Responsibility, Failure, and Punishment

Responsibility, Failure, and Punishment A Constructivist Account Carla Bagnoli Some sort of freedom is a pre-condition for moral responsibility. If we are not free in some such fundamental way, then moral responsibility is unintelligible and its associated categories become inapplicable or dispensable. Not all sorts of practices become dispensable when their key concepts become unintelligible. Arguably, there may be good reasons to keep practices of holding each other responsible even when the pre-conditions of applicability of their key concepts have been denied. The issue of dispensability of concepts may be not thoroughly or completely conceptual; that is, it may be determined by considerations that have nothing to do with the nature of concepts. But I will not be concerned with such external considerations. I am interested in elucidating the concept of moral responsibility, where an agent is responsible for x if he is appropriately held accountable and it is answerable for x. There are three important theoretical constraints that a plausible account of moral responsibility should respect. First, it should have epistemic relevance, that is, be able to explicate the distinctive role of agents in action and in the activities in which they are involved. Second, it should be transparent or congruent with our first personal experience. Third, and finally, it should be compatible with naturalist conceptions of the mind and its relations to the world. University of Warwick Seminar in Political Philosophy 16 October 2012 I will argue for an account of moral responsibility in terms of recognition respect.1 I will argue that respect is an affective concept constitutive of practices of attribution of responsibility, rather than being an emotional response to a valueproperty such as the dignity of humanity. In discussing practices of attribution of responsibility, I choose the generic term “practical subjects” rather than the more specific term “agents” because I take the concept of moral responsibility to coincide only partially with voluntary or intentional agency. On the one hand, it covers a wider range of propositional attitudes and feelings that are attributable to the agent but that it would be misleading to characterize as actions or fully voluntary activities – such as, beliefs and emotions. On the other hand, its scope is narrower in that moral responsibility concerns only actions, activities, emotions, and propositional attitudes for which practical subjects qualify as appropriate objects of moral assessment. (Part of the problem, as we shall see, is to define this relation independently of a pre-fixed metaphysical domain of moral objects and properties). In centering responsibility on recognition-respect, I distance myself from some other prominent accounts of moral responsibility that privilege backwardlooking, reactive or negative attitudes such as blame and resentment. My contention is that respect is conceptually prior to other reactive moral emotions and it plays a governing role on blame and other excluding or redeeming attitudes insofar as these are forms of moral address. This proposal does not offer any novel view regarding the relation between moral responsibility and freedom of the will. It agrees with Peter Strawson’s 1 Darwall defines “recognition respect” as opposed to “appraisal respect.” Darwall 1977: 36-49. “Recognition respect” does not involve any consideration of appraisal of 2 University of Warwick Seminar in Political Philosophy 16 October 2012 argument that the concept of moral responsibility that is central to our practices is independent of the metaphysical issue of free will. The first step in my argument is to show that what is really at work in Strawson’s account of moral responsibility is a broadly Kantian concept of recognition-respect as constitutive of the normative status of practical subjects. Second, I develop a constructivist account of recognition-respect, which does not presume the value of humanity, but it takes a constitutive norm of reasoning that we take others as having the same moral standing. Finally, I consider some advantages of this view regarding responsibility and punishment. 1. Grounds of responsibility and reactive attitudes To emancipate moral responsibility from metaphysical freedom, P.F. Strawson appeals to facts about human psychology that are not only congruent with naturalism, but such that they cannot be disputed by determinism. These are facts about human beings as social animals, that is, capable of social interactions but also capable of emotional attitudes that make sense only from within social practices. The category of reactive attitude is quite broad, and includes both negative and positive feelings, such as gratitude, resentment, moral indignation, love, forgiveness, and blame, among others. What is common to this large variety of reactive attitudes is that they react to harm or benefit done intentionally. It appears that in ordinary morality, the will is an appropriate object of moral appraisal: it is good or ill will; and it is morally assessed under several dimensions (e.g. in terms of consequences, because of the importance in moral relations, for purpose of character assessment, for purpose of strategic rationality, and most crucially for its role in trust relations and cooperative schemes). Strawson’s claim is that to recognize agents as morally responsible is to say that they qualify as appropriate targets of reactive attitudes. The relevant issue, then, is 3 University of Warwick Seminar in Political Philosophy 16 October 2012 under what conditions such reactive attitudes are appropriately addressed. According to Strawson, such conditions are fundamentally related to and fully determine the grounds of moral responsibility. There are two sorts of conditions under which reactive attitudes are not appropriate. First, there are “mitigating circumstances”, that is, conditions that excuse or justify in various degrees the failure and consequently qualify the offence, without undermining the moral responsibility of the agent. Second, there are “conditions of insanity or immaturity”, which cancel or diminish the agent’s moral responsibility. Both these kinds of conditions importantly bear on the moral judgments that the evaluator is entitled to issue. In the former case, the evaluator is entitled to take a personal perspective on the agent, which is marked by reactive emotions, such as resentment or moral indignation (Strawson 1968: 87). In the latter case, the evaluator adopts an objective standpoint, which amounts to withholding moral judgment and silencing the emotions that are normally associated with attribution of moral responsibility, or expressing a range of emotions that are normally not associated with attribution of responsibility, such as pity or compassion (Strawson 1968: 94). Strawson’s point is that no hypothesis on the metaphysics of free will modifies the attitudes of evaluators toward subjects, and thus the category of responsibility applies independently of any of the positions about free will. The question is not what (empirical or metaphysical) capacities and powers agents have, but what kind of normative status the attribution of responsibility implies. Central to the argument that emancipates responsibility from metaphysical freedom is a claim about the personal nature of reactive attitudes. Reactive attitudes are emotions that are typical of and appropriately felt by participants of a community, understood as a practice governed by norms. This is to