Morality makes many demands on us. But few of us believe that we must do, morally speaking, the best that we can. There is a category of acts, the supererogatory, that involve going beyond our obligations. It seems perfectly plausible, for example, that the moral perfectionist would donate her fifty dollars to UNICEF though the rest of us are within our moral prerogative when we buy ourselves shoes instead.Footnote 1 Although donating to UNICEF is considered especially good, it is usually accepted that morality permits us to buy the shoes; buying shoes is not considered a moral failing. When we consider our reactions to examples like this one, the idea that morality does not demand that we do our best is quite intuitive. On its face, however, it is rather curious. If in a given situation it would be morally better to give to UNICEF than to buy shoes, then how is it not a moral failing when one buys the shoes?
I will argue that supererogation poses a serious problem for theories of moral reasoning and that this problem results, at least in part, from our taking too narrow a view of the reasons that can influence an act's deontic status. We tend to focus primarily on those reasons that count directly for and against an act's performance. To adequately account for supererogation, we need to consider also a different class of moral reasons. Aside from those reasons that contribute, for instance, to an act's being morally better or worse than its alternatives, there are moral reasons that govern the ways in which we respond to the performance or non-performance of those acts. These reasons govern the attitudes we express in response to moral acts as well as our practices of issuing demands and of seeking justification in the case of an omission. These sorts of moral reasons are, I will argue, integral to an account of supererogation. Attending to them will allow us to accept that the supererogatory omission does actually involve a moral failure of sorts while denying that it is a failure of obligation. And this will allow us to account for our intuitions regarding supererogation while avoiding what I call the problem of supererogation. I begin with a discussion of what we might mean by ‘supererogation’, and I outline the problem that it raises. I then provide an account of supererogation that avoids this problem.
1. Supererogation: Our Explanandum
In seeking an account of supererogation, it will be helpful to have at the start a good sense of what philosophers typically mean by the term. I offer the following as a kind of target explanandum. If the performance of act A is supererogatory for agent x, then (at least) two conditions are met. The first I call the ‘Preferred Act Condition’:
PAC) x's performing act A would be, morally speaking, better than x's performing some other act(s) B, and performing A (for the right reasons) would, ceteris paribus, make x a better moral agent than would performing B.
The second I call the ‘Permissibility Condition’:
PC) x is morally permitted to perform B instead of A; B is not forbidden and involves no moral failing.
This second condition implies that the supererogatory act is non-obligatory since another act is permitted. The first condition implies that in performing a supererogatory act one is not simply performing a permissible non-duty but rather, to use a common metaphor, one is going above and beyond one's duty. The supererogatory act is not morally obligatory, but it is morally preferable to available alternatives.
Some philosophers have described the supererogatory in ways that do not seem to include PAC. Roderick Chisholm, for instance, describes the supererogatory as “non-obligatory well-doing”. He argues that an act is supererogatory if it is good to do and if the agent is permitted to do it and permitted to not do it.Footnote 2 Others have described the supererogatory similarly as good to do but not wrong not to doFootnote 3 or as a meritorious non-duty.Footnote 4 These descriptions rightly state that the supererogatory is both good and non-obligatory. They do not, however, capture the sense in which the supererogatory is, as it is often put, beyond the call of duty, the sense in which the supererogatory actor is doing something not simply good but better than her merely duty-abiding peers.
We cannot simply say that the supererogatory act is good but not required since for many obligations there will be a variety of particular acts that would each satisfy the obligation. While one is, in such a case, obligated to perform at least one of the various acts that would fulfill the obligation, no particular act (any one of them good) will be required. Of course, rather than saying that the supererogatory act is not required one might just say that it fulfills no obligation. But this would go too far in the other direction, excluding acts that clearly are supererogatory. Some acts of supererogation involve an oversubscription to duty. One may, for instance, have an obligation to give to charity but give even more (or give something more precious) than her obligation demands. This act fulfils her obligation, but it is also supererogatory.
These considerations support including PAC as a necessary condition for an act's being supererogatory. Supererogatory acts are not just good. They are better than some permissible alternative. Along with PC, PAC accurately captures both the non-obligatory nature of the supererogatory and the sense in which the supererogatory is beyond duty – the supererogatory act is better than the minimally permissible act in any given option set.
