Abstract Kantians may be unable to derive all of benevolence from reverence for rational agency, but the remaining lacuna is not as extensive as Arpaly thinks. For while we should take seriously Kantian worries about separating benevolence from reverence, a considerable part of benevolence can be explained in terms of reverence for rational agency on a plausible intepretation of the latter. Furthermore, Kantians have an irreducible role for benevolence within their ethics, which is different from the role of a self-standing virtue. Let me begin by saying where I agree with Nomy Arpaly (2018). I think she is right that benevolence (understood as a concern for another’s well-being) is a virtue and that it cannot be fully reduced to reverence for rational agency. To the extent that Kantians claim that reverence for rational agency is the only genuine moral virtue, they fail to capture something which we normally consider as valuable and their moral theory requires a revision of the pre-theoretical views that inform our everyday life. I do, however, disagree with Arpaly in two important respects. On the one hand, in the assessment of how far Kantians depart from our pre-theoretical views, and on the other hand, on the question of how the relation between benevolence and reverence for rational agency should be understood. Kantians of the sort Arpaly addresses miss something, but when it comes to the question of which actions we ought to perform, they miss considerably less than Arpaly thinks. There are two main reasons for our differing assessments on this point. First, we have a different understanding of what is included in reverence for rational agency; second, I think that Kantians can make a stronger case against divorcing benevolence from reverence than Arpaly is willing to allow. As a result, most of the actions which commonly come under the heading ‘benevolence’ are also actions which we should perform out of reverence for the other’s rational agency. However, as I will suggest, we should not conclude from this that benevolence on its own has no genuine role to play for Kantians, since caring for another’s well-being is crucial when it comes to fulfilling our imperfect duties to help in particular cases. The emerging picture, though, is one on which benevolence and reverence for rational agency are much more intertwined, rather than being self-standing virtues. On Arpaly’s interpretation, reverence for rational agency has two main components: promoting the ends the other person has set herself, and preventing threats to her rational agency itself (2018, pp. 209, 219). As Arpaly points out, neither of these components entails that we should specifically care about another’s well-being, since the other person may not have her own well-being as an end which she ranks particularly highly, and a diminution of her well-being need not interfere with her capacity for rational agency. Nonetheless, we seem to have a weighty moral reason for preventing another’s suffering even when the suffering would not interfere with her rational agency or ends, which strongly suggests that this reason cannot be explained by reverence for rational agency alone. By contrast, I will argue that reverence for rational agency plausibly includes significantly more than the two elements Arpaly discusses, since it must be informed by a proper understanding of the nature of the specific activities in which rational agency is realized. For this reason, reverence must include a concern about the agent’s having the abilities to reach her ends and about her ‘getting things right’. In addition, Kantians can make a strong case that benevolence, at least generally, cannot be separated from reverence for the other as a rational agent. For without looking at the ends the other person has, we risk falsely projecting our own understanding of ‘what would be good for her’, and therefore risk failing to really benefit her. We can only avoid this danger by taking the agent’s own choices of ends to be, normally, authoritative answers to the question of what is good for her. While this means that, in a crucial respect, benevolence ‘depends on’ reverence, there is a dependence in the opposite direction too. Unless we had some concern for the well-being of others and a tendency to be negatively affected by their suffering, we would be practically incapable of fulfilling our moral (reverence-based) duties to help. Benevolence is essential here because it allows us to sufficiently specify our general duty to help so that we can be guided by it in particular cases. In the following, I will flesh out these points in five steps. After briefly setting out Kant’s own views on benevolence in i, I turn to a two-pronged (partial) defence of the Kantian view. In ii, I argue that a plausible understanding of reverence for rational agents covers, at least extensionally, many—if not most—of those things which normally go by the name of benevolence (because in those cases reverence and benevolence require us to perform the same actions). iii presents what I consider as the main Kantian worry about benevolence when it does not extensionally coincide with reverence. In iv, I suggest that while not all of benevolence can be explained in terms of reverence for rational agency, the scope left for benevolence to enlarge the range of actions we ought to perform is fairly limited. However, benevolence can still play an important and irreducible role in the Kantian picture, once we adopt a more plausible view on how benevolence and reverence for rational agency interact (v). Two caveats before I begin. First, I will only be concerned with benevolence towards rational adults. As in Arpaly’s own discussion, benevolence towards small children, animals and people with mental illness will be bracketed. This restriction is not a trivial one, since the boundaries between rational adults and children or people with mental illness are not always clear-cut. One may therefore think that, in so far as our behaviour towards the latter should be governed by a self-standing virtue of benevolence, the same virtue must also govern, to some extent, our relationship to rational adults. But this line of argument against the Kantian view is one I will not go into here. Second, it is a tricky question what factors are relevant to determining a person’s well-being. I take myself to be following Arpaly when