Hanno Sauer University of Leiden, the Netherlands

Transcription

1 The Appropriateness of Emotions: Moral Judgment, Moral Emotions, and the Conflation Problem Hanno Sauer University of Leiden, the Netherlands ABSTRACT. What is the connection between emotions and moral judgments? Neo-sentimentalism maintains that to say that something is morally wrong is to think it appropriate to resent other people for doing it or to feel guilty upon doing it oneself. But intuitively, it seems that there is no way to characterize the content of guilt and resentment independent from the fact that these emotions respond to morally wrong actions. In response to this problem of circularity, modern forms of sentimentalism have favoured a no-priority view, arguing that judgments of moral wrongness cannot be reduced to expressions of feelings of guilt and resentment, but that emotional responses and moral judgments mutually elucidate each other. In the present contribution, I argue that this strategy is not successful: the problem of circularity returns at a deeper level of the account, a level at which the no-priority view can no longer escape it. The concept of appropriateness that is invoked by neo-sentimentalism is liable to the so-called conflation-problem : it fails to distinguish between right and wrong kinds of appropriateness. In order to draw that important distinction, neo-sentimentalism has to presuppose a substantive notion of moral wrongness already. Moreover, I show that the most influential contemporary attempts to achieve an independent, non-circular fix on the emotions fail for one of the following three reasons: they either cease to be sentimentalist, for capture the normative dimension of moral judgment or end up being circular again. KEYWORDS. Moral judgment, moral emotions, meta-ethics, neo-sentimentalism, conflation problem, Justin D Arms, Daniel Jacobson I. INTRODUCTION According to contemporary neo-sentimentalists, to say that X is wrong is to express an attitude towards X that consists in the endorsement ETHICAL PERSPECTIVES 18, no. 1(2011): by European Centre for Ethics, K.U.Leuven. All rights reserved. doi: /EP

2 ETHICAL PERSPECTIVES MARCH 2011 of (a system of) norms that warrants a certain negatively charged emotional response. Neo-sentimentalism claims that making a moral judgment is making a judgment about the appropriateness of reactive attitudes such as guilt or resentment, shame or disdain. With this account in mind, X is wrong simply means: it is appropriate for a person A to feel guilty upon doing X and for other persons to resent A for doing X (Gibbard 1992, 200ff.). In analytic meta-ethics, neo-sentimentalism is the legitimate heir to emotivism s throne; it promises to capture the expressive dimension of moral judgment in a way that avoids the most ruinous objections against its less sophisticated predecessor. Accordingly, neo-sentimentalism does not claim that when persons make moral judgments they simply frown upon a deed (traditionally, this was given an analysis using the notorious boo-operator ). It argues rather that frowning upon the said deed is appropriate, justified or makes sense (Gibbard 1986). Intuitively, it seems that neo-sentimentalism is getting things backwards. It seems that there is no way for us to characterize feelings of guilt and resentment independent from the fact that these feelings respond to moral wrongs. Hence there is no way for us to understand the cognitive content of these emotions independent from an understanding of what a moral judgment is. Neo-sentimentalism, however, goes down precisely that route, arguing that moral wrongness has to be analyzed in terms of justified emotional responses. Confronted with this disagreement about the proper explanatory direction, who then is right about the issue? At this point, the problem boils down to a question of priority and independence. If we want to characterize what makes morally wrong acts morally wrong, where do we have to start? Recent neo-sentimentalists favour a strategy that has been dubbed the no-priority view (McDowell 1998). According to this account, moral judgments have to be understood in terms of judgments about the appropriateness of emotions, but cannot be reduced to them. Rather, moral judgments and emotional reactions mutually elucidate each other, 108

3 SAUER THE APPROPRIATENESS OF EMOTIONS and neither has a priority over the other in constituting moral judgment. The problem of priority is solved, and the threat of circularity disappears. In a series of papers, Justin D Arms and Daniel Jacobson have convincingly argued that neo-sentimentalist accounts of moral judgment face a serious challenge that has not yet been sufficiently met (D Arms and Jacobson 2000a and 2000b). According to my understanding of their argument, the conflation problem, as D Arms and Jacobson call it, primarily illustrates that the no-priority view does not offer a workable solution to the problems mentioned above and that it is not a viable alternative for neo-sentimentalists to endorse. Moreover, I argue that none of the solutions to the conflation problem that have been offered by sentimentalists succeeds, and that neo-sentimentalists run into what I call the sentimentalist s trilemma: in trying to meet the challenge posed by the conflation problem, neo-sentimentalism either (i) ceases to be sentimentalist; (ii) does not offer a distinction between relevant and irrelevant considerations in favor of an emotional response; or (iii) deflates the normativity of appropriate emotional responses to mere actual emotional responses. In the present paper, I shall begin with a brief outline of neo-sentimentalism (ii) and reconstruct how the conflation problem arises in the analysis of sentimentalist concepts in general (iii). I endeavour to demonstrate how different accounts of the connection between the retributive emotions and moral judgment attempt to solve the problem by developing an independent naturalistic fix on these emotions. I distinguish between two types of naturalism: natural kinds-naturalism and evolutionary naturalism. Natural kinds-naturalists hold that the conditions under which a reactive attitude is appropriate has to be understood in terms of the marks these emotions typically respond to (iv) or the features a moral emotion fits due to the way it presents its object (v). Evolutionary naturalists maintain that the appropriateness of emotional reactions is a 109

