our responses to wrongdoing— sanctions, cold shoulders, etc.—become expressions of the way in which we care about the moral quality of others’ conduct and the regard that it expresses. This account furnishes material for a response to Hieronymi’s objection that affect seems “insufficiently robust” to account for the characteristic force of blame. This objection was based on the idea that the kind of affect that was in question was Responsibility and the Emotional Structure of Relationships 140 “the affective accompaniment of a judgment” that would itself be nothing more than “a certain unpleasant emotional disturbance, occasioned by the judgment” (2004: 121). But in Wallace’s view affective responses can be more than this. In the first place, the affect need not merely accompany a judgment, but can be part of an emo- tional reaction that is quite independent of judgment. Second, and more importantly, according to Wallace emotional reactions are not merely “emotional disturbances”— they are manifestations of caring or valuing. And caring matters; what we care about or value, as well as the manner and degree to which we care about it, matters to us and—crucially—to others. Indeed, it often matters a lot. Consider, for example, dis- covering that your friend or your spouse does not care for you in the same way they used to do. The fact that emotional reactions can be manifestations of caring or valu- ing means that they can carry a force that is independent of the force of the judgments or assessments that they might be associated with. Wallace’s proposal thus both delivers the emotional dimension of blame and explains how it is that blame can carry a force that is more than adverse assessment yet less than adverse treatment. These count as important advantages of his view, relative to the other views that we have considered. There remains, however, the issue of accounting for the other elusive dimen- sion of blame—the personal dimension. Though I have been suggesting that the emotional and the personal dimensions are linked, Wallace’s reliance on emotion does not automatically address this problem, as Wallace is aware. He identifies the difficulty by pointing to the fact that his account of blame still owes an explanation of why it is that blame seems to have a force that is different even from that of cer- tain other emotional responses to wrongdoing. After all, not every emotional response that is a manifestation of caring about “the values at the heart of moral- ity”—the characteristic object of care in blame, according to Wallace’s account— carries the force of blame. As Wallace argues: [O]ne can value something without being susceptible to the reactive emotions in particular. For all that has been said so far, one might show that one cares about the values at the basis of morality through a tendency to feelings of sadness or distress when moral values are thwarted, and to elation or pleasure when they are vindicated. . . . [E]motional engagement of this more generic kind would not suffice Chapter 4 | The Emotional Economy of Responsibility 141 for the kind of ascription of responsibility involved in blame (2011: 368). As I would put the point that Wallace seems to be making in this passage: When someone cares about the values at the heart of morality, this caring can be mani- fested by emotional responses that lack the distinctive personal force of blame. Wallace repeatedly characterizes the still-missing element as a “quality of op- probrium” (2011: 349, 368), 21 but this seems to be a misnomer. Wallace never explains what he means by “opprobrium,” and rather than explaining the distinctive- ness of blame in terms of opprobrium, disapproval, censure, or the like, he relies upon what he calls the relational character of blame. This is a sign that the nature of Wallace’s concern about “generic” responses like sadness or distress is not that they lack opprobrium, but rather that they are insufficiently relational or (in my term) per- sonal. 22 If you act badly, my caring about the moral values that you have contravened could manifest itself as a kind of impersonal sadness, distress, or disappointment. 23 (Think of the sadness or distress that a perfectly compassionate moral saint, or someone like Jesus or Gandhi, might feel on learning that someone has wronged them.) This kind of reaction seems to be a way of manifesting genuine care for mo- rality, but it also lacks the distinctive, personal force of blame. What is it, then, that distinguishes blame from other, more “generic” or im- personal kinds of emotional engagement, and that gives blame it’s special force? Wallace’s answer to this question is that, unlike sadness, distress, disappointment, or other impersonal responses, “the disposition to blame is a way of taking to heart the values at the basis of morality that is peculiarly appropriate to the relational character 21 The term also appears in the title of Wallace’s 2011 paper, “Dispassionate Oppro- brium,” which is meant as an ironic rebuke of Scanlon’s account of blame. 22 Wallace uses language more similar to mine in another place, in describing the dif- ference between the objective and participant attitudes. He describes professionals like therapists or psychologists who ordinarily take the objective attitude towards their subjects, but who might lapse on particular occasions and experience reactive attitudes like resentment. Wallace remarks: “As we might colloquially put it, they have started to ‘take it personally’ that their subjects act as they do” (2014: 123). 23 As I will explain in § 4, I actually disagree with Wallace that attitudes other than his reactive attitudes (resentment, indignation, and guilt) are always impersonal in this sense. In my view there are kinds of sadness or distress that do have a force that is at least similar to blame’s; these are what I would call, in accordance with the discussion in chapter 2, participant versions of sadness or distress. Responsibility and the Emotional Structure of Relationships 142 of those values” (2011: 368-69). Or as he says again, blame “has a relational aspect” that is lacking in other reactions like sadness: When we are wronged by another, we are not just saddened by their failure to relate to us on a basis of mutual regard; we resent such treatment, and this emotion has a relational aspect that fits the char- acter of moral norms as constituting a valuable form of human relationship. To resent someone is to feel not merely that they have acted wrongly, but that they have wronged us in particular, violating the norms that constitute relations of mutual regard. (2011: 369) This “relational character” or “relational aspect,” Wallace argues, is what gives blame “a quality of opprobrium that is lacking when one feels mere sadness or melancholy” (2011: 368). This, as I understand it, is Wallace’s answer to the problem of the per- sonal dimension of blame. Altogether, this is perhaps as close as we have come to an account of blame and its force that does justice to both its emotional and personal dimensions. But still, I wish to suggest that Wallace’s strategy of accounting for the personal or rela- tional character of blame in terms of the internalization of relational values is lacking in two respects: first, it is incomplete; second, and more seriously, it misrepresents the personal dimension of blame. The respect in which Wallace’s account is incomplete is that it explores the character of the blaming party’s attitudes, and the expressive significance that those attitudes can lend to their actions, but simply does not address the significance that blame has—independently of its expressive significance—for the agent through her relationship to the blaming party. This is connected to the (somewhat narrow) way in which Wallace conceives of relationships, 24 but for present purposes this is not a se- rious