1. 

I believe in moral obligation, but not in the way many other moral philosophers evidently do. I believe in keeping my word, telling the truth, playing fair, and so forth. And I acknowledge that mustering the motivation to do these kinds of things can be a challenge sometimes. The truth can be embarrassing, for example, which can make it tempting to dissemble. Yet while my record is far from spotless, I often manage to tell the truth, even when it is difficult. It is important to live honestly with others, I believe, and so I strive to do so, even in the face of what are sometimes strong countervailing motives. In that way, truth-telling is something I regard myself bound to do.

At this point, however, some moral philosophers insist, and many more argue in ways that imply, that I have left out something crucial. There has got to be more to the sense that one is bound for it to count as moral. Of course we regard ourselves as bound in all sorts of other ways—we talk, for instance, about the bonds of love, or loyalty, or nationalism, or friendship. Such bonds run deep, and are often expressed in action. For all but the most high-minded moralists, the sense of devotion to much besides morality may move a person to do what she otherwise might prefer not to do. But regarding ourselves bound in that sort of way cannot be what is meant when we talk of being morally bound. We are not merely loyal to our moral values; not simply in love with our moral ideals. Morality must take hold at a different place within us—somewhere even deeper than that.

Or at any rate, so we are told.

This was evidently Kant’s view. In the Groundwork he tells us that if there is to be anything at all that counts as moral obligation, it must be possible for us to be moved purely through the recognition of the moral law’s unconditional right to govern us.¹ In that way, for Kant, the very existence of anything deserving the honorific rests on establishing that we can rightly regard ourselves as bound by morality in a way that is altogether unique. This conviction about the special nature of moral bondage serves as the starting point for Kant’s inquiry in the Groundwork. His aim is to vindicate our sense of moral bondage so-conceived, through identifying the conditions of our being subject to such a requirement, and then arguing that we are entitled to treat those conditions as met.

This paper is part of a larger campaign, which aims to advance a way of thinking about our sense of obligation and how it might move us that is animated by the rejection of what Kant begins by supposing. While few moral -philosophers nowadays are prepared to follow Kant in everything, many do still seem to -follow him in this. They assume, when it comes to obligation, that the “mark of the moral” must be a matter of some truly unconditional, categorical, or otherwise inescapable rational constraint.² A person’s sense that she is morally obliged, then, is to be construed in terms of her registering that she is subject to just such a constraint, and being moved accordingly.

On the sort of view I prefer, a person’s sense of obligation can instead be thought of in terms of her relation to certain socially established rules, which tell her what she must, may, or cannot do in certain sorts of situations, through defining offices, positions, and roles in terms of the privileges, powers, and responsibilities that attach to them.³ These rules help constitute what are often called social practices.⁴ The key question concerns the motivational basis of a good person’s adherence to our practices, so understood—the motivation to follow, or comply with, our practices’ rules. Unlike Kant, and unlike those who follow him, I do not find it appealing to think that a good person is one who is moved to comply with our practices because she regards herself rightly subject to some special, unconditional claim to authority that certain rules are entitled to make upon her.

Rather, I think we do better to locate the roots of our sense of moral bondage in our complex capacity for sensibility or feeling. In particular, the sort of motive I think we should cultivate and prize reflects what we might characterize as a felt sense of the crucial importance to human social flourishing of certain among our practices, as expressed centrally (though by no means exclusively) in a person’s being moved to follow strictly and strenuously the rules that constitute those practices. Throughout this paper, I will talk about this sort of motivation in terms of an agent’s manifesting reverence or regard for, veneration of, or devotion to, certain among our practices. I intend for these and similar expressions to function as terms of art, meant to evoke a certain image what is—and is not—going on motivationally when a decent person acts from a sense that she is morally obliged.

This way of conceiving of our moral obligations, as born of rule-constituted social practices ultimately adhered to out of a form of feeling that reflects or embodies a sense of their paramount human importance, descends from Hume’s account of the “artificial virtues” in the Treatise.⁵ Of course Hume had his own “mark of the moral,” which for him took the form of a particular type of approbatory sentiment, felt in response to someone’s motives when contemplated in a suitably detached way. In order that any particular practice of ours give rise to moral obligations, for Hume, the sorts of motives it characteristically engendered in practitioners had to elicit this specific form of sentimental response.