I focus on one aspect of well-being, namely, the aspect of feeling happy or unhappy, misery, suffering, pain or joy. These are at least the elements of well-being she focuses on in her discussion, with the additional emphasis that benevolence is mainly concerned with protecting people from ‘ill-being’. Plausibly, there is more to well-being than that. But I will stick to this more limited aspect here, because it seems to raise particular difficulties for the Kantian view. I In the Metaphysics of Morals, Kant states that we have a duty to further, within limits, the happiness of others. But when describing the duty to care for the happiness of others in the introduction to the Doctrine of Virtue, Kant characterizes it as a duty to adopt and promote (some of) the ends of the other person: ‘When it comes to my promoting happiness as an end that is also a duty, this must therefore be the happiness of other human beings, whose (permitted) end I thus make my own end as well’ (MM 6:388).1 (Similarly, in MM 6:450, Kant describes it as ‘the duty to make others’ ends my own’.) This does not mean that Kant at this point strictly identifies the other’s happiness (‘Glückseligkeit’) with the attainment of her ends. As Kant makes clear, I may disagree with another person about what she regards as conducive to her happiness, and as a result may, licitly, refuse to help her in those respects: ‘It is for them to decide what they count as belonging to their happiness; but it is open to me to refuse them many things that they think will make them happy but that I do not’ (MM 6:388). Since I could not, without being conceptually confused, disagree with the other person’s judgement in this way if the other’s happiness just (by definition) consisted in the attaining of her ends, Kant must therefore allow the possibility that happiness and attaining one’s ends can come apart, and that someone may be wrong in setting her ends when it comes to what is conducive to her own happiness. But, for Kant, I must submit to the other’s judgement in so far as furthering his happiness (in the sense relevant for virtue of beneficence) consists (only) in adopting and promoting a subset of his ends. I need not further all his ends, but I cannot add new ends, whose attainment I think might be ‘good for him’, but which he hasn’t adopted himself. However, in addition to the duty of beneficence, Kant discusses a range of other ‘sociable’ feelings, tendencies and duties in the Doctrine of Virtue, which might provide candidates for something like the virtue of benevolence which Arpaly is concerned with. Setting aside the natural predisposition to love one’s neighbour, whose possession Kant lists under the necessary preconditions for being receptive to the concept of duty (MM 6:399 ff.), the prima facie most promising candidate is the (indirect) duty we have to sympathetically engage in one another’s joys and pains (MM 6:456–7). This duty might provide the grounds for a much wider duty of benevolence than the duty of beneficence, since the scope of the sympathetic engagement cannot be restricted to the frustration or fulfilment of the other’s ends. For the feelings which we sympathetically engage in (joy and pain) do not all come from the other's rational nature; nor are they restricted to rational agents. That another feels pain can often be realized without first realizing that an end of hers has been thwarted or that her rational agency been impaired, and, on many occasions, pain does not even indirectly involve rational agency, for instance, in small children. This fits well with Kant’s own emphasis that our sympathetic engagement looks beyond our purely rational nature to our animal nature: ‘a human being is regarded here not merely as a rational being but also as an animal endowed with reason’ (MM 6:456, emphasis added). Kant believes we ought to cultivate this disposition for sympathetic engagement with another’s feelings and, in particular, with their pain, entreating us not to ‘shun sickrooms or debtor’s prisons … in order to avoid sharing painful feelings [of sympathy] one may not be able to resist’ (MM 6:457). On a plausible understanding, our sympathetic engagement with the pain of others will regularly give rise to some motivation to help (at least if I don’t shun the situations which give rise to it): it is hard to imagine that I experience the other’s experience of pain as painful (to myself) without having any such motivation. So don’t we have here, even in Kant, something like a virtue of benevolence, which is neither identical nor reducible to reverence for rational agency? For Kant obviously believes that our disposition for sympathetic engagement is a valuable disposition which we ought to cultivate, and this disposition will typically and non-accidentally manifest itself in our caring about others and being motivated to help them when they are suffering. Is this not just the virtue of benevolence that Arpaly is looking for? But I think one can well retort that Kantian benevolence, thus construed, misses out something important. For Kant, the value of sympathetic engagement is merely instrumental: we ought to foster this disposition only because and in order that it helps us fulfil our primary moral duties.2 Nor does Kant regard this sympathetic engagement as valuable in all instances (MM 6:457); as he implies, it is only valuable in so far as it is guided by a proper understanding of moral value. This purely auxiliary and subordinate role of sympathy raises doubts about the status of our sympathetic disposition as a genuine virtue: virtuous activity—many philosophers have thought, following Aristotle—is something we should engage in ‘for its own sake’. If one agrees with this tradition that a virtuous activity should be (also) non-instrumentally valuable, one can thus reasonably deny that the disposition for sympathy constitutes a genuine virtue. II Let us turn from Kant himself to his more recent followers. Following Arpaly, I will focus on versions of Kantianism which take the only ultimate moral demand to be one of reverence for rational agency or autonomy. How much of benevolence will such accounts miss? Arpaly thinks quite a lot, while I am more optimistic on the Kantians’ behalf. In this section, I will explain why, proceeding from two different starting points. First, I will argue that, even when we just look at the element of promoting another’s ends, Kantians can help themselves to some natural assumptions about which ends people pursue and how important they are for them (2.1). Second, I will suggest that a plausible understanding of ‘reverence for rational agency’ includes considerably more than promoting another’s ends (2.2). 