2. The Problem of Supererogation
According to our working definition, an act is supererogatory if it meets the Preferred Act Condition (PAC) and the Permissibility Condition (PC). The problem of supererogation is that when we consider the reasons an agent has for performing or not performing the supererogatory act these two conditions are put in tension.
PAC states that the supererogatory act must be morally better than some permissible alternative. As such, it implies then that we have more moral reason to perform the supererogatory act than to omit it. But this implication threatens to upset PC. The problem is that it seems we ought morally to do what we have the most moral reason to do. One standard account of moral reasoning would have it that such an ought is tantamount to a moral requirement. But even if we do not believe that we are required to act on the balance of moral reasons, supererogation remains problematic since not doing what one morally ought to do seems nonetheless to involve a morally significant failing. When considering two acts, there is a failing involved if we choose the act that we have less reason to do, and it is a moral failing if we choose the act we have less moral reason to do, perhaps especially in cases in which one could perform the better act at a relatively low cost. Indeed in many cases – including probably the UNICEF example we began with, the supererogatory act will be supported by quite weighty moral reasons, and the merely permissible alternatives will be supported by no moral reasons at all or by fairly trivial moral reasons.Footnote 5 But this raises the serious question of how these alternatives can be wholly permissible. How can one avoid a moral failing while performing the act that she has little or no moral reason to perform instead of the act that she has great moral reason to perform? For an act to be supererogatory, it seems one must be able to do just that.
A number of philosophers have tried to devise accounts of supererogation that address this question. Some have appealed to the role of non-moral reasons in determining an act's deontic status.Footnote 6 Others have appealed to agent-relative moral reasonsFootnote 7 or to distinctions between deontic and non-deontic moral reasons.Footnote 8 I argue elsewhere that, for all the differences between these writers, a common root of their shortcomings is their focus on the reasons for and against performing either the supererogatory act or its alternatives.Footnote 9 In this paper, I will focus on defending a different kind of account.
3. Oughts and Obligations: Solving the Problem
To properly account for supererogation, we need to attend to a distinction between ‘ought’ and ‘obligation’. The term ‘obligation’ is sometimes used as a synonym for ‘ought’ such that one is obligated to do whatever one ought to do, and it is sometimes used to identify a subset of duties that result from our voluntary agreements and distinct social roles – like mother, brother, employee, etc. – as opposed to duties that might result, for example, from a general requirement of beneficence or rescue. Clearly, the first sense of ‘obligation’ will not be helpful. Neither will the second. This is because many moral acts that do not satisfy anything like a distinct and assignable duty are not supererogatory either. It is hardly supererogatory, for example, to call an ambulance if you notice someone dying on the street even if the person suffering is a complete stranger. However, there is a third sense of ‘obligation’, and so a different type of distinction between ‘ought’ and ‘obligation’, which will be helpful.
This distinction involves the social nature of obligations, not in the sense of obligations resulting from distinct social roles but in the sense that obligations are the kinds of things that we can properly be held accountable for. The conceptual tie between moral obligations and the practice of holding persons accountable has been recognized by a number of philosophers. Stephen Darwall, for instance, argues for this connection in his recent work on morality and the second-person standpoint.Footnote 10 Darwall mentions a number of other philosophers that have noted this connection as well, notably P.F. Strawson and Bernard Williams but perhaps most notably John Stuart Mill.Footnote 11
Mill famously claims in chapter 5 of Utilitarianism that we ‘do not call anything wrong unless we mean to imply that a person ought to be punished in some way or other for doing it; if not by law, by the opinion of his fellow creatures; if not by opinion, by the reproaches of his own conscience’.Footnote 12 In considering the ways in which we are held accountable for action, there is a tendency to focus on blame. But Mill, following Bentham, brings to light a much broader range of sanctions. Bentham focuses on external sanctions, or those that involve the acts and opinions of others. These include political sanctions like legal penalties and moral, or popular, sanctions like insults and blame.Footnote 13 To these external sanctions, Mill adds the internal sanction of the agent's own conscience.Footnote 14 For his part, Darwall follows Strawson in focusing largely (though not exclusively) on what they term ‘reactive attitudes’, or the attitudes we express in responding to each other as morally responsible agents.Footnote 15 These attitudes include, of course, many of what Mill and Bentham referred to as moral sanctions.