I will, insofar as possible, try to set to the side this aspect of the Humean view, about what makes a motive qualify as a moral one. Hume and Kant are, for all their many and important differences, alike in wanting considerable unity out of morality. Most philosophers still follow them in searching for some special type of reason, or principle, or source of normativity, or type of feeling, or way of harmoniously organizing the self, from which all of what is properly called morality might spring. Unlike these philosophers, I tend to think what we call morality is much messier and less unified—more like a loosely associated constellation of concepts, distinctions, norms, practices, feelings, motives, and so forth, connected to one another by relations of family resemblance and history at best.⁶

What there undeniably are, even when morality is viewed that way, are paradigm instances of moral obligation: the obligation to keep a promise, or to tell the truth, or to respect another’s property, for instance. And when it comes to thinking about these paradigmatic examples, I think Hume has real insight. As he sees it, a person with a strong and healthy sense of these obligations is someone who understands what certain practices require or forbid, and what others are entitled to expect or demand of her in accordance with their defining rules. She is moved to act or refrain accordingly, even in the face of what might otherwise be strong incentives to do otherwise. Because of the way she regards herself and her relation to others as mediated by these rules, her sense of herself will be implicated in her deferring decision to the rules in this way. In any individual instance, she will be moved to keep her word, or tell the truth, or respect another’s property, “just because” that is what a faithful, honest, or judicious person does, and she takes great pride in being that sort of person.

Crucially, none of this is a matter of an agent’s judging herself to be under any truly unconditional rational requirement to act in these ways. Rather, the motivation of the mature ethical agent is ultimately explicable as an expression of her capacity for feeling. Of course, this is not to be understood in any simple or reductive way: the idea is not that we simply happen to “like” certain social practices in some peculiar manner, nor that we merely find ourselves moved to adhere to some of them with unusual vehemence. The sort of motivation we prize when, for instance, a woman of her word tells an uncomfortable truth or keeps a difficult promise, is one that must be acquired, cultivated, and refined.

Moreover, the operation of such motives in us is sensitive to our reflection upon the nature, history, and social value, of living together as people characteristically moved in these familiar ways. Our sense of our moral obligations is necessarily affected by, and refined in light of, the deliverances of the rational faculty, even if is not properly regarded as originating in that department of the soul, nor subject to its jurisdiction.

I will not be able, in this paper, to offer a full-dress defense of this more “Humean” approach to thinking about what moral obligation is and how it moves us. My narrower aim, here, is to mount a response to one familiar type of objection. Some suppose that nothing other than the recognition of unconditional authority can serve as an adequate motive to follow the rules of important social practices, because the direction given by the rules will sometimes chafe against an agent’s other values or interests. In some cases, these may even be the very values or interests which, were she to reflect, a person might cite in explaining why she holds a particular practice in high regard. In those instances, the very values that ordinarily support following the rule might seem better served by setting the rule aside. Yet observance of the rules in precisely these instances is often seen as the height of virtue. Unless it can be established that the rule has a genuine “claim” on the agent—a claim she is not allowed to ignore or set aside for any reason—then any impulse to persevere in observance of the rule in these instances will seem poorly grounded from the agent’s own point of view.

A forceful version of this sort of challenge to the Humean approach has been raised by Christine Korsgaard in The Sources of Normativity.⁷ Korsgaard there imagines an ethical agent who explicitly understands her own obligations in a more Humean way, and finds herself facing a difficult dilemma. The rules of a practice for which she feels a profound sense of allegiance require an action of her that would undermine the very sorts of values or concerns that, when she reflects, seem to her to render that practice deserving of her devotion. An agent who views her obligations in the more Humean way, Korsgaard claims, will be unable to resolve this sort of conflict in any satisfying, ethically coherent way. In situations of precisely this kind, someone who signs on for Humeanism will be condemned to “normative failure” (Korsgaard 1996: 88).

At the heart of Korsgaard’s challenge to the Humean approach one finds a familiar charge: one that is sometimes put in terms of “rule-worship.”⁸ On the one hand, the Humean agent sees her own motivation to follow the rules of a given practice as somehow informed or shaped by the value of that practice, taken as a whole. And yet, her motive to follow the rules in each and every instance is also one she understands as neither wanting nor admitting of rational sanction—it is at bottom her felt sense of the urgency of adherence to the practice that moves her to comply. In precisely the sorts of situations Korsgaard imagines, where the values that shape and inform an agent’s own understanding of her motive to meet her obligations is in tension with the direction given by the rules here and now, meeting her obligations is exposed to the agent herself as simply “worshiping” the rules, when it might be more sensible to set the rules aside or violate them.