4 ETHICAL PERSPECTIVES MARCH 2011 function of the degree to which they contribute to smooth social coordination (vi) or accurately represent the kinds of objects they have been developed to be reliably caused by (vii). I argue that both types of naturalism are impaled on at least one horn of the above trilemma. For methodological reasons, I follow recent sentimentalists in not focusing exclusively on the reactive emotions, but also on more simple examples like amusement, envy, or sadness. Once we have understood the nature of these emotions, neo-sentimentalists hope, the account can be applied to moral emotions as well. In addition, I leave out the question whether some emotions like disgust or amusement are even governed by norms of appropriateness to begin with. I assume that the retributive emotions are governed by these norms, and that more basic emotions might help understand how we can make sense of this fact. I conclude with some remarks about the general prospects of the reactive attitudes account of moral judgment (viii). The issues raised in this paper are of particular importance for questions of moral education and the way in which human agents can be educated to act virtuously. One of the main goals of moral education is to enable persons to act in morally good ways not by mere accident, but reliably. If, however, subjects have not learned how to respond to the right kinds of reasons in arriving at a decision about what to do, it seems that the connection between the considerations subjects pick up on and their morally good behaviour is inevitably going to remain fluky, and will always remain a matter of luck. A virtuous person is a person who reliably does the right thing; and this can only be achieved by a proper understanding of the reasons that genuinely bear on whether something is the morally right thing to do or not. If we do not have a firm grasp of what it is for a consideration to bear on the moral (in)appropriateness of an action, our attempts to communicate to the moral novice what constitutes good character and laudable action are doomed to fail. 110

5 SAUER THE APPROPRIATENESS OF EMOTIONS II. NEO-SENTIMENTALISM AND THE PRIORITY-PROBLEM For neo-sentimentalism, to make a moral judgment is not to have an emotional reaction towards an action but to endorse one: Neo-Sentimentalism (NS) To judge an action A to have some evaluative property P (goodness, badness, rightness, wrongness ) is to judge it appropriate (warranted, justified ) to have an associated emotional response (guilt, resentment, contempt ) towards A. It is important to note that in its classical form, sentimentalism aims to be a reductive doctrine. It promises to decompose the concept of moral wrongness (or rightness, respectively) into concepts that refer essentially to subjects emotional reactions towards certain actions or events. And crucially, this analysis is supposed to be exhaustive, that is, all moral concepts should turn out to be analyzable in a sentimentalist vocabulary. The sentimentalist owes an explanation of specially moral feelings guilt and impartial anger [ ] in terms that do not require a prior understanding of moral judgments (Darwall et al. 1992, 151). In order to see why this task is so difficult to carry out, one has to remember the following problem: if judgments of moral wrongness are supposed to be understood on the basis of a specific kind of disapproval, it is immediately striking that we must be able to differentiate moral disapproval from other types of disapprobation (this is Miller s moral attitude problem [2003, 88ff.]). But in the case of moral disapproval, the only plausible candidate is a cognitive judgment that the thing in question is morally wrong. If so, we need to understand judgments of wrongness before we can understand moral disapproval (Darwall et al. 1992, 149). At the very outset, sentimentalism has a problem of priority. Is it possible, after all, to carve out what is distinctive of moral disapproval, despite its being disapproval of morally wrong acts? Or can sentimentalists show that they do not owe such an explanation to begin 111

6 ETHICAL PERSPECTIVES MARCH 2011 with? In the following section, I shall discuss this problem in more detail. III. THE CONFLATION-PROBLEM AND THE RIGHT KIND OF REASON PRINCIPLE Let me first apply the above account of NS to particular types of moral disapproval. The content of moral judgments varies with respect to the question whether the judgment is first-personal or third-personal (this parallels a well-known distinction within moral psychology between selfdirected and other-directed moral emotions [Haidt 2003]). In the firstpersonal, self-directed case, the judgment It is wrong for me to do X can thus be analyzed as It is appropriate for me to feel guilty about doing X. In the complementary case, It is wrong for that person to do X can be translated into It is appropriate to resent that person for doing X. And the analysis also works for thick normative concepts: It is cowardly to do X turns into It is appropriate to feel ashamed about doing X, X is funny turns into It is appropriate to be amused by X and so on. What is important, and what sets NS apart from simple emotivism, is that the concept of appropriateness reappears in every instance. A major problem for sentimentalist analyses of evaluative concepts arises when examples like the following come up: Consider a wickedly clever joke told at the expense of a socially marginalized person or group. Someone sympathetic to the butt of such a joke might well think it inappropriate to be amused because the joke is cruel or offensive. But does this mean the joke isn t funny? (D Arms and Jacobson 2000a, 731). To affirm this question is to commit what D Arms and Jacobson call the moralistic fallacy : an inference from one type of (in)appropriateness moral (in) appropriateness, for instance to other types of (in)appropriateness. The lesson to be learned from examples like these is that there are different species of appropriateness, and that NS fails to distinguish them 112