say that anybody who actively 4 University of Warwick Seminar in Political Philosophy 16 October 2012 participates in such a normative system is entitled (but also demanded and expected) to feel such reactive attitudes on the relevant occasions. For a member of the moral community, lack of appropriate reactive attitudes is a sign of malfunctioning and malpractice. That is, it indicates that the agent does not properly exercise the requisite moral sensibility or fails to be guided by moral norms. In extreme cases, total lack of moral sensibility is the ground for disputing the status of moral agent.2 According to Strawson, reactive attitudes are moral emotions that are inalienable insofar as they pertain to human nature. It follows that practices centered on such emotions are indispensable. To cancel out the concept of moral responsibility “is barely more than a conceptual possibility; if it is that… For all these types of attitude alike have common roots in our human nature and our membership of human communities” (Strawson 1968: 115). In their characteristic moral function, reactive attitudes are both personal and vicarious. They are personal in two senses. First, they are internal to the participant’s stance as opposed to the objective stance. Second, they are modes of personal address, that is, ways of relating to one another from within a social practice. Reactive attitudes are vicarious when they are felt qua representatives, rather than as the direct object of harm or benefit. (Of course, there are moral emotions that are more appropriate to the victim because of their positions, such as anger, humiliation, vengeance etc. But the point is that these are all attitudes that can be felt vicariously). Strawson’s central insight, as I take it, is that the concept of responsibility cannot be elucidated but in terms of moral emotions directed to good/ill will. To be 2 For Kant these are the paradigmatic cases of “moral death”, Kant 1796 6. Cf. Bagnoli 2011. 5 University of Warwick Seminar in Political Philosophy 16 October 2012 responsible is to qualify as the appropriate object of such moral emotions. Responsibility does not depend on the metaphysics of agency, but it is explained by reference to the fact that we are sensitive to and concerned with whether others are ill or well disposed toward us. For Strawson’s argument to work, it is important to recognize the constitutive role of reactive emotions. The argument is not merely that such reactive emotions are associated with the practice of attributing responsibility. Rather, the radical idea is that such reactive emotions are constitutive of responsibility.3 By feeling such reactive emotions as blame, resentment, anger or moral indignation toward an agent, we take her responsible for wrongdoing. The key function of such reactive emotions, then, is not expressive but constitutive. We direct our blame toward the agent to signify that she is responsible of wrongdoing. Taking her responsible for wrongdoing is something we do by addressing reactive emotions such as blame and resentment. This is to say that such moral reactive emotions are constitutive of responsibility, that is, they are modes of taking and attributing moral responsibility. 2. From reactive attitudes to blame 3 This interpretation differs from Darwall (2006) and Deigh (2011). Deigh is right that the expressive function of moral emotions is most relevant to Strawson’s noncognitivism. However, for Strawson’s argument to work, it is important to recognize that the function of emotions is not only expressive, but also constitutive of responsibility. This more radical claim does not commit to the realist view that takes moral emotions to be emotional responses to value. In this respect, it also differs from Darwall’s interpretation. 6 University of Warwick Seminar in Political Philosophy 16 October 2012 Strawson’s argument has been extraordinarily influential in recent debates on free will and moral responsibility. However, current treatments of responsibility have departed from Strawson in two major respects. First, they have significantly narrowed down the category of reactive attitudes that are relevant to moral responsibility to negative, backward-looking deontic attitudes, that is, attitudes that are associated with past violations of moral obligations rather than those associated with the expression of good/ill will. In all such accounts, blame is elected as the key reactive attitude to measure moral responsibility. Secondly, more recent accounts have significantly discounted or even denied the role of the emotional aspect of reactive attitudes in understanding responsibility. For instance, Jay Wallace starts with Strawson’s model and develops an account of moral responsibility where reactive attitudes are taken to be violations of normative expectations or ‘demands’—that is, to violations of ‘practical requirement[s] or prohibition[s] in a particular situation of action’ (Wallace 1990: 22). In a similar vein, T.M. Scanlon focuses on the normative dimension of blame in personal relations, arguing that it is not a moral emotion. Scanlon denies that Strawson’s central insight consists in highlighting the constitutive and expressive role of reactive attitudes in accounting for responsibility. He takes his view to be in agreement with Strawson in “seeing the human relations as the foundations of blame” (Scanlon 2008: 128). But it differs from Strawson “in placing emphasis on expectations, intentions and other attitudes that constitute these relationships rather than on moral emotions such as resentment and indignation” (Scanlon 2008: 132). Scanlon takes this modification to be an improvement over Strawson’s original account. In my opinion, instead, this departure from Strawson is a mistake and leads to some serious misunderstanding of moral responsibility and the ethics of 7 University of Warwick Seminar in Political Philosophy 16 October 2012 blame. To establish this point, it is useful to consider the wide range of moral functions that are traditionally attributed to blame. Perhaps the most interesting and widely endorsed perspective on blame is J.S. Mill’s. On this view, blame is a sanction that explains the bindingness and special authority of moral obligations (Mill 1863: chapter 3). Through social conditioning and education, our minds become accustomed to associate some types of actions with some emotive reactions. Actions externally sanctioned by society become internally sanctioned thanks to the work of natural emotions, such as guilt, self-reproach and shame. Mill’s conception of blame is meant to explain away the halo or mystic aura of moral obligation that is insisted upon by rationalist metaphysicians such as Kant and Clarke. The authority of moral norms is nothing but “a subjective feeling in our mind, attendant on violation of duty, a pain more or less intense, which in properly cultivated minds rises, in the more serious cases, into shrinking from it as an impossibility” (Mill 1861: 74). Appealing to Mill’s model, Allan Gibbard holds that the function of blame is to mark the bounds of the moral community and reinforce moral obligations. By and large reactive emotions such as blame are “reactions against threats to one’s place in cooperative schemes”. This is because “norms for guilt can attach a bad feeling to things bad feelings can move us to avoid” (Gibbard 1990: 297). Gibbard takes blame to have an immediate motivating power: it serves as an incentive to comply or as a deterrent against future defection. On a further modification of Mill’s account, blame