To the extent that words like regard, veneration, reverence, and devotion are all more or less synonyms for worship, the view I have been sketching may seem like a thinly-veiled attempt at rebranding. That accusation is not entirely off base. Nevertheless, my hope is to persuade you that this sort of objection can be answered. I will not deny that tension—indeed, tension that can sometimes rise to the level of deep and serious conflict—may arise between a person’s veneration for the rules on the one hand, and the very considerations or values she thinks of as making a given practice and its rules worthy of her devotion on the other. My idea is rather that, once the source and nature of this kind of conflict are better understood, its possibility turns out to be no real threat to the Humean view, taken on its own terms. Indeed, far from a strike against that view, the way that it helps us to make sense of this kind of conflict—and of ourselves, insofar as we are subject to it—should be counted as a mark in its favor.

My approach will be circuitous. I will not get around to considering directly this charge of “rule-worship” as it has been levied by Korsgaard against the Humean approach until near the end of the paper. After some more discussion of methodology and background in the next section, I am going to explore, in Sections 3 and 4, two other aspects of the way Hume invites us to see our obligations, focusing on certain crucial points of contrast with the more familiar, Kantian approach. Section 3 concerns what sort of support we can offer on behalf of our obligation-engendering practices, if we take up a more Humean view. Of particular interest is the question of how the support we provide does and does not relate to the motives of individuals acting within those practices. Section 4 concerns deliberation, and the role of obligation within it. As I shall explain, for those who take for granted the more mainstream view of obligation as unconditional constraint, it is natural to construe deliberation as aiming to determine what one is “really” obligated to do. But I will argue that, seen Hume’s way, this is not in general how we should understand the aim of deliberation.

Finally, in the fifth section, I will consider Korsgaard’s challenge. The hope is that, once the Humean position is a bit more fleshed out, I will be able to make my case. Far from a reason to reject the view, the image of ourselves as people whose sense of obligation is the result of our learning to “worship” certain social rules turns out, when seen in the right light, to be a deeply attractive one.

2. 

Chiefly, I mean by this that it is ethically attractive. While some may see inquiry into our sense of obligation and how it moves us as a primarily theoretical or explanatory enterprise, that is not at all how I see it. My discussion of obligation is intended as a contribution to what Rawls labeled “moral theory”, that is, the comparative study of competing or contrasting moral (or ethical) conceptions.⁹ There are at least two important things to stress about this. First, the inquiry should be understood as practical “all the way down,” in that the basis on which claims are advanced or adopted must, in the end, be tethered to our concerns as agents. A good “account” of obligation earns its keep, in other words, by speaking to us as people faced with figuring out how to think, feel, and act where we encounter the concept. And second, the emphasis is on working out some of the differences that would result from actually adopting a given view of obligation as one’s own. Obligation is a concept that we make use of in ethical life, and competing conceptions of obligation can have importantly different implications for a broader understanding of ourselves and what we are doing when we deploy the concept.¹⁰

I take this job-description to be more or less common ground among champions of the more familiar, Kantian conception of obligation as an unconditional constraint, as well as its more (in)famous detractors. Those who advance the familiar view purport to tell us how we are required to think, act, and feel, when we encounter an obligation, or a putative one, embedding their claims within a broader vision of the person and her ethical life as a whole.¹¹ On the other hand, thinkers like Nietzsche, Elizabeth Anscombe, Bernard Williams, Phillipa Foot, and Susan Wolf have all sought to attack that conception of obligation, by showing its ethical deficits. None of these thinkers denies that many people do, as a matter of fact, conceive of morality as making rationally unconditional demands on them. The point is rather to make an ethical case against the prevailing view: the shared aim of these thinkers may therefore be put in terms of discrediting that way of seeing one’s obligations. In that way, I put these philosophers squarely within the tradition of moral theory as Rawls characterized it, though they no doubt represent its insurgent wing.

The arguments of these insurgent moral theorists against mainstream thought about morality and its alleged unique authority to oblige take various forms. Some aim at exposing mainstream moral thinking as rooted in far less savory impulses, like the desire to control, do harm, or seek revenge: Nietzsche’s (1887/1967) ideas about ressentiment, for instance. Others aim to show that way of thinking incoherent or internally confused: Anscombe’s diagnosis in terms of the vestiges of divine command, or Foot’s allegation that it rests on some “fugitive thought” about reasons independent of an agent’s ends (Anscombe 1958; Foot 1972). Others offer a critique from some external perspective: Wolf’s “standpoint of individual perfection,” or the role of aesthetic concerns throughout Nietzsche’s work (Wolf 1982). Still others cast it as issuing from a form of motivated belief. Williams, for instance, insinuates that we may be seduced into accepting morality’s authority, because it offers the reassuring guarantee of an unimpeachably correct procedure for decision, but this can only be magical thinking (Williams 1981).