7 SAUER THE APPROPRIATENESS OF EMOTIONS properly. Take the analysis of funny mentioned above. We said that X is funny iff it is appropriate to be amused by it. It is not appropriate, of course, to be amused by an offensive, politically incorrect joke. But still, the joke is funny, and so the sentimentalist analysis turns out to be inadequate. Consider another example. D Arms and Jacobson ask you to imagine that you have a rich and generous but touchy friend, who is extremely sensitive about his friends attitude toward his wealth. If he suspects you of envying his possessions, he will curtail his largesse. That is a good reason not to envy him, [ ] but surely it doesn t speak to whether his possessions are enviable (2000a, 731). The neo-sentimentalist says that something is enviable just in case you have good reason to envy it. Clearly, you don t have good reason prudential reason, this time to envy your friend, because you don t want to scare him away. But his wealth is still enviable and so the sentimentalist analysis fails again. It does not have the resources to distinguish between kinds of appropriateness when it comes to the normative assessment of an emotional response and is, hence, likely to conflate irrelevant considerations in favour of a feeling with relevant ones. Call this the Conflation Problem. Let me make a short detour. Prichard (1912) has famously argued that moral philosophy rests on a mistake if it tries to answer the question Why be moral? by appealing to the fact that virtuous conduct contributes to a person s individual happiness. This rationale just does not seem to capture adequately the special kind of normativity that is supposed to spring from moral demands, a kind of normativity that assuming that such a thing exists is thoroughly sui generis. People who strive for a morally good life simply because they are convinced that it will make them feel good in this life or the next miss the point of what it means to have a morally good character: they act on the wrong kind of consideration. Recently, this point has been nicely illustrated by Stephen Darwall (2006, 15ff.; also Strawson 1962). Imagine, he asks us, you were offered one million dollars to believe in God. Do you have a reason to believe in 113

8 ETHICAL PERSPECTIVES MARCH 2011 God? It depends. If your main goal is to make as much money as you can, accepting this peculiar offer will move you a step closer towards that goal. But if you think that you ought not to believe anything you are convinced is not true (you happen to be a die-hard atheist, by the way), then you do not have a reason to accept the offer. It seems that what we have in mind when we think about reasons for belief are epistemic reasons: considerations that count in favour of a proposition being true, not merely advantageous to believe. To put this more explicitly, we can propose the following principle: Right Kind of Reason (RKR) For a consideration to genuinely count as a reason to believe, do or feel something, it has to be the right kind of reason. Obviously, the principle is merely formal. What makes a consideration a consideration of the right kind is a question that can only be addressed with reference to particular types of rationally amenable attitudes. Here we can return to the appropriateness-analysis of evaluative terms. D Arms and Jacobson have the following suggestion. If, they argue, cases like the offensive joke and touchy friend examples illustrate that something can be inappropriate to laugh at yet be funny, or inappropriate to envy yet be enviable, what we have to look for if we are searching for a characteristic that distinguishes the right kind of reason in something s favour from the wrong kind are considerations that bear on whether some X has the property P. If I want to know whether something is shameful, I should not ask myself whether some or most people are actually ashamed of it, but whether it is appropriate to have that particular emotional attitude towards it. If you offer me one million dollars for not feeling ashamed about something I did, or the person I am, or what I look like, this, in a sense, renders my reaction inappropriate. But this doesn t bear on my situation being genuinely shameful or not, because it is not the right kind of reason. 114

9 SAUER THE APPROPRIATENESS OF EMOTIONS To sum up the argument so far, the Conflation Problem emerges from a combination of appropriateness analyses of normative concepts that are typical for NS with the RKR principle. In order to solve the problem, sentimentalists acknowledge, they will have to cash out the principle in more detail. Start with belief: Right Kind of Reason for belief: For a consideration to count as the right kind of reason for belief, it must bear on the truth of the belief. So far, so good. But we have to apply this to an example from the area of our interest, namely, evaluative attitudes. Take envy as an example. Here we get: Right Kind of Reason for envy: For a consideration to count as the right kind of reason for feeling envy towards something, it must bear on that something being enviable. Furthermore, and even more head-on, we can apply the principle to our main focus, the concept of moral wrongness, which, according to NS, can be reduced to appropriate retributive attitudes like guilt and resentment. What does a consideration have to bear on for it to be of the right kind? As Darwall, Gibbard and Railton have remarked in the above quote, there seems to be only one plausible candidate: Right Kind of Reason for guilt/resentment: For a consideration to count as the right kind of reason for guilt/resentment towards an action, it must bear on the moral wrongness of the action. This is where things get complicated. Not only does the appropriateness analysis of moral wrongness run into the Conflation Problem, any attempt to solve this problem in a way that respects the RKR principle eventually 115