shortcoming of the account. As I will suggest in the next section, there is no obstacle to remedying this lacuna by combining some of the material from Wallace’s account with a structure like that of Hieronymi’s. The more serious issue with Wallace’s account is that it misrepresents the personal dimension of blame. To see this, consider first another set of relational val- ues: the values around which friendship is organized. Part of being a fully initiated participant in the kinds of communities with which we are familiar is internalizing the values at the heart of friendship, where this is a matter of caring about relating to 24 For discussion see chapter 2, especially § 3. Chapter 4 | The Emotional Economy of Responsibility 143 people on terms of friendship—trust, mutual affection, mutual concern, mutual sup- port and aid, taking pleasure in each other’s company, and so on. It makes sense to care about relating to people on these terms, because this kind of relating constitutes a distinctively valuable form of human connection that is an important element of the good life as we know it. And when we care about this form of relationship, we become susceptible in our own relations with other persons to a range of emotional responses when the terms of friendship are, or fail to be, fulfilled. For instance, we become inclined to be saddened when we grow apart from one of our friends, not just because we cared for them, took pleasure in their company, and so on, but also because we have lost a friend. Similarly, we become inclined to rejoice in a friend’s successes not just because we want them to succeed and consider them deserving, but also simply because they are our friend and it is rewarding to see a friend do well. But it is a crucial fact about friendship that internalization of the values around which friendship is organized is not alone sufficient to account for any of the personal emotional reactions that are grounded in caring for this form of relation- ship. That is because these reactions are parasitic on, and can only be triggered in the presence of, another form of caring, namely care for the particular other persons who are our friends. 25 Caring about the value of friendship can give me a reason to be a good friend, as well as a reason to have and to cultivate the kinds of emotional re- sponses that friends are subject to, but this kind of caring alone gives me no guidance as to which particular persons I should have those responses towards. For that, I need to consult my own attitudes about particular other people (as well as learn about their attitudes towards me). After all, if I cared deeply and obsessively about friendship, but not a whit for any particular other people, I would in fact have no friends. So we need to distinguish between caring about the values around which friendship is organized, and caring about particular other people—namely, our 25 Wallace also considers friendship to be an important example of a relationship whose organizing values we can internalize, but in his discussion of friendship (2011: 356-57) he elides the difference between emotional responses that are grounded in care for the value of friendship, and emotional responses that are grounded in care for the particular other persons who are our friends. As I explain, in my view this is a consequential elision. Responsibility and the Emotional Structure of Relationships 144 friends. Both kinds of caring are important, and play an important role in the pat- terns of emotional investment that constitute the good life as we know it; but the latter form of caring is in one sense more fundamental, at least for purposes of ex- plaining our susceptibility to the emotional responses that friends are distinctively susceptible to. Attention to an analogous distinction exposes an important gap in Wallace’s account of the force of blame. For Wallace, blame is “a way of taking to heart the values at the basis of morality that is peculiarly appropriate to the relational character of those values”; this means that when I blame someone, this is a manifestation of the way in which I “care about relating to people in the distinctive way that is consti- tuted through compliance with basic moral requirements,” for instance, norms of due regard or non-contempt. And to be sure, when I blame you for the contempt that you have shown me, assuming that I have an appropriate regard for morality, my response will involve my deploring the way in which your contempt violates the moral norms that make possible valuable relations of mutual recognition and regard. But in all likelihood that is not all that my response will involve. In addition to caring about the fact that your contempt is contrary to moral values that I have internalized, if I am a normal, well-socialized person and have even the barest modi- cum of respect for you, I will care in a personal way—a way that is unmediated by care for the values around which morality is organized—about the fact that you feel contempt for me. To be the object of another’s contempt, merely in itself, is ordinar- ily a deeply disquieting thing, and no reference to moral values is necessary to explain this. We quite reasonably strongly dislike to be objects of contempt to other mem- bers in good standing of our communities. 26 If I didn’t care one whit about your contempt for me personally, but cared about it only as a manifestation of my care for the values around which morality is 26 Hieronymi makes a similar point: “I don’t resent you because you failed to achieve some standard of human excellence or failed to live up to the norms internal to will- ing. Rather, I resent you because your action or attitude communicated disrespect or disregard for me. . . . I am more concerned with myself, or with the relation in which we stand, or with your concern for me, or with how I figure into your world, than I am concerned with your per- formance, as such—with how you fared against a standard of moral excellence or rectitude or goodness. . . . Talk of assessment against a standard can therefore mislead” (2004: 124, emphasis added). Chapter 4 | The Emotional Economy of Responsibility 145 organized, my reaction to your contempt would—despite perhaps being affectively “hot”—be oddly impersonal. 27 Compare the person who shows up at the hospital to visit a sick friend only because to do otherwise would be a sin against the values of friendship and because “that’s what friends do”—without caring one whit independ- ently about how the friend is faring. (“If she didn’t happen to be my friend, whether she lived or died would be a matter of total indifference.”) 28 Such an act of “friend- ship” is perhaps not devoid of value, but it is lacking in the special warmth and consolatory significance that a genuinely friendly version of this act would carry. Similarly in the case of blame, I want to suggest that the blame of someone who is resentful only in the thin sense that Wallace describes lacks at least much of the dis- tinctive force that we see in many paradigmatic instances of blame. I will aim in the next section to fill out this suggestion by providing a positive account of the force of blame in such cases. First, let’s take stock again. The metaphor with which we started, of blame as debits in a ledger, left out both an essential emotional dimension of blame and an 27 Though he approaches from a different direction, John Deigh seems to be on to a closely related point when he argues that Wallace’s theory “abandons Strawson’s conception of resentment as a distinctly personal attitude” (2011a: 210). Deigh grounds this objection in the observation that whereas Strawson saw the reactive at- titudes as direct responses to the good or ill will of others towards us, Wallace sees them as responses to others’ choices—that is, to whether they have correctly balanced all of the reasons that bear on their actions. The shift from quality of will to quality of choice, Deigh suggests, marks a shift from personal to impersonal evaluation and reaction: “On Wallace’s conception . . . resentment towards someone does not in- clude any sense of that person’s being ill-disposed towards one or contemptuous of one. What it includes is a belief that the reasons on which the person based his choice were weaker than the reasons he had to meet the expectations to which one holds him. When those expectations are based on moral obligations that the person has to one, then what Wallace does is to replace Strawson’s conception of resent- ment 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” (2011a: 210-11). 