Some may be inclined to read these thinkers as advocating for doing away with concepts like morality and moral obligation altogether, perhaps in favor of some nostalgic return to an ethics of Athenian virtue that is supposed to have lacked these concepts, or some not-as-yet defined system of ideas that no longer has any use for them. On the whole, I think this is the wrong conclusion to take away from these philosophers, though I acknowledge that much of the rhetoric in their assault on mainstream thinking can encourage that reading.¹²

At any rate, I will not further rehearse these arguments against the more familiar conception of moral obligation as unconditional, categorical, or inescapable constraint. I do find these arguments compelling, on the whole. But my interest here is not to press that point. I mention them primarily in order to locate and motivate my own discussion: I mean to pick up where the negative arguments leave off. For those who, like me, are distrustful of the idea that moral obligation should be understood in those terms, but not ready to do away with the concept altogether, we need to think through other, more attractive ways of seeing our obligations.¹³ What else can they be, if not demands on us that we are not free, rationally, to disobey? How else should we be moved to fulfill them, if not by the recognition of some uniquely unconditional claim to authority they are supposed to have over us? How should we think in, and about, those parts of life where having some understanding of oneself as morally obliged is vital?

It is in grappling with these sorts of questions that working through some distinctively Humean thoughts about obligation is useful. It is not an accident that many philosophers dissatisfied with mainstream moral thinking have turned to Hume, seeking to appropriate or develop key Humean ideas. It is perhaps worth reminding ourselves that Hume himself was quite the insurgent moralist. Hume’s well-known claims about, for example, the motivational impotence of reason, or the emotional character of ethical judgment, may nowadays seem to some to be of primarily “meta-ethical” import. It is true that when it comes to obligation, just like his rationalist opponents did, Hume maintained that people really are bound in a moral way to keep their promises, or to respect the property rights of others. So perhaps his contribution lies primarily in his pointing to alternative psychological, metaphysical, or semantic foundations for these uncontested “first-order normative claims.”

Rawls’s characterization of moral theory invites us to reject this familiar way of carving things up. While perhaps useful for some purposes, it obscures the many significant implications Hume’s moral philosophy will have for us as ethical agents, insofar as we are attracted to his ideas, or persuaded by his arguments.¹⁴ This is especially true when it comes to Hume’s account of obligation, which does not merely provide the disinterested philosopher with a sideways-on look at what happens when someone feels herself obliged. Hume’s way of thinking about obligation is part and parcel of a different way of thinking about oneself: a different way of looking out onto one’s own ethical world, that is at odds with the prevailing ethical rationalism of his day, and also of ours.

3. 

One way to start developing a more Humean view of our obligations is to consider what Hume has to say about how the social practices that give rise to obligations are supported, ethically speaking. Specifically, there is a familiar (set of) question(s) concerning how what we might say in support of our practices does, and does not, relate to motivation and justification at the level of an individual agent.

In Hume, the issue arises in connection with the just-so stories of moral “artifice” in the Treatise, where he gives us his account of the ‘origins’ of our practices of Property and Promises. Hume’s stories begin in an avowedly fictional “State of Nature,” where agents are initially moved only by weak benevolence and limited partiality, but also come equipped with a capacity for cooperation and intelligent problem solving. Such a world, Hume imagines, would be plagued by coordination problems that would frustrate cooperation, leaving everyone worse off. This is quickly remedied, however, by convening on certain social rules. First, the rules that invent property put material possession on a more stable footing, allowing for the pooling of labor, specialization, and so forth. Rules inventing promise then secure the possibility of non-simultaneous exchange, through creating an effective vehicle of assurance.¹⁵

Initially, each agent in Hume’s State of Nature is motivated to comply with these new rules because she sees how doing so, provided others do likewise, makes it possible to cooperate in ways that serve everyone’s mutually held and recognized, yet thoroughly individual interests.¹⁶ In that way, the rules in Hume’s stories first exist in a kind of embryonic stage. But in light of the principles of his practical psychology, the rules are subject to “annexation” by the moral sentiments, so that what begins as a self-interested and instrumental motive to follow the rules is, over time, transformed into a genuinely new form of motivation. The end products of this process are more or less the motives Hume imagines we have: the motives he labels justice (the capacity to be moved by another’s rights to her property) and fidelity (being moved to honor one’s word), which reflect an utterly un-self-interested reverence for the practices themselves and the relations to others their constitutive rules define.

An obvious question invited by Hume’s discussion of artifice concerns what sort of support for our actual social rules, and our own motives to follow them, these made up stories about their origins can provide. How exactly is a fictional historical explanation of our evolution out of a pre-social condition that, by Hume’s own lights, we never actually existed in, supposed to relate to our own motives as trustworthy promisers, or people moved to respect each other’s belongings?