is a sanction that signals withdrawal of social recognition and membership in a concrete community governed by norms (Skorupski 1999: 42-43, Skorupski 2005). Skorupski does not view blame and other reactive sentiments as playing a chief role in making moral obligations 8 University of Warwick Seminar in Political Philosophy 16 October 2012 authoritative and binding. Rather, he takes blame to have a very specific function of disposing members of the moral community to withdrawal of recognition. This approach has the merit of highlighting the role of institutional morality in shaping identity, but its risks are significant. First, in linking blame to withdrawal of recognition, Skorupski makes moral standards coincide with criteria of social membership. The criteria of social membership thus become almost invulnerable to moral criticism. The account is thus vulnerable to the objection of dogmatism (van Willigenburg 2003; Bagnoli 2007). Second, members of certain groups (e.g. foreigners, women) may fail the conditions of membership in the relevant de facto community, but this does not show that they are not fully responsible practical subjects. To prevent this case, it seems necessary to say that the condition for full responsibility is being held fairly accountable. But then the relevant community is an ideal moral community, not a concrete community. The account is thus vulnerable to the objection of unfairness. Third, one could simply opt out of any concrete community, and yet be still morally responsible. Bennett Helm has recently presented the misanthrope case as an objection to this account (Helm 2011: 234). The account cannot plausibly make sense of this case. Scanlon’s account of blame seems to eschew these objections because it denies that blame is an emotion that negotiates social identity by functioning as a sanction. On Scanlon’s view, the role of blame is more concrete and diffused because it concerns specific personal relations. Blame marks a change in personal relations; it registers that a morally relevant alteration has occurred (Scanlon 2008: 186). To blame the person is to hold the attitude toward him or her that this impairment makes appropriate; “to take your relationship with him to be modified in a way that this judgment of impaired relations hold to be appropriate” (Scanlon 2008: 128). The 9 University of Warwick Seminar in Political Philosophy 16 October 2012 judgment that a person is blameworthy for an action amounts to “a judgment that the action shows that person to hold attitudes that impair his or her relations with others” (Scanlon 2008: 128, 145). The aim of expressions of blame is neither to punish the wrongdoer, nor to reinforce the bounds of moral community or the force of moral obligations. The primary purpose is to register such a modification in a personal relation that has been impaired by another, and thus it does not carry the expectations that it did prior to the impairment. Scanlon emphasizes the personal quality of the moral violation. Moral failure is not merely a violation of a norm. It resembles more a personal reaction to a personal attack: it is a reaction of somebody who has been betrayed or whose expectations have been frustrated. Blame signals the wrongdoer’s failure to stand in the appropriate relationship to us: “Impairment occurs when one party, while standing in the relevant relation to another person, holds attitudes toward that person that are ruled out by the standards of that relationship, thus making it appropriate for the other party to have attitudes other than those that the relationship involves” (Scanlon 2008: 135). Like Skorupski Scanlon takes blame to signal some kind of distancing and withdrawal, but unlike Skorupski he does not take it to govern moral membership. For this reason, Scanlon’s account does not run the same risks as Skorupski’s account of socializing morality. Moreover, it is a virtue of Scanlon’s model that it leaves open the range of the relevant attitudes. The norms of appropriateness for such attitudes are relative to the nature of the relevant ground relationship, the sort of impairment occurred, the position of the respondent,4 and the significance of the impairment. The 4 Scanlon 2008: 138. As for Strawson, this framework of personal relation is 10 University of Warwick Seminar in Political Philosophy 16 October 2012 advantage of this model is that it makes blame more concrete and nuanced according to particular relationships (Scanlon 2008: 189). My general contention is that the strategy of narrowing reactive attitudes to blame and discounting its emotional significance fails to further our understanding of moral responsibility. I side with Strawson in thinking that the emotional aspects of reactive attitudes play a crucial role in the practice of holding each other accountable. However, my general proposal differs from Strawson in two important respects. First, it relates the indispensability of the concept of moral responsibility to the first person experience of autonomy, rather than to social practices and invariant features of human psychology. Second, it centers on the moral emotion of respect as the emotional aspect of autonomy that governs all other moral emotions, including reactive and redeeming emotions that apply in the aftermath of moral failure. 3. The emotional aspect of blame It is a striking feature of Scanlon’s account that it deprives blame of its most intuitive quality, that is, its emotional aspect. This philosophical position is not congruent with the common understanding of blame as an emotion, and it is at odds with traditional philosophical accounts of ordinary morality. 5 The counter-intuitive aspect of Scanlon’s account is not a definitive reason to reject it, however. The real question is something we entertain in friendships as well as in relations with strangers. Moral relation is a relation of mutual concern we have with other rational beings, Scanlon 2008: 140. 5 For criticism that Scanlon misunderstand the emotional nature of blame, see Wolf 2011, Deigh 2011. 11 University of Warwick Seminar in Political Philosophy 16 October 2012 whether lack of consideration of the emotional nature of blame leads to mischaracterize or misunderstand the blaming activities. In other words, the philosophical issue is whether there are some crucial functions of blame that depend on its being an emotion. My guiding hypothesis is that this is the case. To ignore the emotional nature of blame is detrimental for a number of reasons. First, to ignore the emotional nature of blame leads to misunderstand its role in establishing the authority of moral norms. Emotions are largely recognized to play a central role in accounting for the authority of moral norms. Following Mill, many philosophers hold that deontic emotions account for moral authority in the same way sanctions account for the authority of norms.6 Transgressors are the targets of reactive emotions. Morally worth actions are motivated by respect understood as reverence for the law. This determination of respect is a moral feeling. A second reason not to discount the emotional nature of blame is its role in moral motivation. Emotions are generally understood to have motivating force and reference to emotions is part of explanation of action. To deprive blame of its emotional nature leads to 6 Kantians do not disagree that emotions play this role. They think that when natural emotions motivate the action lacks moral worth. Kant acknowledges that emotions are key auxiliaries to moral motives (Kant 1788: 5.152 ff; Kant 1797: 6.402, 456-458). But he denies that emotions are genuine moral motives that account for the efficacy of moral reasons. When they are represented as surrogates for moral motives, emotions do not provide direct normative support for morality (Kant 1788: 5.152). On Kant’s view, emotions that work vicariously as auxiliary motives detract from morality, insofar as they undermine its genuine authority (Kant 1784a; 1788: 5.152 ff. esp. 157). See Bagnoli 2011. 