28 Here I am adapting a classic example from Stocker 1976: 462. There is of course in this case a question about whether friendship is really compatible with this kind of indifference. This issue doesn’t arise (at least in the same way) in the moral case be- cause we stand in the “moral relation” to all members of the moral community, even if we lack a personal attachment to them. It is without question possible to care about someone’s compliance with moral norms even if we do not care, in a direct, unmediated way, about their attitudes towards us. But I am suggesting that reactions that manifest only the former, mediated sort of care are notably impersonal, and lack much of the force that we find in many paradigmatic instances of blame. Responsibility and the Emotional Structure of Relationships 146 essentially personal dimension of blame. Scanlon’s recent work on blame has filled in the picture in important ways but without, I think, significantly advancing our under- standing of these missing elements. Hieronymi’s account supplies a personal dimension to blame, but leaves out the emotional dimension; in this way, her ac- count risks leaving (at least some of) “the blame out of blame.” Wallace’s account supplies an affective or emotional dimension, but on his account blame is rendered strangely impersonal. This combination of results naturally suggests that neither Hi- eronymi nor Wallace succeeds completely, precisely because each of their accounts has something that the other is missing. I believe this is correct, and that this obser- vation puts a solution to the problem within reach. 4 Blame as a participant attitude We have now reached, I hope, a point where we can draw material from the foregoing discussion together with some material from previous chapters to address the gap in our present understanding of blame and its force. I propose that the force of blame consists, in a familiar range of paradigmatic cases, in the combination of blame’s being an affective or emotional manifestation of our caring about others’ attitudes, together with the significance that such caring has in our relations with those others. This proposal combines Wallace’s thought that blame is a manifestation of a kind of caring or valuing, with Hieronymi’s thought that blame is a personal, unmediated response to others’ attitudes, one that itself has significance for relations between the blaming party and the agent. This proposal strikes me as highly plausible and intuitive—indeed almost obvious, once we over- come the natural philosopher’s tendency to intellectualize and de-personalize blame. What lies behind the intuitiveness of this proposal is the simple fact that what other people care about matters greatly to us. In particular, it matters to us whether other people care about us, including whether and how they care about our attitudes towards them. Our caring about whether and how others care about us is, moreover, an instance of our taking up what I called in chapter 2 the participant stance, characterized by the following thesis: Participant Stance Thesis: To take the participant stance towards another is to view the other’s attitudes towards oneself as (non- Chapter 4 | The Emotional Economy of Responsibility 147 derivatively) central to one’s self-understanding (at least in some re- spect and to some degree). This thesis aims to articulate what it means to care about another’s attitudes in the special, personal way that is distinctive of taking it personally how someone is disposed towards you. It thus dovetails with Wallace’s idea that there are patterns of emotional involvement and susceptibility that are distinctive of caring or valuing, and it identi- fies one discrete and significant area of such involvement. The thought underlying the Participant Stance Thesis is that we often— indeed, ordinarily—care directly about others’ attitude towards us, to the extent that many of our fundamental interests and ultimate ends concern such attitudes. As I have previously noted, we strive avidly for the approval, trust, admiration, and love of those who matter to us, and just as zealously seek to avoid disapproval, disrespect, disdain, or disgrace. We care about these things not just deeply and for themselves, but in the distinctively personal way that we care about those of our features or cir- cumstances that are particularly central to our self-understandings. For instance, to repeat an example from chapter 2, a change in fortunes, attributes, or abilities that results from impersonal causes can have a devastating or exalting effect on some- one’s self-understanding—a concert pianist’s loss of the use of her hands through a random accident, say. But taken on their own such impersonal events are not occa- sions for hurt feelings, shame, offense, resentment, or gratitude. What sets these reactions apart, even from other reactions that reflect the importance that we accord to central aspects of our self-understandings, is their essential connection to others’ attitudes towards us. The pianist is devastated by—rather than offended by or re- sentful of—the loss of the use of her hands; however, she resents and her feelings are hurt by the way her musician friends lose interest in her company after the acci- dent. This is an illustration of the way that taking the participant stance manifests itself in the pianist’s reactions to her friends’ attitudes. The ways that others care about us are, needless to say, among the attitudes that are natural objects of our participant attitudes. It is profoundly important to us, for example, that our loved ones care about our feelings for them; if they didn’t care about our feelings for them, our relationships would be utterly transformed—even if they continued to care for us in other ways, for instance by caring about our well- Responsibility and the Emotional Structure of Relationships 148 being. 29 And by the same token, it often matters greatly to us whether and how oth- ers care about our mistreatment of, or lack of regard for, them. When someone we have mistreated becomes angry and blames us for our lack of regard, this trans- forms—even if only locally and temporarily—our relations with them. The attention and concern that we devote to these kinds of transformations are so pervasive and familiar as to become in many instances practically invisible, but they reflect a deep emotional investment. It is thus significant for understanding blame and its force that blame itself is (in many instances) a participant attitude, and moreover that blame is a possible— and indeed a natural and frequent—object of other participant attitudes. Focus first on the idea that blame can itself be a participant attitude. When someone mistreats or disrespects us we typically do not only experience pique at the fact that we have been on the receiving end of a moral norm violation; we typically also (like the pian- ist abandoned by her friends) take it personally. 