One option is to suppose that the appeal to the sorts of interests that motivate the denizens of Hume’s State of Nature to establish certain rules in the first place serves to justify our treating our actual practices as a source of authority for us. On this reading, someone who asks “Why do I have to keep my promises when I wish to break them, or abstain from another’s property when I can get away with taking it for myself?” is offered a reply, in the form of a display of the advantages at the level of an entire society, when more or less everyone observes the rules.

This reading might seem to have been suggested by Rawls, when he cited Hume first among philosophers who relied implicitly on the distinction between justifying a practice, and justifying an action falling under it (Rawls 1955: fn. 2). While perhaps an appealing strategy to some, the idea that it is proper to use the label “justification” for what happens on both sides of Rawls’s distinction must be delicately parsed: failure to take care here leads to serious distortion of Hume’s position.

To try to clear things up, let us stipulate that the justification of a practice would be the sort of thing meaningfully addressed to an agent: for instance, in response to a question of the form why should I? To justify (in this sense) a practice (to an agent) is to offer (her) a (sufficient) reason to obey its rules in the face of this sort of broadly skeptical challenge to its authority. It then seems that someone attracted to Hume’s thought should insist that our practices are not to be justified to us in that sense at all.

Indeed, what we are in shown is why there is no pressing need for any legitimating defense of our practice’s authority for us, insofar as we are well-adjusted people who have acquired the very motives the discussion is intended to help us make better sense of. The need for such a justification of the rules to the agent may have arisen for the denizens of Hume’s State of Nature. And it might still press for a certain sort of person who has, for whatever reason, failed to acquire the appropriate or expected attitudes concerning the practices in question. These abnormal individuals might still need to be shown they “have a reason” to follow the rules: for instance, one that stems from self-interest (perhaps this would take the form of reminding them of the penalties that attend their violation).

The point is not, of course, that the relation to the rules of the rest of us is unthinking acquiescence. Rather, what we are shown is how and why the motives in question are precisely the sort of thing that we cannot but see as integral to our own ethical outlook. And it is for this reason that asking, Why should I keep my promises? and meaning by the question that one would be satisfied only by a showing of the legitimacy of the rules, so as to establish the ultimate correctness of one’s own motive to comply with them makes little sense. We would have to imagine that a person might somehow “distance” herself completely from her veneration for the rules in question, and then query the basis of their right to play the role they do from some point of view outside of an ethical life defined partly in their terms. But this assumes what simply is not true on this view—even supposing that any meaningful question of that form might be raised, it is fundamentally unclear what relevance any answer could have for us as we in fact are viz. people who pride ourselves in being (or striving to be) people of our word, and who therefore would be distasteful, if not altogether unrecognizable, to ourselves if we lacked such a motive.¹⁷

The denial that it is in this way a matter of their justification to the agent, however, need not entail that what Hume does provide at the level of the practice is “merely explanatory” (whatever exactly that might mean). What his stories aim to illuminate is something we might instead label our practices’ rationales. What I have in mind by talk of a given practice’s rationale is something that provides an answer to the question: What is our practice for? What is its point, or purpose?

These are important questions, but they are clearly not the questions of individual agents seeking vindication of some claim to authority of the rules of our practices. As I have just argued, that is not the relation of good agents to the rules on this view. Rather, through comprehending what our practices are for, the hope is that we might render the familiar, urgently felt, estimable motive to adhere strictly to them both more intelligible and more robust for us as people reliably subject to being moved in that way. In other words, Hume’s stories aim to illuminate the attraction for us, from the inside, of the way of life that becomes possible through our exhibiting precisely the sort of devotion to these rules we cultivate and prize.

As with Rawls’s own recounting of the emergence from his “Original Position,” Hume’s stories of the transition out of his fictive State of Nature are in this way read as an “expository device,” in which the temporal relations among the stages of the story help to reveal to us certain structural relations among our own various motives, ideals, and values that might otherwise be opaque for us (Rawls 1971: 21). In either case, ethical work gets done if and because, having come to see certain features of ourselves more clearly, we like what we see. Among other things, Hume’s own stories make it clearer how an otherwise powerful tendency to take up a narrowly instrumental or self-interested view, especially when dealing with those for whom we have little or no special regard, does not come to dominate in situations where certain practices and the motives to which they give rise hold sway.