There are a number of reasons one might highlight the relationship between obligations and accountability. But for our purposes this connection is important because it allows us to distinguish those acts we ought to perform from the narrower range of acts that count as obligations. As such, it allows us to make room for supererogation. By attending to the connection between obligation and accountability, we can make space for a range of behavior that we have good moral reason to perform but that is not obligatory. While we ought to do what we have the most reason to do, we may not always be obligated to do so. There are reasons for the community to allow a sort of latitude in certain kinds of cases.
If we are going to make use of a notion of accountability in order to set supererogation apart from obligation, some distinctions are in order. Darwall identifies a distinction between holding someone responsible and finding someone responsible.Footnote 16 Finding someone responsible might involve, in large part, believing that she is causally responsible for the act – believing, that is, that she did it. Finding someone morally responsible will likely require that further conditions are met. It probably involves at least believing that her act was morally significant, intentional, and uncoerced. Holding someone morally responsible involves not a mere finding of moral responsibility, but rather it involves relating to the agent as one who is morally responsible for the act.
In addition to this distinction between holding responsible and finding responsible, there is a distinction between holding persons responsible and the more specific practice of holding persons accountable. To be morally accountable for an act involves being subject to legitimate moral demands, authoritative claims, and punitive reactions in case of failure. Being held accountable involves, at the very least, being held to account; that is, when one is accountable for an act, one owes an account in the event of an omission. There are, then, many ways in which we can hold each other responsible that are simply not constitutive of holding each other accountable. We can reward behavior, or we can express attitudes like gratitude and admiration, for example. Thus we should distinguish between the punitive sanctions that are ways of holding someone accountable and these other reactions to moral behavior that are not. This puts us on the way to an account of supererogation. Part of what is involved in an act's being supererogatory is that its omission should not be met with punitive sanctions.Footnote 17
In addition to ruling out punitive sanctions in response to the omission of a supererogatory act, we should also note that the moral community ought not normatively expect the performance of supererogatory acts. The term ‘expect’ is ambiguous between a normative and a predictive sense. Someone who says he expects his children to behave in church means something very different by ‘expects’ than someone who says he expects it to rain tomorrow. We can often appropriately expect supererogatory acts in the predictive sense but not in the normative sense. Normative expectations are tied up with holdings accountable. To normatively expect behavior involves believing both that one ought to behave that way and that one is accountable for doing so.
The inappropriateness of both a normative expectation of supererogatory behavior and punitive sanctions in response to the omission of a supererogatory act are important aspects of what distinguishes supererogation from obligation. But this is not the whole story. These may also be inappropriate in the case of an easily excusable failure of obligation. So we will need to look further to fill in our account.
On Uttering ‘Ought’
As J.L. Austin famously noted, we do much more with our speech than utter words and express propositions.Footnote 18 In addition, we perform what Austin terms ‘illocutionary acts’. We assert propositions, but we do a number of other things as well. We may, for instance, express attitudes, issue directives, or make promises.Footnote 19 And these acts can have both illocutionary force and perlocutionary effects. For example, the illocutionary force of a promise involves the speaker's taking up a commitment to do as promised (and being correctly understood by others as having taken up that commitment), and the perlocutionary effect of a promise involves the speaker actually performing the act promised.
Speech acts, like other acts, are governed by norms. Indeed there are a variety of norms involved. Some, for instance, involve the linguistic competency of the listener; the speech act will not have its intended illocutionary (or perlocutionary for that matter) force if directed at one who does not understand the language in which it is delivered. But for our purpose, it will be useful to highlight in particular those norms that are involved in establishing an entitlement, on the speaker's part, to perform the speech act in question.