30 We do not like to be mistreated or disrespected; to be so treated touches on our sense of ourselves and our place in the world; and we quite naturally take such treatment personally and react with a certain degree of heat and defiance. This heated reaction is a participant form of anger, and this participant anger just is blame, according to one familiar use of that term. This proposal about the nature of (one familiar kind of) blame captures the emotional dimension of blame in a way that is similar to Wallace’s account, but does so in a way that I think is more faithful to the personal dimension of blame. Blame, according to this proposal, is a manifestation of our capacity to care or value—in this case, to care about others’ attitudes towards us. But unlike in Wallace’s account, the relevant kind of caring is not grounded in the internalization of moral (or other rela- tional) values; instead it is a function of the way in which our self-understandings 29 Seth Shabo has made this point: “caring about how one’s partner in a romantic relationship treats one, regards one, and feels towards one is partly constitutive of mature, reciprocal love. To see this, suppose that Ben, after spending several years in a romantic relationship with Chris, comes to realize that he genuinely doesn’t care now what Chris thinks or feels about him . . . . Although he may still care deeply about Chris’s well-being, his feelings towards Chris don’t count as feelings of roman- tic love” (2011: 111-12). 30 Indeed, we may take it personally without considering the conduct’s moral status at all. For an exploration of this possibility, see John Deigh’s “Psychopathic Resent- ment” (2011b). Chapter 4 | The Emotional Economy of Responsibility 149 incorporate others’ attitudes towards us. In this sense, the significance that blame accords to others’ attitudes is thus not (only) moral, but personal. It is also worth noting that this proposal provides an independently moti- vated way, distinct from Wallace’s, to explain the special force that blame has in comparison to other, more “generic” or impersonal emotional responses. The differ- ence is not that blame—unlike “generic” sadness, distress, or disappointment—“is particularly appropriate to” (2011: 368) or “uniquely answers to” (2011: 369) the rela- tional character of moral values, but rather that blame is a participant attitude that manifests the way in which we treat others’ attitudes as central to our own self- understandings. In this respect blame is unlike impersonal or non-participant forms of sadness, distress, disappointment, or indeed anger; but conversely in this respect blame is akin to participant forms of sadness, distress, and disappointment. (This ex- plains, as Wallace’s account cannot, why it is that for instance the disappointment of others can have a very similar force to blame.) In addition to this, the current proposal also allows for an integration of the idea of caring and emotional susceptibility from Wallace’s account with something like the structure of Hieronymi’s account, to achieve a more complete understanding of the force of blame. We have seen that Hieronymi takes the view that emotion or affect cannot carry the distinctive force of blame. I have sided with Wallace in think- ing that Hieronymi’s reasons for thinking this are not decisive. But her rejection of affect is not essential to the structure of her account; the basic structure of her posi- tive view of blame is open-textured enough to allow for attitudes other than descriptive judgments or assessments to carry the force of blame. She holds that merely descriptive judgments of ill will have a distinctive force because others’ judg- ments on this matter are important to us, in virtue of their role in constituting relations in which mutual regard is recognized. This suggests that other types of atti- tudes might have a distinctive force in virtue of the constitutive role that those attitudes play in interpersonal relations that matter to us. In particular, the possibility lies open that interpersonal emotional responses, and particularly participant responses, could play a significant role in constituting rela- tions that are important to us. And we have already seen examples suggesting that they in fact do this. It is part of reciprocal, intimate relationships, for example, that Responsibility and the Emotional Structure of Relationships 150 each party cares—in the personal way that is distinctive of the participant attitudes— about how the other feels about them. Real friendship, to take one instance of this dynamic, involves not only an appreciation of the way in which the other person wishes you well, enjoys your company, and so on, but also an appreciation of the way in which the other person appreciates that you wish them well, enjoy their company, and so on. These lattices of reciprocal participant attitudes are an indispensible part of the structure of our most significant interpersonal relationships. And by the same token, a change in one of the participant attitudes that constitutes these structures is significant for the character of the relationship. If my friend ceases to care in a per- sonal way about how I feel about her, then our relationship has become, at least to some meaningful degree, estranged. Blame is one of the participant attitudes that plays this sort of role. It is a manifestation of the personal way in which the blaming party cares about the agent’s lack of regard, and as such blame itself can constitute a significant change in the rela- tions between the parties. In a sense, blame puts the parties into a new or different relationship than they were before. In explaining the force of blame, we can thus avail ourselves of the structure outlined by Hieronymi’s account. We can say that the force of blame consists in the significance that blame itself has for relations between the parties, and in the fact that, at least ordinarily, the agent has good reason to regret this significance. This latter point is one of the main lessons to be learned from Hi- eronymi—that the force of blame involves the significance of blame for the blamed person. 31 Again, I offer this proposal as a natural and intuitive gloss on our common experience of blaming and suffering blame. Most of us are, I think, deeply emotion- ally invested in being seen as good willed, trustworthy, kind, reliable, and so on by other members of our community. When one of them perceives that we have mis- treated or disrespected them, takes this personally, and responds with heat and 31 In an in-progress paper titled “I’ll Bet You Think This Blame Is About You,” Hi- eronymi argues that blame is not about the target of blame in a way that might give the target of accurate blame (that is, blame constituted by accurate judgments) grounds for complaint that the force of blame is unfairly imposed. But I do not think that this point is in conflict with the lesson that I have derived from her 2004 pa- per—indeed, the in-progress paper expands upon and defends a position adopted in the 2004 paper. Chapter 4 | The Emotional Economy of Responsibility 151 defiance, this reaction in itself represents a significant change in our relations with them. The fact that they care in this personal way about how we have treated them tends to go straight to our expectation and aspiration to be seen (and valued) as a good-willed participant in the relationship. It is quite natural in such a case for us, in turn, to take their blame personally. When we do, this is a way of recognizing and re- sponding to the force of blame. In other words, given the force of blame and its significance for relations that are important to us, we should expect to find that blame itself is a natural object of participant attitudes. And indeed this is the case. 5 Shame and the force of blame The foregoing account of blame and its force puts us in a position to better appreciate blame’s role in the interpersonal emotional economy of responsibility. Blame’s force consists in the significance that it has for the agent and for relations between the blaming party