Hume’s own understanding of why this is something to feel good about tends to highlight our needs and interests as homo economicus. But there is another current in Hume’s discussion, which has been emphasized by Annette Baier (1988: 769–74). Hume’s stories of moral invention also invite us to see our feelings of bondage to one another in accordance with these social rules as a kind of artificial simulacrum, of the more partial feelings we share with our intimates—for instance, the feelings of love, loyalty, and affection that bind together the members of a family. In the case of promises, for instance, we see how we are not condemned to distrust in our dealings with strangers. Rather, through life together within our practice, we realize the possibility of living together as people of our word, who can choose to show one another substantial forms of trust and reliance that resemble the more “natural” trust forged within loving families.¹⁸

At any rate, the important point for now is less about any specifics of Hume’s own proposed “rationales” as they might be gleaned from his discussion in the Treatise, and more about their intended effect. The result of considering what sort of rationale can be given in support of our practices is greater clarity concerning the values and interests served by our practice, as it is broadly realized in its ethically mature form. This is achieved through reflecting on the practice’s role in shaping the feelings, motives, and identities of those of us who live inside of it. The result is not merely idle or disinterested comprehension of ourselves, for the ultimate purpose of the investigation is to turn our ethical faculty in on itself and see what it makes of its own operation in exhibiting these special, sophisticated motives.

Hume’s own rationales are, he thinks, bound to be edifying—our moral faculty, turned in on itself, approves of its own operation within the settings of our practices. As a result, the motives on which those practices run are likely to be strengthened in us: learning even more about what we are like, we like what we are even more.

Whether that is always the case is a question to which we will return.

4. 

Taking a different, more Humean view of moral obligation will, in turn, make a difference to how one reasons about substantive ethical issues in regions of life where obligations are in view. One central difference, which I now want to begin to explore, concerns whether one conceives of the task of “everyday” ethical thought, as well as its more “rigorous” continuation via the methods of ethics, in terms of figuring out which our moral obligations “really” are.

This is the natural approach, when one accepts the mainstream view of obligation as inescapable constraint. The idea is familiar. An agent encounters a putative, or prima facie, moral obligation. But appearances can be deceiving. She must therefore determine—through intuition, or pure practical reason, or some appropriate procedure of construction—whether what at first looks like a moral obligation actually is one. If it is, then deliberation is brought to a close, for one must do whatever one has a real moral obligation to do.

An alternative approach, implied by the more Humean picture of obligation I am advocating, maintains that it is for the most part easy to know which our “real” moral obligations are. But figuring out that one indeed has an obligation is, at least sometimes, just the beginning of the task of deliberation. It certainly need not mark its successful completion. One may readily acknowledge that one has a bona fide moral obligation, yet feel unsure about what to do.

The more familiar picture is well-illustrated in a pair of complementary arguments recently offered by Elizabeth Brake, for the surprising conclusion that the taking of marriage vows is not, despite what most people tend to believe, the making of a valid promise. Rather, according to Brake, we have reason to conclude that marriage vows generally constitute a confused or failed attempt at promising.¹⁹ The aim of my discussion of Brake’s arguments is not so much to refute her (I have at least some sympathy for some nearby ideas). Rather, I am interested in the method by which she reaches her conclusion, and the significance it is supposed to have. Brake’s arguments serve as an especially vivid example of a type of moral argument that, on the view of obligation I am advancing, one should want to resist.

Each of Brake’s arguments is straightforward. Both start from the assumption that a valid promise must generate a moral obligation for the person who made it. Certain ideas about the conditions under which a person can have a bona fide moral obligation are then introduced as premises, in order to show that a person who takes marriage vows, no matter how earnestly and sincerely, cannot be understood to be morally obliged as a result, and so cannot have succeeded in promising.

In the first of Brake’s arguments it is treated as a premise that, if one has a valid moral obligation to $\varphi$, then one will be subject to moral blame if one does not $\varphi$. But unilateral divorce, according to Brake, is generally not seen as grounds for blame—at least not as such, by us, nowadays. Yet opting for divorce is presumably contrary to whatever content is specified in one’s vows. So, we should conclude the marriage vows generate no genuine moral obligation, and therefore cannot constitute valid promises.

In the second argument, the crucial premise is that one can only have a moral obligation to do what is under one’s direct, voluntary control. However, on any realistic construal, the content of the marriage vow is sure to include such things as affection, loyalty, and—especially—love. These matters of the heart, though often explicitly mentioned in the exchange of vows, are not the sorts of things we take ourselves to be able to do “on command,” in anything like the manner of the actions specified in our “typical” promises (which, we may suppose, involve giving rides to the airport and the like). It once again follows there can be no moral obligation to do the sorts of things we vow when we marry, which implies that we cannot succeed in promising to do them.