The entitlement to assert a true belief is, in principle, open to everyone. But contrast this with the case of a demand. When I tell my daughter to put away her toys, the legitimacy of the demand depends on my having a certain normative authority. Of course, this authority is not held uniquely by me - her mother could legitimately make the same demand, but nor is it open to everyone – her brother, for instance, cannot.Footnote 20 The entitlement to assert does involve a normative status of sorts (epistemic justification), but it does not involve, as does the entitlement to demand, the possession of normative authority. The entitlement to assert a normative belief (‘Jane ought to help her sister with the move’.) is, then, quite distinct from the entitlement to issue a related demand (‘Jane, help your sister move!’).Footnote 21 To demand that one do as she ought is a way of holding her accountable for that act, and we are not entitled to hold each other accountable except where we possess the authority to apply the relevant norm.Footnote 22
Who exactly does possess that authority will depend on the norm in question and on the reasons one might have to enforce it. Some obligations are such that adherence to them can be ordered only by those occupying a narrow authoritative role – e.g., drill sergeant, boss, mother, etc. But other obligations allow a more universal type of enforcement. The authority to enforce these obligations is held broadly by the members of a moral community. If you come upon a mother abusing her child, you need not occupy any special authoritative relationship over the mother in order to tell her to stop. Not beating one's children is an obligation such that the authority to enforce it is broad. One of the things that members of a moral community do is enforce obligations, either by ordering adherence to those obligations or by enacting some type of punitive sanction when the obligatory act is omitted (or both if the order is not obeyed).
In some cases a legitimate demand will actually generate a new reason for the hearer to perform the act commanded – a reason to follow the command – where there was no reason prior. When a sergeant issues a reasonable order, like an order that a private tuck in his shirt, the sergeant makes it the case that the private is obliged to do so. The sergeant's authority to issue the demand derives from his role in a hierarchy and not from any preexisting obligation that the private has to tuck in his shirt. In other cases, the authority to issue a demand is premised on a preexisting norm. When we order someone to stop abusing her children, the order is legitimate because she is subject to an existing norm that tells against such behavior and because it is the kind of norm to which we can properly hold her accountable. Here, her reasons to act as demanded are not primarily reasons to follow the command. She should refrain from the abusing primarily because of those reasons that underlie the preexisting norm – because of the damage it would do to the children – and not simply because we told her to. Our issuing the demand serves to make occurrent for her those preexisting reasons and to highlight that this is a case in which she can be held accountable for doing as she ought. The authority to issue the demand rests on the reasons the community has to enforce the norm in question, to expect compliance with it. And the demand itself underscores that reasonable expectation and any related threat of punitive sanctions. The threat of sanctions and the reasonable expectations in her community may provide reasons (in addition to those reasons that undergird the existing norm itself) for the agent to act on her obligation.Footnote 23
One of the things that marks off an obligatory ought from a supererogatory ought is that where an act is obligatory there is at least someone with the normative authority to demand that the agent perform it. When an act is supererogatory, there is no such authority. Just as we should not hold agents accountable for supererogatory behavior by issuing punitive sanctions in response to omissions, we should also not hold agents accountable for supererogatory action by demanding its performance. And notice that it is not just that a demand for supererogatory behavior is inappropriate, but it is inappropriate in a specific way. It is illegitimate because the speaker lacks the entitlement that is proper to that kind of speech act; the speaker lacks the authority to issue a demand. Here, another distinction is in order.
Even if a speech act is legitimate by virtue of the speaker holding the proper entitlement, it may be defective or inappropriate in various ways. For example, a fully justified assertion can be inappropriate if rude or trite. A statement like ‘Wow, you've really gained a lot of weight’ will usually be quite inappropriate even if the evidence suggests the listener has, indeed, gained a lot of weight. Despite the potential cruelty of such a remark, the assertion may still have its illocutionary force. It can generate (or reinforce) justified beliefs on the part of the listener and so probably license various inferences.Footnote 24 On the other hand, where an assertion is not warranted it will not have this effect. Coleen Macnamara makes this point with respect to demands in particular by distinguishing between legitimacy and permissibility.Footnote 25 Even where a speaker is entitled to make a demand, it may be impermissible that she do so.
Take the case of demanding that a mother not beat her children. Where there is some good reason to believe that the mother will do so, such a demand may be very appropriate. But it is deeply inappropriate to demand of a particular mother that she not beat her children if we have no reason at all to believe that she would. This is not, I would argue, because we lack the normative authority to make this demand – not beating one's children is a norm that we have the authority to enforce. The trouble in this case is that the norm does not need enforcing. And moreover, to act as though it does need enforcing suggests the (probably) false and offensive belief that this mother would hurt her child. But the demand may, nonetheless, have the illocutionary force of a demand. It will serve to highlight those reasons that make it such that one ought not hurt her children, and it will highlight also those reasons that are tied to the proper normative expectations of the moral community; it will highlight the fact that one is accountable for following the norm in question. Again, the problem here is that these reasons are not in need of highlighting.