2.1. The minimal Kantian response to Arpaly is that even if reverence did only include promoting the other’s ends, it would still require us to do a lot of the things demanded by benevolence, since we should not only promote ends we positively know the agent to have, but also those ends we can naturally presume her to have. Similarly, a Kantian can argue, we can make some natural assumptions about what importance agents ascribe to certain ends. Arpaly herself acknowledges the possibility of this kind of move (2018, p. 215), allowing that Kantians can presuppose that (almost) everyone has her own well-being as a major end. While I am sceptical that just this assumption will be very helpful for the Kantian here, I am much more optimistic when it comes to presupposing some more specific ends, such as the avoidance of pain. It is indeed natural to assume that everyone has her own well-being or happiness as an end. However, just knowing you have this general end gives me no real practical guidance in helping you. The notion of ‘well-being’ on its own is far too abstract and could be filled out in far too many ways (since different people enjoy and dislike wildly different things). It needs much more specification in order for me to have a sufficient grip on what is likely to make you happy or suffer, so that I can have some idea of how to benefit you. If natural assumptions about agents’ ends are going to help the Kantian, the end in question cannot therefore just be ‘well-being’: more specific ends must be in play, ends which determine how you want to pursue and realize your well-being. There are several such more specific ends that we can naturally assume people to have, though. The most obvious case is the avoidance of pain and suffering (though similar arguments can be developed for other basic needs). The natural presumption is that agents capable of suffering have the end to avoid (further) suffering, both as a general standing end to avoid pain and for particular instances of pain (that is, to end or alleviate these specific instances). This presumption can be defeated in particular cases—think of a boxer who is willing to undergo major pain in his training in order to learn to take punishment. But without specific evidence to the contrary, we can safely assume that in preventing or alleviating a person’s pain we are also promoting ends she actually has. However, Arpaly argues, this assumption does not yet explain why protecting the agent’s well-being is usually morally more urgent than helping her to achieve another important aim. Protecting someone from the threat of being beaten up seems morally more urgent than helping her with another positive project which is important to her, even when she herself does not value her well-being more than realizing this project (Arpaly 2018, p. 218). But it does not seem correct to me that preventing threats to someone’s well-being always has greater moral urgency than supporting her other projects. And to the extent it does have greater urgency, this can be given a reverence-based explanation, since it can be seen as a reflection of the fact that, usually, suffering and the thwarting of important ends of the person go together. Does the prevention of suffering or pain always have higher urgency? No. Pain and suffering come in degrees, and even the acute threat of mild pain need not provide a particularly weighty moral reason to prevent it. Imagine that you feel somewhat physically uncomfortable on your seat in the theatre and ask me to change places, and I refuse. I do so because I want to give my seat to someone who wants a better view of the stage in order to follow the play closely, because he is especially interested in this production. In this scenario, my choice is not necessarily morally objectionable: in order to determine whether it is, we would have to know how much physical discomfort and pain you are feeling and to compare that to how important getting a good view is to the other person. Even when we look at more drastic threats to well-being and the attaining of important ends for the person (which Arpaly is primarily concerned with), it is not true that the former has a kind of moral urgency which the latter could never have. Refusing to help with the latter can be genuinely malevolent, even when no expectation of suffering is involved. One case in point is the callous failure to respect the wishes of a dying person to help finish some crucial project of theirs. Imagine that a philosopher friend of yours on his deathbed gives you the manuscript he has been working on for decades and which was of immense importance to him. You know that it was one of his chief projects to get this research published, a project for which he has been, and is still, willing to make major sacrifices. Just before falling into terminal coma, he asks you to ensure that it gets to the publisher. But even though this would involve no relevant cost for you, you just cannot be bothered to do so once you have left the hospital. This does show a lack of basic decency, given that you know how important publishing his research has been to your friend. At the same time, that your action is deeply wrong cannot be explained in terms of your failing to prevent a threat to your friend’s well-being, for, given his falling into terminal coma, your friend’s happiness can no longer be affected by your behaviour. So preventing serious threats to an agent’s well-being is not always more morally urgent than preventing threats to the fulfilment of her other ends. Nonetheless, Arpaly is right that very often it is more urgent. But Kantians can interpret this as a reflection of the fact that suffering and the thwarting of an important end of the person are normally linked. This link has two components. First, we usually assume that avoiding pain and suffering is an end of the person which ranks particularly highly, at least in the particular situation where suffering is imminent or has already begun. This holds true at least for instances of pain and suffering which are severe and have some non-negligible