The crucial premise of each argument is of course a familiar commitment of the standard view of what moral obligation is and how it moves us. (They are, moreover, conceptually linked, through the idea that it would not be fair if we were subject to blame for doing, or failing to do, that which we could not voluntarily command ourselves to do.) These claims about the conditions on bona fide obligation are, it would seem, the sort of thing that must be justified a priori; or at any rate, they are already presupposed in the understanding of the pleonasm that promises generate moral obligations, since they are used by Brake to deny that a genuine promise has here been made.²⁰

Brake acknowledges that her claim that marriage vows are failed attempts at promising will come as a shock to many people, including especially those who have made marriage vows themselves. It seems safe to assume that people do in fact believe themselves bound by their vows in some way that they take to be morally meaningful. They see those vows as reason to do or refrain from doing various things. They see certain transgressions as a wrong done to their spouses. Often the term “promise” is explicitly used in taking the vows. Those who take them are apt to feel various reactive attitudes towards themselves and others if they do not live up to them, whether through overtly voluntary acts or omissions like sexual infidelity, or in less voluntary ways, like gradually falling out of love or growing apart. This is all acknowledged by Brake (much of it explicitly; all, at least tacitly). Yet according to her, these and any related beliefs, expectations, and other attitudes do not add up not to any real moral obligation. Rather, to borrow Kant’s expression, Brake’s arguments purport to show how the very idea of a marriage promise is a “chimerical concept” or an “empty delusion.”

An alternative, more Humean approach might start with the fact that people are, as a matter of fact, bound by their marriage promises, in a way that counts as a moral obligation if anything does. Taking that as uncontestable, we might then approach the issue from the other way around: given that this is a central instance of undertaking to bind ourselves in an especially meaningful way, through a central organ of our institution of promising, what might be learned from carefully considering this instance, about what it can mean to be morally bound or obliged?

Returning to the first argument, Brake is of course correct to note that certain forms of disinterested rebuke seem out of place to us nowadays, though it is perhaps harder to imagine a completely “faultless” unilateral divorce. Certainly, if it is truly unilateral, one’s spouse would ordinarily seem entitled to find some fault (though this may be attenuated by a disposition towards forgiveness or grace). At any rate, Brake must reject the idea that the familiar sorts of hurt feelings we would find warranted if one’s spouse unilaterally fell out of love could ever constitute holding her responsible for having “wronged” one in any distinctively moral sense.

But why suppose that? Part of what is at issue in Brake’s premise about blameless unilateral divorce would seem to concern whether there exists some special, fully impartial form of recriminative attitude, of the sort often assumed in talk of moral blame—a negative attitude appropriately held by anyone at all, towards a moral malefactor “on behalf of” the community at large. Even if there does exist that attitude, there is the further question of why we would have to take its appropriateness as a necessary condition for some transgression to count as a moral one. One might instead think that the lesson of considering such divorces is that at least some of a person’s serious failings can be at once considered moral, and yet not everyone else’s business.

With respect to the second argument, Brake is surely right that the norms and expectations that make up our practice of marriage are not exclusively or even primarily a matter of permitted or required conduct. Rather, they are as much as anything organized around experiencing and exhibiting appropriate attitudes, including especially forms of feeling and emotion, the relation of which to the will is, to say the least, obscure. While these may not all be the sorts of things we can reliably produce “on command,” they nevertheless do seem to constitute central examples of the sort of thing for which we regularly hold ourselves and others responsible.

Indeed, carried through to its logical conclusion, the sort of policing of the boundary between the voluntary and the non-voluntary on which Brake’s second argument depends threatens to invalidate all sorts of promises we are otherwise apt to make and take seriously: promises to be more patient, for instance, or more considerate, or more observant. Consistently applied, the condition that valid promises can concern only what an agent can do “on command” would shrink promising to little more than a device for controlling one another’s behavior, useful primarily in contexts characterized by considerable distrust, where we want the power to order one another around. Taking that upshot as a reductio, we might instead opt for a vision of the activity more governed by what we can credibly choose to take responsibility for when it comes to ourselves, which on any reasonable construal goes well beyond what we think ourselves able to control voluntarily (Williams 1995).

The basic point is this. On the alternative approach I am now sketching, it is perfectly sensible to wonder if this or that particular promise was “real” or “valid” on the grounds that, for example, it might have involved a confusion or reasonable dispute about the conventional conditions for giving one’s word. But the idea that there is some widespread, wholesale misapprehension about an entire class of promissory obligations, which people nevertheless earnestly and sincerely undertake, and are prepared to regard as a real force in shaping how they think, feel, act, and react to one another, veers toward the incoherent. This is the force of the Humean slogan put into action, that what affects us can never be a chimera.