12 University of Warwick Seminar in Political Philosophy 16 October 2012 mischaracterize its role in the generation of action. Finally, the phenomenology of blame and self-blame fits the common experience of an emotional deterrent against moral failures and violation of moral norms. These considerations support the ordinary understanding of social morality as a “blame-system”. Arguably, to ignore or deny the emotional aspect of respect is to understate or misunderstand its punitive and coercive role in ordinary morality. But the consequences of discounting the emotional aspect of blame can be even more significant. The general problem with an approach that denies the emotional nature of reactive attitudes is that it takes emotions to be either completely separable from or only contingently related to practical reasoning. By contrast, I argue that the experience of moral emotions such as recognition-respect is constitutive of the exercise of practical reasoning.7 4. A constructivist proposal about reactive attitudes A route to this conclusion is Strawson’s remark on the personal nature of reactive attitudes. This large category captures the perspective of subjects engaged in the practice. Importantly, reactive attitudes are said to be personal even when they are felt vicariously, that is, on behalf of somebody else. We feel vicarious feelings when we identify with the victim of some moral wrong, e.g. because of proximity or via some more or less complex empathic mechanism. More generally, however, we feel such 7 I argue for the constitutive role of moral emotions such as respect in Bagnoli 2011. I do not suggest that “reverence for the law” is a deontic emotion, though. In fact, this moral emotion is also related to character insofar as it accounts for moral motivation and it qualifies as the only moral incentive 13 University of Warwick Seminar in Political Philosophy 16 October 2012 moral emotions independently of proximity, e.g. toward strangers, lacking any form of empathy, or toward people for whom we feel no sympathy. In such cases, we feel vicarious moral emotions qua representative of the moral community. This gives us some grounds for rejecting the view that blame is mainly or solely a form of distancing and withdrawing. The role of negative emotions such as resentment, indignation, guilt, and righteous anger is not merely constraining and punitive. They play a larger and more pervasive role in moral relations. They are characteristic forms of moral address. Even when it is harshly punitive, blame may not address the transgressor with the intention of parting ways with him, but in order to voice a moral concern, protest, make an offence explicit, express the need of reparation or demand compensation. These various “blaming activities” are certainly acts of disapproval but they are meant to relate to the other and face a moral problem. Sometimes these are attempts to repair and mend existing relations. Other times, they are simply attempts to come to closure, or to show the transgressors that they are already in a relation with their victims. But they can also be moral requests that help reinforce or develop the moral relations further. If blame is understood to signal withdrawal of recognition, then it aims only at separating. But separation from the community is not the only or the most appropriate moral response to moral wrongs. The relation between victim and wrongdoer appears to be complex and dynamic. This indicates that blame and similar reactive emotions may have different functions in addition to punishment, but also that we need some normative criteria for deciding when different functions of blame are appropriately activated. Responding to wrongs is a normative practice. Insofar as they are members of the moral community, moral evaluators entertain a moral relation with the 14 University of Warwick Seminar in Political Philosophy 16 October 2012 wrongdoer. These remarks may seem to go in the direction of situating blame squarely in the retributivist model of sanction. However, my point here is that blame does not work primarily as a means to separate the moral offender from the community. Obvious examples belong to the experience of parenting, where the key function of blame is educational and inclusive more than punitive or corrective. In such cases, blame marks and thereby teaches the difference between right and wrong, and it is mainly directed to integrate the child in the moral community. Restorative justice offers more complex examples of the constructive and transformational functions that reactive emotions such as blame may play. Blame can be the input of complex processes of repairing the harm caused by crimes. As Susan Wolf remarks, restorative justice aims “to address crime in a way that encourages victims to express and work through their anger, rather than to separate the criminal from the victim and her community so as to bypass their anger altogether” (Wolf 2011). Furthermore, punishment may be intended as a means to mend impaired moral relations. It can be a necessary step in the communal practices of moral repair (Walker 2006: 29 ff.). Communities are responsible for adequately addressing moral wrongs. They are responsible for reiterating and reasserting the standards and norms that the transgressor violated. But they are also responsible for the legitimization and enforcement of the wrongdoer’s acceptance of responsibility. In programs of restorative justice, the first step is to create opportunities for victims, offenders and community members to discuss the crime and its aftermath. Understandably, the emotion of blame takes center stage in these meetings. The role of the educators is to direct such emotion so as to elicit an appropriate moral response from the side of the transgressor and the victim. Through appreciation of blame the wrongdoer is 15 University of Warwick Seminar in Political Philosophy 16 October 2012 encouraged to recognize the wrong done, amend, be disposed to remedy, and prepared to restore the impaired relation. From the perspective of victims, it is the reassurance that the wrong done is appropriately addressed. Thus, blame aims at reintegration of wrongdoers and protection of victims, not at withdrawal of recognition. It is not an act that separates wrongdoers from their original community, but it is a form of moral address that aims at increasing moral awareness and, eventually, at healing, integration and inclusion. Finally, emotions play a pervasive role in structuring moral conflicts. Hate and fear are important drives of all kinds conflicts, including conflicts of interests, identity, and ideology. Working on emotions such as blame and anger is thus a crucial resource in devising strategies of conflict-resolution as well as grounding practices of moral repair. As Margaret Walker writes “reproving wrongdoers confirms that membership confirms that membership is conditional upon certain responsibilities entails forms of recognition and protection Assuring or satisfying victims confirms that membership entails forms of recognition and protection” (Walker 2006: 31). One important