duration. Even when, in a cool reflective moment someone might be willing to accept such pain for attaining some other end, things tend to change when significant pain is actually experienced. For pain is then usually experienced as something to be avoided with special urgency, which, the worse the pain becomes, crowds out other aims the agent is pursuing and makes her change her priorities. It is therefore a natural assumption that when people suffer pain or are about to do so, the end of avoiding this pain, or ending it, is of major importance to them, at least in the situation at hand. (Just think of people who suffer excruciating pain and whom we naturally assume to most want this pain to end, no matter what the costs.) Of course, this assumption can be defeated in special cases: for example, when the agent has made it clear that on this occasion avoiding the pain is less important to her than achieving another end and we can suppose that she hasn’t changed her mind. But then the moral urgency of alleviating the pain significantly diminishes too. There is a second aspect to the link between suffering and the thwarting of important ends. The intensity of a particular kind of suffering, namely, the suffering of disappointment, is a key criterion for how important the end was to the agent. When the agent easily gets over the thwarting and does not feel intense disappointment, we usually assume that the project was not so important to him after all (which makes cases like Arpaly’s Leonard case, pp. 217--8, so hard to assess: if Leonard ‘easily gets over’ the project’s failure, this will raise doubts about how much he cared about this project in the first place). The Kantian can therefore argue that in helping the agent to avoid a disappointment which would make her suffer significantly, we are also, ipso facto, helping her to achieve an end which is especially important to her. 2.2. In addition to relying on some natural assumptions about the ends persons have, Kantians can also insist that reverence for rational agency, on a plausible understanding, includes more than promoting (some of) the other agent’s ends. At its core, this point is an old one, since many Kantians believe that threats to my rational agency itself can provide you with urgent reasons to help me (for example, Herman 1993, p. 67; see Arpaly 2018, p. 219). For such Kantians, if the conditions on which my rational agency depends are threatened, you cannot remain indifferent if you truly respect me as a rational agent. But reverence for rational agency must go even further. For it must be informed by an adequate understanding of what rational agency is; otherwise rational agency would be little more than a ‘fetish’ that is venerated blindly. As it is, our rational agency is not realized in abstracto; we are never ‘simply rational’ while doing nothing else. Rational agency is realized in more specific activities such as forming a belief, setting oneself an end, or deciding against adopting one. It is in these activities that the exercise of our autonomy consists, and our reverence for rational agency must be informed by a proper understanding of the nature of these more specific activities. Many of these activities have constitutive aims or features, or are essentially directed at other things, and these features and connections must be adequately reflected in the content of reverence. In particular, in respecting your rational agency I cannot completely ignore or be indifferent to whether you reach the constitutive aims of these activities. Let me illustrate this for two kinds of rational activity, setting oneself ends and forming beliefs. 2.2.1. My activity of setting ends is not something that is ‘complete in itself’ or that it would even make sense to engage in for its own sake. Instead, it is something which necessarily points towards a further activity, namely, pursuing and trying to achieve the ends I have set. Without its reaching towards this further activity we could not even understand what ‘end-setting’ was: in setting myself an end, I must also have pursuing and achieving this end in mind; otherwise end-setting would not just be crazy, but impossible. This connection between end-setting and end-pursuit must be reflected in what my respect for you as an end-setter encompasses and requires of me. I cannot value you as an end-setter without having any concern at all about your being able to reach your ends—doing so would just be perverse. (This connection also explains what is truly tragic about Sisyphus: while Sisyphus can set himself ends—rolling the rock up the hill—,he is doomed to fail. The pointlessness of his endeavour already makes his end-setting pointless,3 and this makes him a tragic figure for us already when he is setting those ends, and not just later when the rock is going downhill.) This is not to deny that the ability to set ends and the ability to achieve ends are different abilities, and that, in special cases, I may consider your choosing some end to be more valuable than your achieving it. But we cannot generally divorce your end-setting and end-pursuing nature. I must, in caring about the former, also care about the latter and about your having, at least to some degree, means to pursue your ends (even those that you do not yet have, but may choose to pursue later on). An agent who is so physically disabled that he can no longer pursue his ends may still be an autonomous agent, in Kant’s sense. But I cannot be indifferent to his (in)capacity to pursue his ends as long as I care about his capacity as an end-setter. Similarly, it cannot leave me completely cold if the agent’s ability to pursue his ends is, for a certain time, impaired, for example, by illness. These considerations already significantly enlarge the range of things reverence for another’s rational agency gives us reason to care about. (This point, I take it, in part underlies Kant’s argument in the Groundwork that we must cultivate at least some of our talents. As end-setters we cannot help caring about having means to achieve our ends, even if we do not yet have the particular ends for which they might be useful. This general connection will not only hold in the first-person, but also in the second-person case.) 2.2.2. Let us turn to judging, or belief-formation. According to what I take to be a plausible view, this rational activity has a constitutive