and the agent; this is an aspect of the personal dimension of blame. And because blame’s force consists in its significance for the agent, we should expect blame to be associated with a reaction on the part of the agent that registers and responds to this significance. Indeed, we might expect to find a range of such responses, depending on the particulars of different blaming transactions (the standing of the parties, the nature of their reciprocal emotional investment, etc.). But here it will be useful to focus on one fundamental response that is I think a natu- ral and typical reaction to blame: shame. The idea of focusing on shame may seem surprising; blame is usually associ- ated in the theoretical literature not with shame but with guilt. Wallace, for example, names resentment, indignation, and guilt as the reactive attitudes that constitute blame, and conceives of guilt as the self-directed version of blame. Similarly in Hi- eronymi’s account, just as another’s judgment that you have shown ill will carries much of the force of blame, “[i]f you acknowledge it’s truth—if you make the judg- ment of yourself—that acknowledgement carries the corresponding sense of guilt” (2004: 124). Here guilt is again the self-directed version of blame. This appears to be the standard view of guilt in the responsibility literature, which makes it perhaps natural to suppose that guilt would be the most fitting or appropriate response to Responsibility and the Emotional Structure of Relationships 152 blame (though to be clear, as far as I am aware neither Wallace nor Hieronymi makes this further claim). But this perhaps initially natural thought does not, on inspection, do justice to the force of blame as we have elaborated it. According to the standard view, guilt carries the same force as blame but from, as it were, a different direction. The force of guilt thus multiplies or repeats the force of blame—guilt is not on this view a re- sponse to the force of blame. While we should expect that where there is blame (particularly justified blame) there will often also be self-blame or guilt, our observa- tions about the force of blame also lead us to expect something more. Given the significance that blame has for the blamed party and the relationship that she stands in to the blamer, we should expect to find a response that registers and answers to blame, rather than merely repeating it. Shame is one such response. 32 Chapter 3 developed a participant-attitude conception of shame according to which shame is essentially a response to attitudes towards the subject of shame. This makes blame or resentment 33 a potential object of shame. Moreover, though chapter 3 focused on the structure of shame rather than on its valence or general character, it is clear that shame is a negative reaction; shame registers that something is, in some sense, not as the subject would wish it to be. Given the ordinarily adverse significance of blame for the blamed party, this makes shame a potentially fitting reaction to blame. Now return to guilt. Whereas in guilt you hold yourself responsible, shame is a way of registering the significance of the fact that someone is holding you respon- sible. This makes shame responsive to blame in a direct way that guilt is not. Guilt responds to one’s own misconduct or one’s own attitudes (e.g., lack of regard) to- wards others, and it registers the importance that one accords to one’s own good will towards others, but it is not itself a response to blame; guilt works rather in parallel to blame. But shame can serve as a response to blame that registers the importance 32 There may be other attitudes that can play this role; I only claim that shame is one of them. Shame can of course also be a response to a very broad range of other atti- tudes towards the subject, in addition to blame. 33 Or indeed guilt. According to my account of shame, guilt is—along with resent- ment, contempt, derision, etc.—one of the possible objects of shame. Chapter 4 | The Emotional Economy of Responsibility 153 that the subject of shame accords to the blaming person’s attitudes towards her. In a slogan: Guilt takes misconduct seriously, shame takes blame seriously. Not only can shame respond to blame, it can also answer to blame in a way that makes blame and shame a kind of complementary or matched pair. 34 Shame, felt in response to blame, is an affective manifestation of the fact that the subject takes the other’s blame seriously—that it matters adversely to the subject, on a personal level, that she is blamed. This means that shame is a reflection of the fact that the subject accords the blaming party’s attitudes towards her a certain weight or impor- tance. As Heidi Maibom has described in her discussion of the evolutionary descent and significance of shame, this tends naturally to mollify the blaming party’s concern that she has not figured into the subject’s attitudes in the way she would have wished, precisely because it shows how the blaming person’s attitudes matter, per- sonally and emotionally, to the agent. 35 This represents at best a first step towards a full picture of blame’s role in the emotional economy of responsibility. But it starts to put some of the missing flesh onto the bones of the now-common idea that blame has a distinctive force. Blame has a force that is independent of its further effects or consequences; but at the same time, if blame’s force is real we should expect to be able to trace its ordinary effects 34 For the idea of a matched pair of emotional responses, see Gibbard 1990: 133. Gibbard argues that shame and disdain, and guilt and anger, form matched pairs of responses (1990: 139-40), but Gibbard thinks of guilt in a different way than is stan- dard in the reactive attitude literature. Guilt, Gibbard says, is not “self-directed anger; feeling guilty is different from feeling you could kick yourself. Rather, guilt is coordi- nated with anger in a special way: it aims to placate anger, and is governed by the same norms as govern anger” (1990: 139). This description is highly abbreviated, but if we can read Gibbard as using “guilt” to refer to a particular class of (what I would call) shame reactions (namely, those that are suited to placate anger, and are governed by the same norms as anger), then these remarks would be consistent with my view of shame, guilt, and blame. 35 Compare Maibom’s description of the display function of shame: “The person who is ashamed shows to others—through the shame display—not just a recognition that they have failed to live up to public expectations, but also that they have an ad- verse emotional reaction to it. The experience of shame confirms the social grounding of the subject. She demonstrates to others that their opinion matters to how she conducts herself, how she decides to live her life. Her shame indicates that she can be counted on to live a life with others within the constraints set by the community” (2010: 587-88). Responsibility and the Emotional Structure of Relationships 154 and consequences and describe their significance in human relations. Conceiving of blame as a participant attitude suggests how we might go about doing this. Conclusion I have argued that, despite recent advances in our understanding of blame, there remains a familiar and important dimension of blame’s force that has not yet been adequately reckoned with, largely due to a tendency to intellectualize and de- personalize blame. As illustrations both of the important progress that has been made and of the theoretical tendencies that have left us with this persistent gap, I have focused on Pamela Hieronymi’s and R. Jay Wallace’s accounts of blame. In contrast to their views, I have suggested that the special force of many paradigmatic instances of blame is precisely a function of blame’s emotional, personal character, which we can understand as aspects of its status as a participant attitude. Finally, I have suggested that there is a fit between blame and shame that illustrates blame’s role in an interpersonal emotional economy of responsibility. I hope that this helps to situate both blame and shame within the broader emotional structure of our rela- tionships. 