implication of Scanlon’s account of responsibility is that not everybody has the same standing to blame. There are different kinds of blame depending on the sort of relevant personal relation. But there is an asymmetry between personal and moral relations. In the personal case of friendship, reciprocation plays a large role. Not so for the general moral case: “the basic forms of moral concern are not conditional on this kind of reciprocation. Even those who have no regard for the justifiability of their actions toward others retain their basic moral rights – they still have claims on us not to be hurt or killed, to be helped when they are in dire need, and to have us honor promises we have made to them. … Moral deficiencies do not justify their general suspension” (Scanlon 2008: 142). This 16 University of Warwick Seminar in Political Philosophy 16 October 2012 passage indicates that the there are some fundamental moral claims and demands that others can legitimately place on us even when other more special bonds are severed or damaged. The question arises who is entitled to blame and on which grounds. For Strawson all moral agents are entitled to blame, irrespectively of the special position they hold. We can blame the wrongdoer as victims on the ground that we have been personally wronged, but we can also blame vicariously, qua representative of the moral community. The capacity for vicarious reactive attitudes tells something important about the bonds of moral community and human fellowship. Moral emotions are at the same time personal and impartial, that is, not fully determined and qualified by the special position we occupy in any given situation. Of course, special positions and relations are sources of special obligations and ground a large variety of moral emotions. However, everybody is entitled to blame, as much as they are entitled to judge others and their actions. The claim that the capacity for reactive attitudes and moral judgment does not rest on holding special positions is based on the presumption that we all have the same moral standing. (This is not an assessment of egalitarianism based on the value of humanity, but a normative condition of epistemic and moral parity). The concept of recognition-respect captures this insight. It is an emotional mode of recognition of normative status. Specifically, it is the key mode of recognition of moral autonomy. The qualification “recognitional” is ambiguous, though, because it may seem to suggest that respect is a form of emotional discernment of a metaphysical property. On a realist reading, respect is recognitional in this sense, because it responds to the value of dignity. This view is typically attributed to Kant, as he claims that respect is owed to persons because of their dignity, which is an intrinsic kind of value. In the 17 University of Warwick Seminar in Political Philosophy 16 October 2012 realist interpretation of Kant’s ethics, dignity is a value property that depends on autonomy, which is a metaphysical property of the will (Kant G 4.412). If there is no such metaphysical property, respect is simply misplaced because the ground upon which respect is addressed is found lacking. The realist view identifies autonomy as the metaphysical condition of the appropriateness of respect, and it is implied in many recent accounts of moral responsibility. For instance, Bennett Helm writes: “being morally responsible is a matter of being susceptible, out of respect, to the call of another’s dignity” (Helm 2011: 234). Stephen Darwall remarks that respect is “the fitting response” to the distinctive value that persons equally have (Darwall 2006: 119-120). While Darwall does not appear to side with realism it is an outstanding question which place his theory occupies in meta-ethics. 8 Helm is not explicitly committed to realism. In fact, he proposes this view tentatively, as a working hypothesis, exactly because he is aware of the fact that a convincing defense of recognition-respect requires that we take a decisive step in meta-ethics. I agree with Helm that an account of moral responsibility in terms of dignity raises issues about the nature of value and the boundaries of the moral community that can be answered only from a full-fledged meta-ethical theory. My point here is that to talk of respect as the morally appropriate response to a value property – such as dignity or humanity – is not an innocent meta-ethical claim. 8 Deigh convincingly argues that Darwall and Wallace misunderstand Strawson exactly on this issue, see Deigh 2011. Deigh objects that Darwall’s meta-ethics of respect is very distant from Strawson’s non-cognitivism. 18 University of Warwick Seminar in Political Philosophy 16 October 2012 Interpretative issues aside the realist model faces a significant problem. It is a fact that human agents appear to be more or less autonomous. But it is unclear what it takes to measure autonomy understood as a metaphysical property or as a mere capacity. As Bernard Williams has pointed out “though secular [the Kantian view] is equally metaphysical: in neither case is anything empirical about men that constitutes the ground of equal respect” (Williams 1973: 235). The conclusion is troublesome: “to hold a man responsible for his actions is presumably the central case of treating him as equally responsible, there is no much left to their equality as moral agents. If, without its transcendental basis, there is not much left to men’s equality as moral agents, is there anything left to the notion of respect owed to men?” (Williams 1973: 235). More recently, John Skorupski voices a similar worry: “Kant’s doctrine is that every rational being, through its own inherent spontaneity—through the causality of freedom—has insight into all pure reason relations. Quite clearly, however, what especially matters to Kant is an egalitarianism of pure moral insight. It drives Kant’s insistence that morality is accessible to everyone, that arriving personally at a correct moral judgment is open to the simplest human being. Every human being merits absolute respect, because every human being has absolute moral insight” (Skorupski 2010: 175). The constructivist proposal promises some resources to address this question. There is no presumption of equality in terms of equal moral insight, as Skorupski assumes. There is no presumption of equal moral competence, as Williams assumes. By contrast, the constructivist claim is that in order to reason correctly our reasons must address our peers – however we construe this category. This is not because we are as a matter of fact all similarly situated relatively to the moral object, or because 19 University of Warwick Seminar in Political Philosophy 16 October 2012 we have equal access to moral reality. The claim is that to rationally form a normative judgment, we need to take into account others as epistemic and moral peers. This is a constitutive requirement of reasoning, not a moral value from which moral duties are derived. The condition of parity is a requirement of rationality, that is, a constitutive norm about forming reasons. It does not constrain the content that can be shared, but the considerations that could count as reasons in the rational exchange with others. To qualify as reasons considerations must be such that they address others as peers, and thus exhibit a law-like form. This formal constraint grounds the genuine authority of reasons. If reasons appealed only to some groups or other, on the basis of shared interests or properties, instead of been sharable by all peers, then their authority