aim, namely, ‘getting things right’. Beliefs qua beliefs—to use Velleman’s phrase—‘aim at the truth’ (2000, ch. 11), and when I engage in the activity of forming a belief on some matter, forming a true belief must be the constitutive aim my activity is directed at. Otherwise it would not even qualify as a belief-forming activity. As a consequence, in so far as you care about me as a rational belief-former, this cannot just mean that you care about my having the capacity for forming beliefs. You also cannot be fully indifferent as to whether I get things right or wrong: if I don’t reach the constitutive end of my activity, this cannot leave you cold, if you have really understood what the exercises of the capacity you are valuing in me are about. That my reverence for you as a rational belief-former should include that I care, to some extent, about whether you get things right fits nicely with our ordinary understanding of what respecting another person requires. Imagine that we are engaged in a discussion about the shape of our planet and you seriously propose that the world is a disc. Would it be showing respect if I just left your opinion standing, not pointing out the many reasons that speak against it or trying to convince you that you are wrong? Clearly not. Such indifference to whether you hold absurd views is not respecting you—it shows, on the contrary, that I consider you to be beyond the pale or ‘beneath contempt’ when it comes to such discussions. If I properly respect you as a person, I will criticize your view and try to convince you of its falsity. This point applies to all cases of judgement: in so far as you judge falsely, I have a reason to try to get you to see things correctly. Of course, I cannot try to do so by manipulation or by force, since reverence for your rational agency doesn’t include concern for true beliefs as such, but only for true beliefs which you form autonomously. Therefore, if you remain stubborn in your false belief, I will have to leave matters at that. Even if you agree with what I have said so far about judging, you may wonder why this case deserves such detailed treatment in our discussion of benevolence. Trying to correct your errors is sometimes what benevolence requires, but it seems to be a fairly special case. But there is a possibility for Kantians to significantly broaden the scope of this point. Several contemporary Kantians believe that the decision and adoption of ends generally embodies a practical judgement about what is good (for example, Engstrom 2009, p. 49; Korsgaard 2012, p. 7). When I choose a certain end, I must also judge it to be good. If this is true, then in so far as I believe your judgement ‘not to get things right’, the earlier argument shows that I have a reverence-based reason to try to set you right. This will be so even when your error lies in not properly taking into account your own interests or needs. And even for a Kantian there is considerable scope for error here: you may, for example, not pay proper attention to your own ‘true needs’ (Herman 1993, p. 55), misjudge how the effects of an action ‘will feel for you’, forget about other conflicting aims you have, or risk undermining your effective pursuit of ends in general (for example, by ruining your health). When you do so, I have a reverence-based reason to try to set you right. This reason is the same as the one I would have in cases of false judgements that have nothing to do with your well-being; so it is not ipso facto connected to benevolence. But what I have reason to do in such cases is to try to get you to accord proper weight to your own needs and interests—that is, to do what benevolence would tell me to do as well. Of course, trying to get you to look after your own interests and needs is not the whole of benevolence. But it covers a lot of cases of benevolence which are often held to be problematic for Kantians (including, I take it, Arpaly’s case of Mirja and Alaitz, p. 215), since it is often thought that Kantians must disapprove of this kind of ‘meddling’ as overly paternalistic. From the considerations presented in this section, we can expect that reverence will already demand of us a considerable number of the things which usually come under the heading of ‘benevolence’. But I think a Kantian should admit that reverence won’t explain all items we pre-theoretically call ‘benevolent’.4 In particular, it doesn’t seem to explain many instances of low-level everyday kindness. Imagine, for instance, that, being your housemate, I believe that you would enjoy listening to your favourite Mozart opera, but that you are very much focused on your work nowadays and don’t give enough thought to what kind of things will make you feel good. While you are working in your room, I therefore put on a cd of your favourite opera, in order that you will listen to it (knowing that this will not interfere adversely with your work, but will not make you work better either). This action of mine is ‘nice’, benevolent and morally good (though of a pretty low-level category). But it does not seem amenable to an explanation in terms of reverence for your rational agency (even on a properly expanded understanding of reverence), since it doesn’t have anything to do with protecting your rational agency, or promoting the ends you have set yourself or which are constitutive of your rational activities. III Is this not enough for Arpaly’s critique to succeed, since it shows that, as a complement to the virtue of reverence, we need to accept benevolence as an additional virtue? But to see whether that is true, we have to look at the Kantian ‘case against benevolence’; for if this case is sound, the reverence view may not miss anything which is of genuine moral value. I do agree with Arpaly that benevolence is neither in essential tension with respect nor necessarily includes treating the other person as weak or incompetent in a morally objectionable way. But there is another problem with benevolence, or rather with benevolence unless it extensionally coincides with reverence for rational agency, and this seems to have been one of the main problems which troubled Kant himself. At first, this problem can be seen as a purely epistemic one: benevolence is concerned with another’s good or well-being; but, as I pointed out earlier, the other’s ‘well-being’ is far too abstract a notion on its own, and needs