No Excuses
Just as legitimately demanding that one perform an act requires that the speaker have the proper authority to issue that demand, demanding that one provide an excuse for an act's omission also requires that the speaker have the proper authority. If a speaker attempts to demand that a listener provide some excuse for not performing a supererogatory act, then the demand will be illegitimate because the speaker is not entitled to make it. An excuse is not merely an attempt to explain our reasons; it is an attempt to minimize our accountability in the case of failure. But an act that is supererogatory is an act that the moral community cannot reasonably expect or demand.Footnote 26 It is an act for which one is not accountable. And if one is not accountable for an action, one should not be held to account for its omission.
That does not mean that we do not sometimes attempt to at least explain our reasons for failing to perform acts that we deem supererogatory.Footnote 27 In fact, we do this all the time. When faced with a request to perform a supererogatory act, we do not usually dismiss it with a simple ‘No, I don't want to’.Footnote 28 Instead we tend to offer what look a lot like excuses – ‘I gave at the office’. But these ‘excuses’ are notoriously insincere. They may at times be non-defective as assertions regarding our motivating reasons for not performing the act in question. But they are defective as excuses precisely because we do not believe that the failure is one we need to justify.Footnote 29
The Special Worth of Supererogatory Acts
David Heyd, who provides one of the few sustained treatments of supererogation, writes that ‘the value of a supererogatory act consists not just in the increase of the net amount of goodness … in the world, but also in their being totally optional and voluntary’.Footnote 30 Heyd claims that if we allow that one ought to perform the supererogatory act – a claim Heyd says amounts to ‘qualified’ supererogation, then we will fail to account for this value. Heyd argues that ‘acts of supererogation are characterized as purely voluntary, optional, and in a sense arbitrary, that is, not determined by universal standards or rules’.Footnote 31 Heyd is right that the voluntariness of the supererogatory contributes to its value. Indeed, the fact that an act's value may depend, at least in part, on its being performed voluntarily may be one important reason for the moral community to treat those acts as non-obligatory. But Heyd also suggests that the voluntariness of the supererogatory is derived from the fact that the supererogatory act is not one that the agent ought to perform, that its omission involves no moral failure. I fail to see why an act's being one we ought to do prevents its performance from being optional and voluntary. Doing less than we ought is, after all, very often an option. And on my account, the performance of the supererogatory act cannot properly be coerced, not even by the subtle pressures of a community's demands for excuses.
One important feature of a supererogatory act is that punitive sanctions do not appropriately attach to its omission. Sanctions such as punishment, blame, and disapprobation are inappropriate. Nor ought one feel guilt, shame, and the like.Footnote 32 The supererogatory act has special worth because the positive moral motives of the agent are not colored by any threat of penalty. The agent will not be performing the act (even in part) because he wants to avoid discipline or because he wants to avoid having to justify the act's omission. One might, of course, be motivated to perform one's obligations by the moral reasons in their favor and not by a threat of punitive sanctions. And supererogatory acts may often be motivated, in part, by considerations other than the moral reasons that count in their favor. One might for instance be trying to impress someone or to acquire accolades. But it is nonetheless important that he will be acting in the absence of any threat or coercion, and so the act will be optional and voluntary. He will have done some good when, in an important sense, he did not have to.
Moreover, the supererogatory has a distinct normative status. Consider an agent faced with performing a supererogatory act and an agent faced with fulfilling an obligation. Both ought to perform the act. But the agent that is under an obligation is thus in an importantly different normative position. She will have reasons to act generated by the reasonable demands of her moral community and by the prospect of deserving punitive sanctions, including those of her own conscience. These reasons are not in play where an act is not obligatory. The prospect of falling out of harmony with one's community, of failing to live up to reasonable expectations creates a different kind of reason to perform the obligatory act. That someone does, in fact, properly demand the act of me may in some cases give me an additional reason to perform the act, a reason to respond to that demand. But even when no demand is actually expressed, I can be motivated to perform the act in part because I do not want to fall short of what others may reasonably expect and demand of me.