would be also limited and local. The subjective counterpart of this claim about the constitutive requirement of parity is recognition-respect. On this view, respect is not a mere “response” to a preconstituted moral object. In fact, to call respect a moral feeling indicates exactly that it is not a mere emotional response, but it is expressive and constitutive of moral activity (Kant CrPR 5.73). On the Kantian view, moral feelings are not merely emotional reactions to events or responses to value. Their role is active, not receptive. This claim might be perplexing because emotions seem to have some passive component in that they are never fully under the agent’s control, unless they are fake or simulated – as in the actor’s studio. The puzzlement does not vanish if we restate the Kantian definition that the moral feeling is peculiar in that it is sensitive and yet a priori. But the ultimate point here is that some emotions and, specifically moral feelings, are a priori with respect to empirical reason. They are not generated by or directed to desirable features of objects, but they are directly aroused in reasoning. The basic idea is that respect is a subjective feeling peculiar to the activity of 20 University of Warwick Seminar in Political Philosophy 16 October 2012 agency. It is a priori because it can be aroused by the mere thought that something is right. This is not to say that it precedes our moral activity of reasoning in conformity with moral norms. Instead, respect is the subjective awareness of the operations of such norms in us. It constitutes what it feels like for us to perform practical reasoning. The argument thus points to the phenomenon of subjective moral authority. The claim is that, in deliberating, the agent experiences her freedom immediately, that is, as a fact that she cannot derive from higher principles or justify by introspection or further reflection on the nature of agency (Kant CrPR 5.31). The moral feeling of respect or “reverence for the law” conveys our awareness of rational agency and shows our responsiveness to the demands of practical reason. It is an emotional mode of practical knowledge of oneself as an agent, not a mode of moral discernment of a value-property. Practical knowledge is the immediate awareness of one’s doing that action.9 It identifies the relationship of ownership and authorship that is constitutive of doing that very action intentionally. Recognition-respect is thus a constitutive mode of autonomy as the ground of responsibility. On this view, autonomy is not a monadic metaphysical property of the will, but a normative relation of authority that the agent establishes with herself as a representative of the moral community. This is mainly the capacity to critically assess and reflectively endorse one’s actions and attitudes as one’s own. The critical 9 Practical knowledge is not observational. Observation may provide useful information and thus play a role in the specification of the circumstances of action, but it does not play any foundational role in the understanding of the action as such. It differs from observational knowledge of the facts that some movements result in a certain sort of action. I develop this view in Bagnoli 2013. 21 University of Warwick Seminar in Political Philosophy 16 October 2012 standpoint is the standpoint of the moral community, understood as the community of agents capable of critically assessing emotions, beliefs, actions and relations. This sense of “moral” community is minimal in that it does not fully determine the contents of actions and attitudes that the agent can legitimately authorize. It is basic in that it is the standpoint of rational assessment. Autonomy grants that there is a special normative relation between the agent and her action (belief, or attitude). Such a relation is special because it cannot be alienated without the very actions being radically different or even cancelled. Since I am construing the special relation as normative (rather than in terms of attribution of states or actions). Such relation is epistemic and normative in that it constrains beliefs and actions that the agent authorizes and for which she claims authorship. The qualification “epistemic” does not indicate that the agent is infallible or incorrigible regarding the attribution of mental states to herself. The point is not that the agent has privileged access to her mental contents. Nor does it imply that the agent is in full control of her mental contents. While agents are not fully in control of all their own mental contents, they can exercise some distinctive authority over them. In some cases, this is the authority to shape them, direct them, or make them inefficacious. In other cases, it is the authority to disown them, and treat them as obstacles. These are all forms of reflective self-government at which we can fail or succeed. As a normative relation, autonomy accounts for the special authority that the agent claims on her mental states. This claim is congruent with the common subjective experience that the agent normally represents herself as in a special position regarding her own actions, attitudes, and beliefs, e.g. she represents herself as the author of her actions. There are cases in which the agent does not feel her actions 22 University of Warwick Seminar in Political Philosophy 16 October 2012 or beliefs as fully hers, but these are pathological or borderline cases. Normally, agents recognize actions they do, attitudes they feel and beliefs they hold as their own. The agent is in the privileged position of being the only one entitled to authorize her actions, attitudes, and beliefs as hers. However, the criteria of authorization are not special to the agent’s own position, but the same that holds for the moral community. In other words, the agent’s criteria of critical assessment are the same as those valid for the moral community. The accordance with the standpoint of the moral community does not have to be explicit. Typically, the need to authorize actions is manifest in response to a request of intelligibility and justification. The practice serves as a model for understanding the reflective standpoint. In claiming authorship of her action, the agent represents herself as a member of an ideal community of agents having equal standing. This ideal community not only constrains what agents can do, but it also shapes the way in which agents conceive of their action. It requires that our actions make sense to others. Others have the authority to ask for reasons insofar as they accept the burden of offering reasons to us as well. This is to say that they stand in a relation of mutual recognition with us. Mutual recognition does not guarantee that we respond to the demands of others and exchange reasons with them. It does not imply that others provide us with external normative standards or that we ought to recognize them as independent sources of our standards for acting. As the grounding relation of authority, mutual recognition governs not only the exchange, but also the formation and the conception of reasons for action. Agents have authority on themselves insofar as they respect others as having equal standing. This is called requirement of publicity. Reasons are not public because they are addressed to others or modeled on their requests. They are 23 University of Warwick Seminar in Political Philosophy 16 October 2012 public in the deeper sense that the need of them arises only if and at the same time as we relate to others as having equal standing. The upshot is that reflective