further specification in order to guide my actions. And how can I know what will be good for the other, or make him happy, except by looking at his ends? Otherwise I run the risk of falsely projecting my own conception of happiness onto him, by assuming that those things which bring pleasure to me will also bring pleasure to him, and that their relative contributions to his happiness will be the same as they are in my case. And doing someone ‘good’ on the basis of such a false projection is not really doing him good at all: I falsely think I am helping him or doing him a favour, but really I am forcing something onto him that he does not like. This seems to be precisely Kant’s point when he says, ‘I cannot do any good to anyone in accordance with my concepts of happiness (except to young children and the insane), thinking to benefit him by forcing a gift upon him; rather I can benefit him only in accordance with his concept of happiness’ (MM 6:454). This point is not merely an epistemic one, though: it is not just that every person knows best what will make her happy, where the latter would be something that is already settled independently from what her ends are. In many cases, what makes us happy (constitutively) depends on which ends we have and how important these ends are to us. One reason for this is the constitutive connection I have mentioned in 2.1: how unhappy a particular failure makes you depends, inter alia, on how important reaching the aim in question was to you. Another reason is the satisfaction connected to achieving one’s goals itself. To a considerable extent, our choices thus determine what makes us happy. Taking these two points together—that you know better than me what will make you happy, and that your choices partly determine what makes you happy—I have good reason to go by your own choices in determining what is good for you. I should take them to be authoritative in this respect—meaning that I have good reason to believe that I will be better in figuring out what is good for you and makes you happy when I follow your own choices than when I try to figure it out by myself;5 for in the latter case I will be liable to the error of projection. If this is true, then we cannot in general be truly benevolent towards another person without showing respect to her as an end-setter, not only because this would be an inadmissible and immoral way to pursue the constitutive end of benevolence (i.e. because unless we show the other person respect we would pick inadmissible means to what is otherwise a good end, namely, promoting her well-being). We also cannot do so because we couldn’t otherwise realistically determine what the other’s well-being consisted in. (This does not mean, of course, that we always have to ask the other explicitly what her ends are: as I pointed out in 2.1, we can often make some natural, though defeasible, assumptions about them.) You might respond to this argument that its key assumption, that an agent generally knows best what is good for him, is simply unwarranted. On the one hand, often close acquaintances know what is good for the agent better than he does himself. On the other hand, even if I don’t know you particularly well, I may already have had certain experiences you haven’t had yet, and therefore know better than you whether pursuing some end is worthwhile. I think neither of these two points should be denied: we are not infallible about what is good for us (as already pointed out in 2.2.2), and we would be foolish not to profit from the advice of others on how good certain ends are for us. There are even certain kinds of situations in which our judgement is typically bad (for example, when someone is particularly stressed out by his work, he tends to be bad in judging whether he needs a rest). But there are two reasons why these points merely limit, but do not undermine, the agent’s authority on the question of what is good for him. The first reason is epistemic. Of course, the agent doesn’t always know what some experience will be like, and on this point, another person may well be more knowledgeable. But given that there is no intersubjective convergence in how we subjectively experience things, this will not happen too often. Furthermore, knowing how something will be like for you isn’t enough to determine whether something will make you happy. What is usually the key issue is a comparative cost-benefit analysis, along the lines of: how good for you is getting A, compared to not getting B? Not all agents make these comparative evaluations in the same way—and on the question of how he makes them, the agent himself is normally the best judge.6 Furthermore, there is a practical reason why we cannot generally deny an agent’s special competence in determining what is good for her. For imagine that I thought that I could generally determine what was good for you better than you could yourself. Given that you are especially well placed to answer this question—already the much greater amount of relevant experience you have had on this point ensures that—and that you have a much greater interest in answering the question correctly than I do, I would thereby imply that you have a particularly bad judgement, or conversely, that I have an extremely good one. In either case, I would deny that you are to be taken seriously as my epistemic peer. And that would be a sign of significant disrespect. This disrespect can only be avoided if, in general, I accept you as an authority on what is good for you. This does not exclude that you can be wrong sometimes, and as I argued in 2.2.2, I may then have a reverence-based reason to try to convince you of your error. But I must not assume that you are generally wrong or a worse judge on the matter than I am.7 Kantians can therefore argue that benevolence must normally be informed by reverence for the other as an end-setter, not just because pursuit of the other’s good must be ‘hedged’ by side-constraints provided by respect, but because the most reliable way to determine the other’s good is (normally) by looking at the ends she sets herself. IV If this is correct, then what is demanded by benevolence and what is demanded by reverence can only come apart in two kinds of situations: when I think the other is (exceptionally) mistaken in what is good for him, or when I think the other’s assessments are ‘incomplete’. In the first kind