155 Conclusion I would like to conclude by mentioning a set of topics that have been nota- ble, perhaps, by their absence. I have used responsibility as a point of entry into broader questions about interpersonal relationships; I have sought to develop an ac- count of a distinct class of attitudes that play a constitutive, structuring role in our relationships; and I have demonstrated the fruitfulness of this account by proposing resolutions to particular theoretical issues that arise in connection with social phe- nomena like shame and blame. The larger goals of this inquiry have been twofold: first, to show that a better understanding of the underlying emotional structures of our relationships can yield a more complete, true-to-life, and fully integrated under- standing of our responsibility practices; and second, to develop a general account of those underlying emotional structures that puts us in a position to appreciate the quite fundamental role that they play in organizing our lives. But for all the talk of relationships, there has been very little direct discussion of actual, lived relationships, or of the relationship types—romantic, friendly, famil- ial, professional, civic, moral, and so on—into which the most significant of our actual relationships generally fall. Nor has there been discussion of the value or point or significance of these various relationships, beyond the bromide that some of them are very important to us and form a substantial part of what makes life worth living. Nor, moreover, has there been discussion of what distinguishes a good or healthy relationship from a bad or dysfunctional one; nor of how we, as individuals, ought to approach our existing relationships; nor of how many and what kinds of relation- ships we ought to try to cultivate, and in what combination and with whom; nor of how the value of our relationships and the reasons that are grounded in them inter- act with other values, reasons, and obligations, including moral ones. Indeed, the word “moral” has hardly appeared, other than as necessary to discuss others’ work. Responsibility and the Emotional Structure of Relationships 156 I wish to be upfront about these omissions, and to concede their signifi- cance—the omitted topics are of the greatest interest and theoretical importance. But I wish also to insist on the indispensible importance of the spadework undertaken here. I have focused not on full-blown relationships and their elaborate structures, but on the most basic emotional building blocks out of which those relationships are made, and on how those building blocks fit together. This has not been out of an effort to avoid richness or complexity, but rather out of a suspicion that we will only fully appreciate the richness and complexity (or, seen in another way, the messiness ) of our actual relationships by appreciating the specific contributions of their various constituent elements—rather in the way that the richness of an orchestra’s sound can be more fully appreciated by listening to the part of each instrument separately, and then listening again to the whole. Listening to music in this way can produce a breakthrough—an epiphany of appreciation for the organic structure, unity, and significance of pieces that previ- ously sounded impenetrably dense and complex. I think similar breakthroughs are possible in understanding the vibrant complexity of social phenomena. For this rea- son, I am not at all pessimistic about our prospects for arriving at satisfying answers to the formidable questions that I have deferred here concerning the constitution, character, variety, value, health, maintenance, and morality of our actual relation- ships. Indeed, I hope to have prepared the ground to proceed directly to some of those questions, and I am eager to start work on that next step. 157 References Abramson, Kate & Adam Leite. 2011. “Love as a Reactive Emotion.” The Philosophical Quarterly 61(245): 673-99. Allais, Lucy. 2008. “Wiping the Slate Clean: The Heart of Forgiveness.” Philosophy & Public Affairs 36(1): 33-68. Baldwin, Mark W. & Jodene R. Baccus. 2004. “Maintaining a Focus on the Social Goals Underlying Self-Conscious Emotions.” Psychological Inquiry 15(2): 139-44. Benedict, Ruth. 1946. The Chrysanthemum and the Sword: Patterns of Japanese Culture. The New American Library. Bennett, Christopher. 2008. The Apology Ritual: A Philosophical Theory of Punishment. Cambridge University Press. Bennett, Christopher. 2013. “The Expressive Function of Blame,” in D. Justin Coates & Neal A. Tognazzini (eds.), Blame: Its Nature and Norms: 66-83. Oxford University Press. Bennett, Jonathan. 1980. “Accountability,” in Zak van Straaten (ed.), Philosophical Subjects: Essays Presented to P.F. Strawson: 14-47. Clarendon Press, Oxford University Press. Buss, Sarah. 1999. “Respect for Persons.” Canadian Journal of Philosophy 29(4): 517-50. Calhoun, Cheshire. 2004. “An Apology for Moral Shame.” Journal of Political Philosophy 12(2): 127-46. Coates, D. Justin & Neal Tognazzini. 2012. “The Nature and Ethics of Blame.” Philosophy Compass 7(3): 197-207. Darwall, Stephen. 2006. The Second-Person Standpoint: Morality, Respect, and Accountability. Harvard University Press. Deigh, John. 1983. “Shame and Self-Esteem: A Critique,” as reprinted in his The Sources of Moral Agency: 226-47 (1996). Cambridge University Press. Deigh, John. 2011a. “Reactive Attitudes Revisited,” in Carla Bagnoli (ed.), Morality and the Emotions: 197-216. Oxford University Press. Deigh, John. 2011b. “Psychopathic Resentment,” in Samantha Konzelmann Ziv, Keith Lehrer & Hans Bernhard Schmid (eds.), Self-Evaluation: Affective and Social Grounds of Intentionality, Philosophical Studies Series 116: 155-69. Deonna, Julien & Fabrice Teroni. 2011. “Is Shame a Social Emotion?”, in Samantha Konzelmann Ziv, Keith Lehrer & Hans Bernhard Schmid (eds.), Self-Evaluation: Responsibility and the Emotional Structure of Relationships 158 Affective and Social Grounds of Intentionality, Philosophical Studies Series 116: 193- 212. Deonna, Julien A., Raffaele Rodogno & Fabrice Teroni. 2012. In Defense of Shame: The Faces of an Emotion. Oxford University Press. Duff, R.A. 2001. Punishment, Communication, and Community. Oxford University Press. Duff, R.A. 2003. “Penance, Punishment and the Limits of Community.” Punishment and Society 5(3):295-312. Elison, Jeff. 2005. “Shame and Guilt: A Hundred Years of Apples and Oranges.” New Ideas in Psychology 23: 5-32. Enoch, David & Andrei Marmor. 2007. “The Case Against Moral Luck.” Law and Philosophy 26: 405-36. Feinberg, Joel. 1970. Doing and Deserving: Essays in the Theory of Responsibility. Princeton University Press. FitzGerald, Chloë. 2015. “Defending Shame.” Journal of Moral Philosophy 12: 231-54. Galligan, Phillip. 2014. “Shame, Publicity, and Self-Esteem.” Ratio (new series) (March 2016) 29(1): 57-72. Gibbard, Allan. 1990. Apt Choices, Wise Feelings: A Theory of Normative Judgment. Clarendon Press, Oxford University Press. Glover, Jonathan. 1970. Responsibility. Routledge & Kegan Paul. Hieronymi, Pamela. 2004. “The Force and Fairness of Blame.” Philosophical Perspectives (Ethics) 18: 115-48. Hieronymi, Pamela. MS. “Freedom, Resentment, and the Metaphysics of Morals” (dated September 21, 2015; available at https://dl.dropboxusercontent.com/u /41585851/In%20Progress/Hieronymi%20on%20Strawson.pdf). Hurley, Elisa A. & Coleen Macnamara. 2010. “Beyond Belief: Toward a Theory of the Reactive Attitudes.” Philosophical Papers 39(3): 373-99. Kekes, John. 1988. “Shame and Moral Progress.” Midwest Studies in Philosophy 13: 282- 96. Kemeny, Margaret E., Tara L. Gruenewald & Sally S. Dickerson. 