10 ETHICAL PERSPECTIVES MARCH 2011 seems to end up being circular. Remember that, as was noted above, sentimentalism originally aims to be a reductive doctrine: it promises to facilitate exhaustive analyses of evaluative concepts in emotionist terms and the conditions under which their application is warranted. But there are different kinds of warrant, wrong and right kinds, and NS does not have the resources to distinguish the former from the latter without having to give up its reductive aspirations. It cannot provide an analysis of moral wrongness that does not in some way rely on a prior understanding of the very notion in question, namely, moral wrongness. It turns out that while guilt, resentment and other reactive attitudes might be crucial to understand the concept of moral wrongness, they are not thoroughly constitutive of it. A natural response to this line of argument might read something like this: Granted, the appropriateness analysis of normative concepts tends to conflate different kinds of appropriateness. An adequate analysis has to do justice to the RKR principle, and it seems that this cannot be done in a reductive, non-circular manner. But this whole problem hinges on the claim that sentimentalist accounts of moral judgment necessarily have to be reductive. The above argument, it seems, is attacking a straw man. One might think that both directions of explanation are correct: Disapproving something must be explained as feeling that it is wrong, and conversely, to judge something wrong is to judge that it merits disapproval (Darwall et al. 1992, 149). This is exactly what the no-priority view suggests. It holds that one can think about the relation between the concept of moral wrongness and, say, guilt, as a relation of mutual elucidation, rather than reduction of the first concept to the second. But interestingly, this is not what sentimentalists actually aim for. They acknowledge that in order to analyze wrongness in terms of the appropriateness of guilt and impartial anger without circularity [ ] one needs an independent fix on these emotions (D Arms and Jacobson 2000a, 732). The Conflation Problem shows that the threat of circularity sentimentalists try to avoid by subscribing to the no-priority view does return on a deeper level, and hence cannot be avoided. The problem is 116

11 SAUER THE APPROPRIATENESS OF EMOTIONS not so much that the sentimentalist cannot account for specifically moral emotions in a non-circular way: the problem is that sentimentalism cannot do justice to its normative intuitions, because it necessarily conflates relevant and irrelevant considerations in favour of the appropriateness of emotions. IV. NATURAL KINDS-NATURALISM I: EMOTIONS AND THEIR MARKS In order to see that sentimentalists really are in the business of finding an independent fix on the appropriateness of emotions, one has to detect the traces of naturalism that figure in the account. Why is naturalism attractive for neo-sentimentalism? NS gives an account of moral judgment in terms of the conditions under which certain emotions make sense. We have seen that emotions can make sense and nonsense in many different ways, some of them being utterly extraneous to the question of whether a particular emotion really is appropriate to have. Some considerations of appropriateness simply do not bear on whether something is the legitimate object of a particular response. Imagine, however, one could analyze emotions as natural kinds: mental events, typically associated with patterns of bodily reactions, that have an underlying microstructure or a set of objects and properties to which by their very nature they typically respond. When it comes to responding to the conflation problem, this would enable NS to directly refer to that set of objects and properties and thus to carve out what makes an emotion genuinely appropriate to have. In what follows, I deal with the conflation problem insofar as it poses a challenge to neo-sentimentalist accounts of moral judgment. I draw on authors as diverse as Allan Gibbard, David Wiggins, and Jesse Prinz. It should be noted, however, that I do not do so because I take these authors to be proponents of the same theory. Rather, I use their respective accounts as theories that potentially provide neo-sentimentalism with the conceptual resources to answer the challenges it faces. 117

12 ETHICAL PERSPECTIVES MARCH 2011 These theories need not subscribe to all of the main tenets of neo-sentimentalism as outlined above. The sentimentalist account of moral wrongness and rightness hinges essentially on the connection between judgments of wrongness and human emotional response. But not just any old response will do, the sentimentalist reminds us, and so the focus is on appropriate responses. As we have seen, this strategy falls prey to the Conflation Problem: at first glance, there seems to be no way to distinguish between different kinds of considerations in favour of the appropriateness of an emotion that does not already refer to the very concept in question, namely, moral wrongness. But why is it that this circularity bothers us when it comes to the analysis of moral concepts, while the same circularity seems perfectly in order in the case of, say, colour concepts? This line of thought is one of the main motivations for the no-priority view : In all these matters, an analogy with colour is suggestive. x is red if and only if x is such as to give, under certain conditions specifiable as normal, a certain visual impression naturally raises the question which visual impression? And that question attracts the answer an impression as of seeing something red, which reintroduces red (Wiggins 1998, 189). It seems that this is simply the way it works. Colour concepts and moral concepts belong to the same class of concepts in that they are response-dependent concepts. It is typical for concepts like these that their content or the conditions of their correct application is specified with reference to the objects they apply to and the subjective response these objects elicit. If we want to know what red refers to, we have to look at the things that cause a certain visual impression under certain canonical conditions, and if we want to know what wrong or cruel means we have to look at the things that tend to elicit reactive attitudes such as outrage or indignation under certain canonical conditions. In carrying out this task, however, the sentimentalist reminds us that we must not forget the normative dimension NS promises to incorporate: We are not simply to fire off at random in our responses to things. A feeble jest or infantile practical joke does not 118