of case, the agent judges that A is bad for him (for instance, sleeping longer in the morning), while I judge that it is good for him and ensure that he gets A anyway (by switching off his alarm clock). In the second kind of case, I judge that A is good for the agent (for example, taking an evening off watching a movie), while he has no opinion on the matter (because he has been too stressed by work to think about the issue), and, on my own, I ensure, or contribute to his getting A (buy him a cinema ticket). For Kantians, the moral evaluation of both scenarios will clearly be very different. In the first scenario, for the reasons presented in 2.2.2, I may try to convince the agent of the falsity of his view. But if he sticks to his belief, I cannot take any further steps, if his aim is morally licit. I am not allowed to go against his will, because this would be disrespectful. (I take it that Arpaly would agree with this result.) In the second scenario, since I am not acting against the agent’s will, the Kantian can allow that my action is sometimes morally licit. But can a Kantian also explain why my action can be morally good or virtuous? This will depend on the specific circumstances of the case, and we can distinguish three different sub-scenarios here: The agent has a (standing) aim to get B (read the morning newspapers) and it is clear that, even according to his own standards, getting A (getting them from the shop) would be the right way to do that; he has just failed to figure that out (e.g. he falsely thought the newspaper was being delivered to his house), or forgot to get A himself. Then a Kantian can easily explain why my action (buying the newspaper for him when I am at the shop) can be virtuous, since I am promoting the other’s aim (i.e. getting B), though the Kantian will insist that I proceed non-intrusively and give the other the opportunity to avoid my beneficence. The agent has a (standing) aim to get B (get relaxation), but it is not clear that, according to his own standards, getting A would be the right way for that. For instance, aim B may be capable of being specified in different ways, and getting A need not be the specification that the agent himself would adopt (maybe he would rather go for relaxing-by-going-running than for relaxing-by-going-to-the-cinema). Then the way to specify aim B must be left to the agent; at best, I can make suggestions for how this could be done. When I do—as in the movie ticket case, where my buying a ticket is only a suggestion, since it remains up to the other person whether he actually goes or not—then, again, a Kantian can explain why such an action can be morally valuable. For then I help the other person to pursue his aim B by helping him to take a necessary preliminary step, namely, deciding in which specific way to realize B. But again, the Kantian will warn that this must not be done in a way which is intrusive, or forces the other to take a decision when he doesn’t want to. If, by contrast, in cases like (b) the help is provided in such a way that the agent does not need to take any decision, but will just enjoy your help or profit from it without a decision on his part, it seems to me that the Kantian resources for explaining how your action has moral value have finally run out. For your action is linked neither directly nor indirectly to reverence for the other person's rational agency. Just remember the Mozart opera cd case from ii, where your housemate could merely passively listen to the music, without making it his end to relax-by-listening-to-opera. And there are many cases of everyday kindness like this, where you make another person more comfortable while he just passively enjoys this comfort—for example, when you switch on the heating in order to make me more comfortable while I am writing an exam which I am too exclusively focused on to think about the cold. Since I would agree that in cases of type (c), helping or providing comfort can be virtuous, I agree with Arpaly that a gap remains for the Kantian account. But, note, the cases which remain unaccounted for will hardly be cases where something particularly important for the other person is at stake or where a refusal to help would not be ‘decent’. For remember, the Kantian can help himself to the natural assumption that things like avoiding suffering are among everyone’s standing general aims. And with respect to most specific pleasures and potential trade-offs between desires of the agent, the Kantian can argue that we must look at the agent’s own ends in order to figure out what he would like. The remaining cases will thus be ones where helping is just ‘nice’ or ‘accommodating’ and do not impinge on the agent’s other ends or desires. I take it that there is virtue of being nice in that way—we might call it the virtue of ‘amiability’. In so far as they cannot explain the value of amiability, Kantians miss something. But what they miss has only a fairly limited importance as far as general morality is concerned, and that is what Kantians want to explain in the first place. Amiability is much more important in close personal relationships, but Kantians can allow that the setting there changes, because by voluntarily entering or staying in such relationships persons create new normative expectations, for example, to the effect that they behave towards one another ‘as friends’ or ‘as spouses’. Within general morality, though, which regards the duties we have towards anyone, distant or close to us, amiability only comes into play in a very limited range of circumstances, and is obviously of much less importance than fulfilling our reverence-based duties is. What Kantians fail to explain here is certainly not a virtue which would be on a par with reverence for rational agency. V From what I have said so far, it may appear that benevolence on its own has no real role to play within a Kantian ethics. But we have only arrived at this result because (following Arpaly’s discussion) we have framed our original question in a particular way, namely, as the question of whether benevolence can expand the range of things we ought to do beyond those we ought to do out of reverence for rational agency. In this respect, benevolence does not make a major contribution. But I think Kantians should accord to care for another’s well-being a different role, the role of helping us specify our general duties