2004. “Shame as the Emotional Response to Threat to the Social Self: Implications for Behavior, Physiology, and Health.” Psychological Inquiry 15(2): 153-60. Kolodny, Niko. 2003. “Love as Valuing a Relationship.” Philosophical Review 112(2): 135-89. Kolodny, Niko. 2010. “Which Relationships Justify Partiality? The Case of Parents and Children.” Philosophy & Public Affairs 38(1): 37-75. Leary, Mark R. 2007. “How the Self Became Involved in Affective Experience; Three Sources of Self-Reflective Emotions,” in Jessica Tracy, Richard W. Robins & June Price Tangney (eds.), The Self-Conscious Emotions: Theory and Research: 38-52. The Guilford Press. References 159 Lee, Ambrose. 2017. “Defending a Communicative Theory of Punishment: The Relationships between Hard Treatment and Amends.” Oxford Journal of Legal Studies 37(1): 217-37. Lewis, Helen B. 1971. “Shame and Guilt in Neurosis.” Psychoanalytic Review 58(3): 419-38. Macnamara, Coleen. 2011. “Holding Others Responsible.” Philosophical Studies 152: 81-102. Macnamara, Coleen. 2013. “Taking Demands Out of Blame,” in D. Justin Coates & Neal A. Tognazzini, eds., Blame: Its Nature and Norms: 141-61. Oxford University Press. Maibom, Heidi L. 2010. “The Descent of Shame.” Philosophy and Phenomenological Research 80(3): 566-94. Mason, Michelle. 2010. “On Shamelessness.” Philosophical Papers 39(3): 401-25. McGeer, Victoria. 2012. “Co-reactive Attitudes and the Making of Moral Community,” in Robyn Langdon & Catriona Mackenzie (eds.), Emotions, Imagination, and Moral Reasoning: 299-325. Taylor and Francis. McKenna, Michael. 2012. Conversation and Responsibility. Oxford University Press. McKenna, Michael. 2013. “Directed Blame and Conversation,” in D. Justin Coates & Neal A. Tognazzini (eds.), Blame: Its Nature and Norms: 119-43. Oxford University Press. Nagel, Thomas. 1998. “Concealment and Exposure.” Philosophy & Public Affairs 27(1): 3-30. Nussbaum, Martha. 2004. Hiding from Humanity: Disgust, Shame, and the Law. Princeton University Press. O’Hear, Anthony. 1976. “Guilt and Shame as Moral Concepts.” Proceedings of the Aristotelian Society (New Series) 77: 73-86. Pickard, Hannah. 2013. “Irrational Blame.” Analysis 73(4): 613-26. Piers, Gerhart & Milton B. Singer. 1953. Shame and Guilt: A Psychoanalytic and a Cultural Study. Charles C. Thomas. Radzik, Linda. 2009. Making Amends: Atonement in Morality, Law, and Politics. Oxford University Press. Railton, Peter. 2009. “Practical Competence and Fluent Agency,” in David Sobel & Steven Wall (eds.), Reasons for Action: 81-115. Cambridge University Press. Rawls, John. 1971. A Theory of Justice. The Belknap Press of Harvard University Press. Raz, Joseph. 2011. From Normativity to Responsibility. Oxford University Press. Rosen, Gideon. 2004. “Skepticism About Moral Responsibility.” Philosophical Perspectives 18(1): 295-313. Russell, Paul. 2013. “Responsibility, Naturalism, and the ‘Morality System,’” in David Shoemaker (ed.), Oxford Studies in Agency and Responsibility 1: 184-204. Responsibility and the Emotional Structure of Relationships 160 Scanlon, T.M. 1998. What We Owe to Each Other. The Belknap Press of Harvard University Press. Scanlon, T.M. 2008. Moral Dimensions: Permissibility, Meaning, Blame. The Belknap Press of Harvard University Press. Scanlon, T.M. 2013. “Interpreting Blame,” in D. Justin Coates & Neal A. Tognazzini (eds.), Blame: Its Nature and Norms: 84-99. Oxford University Press. Scanlon, T.M. 2015. “Forms and Conditions of Responsibility,” in Randolph Clarke, Michael McKenna, and Angela M. Smith (eds.), The Nature of Moral Responsibility: 89-111. Oxford University Press. Scheffler, Samuel. 1997. “Relationships and Responsibilities.” Philosophy and Public Affairs 26(3): 189-209. Shabo, Seth. 2011. “Where Love and Resentment Meet: Strawson’s Interpersonal Defense of Compatibilism.” Philosophical Review 121(1): 95-124. Shabo, Seth. 2012. “Incompatibilism and Personal Relationships: Another Look at Strawson’s Objective Attitude.” Australasian Journal of Philosophy 90(1): 131-47. Sher, George. 2006. In Praise of Blame. Oxford University Press. Sher, George. 2009. Who Knew? Responsibility Without Awareness. Oxford University Press. Sher, George. 2013. “Wrongdoing and Relationships: The Problem of the Stranger,” in D. Justin Coates & Neal A. Tognazzini (eds.), Blame: Its Nature and Norms: 49- 65. Oxford University Press. Smith, Angela M. 2005. “Responsibility for Attitudes: Activity and Passivity in Mental Life.” Ethics 115 (2): 236-71. Smith, Angela. 2013. “Moral Blame and Moral Protest,” in D. Justin Coates & Neal A. Tognazzini (eds.), Blame: Its Nature and Norms: 27-48. Oxford University Press. Stocker, Michael. 1976. “The Schizophrenia of Modern Ethical Theories.” Journal of Philosophy 73(14): 453-66. Strawson, Peter. 1962. “Freedom and Resentment,” as reprinted in Gary Watson (ed.), Free Will: 72-93 (2003, 2nd ed.). Oxford University Press. Talbert, Matthew. 2012. “Moral Competence, Moral Blame, and Protest.” Journal of Ethics 16: 89-109. Tannenbaum, Julie. 2015. “Mere Moral Failure.” Canadian Journal of Philosophy 45(1): 58-84. Tasioulas, John. 2006. Punishment and Repentance. Philosophy 81(2): 279-322. Tasioulas, John. 2007. Repentance and the Liberal State. Ohio St. J. Crim. L. 4: 487- 521. Taylor, Gabriele. 1985. Pride, Shame, and Guilt: Emotions of Self-Assessment. Clarendon Press, Oxford University Press. References 161 Tognazzini, Neal & D. Justin Coates. 2016. “Blame,” in Edward N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Spring 2016 Edition) URL = <https://plato.stanford.edu/archives/spr2016/entries/blame/>. Tracy, Jessica L. & Richard W. Robins. 2004. “Keeping the Self in Self-Conscious Emotions: Further Arguments for a Theoretical Model.” Psychological Inquiry 15(2): 171-77. Tracy, Jessica L. & Richard W. Robins. 2007. “The Self in Self-Conscious Emotions: A Cognitive Appraisal Approach,” in Jessica Tracy, Richard W. Robins & June Price Tangney (eds.), The Self-Conscious Emotions: Theory and Research: 3-20. The Guilford Press. Velleman, J. David. 2001. “The Genesis of Shame.” Philosophy & Public Affairs 30(1): 27-52. Walker, Margaret Urban. 2006. Moral Repair: Reconstructing Moral Relations After Wrongdoing. Cambridge University Press. Wallace, R. Jay. 1994. Responsibility and the Moral Sentiments. Harvard University Press. Wallace, R. Jay. 2011. “Dispassionate Opprobrium: On Blame and the Reactive Sentiments,” in R. Jay Wallace, Rahul Kumar & Samuel Freeman (eds.), Reasons and Recognition: Essays on the Philosophy of T.M. Scanlon: 348-72. Oxford University Press. Wallace, R. Jay. 2014. “Emotions and Relationships: On a Theme from Strawson,” in David Shoemaker and Neal A. Tognazzini (eds.), Oxford Studies in Agency and Responsibility 2: 119-42. Watson, Gary. 1996. “Two Faces of Responsibility.” Philosophical Topics 24(2): 227-48. Watson, Gary. 1987. “Responsibility and the Limits of Evil: Variations on a Strawsonian Theme,” as reprinted in his Agency and Answerability 219-59 (2004). Clarendon Press, Oxford University Press. Watson, Gary. 2014a. “Peter Strawson on Responsibility and Sociality,” in David Shoemaker & Neil Tognazzini (eds.), Oxford Studies in Agency and Responsibility 2: 15-32. Watson, Gary. 2014b. “Raz on Responsibility.” Criminal Law and Philosophy (2016) 10: 395-409. Williams, Bernard. 1993. Shame and Necessity. University of California Press. Wolf, Susan. 2011. “Blame, Italian Style,” in R. Jay Wallace, Rahul Kumar & Samuel Freeman (eds.), Reasons and Recognition: Essays on the Philosophy of T.M. Scanlon: 332- 47. Oxford University Press. Wolf, Susan. 2015. “Responsibility, Moral and Otherwise.” Inquiry 58(2): 127-42. Wollheim, Richard. 1999. On the Emotions. Yale University Press. Zimmerman, Michael. 1988. An Essay on Moral Responsibility. Rowman & Littlefield.