13 SAUER THE APPROPRIATENESS OF EMOTIONS deserve to be grouped with the class of things that a true judge would find genuinely funny (Wiggins 1998, 193). That said, we can start to look for properties of objects that, were they to figure as considerations in favour of a moral judgment, would genuinely bear on the appropriateness of moral emotions such as guilt and resentment. Keep in mind that according to NS s own standards, the account fails if it does not explain the appropriateness of emotions in a way that allows us to discriminate between warranted and actual responses. One thing NS has to offer in fleshing out this suggestion in more detail is what could be called a speculative genealogy, a story that tells us how we arrived at our collectively scrutinized (Wiggins 1998, 210) emotional responses. This is the crucial part of the story: Suppose that objects that regularly please or help or amuse us [ ] in various ways come to be grouped together by us under various categories or classifications [ ]; and suppose they come to be grouped together as they are precisely because they are such as to please, help, amuse us (Wiggins 1998, 195). In this way, ordinary moral judges come to construct pairs of objects and responses that are typically associated with those objects; these serve as a starting point for further criticism and recalibration. The materials we start with are the objects we regularly are pleased or outraged by. What Wiggins hopes to establish is a method that helps to find out whether there is something in the object that is made for the sentiment it would occasion in a qualified judge (1998, 194). By looking at the history of the responses of a community of judging subjects towards relevant objects we can hope to find out for each sentiment at a time what the marks of situations and objects are that typically tend to elicit amusement, anger, shame or guilt. These marks can be said to be the genuine considerations in favour of a response, because the response is made for them and vice versa. They are the features that really bear on whether an X has the property P and thus enable us to solve the Conflation Problem. This account has two major problems: first, it fails to distinguish actual from justified responses, and is thus impaled on the third horn of 119

14 ETHICAL PERSPECTIVES MARCH 2011 the trilemma. The piecemeal improvement of our sensibilities towards objects and situations cannot be sufficient to explain why some things merit approval and disapproval and do not just regularly elicit it, nor can it account for appropriate sensibilities that simply lack that kind of history, that is, the enormous amount of recent moral issues that have to be dealt with from a modern point of view. Second, and more importantly, it has a problem with the specification of the marks that emotions purportedly are made for: there are no such marks. As a matter of fact, at no point in his account does Wiggins point out what the marks are that count in favour of something being funny, disgusting, fearsome or any other suitable response-dependent property. And even if he could have done so, it remains questionable whether the same would be possible for more abstract, reactive emotions such as guilt. Think about the situations and actions that cause guilt in normal subjects and group them together in the way Wiggins recommends. What are the marks that all these objects you forgot to call a friend, a man in the newspaper killed his wife, a politician broke his promise share? Again, the only plausible candidate seems to be that all these acts share the feature of moral wrongness. But that is what we started from: what is moral wrongness? As a neo-sentimentalist, Wiggins answers this question by referring to our emotional responses to certain objects, and the conditions of their appropriateness. They are appropriate, we are told, if these responses are caused by what they are made for, the marks they have been associated with over a long period of time by judging agents who take an interest in whether their responses make sense: Instead of fixing on an object or class of objects and arguing about what response or responses they are such as to evoke, we can fix on a response [ ] and then argue about what the marks are of the property that the response itself is made for (Wiggins 1998, 198). For present day moral judges who do not want to fire off at random with their judgments, but want to pick up on features of actions and persons that bear on whether something is the appropriate object of guilt or resentment, the only mark that is left is the moral wrongness of the 120

15 SAUER THE APPROPRIATENESS OF EMOTIONS objects that have been put into the same group. What else is there to say about the class of things guilt is made for? Moral wrongness is identified with warranted guilt or resentment, which in turn are said to be warranted if they pick up on the relevant marks of the matter at hand. If no such mark can be found other than moral wrongness itself, the detour (Wiggins 1998, 189) through the sentiments has made us no smarter. Returning to the analogy with colour concepts from above, we can now see that there is a crucial difference between these two subclasses of response-dependent concepts. Colours, as subjective experiences, are phenomenally distinct. It is possible to find out about the essential features of redness via introspection, which is not possible, or at least not in the same way, with emotions like guilt or shame. Despite the phenomenal quality of what it is like to have a subjective experience of red objects, there is nothing that can be said about what bears on an object being red. Moreover, if no agreement can be reached as to whether an object really is red or not by perceptual means alone, the problem can be tackled with reference to wavelengths and light rays. Neither of these two options are open to us in the case of moral emotions: introspectively, there is no distinctive feel to shame that separates it from guilt or a feeling of social insecurity (this is what Jesse Prinz calls the somatic similarity problem [2007, 65ff.]); nor can we investigate the objective properties of objects and actions in the way we can investigate the texture of surfaces in order to find out whether a particular action warrants reactions of guilt or indignation. V. NATURAL KINDS-NATURALISM II: MORAL JUDGMENTS, FITTING EMOTIONS Take one typical example of a concept that seems to be suitable for a sentimentalist analysis, say, enviable. According to the neo-sentimentalist analysis, X is enviable iff it is appropriate to envy X. If we combine 121