to help which would otherwise be too abstract to guide us in our actions. As Kantians like Marcia Baron have pointed out, our duties to help, as imperfect duties, cannot lead us to action ‘on their own’, since no particular action of helping another is strictly required (emergency cases aside). So there is a considerable gap between the abstract duty to help and its application to particular instances. As a consequence, Baron has argued, we need ‘our sympathetic feelings … to prompt us to perform specific acts of helping others. They help us direct our interest and attention to the needs of particular others and to ways we might help’ (1995, p. 220). After what I said in i, it will not come as a great surprise when I propose ascribing this role to benevolence, since I take our disposition for sympathetic engagement with others to be very similar to what we ordinarily describe as benevolence. If a disposition to care for the well-being of others is needed, in the way described by Baron, in order to fulfil our duties to help them, then this disposition plays a role which is not reducible to the role of reverence for rational agency. For the latter only gives us a general duty to help others (emergency cases aside), and without caring about others’ well-being, we would not be responsive to what (and when) help is needed in particular cases. If Baron is right, then Kantians should embrace benevolence as a morally crucial disposition in its own right—though they may not regard it, strictly speaking, as a virtue, given its subordinate role in fulfilling our reverence-based duties. (This might even give the Kantian some handle on explaining ‘amiability’: even though they cannot explain why amiability is a virtue, being amiable may turn out to be a side effect of having a morally valuable disposition. For sympathetic engagement will not be restricted to those cases which fall under the moral duty to help.) In adopting such an account, though, we will have to rethink the respective contributions reverence and benevolence make to moral agency, and the way they interact. Benevolence will not primarily play the role of enlarging the range of morally prescribed or recommended actions, since when benevolence contributes to our fulfilling our duties to help, what we do also counts as fulfilling our reverence-based duty to help. Nor, for the reasons discussed in iii, will the relation of benevolence to reverence be the relation of a virtue of having a good end to a virtue which constrains pursuit of this end to the choice of legitimate means. Instead, reverence and benevolence will be necessarily interconnected: just as reverence must inform benevolence, if we are to really benefit the other person, benevolence is necessary to make our general duty to help more specific and apply it to particular situations.8 Footnotes 1 Kant’s Metaphysics of Morals (1797) is cited as MM followed by volume and page number of the Akademie edition. 2 See Wehofsits (2016, pp. 132–4). 3 PaceCamus (1942). 4 Korsgaard may be more optimistic here than I am. See her ‘Kantian’ account of our duties towards (non-human) animals (Korsgaard 2012), which would imply that humans must care about their ‘animal’ well-being and ‘natural’ good, and pari passu about the ‘animal’ well-being of others. 5 This understanding of ‘authoritative’ takes its inspiration from Joseph Raz’s ‘normal justification thesis’ for authority (1986, p. 53), though I do not want to claim here that we should consider the other’s choices also to be preemptive in the way Raz takes authoritative demands to be. 6 This holds even for the case of pain: the boxer who decides to suffer in order to learn to take punishment during a fight decides that it is ‘better for him’ to suffer the pain than not to learn how to fight a full round. 7 One might object that even if we treat the other as an authority on what is good for her, there are ways to determine what is good for her other than looking at her ends, for instance, looking at her explicit utterances on what she will enjoy. But the agent’s explicit utterances will be of very limited value unless we suppose that they tally with her actual ends (at least low-ranking ones). Otherwise the revealed preferences in her ends will diverge from her explicitly stated ones, and then the former will carry more evidential weight. 8 For very helpful discussions on the topic and valuable comments on earlier drafts of the paper, I am indebted to Stefan Brandt, John Hyman, Christian Kietzmann, Guy Longworth, Angela Matthies and Anna Wehofsits. References Arpaly Nomy 2018 : ‘On Benevolence’. Proceedings of the Aristotelian Society Supplementary Volume 92 , pp. 207–23. Baron Marcia 1995 : Kantian Ethics Almost without Apology . Ithaca, NY and London : Cornell University Press . Camus Albert 1942 : Le Mythe de Sisyphe . Paris : Gallimard . Engstrom Stephen 2009 : The Form of Practical Knowledge: A Study of the Categorical Imperative . Cambridge, MA : Harvard University Press . Google Scholar CrossRef Search ADS Herman Barbara 1993 : The Practice of Moral Judgment . Cambridge, MA : Harvard University Press . Kant Immanuel 1797 : The Metaphysics of Morals . Translated and edited by Gregor Mary . Cambridge : Cambridge University Press, 1996 . Korsgaard Christine M. 2012 : ‘A Kantian Case for Animal Rights’. In Michel Margot , Kühne Daniela , Hänni Julia (eds.), Animal Law: Developments and Perspectives in the 21st Century , pp. 3 – 27 . Zurich : Dike . Raz Joseph 1986 : The Morality of Freedom . Oxford : Oxford University Press . Velleman David 2000 : The Possibility of Practical Reason . Oxford : Oxford University Press . Wehofsits Anna 2016 : Anthropologie und Moral: Affekte, Leidenschaften und Mitgefühl in Kants Ethik . Berlin : De Gruyter . Google Scholar CrossRef Search ADS © 2018 The Aristotelian Society This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)
Aristotelian Society Supplementary Volume – Oxford University Press
Published: Jun 4, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.
Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.
All the latest content is available, no embargo periods.
“Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”Daniel C.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud
“I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”@deepthiw
“My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”@JoseServera