Asset Metadata

Creator Bero, Stephen Leonard (author)

Core Title Responsibility and the emotional structure of relationships

Contributor Electronically uploaded by the author (provenance)

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Philosophy

Publication Date 09/12/2019

Defense Date 07/06/2017

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag Blame,Emotions,guilt,oai:digitallibrary.usc.edu:usctheses,OAI-PMH Harvest,reactive attitudes,Relationships,responsibility,Shame,Strawson

Language English

Advisor Watson, Gary (committee chair), Keating, Gregory (committee member), Quong, Jonathan (committee member), Schroeder, Mark (committee member)

Creator Email bero@usc.edu,stephenbero@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c40-428565

Unique identifier UC11263634

Identifier etd-BeroStephe-5713.pdf (filename),usctheses-c40-428565 (legacy record id)

Legacy Identifier etd-BeroStephe-5713.pdf

Dmrecord 428565

Document Type Dissertation

Rights Bero, Stephen Leonard

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA

Abstract (if available)

Abstract The relational character of our moral practices has featured prominently in recent normative theory, including the work of Stephen Darwall, T.M. Scanlon, R. Jay Wallace, and others. A central theme of this work is the significance of the interpersonal emotional responses known, after P.F. Strawson, as the “reactive attitudes.” For Strawson the reactive attitudes encompassed the broad range of responses that we are susceptible to as participants in interpersonal relationships—resentment, gratitude, hurt feelings, love, shame, and so on (as distinct from non-relational responses like impersonal fear, sadness, or disgust). ❧ But Strawson never tried to say exactly what all of these diverse attitudes have in common, and his classification has widely been deemed unduly broad. The reactive attitudes are now generally conceived as a narrow class of moralized responses closely associated with responsibility and blame. ❧ This dissertation is motivated by the sense that something important has been lost in this shift, and that Strawson’s original idea—that our relationships are structured and regulated by a broad but unified set of emotional responses—represents a rich and fundamental insight into the nature of our sociality. But at the same time, developing this insight in its full generality should also expand our resources for understanding responsibility in particular. The dissertation accordingly develops an account of a distinct class of attitudes that play a constitutive, structuring role in our relationships, and then use this account to shed light on the nature of responsibility and our responsibility practices.