16 ETHICAL PERSPECTIVES MARCH 2011 this with the RKR principle, we get: X is enviable iff it is appropriate to envy X in a sense that bears on whether X is enviable. This additional element is supposed to rule out the conflation of relevant and irrelevant considerations in favour of a particular emotion, because it might be inappropriate, because mean-spirited, to envy your friend s well-deserved success. But to think so is surely not yet to deny that the success is enviable (D Arms 2005, 3). Plenty of similar examples could be given here. That an emotion is mean-spirited, and in that sense inappropriate, does not bear on whether having the emotion would be appropriate in the right kind of (namely fitting) way: What is needed here is the notion of a kind of appropriateness that restricts the range of considerations about whether to feel F to just those that speak to whether or not the circumstance is F. D Arms calls the required notion fittingness : Sentimentalists owe some kind of account of how to distinguish considerations of fittingness from other reasons to feel (2005, 4). One possible suggestion might be that in order to understand this specific kind of appropriateness, one has to look at features of the emotions themselves, particularly because each emotion is concerned with its very own range of relevant considerations. Envy, for instance, is concerned with features of actions or situations that have nothing to do with whether it would be nice or kind or cruel of you to respond to the object of your envy here: your friend s well-deserved success with that nasty feeling: It is because it is not part of the nature of envy to present its object as undeserved that the fact that your friend deserves his success is irrelevant to whether it is enviable (D Arms 2005, 4). The question, then, is how an emotion presents the object of approval or disapproval to the subject who has the emotion. Perhaps emotions have, as D Arms and Jacobson suggest, a specific nature, a way in which each emotion is tailored to a set of features of situations and actions. So it may be inappropriate to laugh about a wicked joke for moral reasons; but when it comes to the question of whether the joke is funny, we have to look at the sentiment of amusement and the feelings that are 122

17 SAUER THE APPROPRIATENESS OF EMOTIONS associated with laughter and try to detect how this emotion and this emotion alone presents its objects. That way we can see what the features are that genuinely count in favour of the funniness of the joke: We think that in order to provide any substantive grounding for the distinction between reasons that are and are not relevant to the fittingness of an emotion, it is necessary to examine our actual emotions piecemeal, in order to articulate differences in how each emotion presents some feature of the world to us when we are in its grip (D Arms and Jacobson 2000a, 746; also Mason 2003, 238ff.). Imagine that you are in the grip of an emotion: there is a politician about whom you have been undecided before, although open-minded and overall quite sympathetic. Now you have heard a speech of his on a square in your neighbourhood, and you feel excited and inspired by his words. You turn to a friend who is standing right next to you and say: I never knew just how charismatic this guy could be! According to fittingness accounts of normative judgment, what you have said can be translated into It is fitting to be excited and inspired by X (the politician s speech, or the person himself) because this emotional state picks up on features of X [ ] that bear on X being charismatic. The features of X that bear on it being charismatic are variable: maybe people are fascinated by the man s eloquence and his solemn pronunciation, his elegant but determined gesturing or simply by his acuteness and factual competence (corresponding properties could be pointed out for the funniness of a joke, or the disgustingness of a picture). These are the features of the world that properly speak to whether some X really has the evaluative property P. Emotional responses can be criticized as unfitting or endorsed as fitting in two different ways: according to their size, and according to their shape (Pugmire 2005, 4ff.; D Arms and Jacobson 2000b, 73ff.). Suppose, after you have heard the speech, you are prepared to turn your life upside down. You start reading about the politician obsessively, you want to join his campaign, donate your money to his supporters, dress like him, talk like him, and, yes, you would sacrifice your life for him if necessary, 123