agency has a dialogical structure, as the normative relation of authorship that one has with one’s actions is made possible by public criteria that constitutively involve others. I am sketching a model of rational agency as self-government, which is distinctive in two respects. First, in contrast to the models advocated by Harry Frankfurt and Christine Korsgaard, I argue for the constitutive role of others in reflection. I hold that agential authority is a normative relation established by appealing to public standards that are made authoritative by mutual respect. There must be a public context for the question of rational agency and responsibility to arise. Second, in contrast to social models of recognition, such as e.g. Honneth’s and Skorupski’s, to appeal to public standards does not amount to reducing moral membership to membership in concrete community. Rather, it is an attempt to clarify the dynamics of reflection that underlie autonomy in dialogical terms. To frame the question of agential authority in terms of public standards is to offer an alternative explanation of the claim that reflective agency is principled. The basic idea that motivates the dialogical model is that in offering reasons for action we are responding to problems that are conceived as such from within a public space. In this sense, agential authority is fundamentally public. 5. Punishment and respect On the view I defend, reactive attitudes are forms of moral address, hence modifications of recognition-respect. Recognition-respect plays a governing role in relation to other moral attitudes directed to practical subjects. In this sense, respect has priority over blame. This means, first, that blame is appropriately directed only 24 University of Warwick Seminar in Political Philosophy 16 October 2012 toward those subjects who are the appropriate object of respect. More generally, when emotions play the role of moral address, their conditions of applicability are governed by recognition respect. And, second, it means that blame and other moral emotions are morally justified insofar as they are governed by respect. The first point concerns the scope of application; the second concerns the normative criteria of justification. Perhaps the most important normative implication of refocusing the debate on respect concerns the issue of withdrawal. While the attitude of blame expresses withdrawal of recognition, it appropriately does so only when it is governed by respect. When we make explicit the reasons why a member of a given community is appropriately found blameworthy, such reasons refer fundamentally to respect. The community finds a criminal guilty on the presumption that he has moral standing and is morally competent. This is not to deny that moral offenders are sometimes under the moral threat of withdrawal of recognition. Nor does it imply that moral competence is the only feature at stake in deciding whether to undertake the objective or participant’s perspective. As Strawson points out, there may be political reasons for taking up the objective perspective instead of the participant’s perspective and thus refrain from moral engagement. But then, again, membership is accorded on the basis of considerations that have nothing to do with the individual capacities for moral standing. In the fundamental sense in which respect is the concept governing moral status, respect cannot be withdrawn – unless moral standing is withdrawn. There are moral wrongs (such as violation of persons) that call for heavy moral (and juridical) sanctions, such as coercion. But when we coerce to stop wrongdoing we do not deny the moral standing of the wrongdoer. Rather, we judge her punishable because we 25 University of Warwick Seminar in Political Philosophy 16 October 2012 consider her as having full moral standing. On this view, then, punishment is not a withdrawal of recognition, but a form of moral address. Likewise blame is a moral sanction that addresses a member of the moral community: it presupposes membership and it is justified on the presumption of membership. We do not punish people if they do not have moral standing. Those standing outside the moral community of peers are simply outside the moral picture: they do not qualify as interlocutors in our moral discourse. In Strawson’s words, we are legitimated (and sometimes required) to undertake an objective stance toward such people, and not resent and react against their apparent ill will. Boundaries are blurred, and some rational and emotional incapacities are also and at the same time character flaws. But the difficulty in drawing the line does not undermine the main point, which is that the insistence on the withdrawal of recognition is misleading. The right to use moral sanctions (such as blame) springs from the recognition of the wrongdoer as having equal standing in the moral community, and therefore as somebody to be addressed as a wrongdoer, rather than obstructed as morally incompetent. Blame does not signal withdrawal of full membership in a community or in a relation. Blame tells the wrongdoer of her wrong, thereby judging her morally competent and thus capable of understanding where the wrongness lies. By being blamed she is not extruded as an outlaw but judged according to the internal standards of the community to which she belongs. A failure to hold a person accountable for wrongdoing is tantamount to a failure to respect her autonomy. Correspondingly, so-called “redeeming emotions” – such as remorse, regret, shame, and guilt, are not meant to restore the transgressor to a full membership. Rather, they are reparative in more complex and pervasive ways; for instance, they 26 University of Warwick Seminar in Political Philosophy 16 October 2012 have complex functions of mending the relations with the victims and the representatives of the community, but also serve the purpose of restoring one’s own integrity. They make sense on the basis of the agent’s self-representation as responsible. 6. Some advantages of the constructivist view In these closing remarks, I would like to point out some advantages of the constructivist proposal. First, there is the advantage of rescinding the (practical) issue of responsibility from the (metaphysical) issue of metaphysical freedom. Moral membership is not decided on the basis of some metaphysical features (e.g. freedom as a property of the will). It depends on rational practices of mutual recognition. To this extent, the question is neither straightforwardly empirical, nor does it pertain to descriptive metaphysics (as it does in the realist models of respect as a response to a value property). As it appears, this is a feature that this model retains from Strawson’s original model in terms of the general category of reactive attitudes. But there are additional features that make the constructivist proposal distinctive and preferable to its competitors. The constitutive account has the merit of refocusing the issue of responsibility on agential authority. To take responsibility for one’s action one is required to vindicate an action as one’s own, authored rather than caused by means of one’s limbs. Agential authority is a normative relation that one establishes with oneself in reflection and in virtue of the capacity of self-reflection. The constructivist model accounts for such relations without assuming unargued values, such as the value of humanity. Furthermore, it attributes a constitutive and structural role to moral emotions in the practices of responsibility. It thus vindicates the experience of moral 27 University