4. Revising the Permissibility Condition
At the beginning of this paper, I outlined a challenge that accounts of supererogation must meet. They need to provide an adequate answer to the question of how the supererogatory omission can be permissible despite the supererogatory act's being (often strongly) recommended by moral reasons. As I noted, a number of philosophers have attempted to meet this challenge by attending to possible distinctions among the kinds of reasons that might favor or oppose a supererogatory act. In this paper, I have attempted to articulate and defend a rather different course. On my view, the permission that makes the act supererogatory rather than obligatory is generated not by the moral reasons that favor or oppose the act itself but rather by those moral reasons that govern our practices of holding each other accountable. While one is permitted by the members of the moral community to omit the supererogatory act (she is not accountable for its performance), the omission may still be seen as a kind of moral failure from the perspective of those reasons that directly recommend the act.
This will require us to revise the condition PC, which as presented in section 1, states that the supererogatory omission involves no moral failure. We should instead say that the supererogatory omission does not involve a failure of obligation. This revision will still allow us to capture the intuitions that underlie our concept of supererogation. But it does represent a significant departure from a standard view of supererogation.
On my view, there are moral reasons to adopt a permissive attitude with respect to the omission of the supererogatory, but morality itself does not provide directly, by way of those reasons that count for or against the supererogatory act, a permission to omit the act. As such, some philosophers will claim that the kind of permission my view appeals to is not of the right sort. My view states that the supererogatory act cannot be normatively expected or demanded and that its omission should not be sanctioned and does not require justification. Thus, the omission is permitted by the moral community and not forbidden. But while the omission of the supererogatory does not involve a failure of obligation, I see no way around the conclusion that it can involve a significant kind of moral failure nonetheless. When we omit the supererogatory act, we do something morally worse than what we really ought to have done.
Jonathan Dancy has described views on which one ought to do what is supererogatory as, at best, weak supererogation; ‘weak’ here is meant to be disparaging.Footnote 33 But we must keep in mind that our intuitions about supererogation are not really intuitions about the meaning of the concept, but rather they are intuitions about particular cases. In our common moral discourse there is not even a word for this concept let alone much precise thought about its meaning.Footnote 34 ‘Supererogation’ is a philosophers’ (and theologians’) term, and the standard account of supererogation is a philosophers’ construction. My revision of that construction is not a retreat from common sense but rather a different, and better, way of capturing it.
The term ‘supererogation’ is used to account for the strong, common intuition that in certain cases it is acceptable not to act on a moral reason in a way that it is not acceptable in other cases. In some cases it looks like we ought to be free to choose either to act or not act on the balance of moral reasons, and our choice in these cases should not be condemned even if we take the less admirable course. We are driven to discuss the supererogatory as a moral category in large part by the intuition that there are cases in which one needs no excuse to not act on the balance of moral reasons. But these intuitions are well accounted for on the view I have described.
In fact, there is strong evidence in our moral practice of a distinction between obligatory and non-obligatory oughts. There is nothing odd, let alone incoherent, about saying, ‘Of course, you probably ought to give even more, but I don't see that you're obligated to’. And there are good reasons for this distinction as well. As noted above, the voluntariness of supererogatory acts can contribute to their moral value, and this may be one reason against our treating those acts as obligations. More importantly, if every moral ought were properly enforceable by means of demands and punitive moral sanctions, very few of our significant decisions would be off-limits, and our range of truly free choice would be severely restricted. In some cases, a concern for the fair distribution of moral burdens may also come into play. When others do not do their share in promoting some good, it may often be possible for us to pick up the slack by doing more than we would otherwise have reason to do. But even where we have good reason to do so, it may nonetheless be wrong to expect or demand that we do more than our share.
In any case, there is nothing strange about the idea that morality asks more of us than we should demand of each other. Indeed, we commonly think of moralism as a vice not because (or at least not only when) the moralist is confused with respect to what we ought to do, but rather because the moralist responds inappropriately to those oughts by treating too many of them as obligations. A strong and well-tuned prior conscience is considered a moral virtue; a punitive, moralistic response to every failing is not. To understand supererogation, we must be attentive to this feature of our ordinary moral practice. If we recognize that not all conclusive moral reasons carry with them legitimate moral demands, then we can also recognize the limits of duty without dismissing those moral reasons that would take us beyond our obligations.