18 ETHICAL PERSPECTIVES MARCH 2011 all because of your passionate feelings about this guy, and the excitement and inspiration that accompany them. Here, your response can be said to be unfitting with respect to its size. You are simply overreacting. Second, your emotional engagement may be unfitting with respect to its shape. Your friend finds your reaction hardly intelligible. He thinks the guy s speech was not that big of a deal, after all, his gesturing was old-fashioned and ridiculous, his voice shrill and unpleasant, and the points he made were not really that smart (or so your friend thinks). Even though these distinctions prove to be helpful, there are problems with the fittingness account. The first one is that it relies on an essentialistic view of the emotions. This point is connected to the naturalistic outlook shared by neo-sentimentalists in general. Emotions are said to have a nature; if we want to know what the features of the world are that particular emotions like amusement, envy, disgust or excitement pick up on, we have to look at the emotions themselves and the features they present to us as counting in their favour. Emotions involve evaluative presentations: they purport to be perceptions of such properties as the funny, the shameful, the fearsome, the pitiable [ ] Envy, for instance, involves a complex set of evaluations in presenting its object as enviable. Very roughly, one s envy portrays a rival as having a desirable possession that one lacks, and it casts this circumstance in a specific negative light (D Arms and Jacobson 2000b, 66). Now this rough characterization of the features that bear on whether something is genuinely enviable may seem plausible at first glance, but some questions remain: does it really have to be a rival that is the object of your envy? Does the person whose envy it is necessarily have to think that the rival s possession be desirable? And what about the negative light? Is that everything there is to say about this? And does envy, for many people at least, not also feel good, in a weird, twisted way? Do people not sometimes become absorbed with their envy and like it (La Caze 2001)? But even if we grant, for the sake of the argument, that emotions involve evaluative representations and that an emotion can be said to fit 124

19 SAUER THE APPROPRIATENESS OF EMOTIONS its object in case the object is as the emotion presents it to be, one runs into serious problems if one tries to specify what exactly the features are that a particular emotion represents. Think about a profoundly basic emotion like disgust. In general, disgust presents its objects as contaminating, and so it seems that according to the fitness account of disgustingness, objects can be disgusting only if they actually are contaminating. But this just is not so: plenty of things that are quite patently non-contaminating nonetheless remain disgusting. Drinking a glass of mucous, eating raw worms, handling human feces, and bathing in urine are all disgusting whether or not the mucous is your own, whether or not the worm has been sterilized, whether or not you are wearing surgical gloves, and whether or not the urine has been distilled into pure water (Knapp 2003, 272). For some emotions, we have a hard time saying what the features are that a particular emotion presents its objects to have, and for some emotions, these features do not actually have to be found in the object of the emotion for it to be fitting. Also, it seems that the standards for when a particular emotion can be criticized as unfitting with respect to its size or its shape are not sentimentalist standards themselves (this is the first horn of the trilemma). Especially in the case of unfitting size, these standards simply seem to come from common sense or cultural conventions. They cannot, at any rate, be read off the emotion directly. It would just be strange to change your whole life just because you found the politician s speech so amazing, or to commit suicide out of guilt for running over a deer. In the case of shape, the criteria for fittingness also cannot be sentimentalist in nature, because that would simply beg the question. In fact, the theory is supposed to explain what makes the shape of an emotion fitting or unfitting, so the concept of shape does not do the job. More importantly, there is a problem with thin moral emotions. Thin moral emotions (like guilt or resentment) correspond to thin moral concepts (like wrong or permitted) like thick moral emotions (like elevation or Schadenfreude) correspond to thick moral concepts (like cruel or noble). Even 125

20 ETHICAL PERSPECTIVES MARCH 2011 with the latter category, it is hard to point out the features that the emotions which typically are associated with judgments of cruelty or nobleness being horrified, or feeling elevated pick up on. Do cruel acts always involve harm, or a wicked motive? And do noble acts have to involve selflessness? With thin moral emotions, all the more so. We have already seen in the discussion of Wiggins proposal that there are no features of actions or persons that directly bear on whether guilt or resentment are warranted apart from the moral wrongness of the action or blameworthiness of the person. Even if we grant, for the sake of the argument, that NS can provide convincing analyses of one subclass of response-dependent concepts like funny or disgusting this does not mean that it can also account for thin concepts like moral wrongness. But this is the subclass of normative concepts that matters most in meta-ethics. VI. EVOLUTIONARY NATURALISM I: EMOTIONS THAT MAKE SENSE Evolutionary naturalists hold that the key to an understanding of the conditions under which reactive emotions are appropriate is an understanding of the function these emotions play within a subject s overall motivational economy. If we want to know, they argue, what guilt or resentment are tailored to, we have to look at the natural history of these emotions and the conditions of selective pressure under which they have emerged: Moral norms coordinate the anger and the guilt it makes sense to feel from these special standpoints [of full, impartial engagement], and in turn these standpoints will have something to do with the feelings it makes sense to have in the flux of life. [ ] The ways he feels about himself, about his qualities and his actions, will need to be coordinated to some degree with the ways others feel about him, or his feelings may prompt him to act to his detriment (Gibbard 1990, 127). This perspective has an openly evolutionary flavour. The bottom line is that in order to understand the nature of emotions, we have to understand their 126