of Warwick Seminar in Political Philosophy 16 October 2012 failing and feeling responsible. A third advantage of this proposal concerns the scope of moral responsibility. In traditional retributivist models, agents are responsible for what they do. On the view I advocate, responsibility is connected both to character and agential integrity. Moral responsibility does not narrowly coincide with voluntariness; it extends to propositional attitudes, and to reactive, reparative, and redeeming attitudes, insofar as they are modes of one’s autonomy, and even though they are not “the object of choice” in the same way actions can be. In this sense, we are morally responsible not only for actions, but also for all attitudes for which we are rationally criticizable. Moreover, in some cases, the distinction between being capable and not being capable is normative: some incapacities are moral; others are subject to moral assessment because they can be re-directed or modified by training. It may be objected that far from being an advantage of the model the extension is problematic. For instance, Thomas Pink finds that this extension implausibly departs from the ordinary view of moral responsibility (Pink 2004). This position is widely shared, mainly because it is assumed that moral assessment and moral demands appropriately apply to mental states that we can control. The objection has some plausibility, since it is does sound unfair to be blamed and punished for things that we cannot avoid or choose. It seems that moral agents can appropriately incur moral judgment and be subjected to moral obligations only regarding acts and states that they can choose. For instance, Jay R. Wallace equates choice and control: “Particular states of emotion or feeling, however, are not the sorts of states that can directly be controlled by the reasons expressed in moral principles: such states as love, esteem, and goodwill are generally not states that could be produced simply by the belief that there are moral considerations that make them obligatory. This is why 28 University of Warwick Seminar in Political Philosophy 16 October 2012 we cannot plausibly interpret moral obligations as governing the quality of people’s will, where such qualities are construed broadly, to encompass emotions and feelings quite generally” (Wallace 1996: 131-132). It is true that emotions differ from actions in that we do not summon them at will. However, this is not a definitive reason for claiming that we are not responsible for emotions and all sorts of propositional attitudes that escape our direct volitional control.10 Ordinarily, we hold people accountable not only for what they do, but also for how they do it, and for the sort of attitudes and emotions they feel in relevant situations. Ordinary moral judgment targets people because they lack the appropriate emotional sensibility to others, not only because they have done something wrong in violating a norm. In fact, some interesting cases of moral perplexity concern the right attitude to undertake, and more fundamentally, whether to undertake a personal and/or objective standpoint in assessing some harm suffered or done. Agents are held responsible not merely for weighting pros and cons in deliberation, but also for their efforts to overcome, repent, mend, repair, forgive, and change attitudes in due time. 7. A qualification about moral consideration Reactive attitudes are part of the vocabulary of respect and recognition, which is the key concept in the practices of attributing and taking responsibility. While recognition-respect governs the attribution of moral status it does not similarly govern moral relevance. That is to say that to define the bounds of the moral community by the concepts of autonomy and respect does not support the view that only autonomous 10 Smith rightly notices that this is a feature of emotions as well as of beliefs, see Smith 2011: 239. 29 University of Warwick Seminar in Political Philosophy 16 October 2012 agents are morally relevant. On the contrary, to undertake the objective stance and withhold judgment and reactive attitudes (such as hatred and resentment) is a mode constitutive and expressive of moral consideration. It would not be simply mistaken to hold a grudge against an insane person; it would be morally inconsiderate; it would show lack of moral understanding. It would be a moral failure. Bibliography Bagnoli C. (2007) “Respect and Membership in the Moral Community”, Ethical Theory and Moral Practice, 10/2: 113-128. Bagnoli C. (2011) “Emotions and the Authority of Moral Reasons”, in Bagnoli 2011a: 62-81. Bagnoli C. ed. (2011a) Morality and the Emotions, Oxford University Press. Bagnoli C. (2013) “Constructivism about Practical Knowledge”, in Bagnoli 2013a: chapter 7. Bagnoli C. ed. (2013a) Constructivism in Ethics, Cambridge University Press, forthcoming. Darwall S. (1977) “Two Kinds of Respect”, Ethics 88: 36-49. Darwall S. (2006) The Second-Person Standpoint. Morality, Respect and Accountability, Cambridge Mass.: Harvard University Press. Deigh J. (2011) “Cognitivism about Reactive Attitudes”, in Bagnoli 2011a: 197-216. Feinberg J. (1971) “The Expressive Function of Punishment”, in Doing and Deserving, Princeton University Press, 1971: 95-118. Fischer J.M. & M. Ravizza (1998) Responsibility and Control, Cambridge: Cambridge University Press. 30 University of Warwick Seminar in Political Philosophy 16 October 2012 Freeman S., Kumar R., Wallace J.R. eds. (2011) Reasons and Recognition: Essays on the Philosophy of T.M. Scanlon, Oxford University Press. Frye M (1983) “A Note on Anger,” in Frye 1983a: 84-94. Frye M. (1983a), The Politics of Reality, Freedom: The Crossing Press. Gibbard A. (1990) Wise Choices, Apt Feelings, Oxford: Clarendon. Helm B. (2011) “Responsibility and Dignity”, in Bagnoli 2011a: pp. 217-235. Honneth A. (1992) “Integrity and Respect. Principle of a Conception of Morality Based on the Theory of Recognition”, Political Theory, 20: 187-201. Mill J.S. (1863) Utilitarianism, ed. Roger Crisp, New York: Oxford University Press, 1998. Pink Th. (2009) “Reason, voluntariness and moral responsibility”, in O'Brien & Soteriou 2009: 95-120. O'Brien L., and M. Soteriou eds. (2009) Mental Actions, New York: Oxford University Press. Raz J. (2001) Value, Respect, and Attachment. Cambridge: Cambridge University Press. Scanlon T. (2008) Moral Dimensions: Permissibility, Meaning, Blame, Belknap Press of Harvard University Press. Smith A. (2005) “Responsibility for Attitudes: Activity and Passivity in Mental Life”. Ethics (2): 236-271 Smith A. (2011) “Guilty Thoughts”, Bagnoli 2011a: 235-256. Skorupski J. (2006) “Blame, Respect and Recognition: A Reply to Theo van Willigenburg”, Utilitas: 17: 333-347. Strawson G. (1987) “The Impossibility of Moral Responsibility”, Philosophical Studies, 75: 5-24. 31 University of Warwick Seminar in Political Philosophy 16 October 2012 Strawson P. (1968) “Freedom and Resentment”, in Freedom and Resentment, London: Methuen, 1974: 1-25. Van Willigenburg Th. (2003). “Shaping the Arrow of the Will: Skorupski on Moral Feeling and Rationality”, Utilitas, 15: 353-368. Wallace R.J. (2011) “Dispassionate Opprobrium”, in Freeman et al. 2011: 348-372. Wallace R.J. (1994) Responsibility and the Moral Sentiments, Harvard University Press. Walker M.U. (2006) Moral Repair, Cambridge University Press. Watson G. (2004) Agency and Answerability, New York: Oxford University Press. Wolf S. (1990) Freedom within Reason, New York: Oxford University Press. Wolf S. (2011) “Blame, Italian Style”, Freeman et al. 2011: 332-348. 32