21 SAUER THE APPROPRIATENESS OF EMOTIONS functional role; and in order to understand the functional role of a psychological pattern, we will have to see to what extent it contributes to an individual s benefit or detriment. Since an individual s flourishing especially under evolutionarily significant conditions millennia ago largely depends on his or her successful interaction with fellow human beings, we have to see how moral emotions like guilt, shame and resentment and the norms they are governed by contribute to smooth social coordination: Human emotions are above all social. A person invariably depends on intricate systems of cooperation and reciprocity if he is to have any decent chance of survival, reproduction, and the fostering of his children. Negative human emotions respond pre-eminently to threats to one s place in cooperative schemes [ ] (Gibbard 1990, 138). The question that has to be addressed, from this perspective, is: are moral emotions adaptive? And just as important: is it adaptive to have norms governing these emotions? On that account, the appropriateness of emotions will turn out to be a feature of their interactive function. A particular emotion is appropriate if it makes sense to have it according to the norms it makes sense to accept for the sake of adaptiveness and social coordination. This is not to say that Gibbard wants to reduce genuine moral appropriateness to evolutionary adaptiveness. He does distinguish the meta-ethical, explanatory question from which patterns of emotional responses come and the normative question in which reasons bear on what it does or does not make sense to feel or do. The problem that remains, however, is that his norm-expressivism tries to understand judgments of moral wrongness on the basis of the moral emotions. The cognitive content of those emotions, in turn and thus the features to which guilt as such, resentment as such, and shame as such respond is largely determined by their functional role in the above sense. For an account that aims to be sentimentalist all the way down, there seems to be little else to do than to identify the norms that govern the reactive attitudes in the normative sense with the reasons that made them useful to have in the explanatory sense. 127

22 ETHICAL PERSPECTIVES MARCH 2011 In principle, there are two ways in which an individual can fail to be a useful member of a social context. If we conceptualize social life under meagre conditions as a struggle for scarce resources, one can either lack personal resources and abilities, or fail to bring them in properly. Shame is tailored to the first case, guilt to the second: On the adaptive syndrome story, guilt is tied genetically to poor cooperative will to a special way a social being can fail to be a good candidate for inclusion in cooperative schemes. It is tied to insufficient will to play one s part in a scheme and share its fruits [ ] Guilt placates anger, and the threat of guilt averts acts that would evoke anger (Gibbard 1990, 296). According to this story, guilt indicates insufficient motivation, and the most disastrous way in which an individual can be insufficiently motivated is by undermining the social chain of continuous reciprocation, either by not sharing to an adequate extent or by not taking part in the acquisition of goods to begin with. Moral norms, which, narrowly understood, simply are the norms directing guilt, shame and impartial anger, work exactly that way, because being capable of guilt and governing it by norms can pay. Indeed the tie of guilt to anger makes guilt an especially fine candidate for governance by norms. Guilt and anger together can help regulate social life. If norms of guilt and anger are well chosen, they will motivate people in desirable ways, and they will diminish the conflicts that arise from anger (Gibbard 1990, 298). Insufficient motivation to participate in social life, therefore, will typically cause guilt and motivate individuals to do things that are not followed by guilt, that is, socially desirable things. At this point, we can see how the naturalistic fix on moral emotions seems to assist NS in solving the Conflation Problem. Guilt is tied to poor cooperative will: these are the considerations that bear on the appropriateness of guilt and this is what specifies the conditions under which guilt makes genuine sense. At first, this proposal seems to be able to cope with the conflation problem. Guilt can be inappropriate in a prudential sense, for example, because feeling guilty undermines my very motivation to participate in social exchange in the future. But this clearly does not speak to whether 128

23 SAUER THE APPROPRIATENESS OF EMOTIONS an action genuinely warrants guilt because, as we have seen, guilt only makes sense as a remedy, as it were, for insufficient motivation. However, some doubts are in order here as to whether the account can rule out irrelevant considerations in a way that satisfies the RKR principle. At the end of the day, evolutionary naturalism tells a story that is framed in terms of efficient functions and advantageous results. How can such a story not conflate relevant and irrelevant, genuinely moral and prudential reasons? Let us examine just two examples and see how the account deals with them. Most gay teenagers (at the very least, a significant number of them) are ashamed of their sexuality, and it is not an alien phenomenon that women who have been raped feel guilty about what happened to them. Cases like this may strike us as understandable; yet these reactions are surely inappropriate emotional attitudes towards one s personality and the situations one sometimes has to cope with. Remember that according to the evolutionary account, guilt and the norms it is governed by make sense if and because it improves social cooperation insofar as it motivates individuals to conduct themselves in a way that is beneficial for a group of fellow reciprocators. Does it make sense to feel guilty upon being raped? At first sight, it is hard to see any reason why the answer shouldn t be yes, and the simple rationale is that even in this deviant case, guilt motivates people to behave in a socially desirable way. To feel guilty about having been raped will motivate people even more to avoid getting raped, and not getting raped is socially desirable in more than one respect. It can be expected that (or at least it is not inconceivable that) norms that prescribe guilt upon being raped would pay too, and that they would help regulate social life in a way that is beneficial for anybody who is affected. But this does not mean that, normatively speaking, it genuinely makes sense for women who have been raped to feel guilty about what happened to them. The same holds for shame. There are many ways in which it is advantageous to have strong feelings about what is perceived as social inadequacy, but none of them is more than fortuitously tied to the appropriateness of shame. This illustrates a 129