Some may object that concluding that we are genuinely bound by some significant social practice (in this case, the practice of marriage promises) just in case we sincerely and earnestly regard ourselves to be implies acquiescence, or a conservative embrace of the status quo. This is a mistake. What it does imply is that serious ethical criticism must find a different purchase.

One could, for instance, argue that it is foolhardy to subject oneself to the demands and expectations that result from the marriage promise, by appeal to other important values that it undermines. Certain ideals of diachronic autonomy, or even a (re)consideration of the conditions of identity over time, could be invoked to argue that nobody should undertake to bind herself in such a lasting a way, for instance. Another familiar kind of critique might appeal to an ideal of love, arguing that marriage promises necessarily change love’s character by, for instance, hindering its spontaneity, in a way there is reason to resist. Notably, both these arguments begin with the fact that marrying does involve making a binding promise, and urge against marrying on precisely that basis.

A second, more radical strategy, which may come closer to Brake’s own in spirit, is to argue for more sweeping reform, or even abandonment of (some part of) our practice, precisely through considering what I above labeled (that part of) our practice’s rationale. What sort of “story” can be told to those who live within the practice of making and striving to honor marriage promises, that might serve either to reinforce or else destabilize their sense of the significance and importance of upholding the rules, norms, and ideals that constitute that practice? If one were to ask these questions, one might find that no fully satisfactory rationale for the practice could be given; that what rationale there is turns out to be outdated, or rests on a view of the relation between the sexes or a vision of the nuclear family that one rejects; that what rationale there is is in fact sinister.

I shall not explore these options further, though they seem to me worth taking seriously.²¹ I simply note the “mechanism” by which they can be understood to operate. In the present instance, they serve not to “invalidate” anyone’s marriage promise by undermining our practice’s justification from some point of view further outside it, from which it supposedly derives the moral right to govern the lives of those within. Rather, criticism at the level of rationale serves to destabilize the sense of bondage “from the inside,” through cultivating a certain kind of alienation or disillusionment that erodes the force of feeling. One goes searching for an integrated vision of where devotion to this practice fits, within a broader view of one’s own ethical life and concerns, and is left wanting.

The force of successful critique on this view cannot be, as philosophers like Brake imagine, a matter of discovering that one was never “really” bound. Rather, its effect is to (begin to) unbind us.

5. 

For some, of course, this sort of picture of our moral obligations will seem manifestly inadequate, particularly when deployed with an eye toward guiding individual deliberation, especially in hard cases. It is in those cases that the specter of “rule-worship” may seem to arise. We are now in a better position to consider that objection head on.

To raise the objection against the Humean view, Christine Korsgaard invites us to imagine a lawyer who finds a secret will and confronts a hard question. The will leaves a deceased client’s fortune to his profligate nephew, and supersedes an earlier will that would have left the money to valuable medical research. The lawyer is the only one who knows about the later will and finds herself contemplating its destruction. This would clearly be against the rules, but it would also do a lot of good. Korsgaard imagines that the lawyer has read Hume and accepts something like his account of the motives by way of which she feels herself bound to follow morally important social rules—in this case, presumably, the rules that define property and its posthumous transfer. Reflecting on this way of accounting for her own sense that the rules deserve her allegiance will, according to Korsgaard, undermine the operation of the motive in question. Here is Korsgaard, imagining a bit of the lawyer’s inner monologue:

She says to herself that she would disapprove of herself if she did this. She hates unjust actions and the people who perform them. But since the lawyer knows Hume’s theory she also knows why she would disapprove of herself. She would disapprove of herself because unjust actions have a general tendency to bring down the system of justice. But she also knows that her distaste for such actions is caused by their general tendency, not their actual effects. As Hume has shown, our moral sentiments are influenced by “general rules.” And our lawyer knows that this particular unjust action will have no actual effects but good ones. It will not bring down the system of justice, and it will bring much-needed money to medical research.

The lawyer believes that her disapproval of this action depends on the fact that actions of this kind usually have bad effects that this one does not have. It is almost inconceivable that believing this will have no effect on her disapproval itself. Her own feeling of disapproval may seem to her to be, in this case, poorly grounded and therefore in a sense irrational. And this may lead her to set it aside or, if she can’t, to resist its motivational force. (Korsgaard 1996: 87)

Korsgaard is clearly making a version of the charge of “rule-worship,” in a way that rightly emphasizes what it would be like from the inside to regard oneself bound to a rule, even though the sort of value one takes the rule to serve would seem to be best served by saying “to hell with the rule” in this instance. But it seems to me that there is much for the Humean to say here.

Perhaps the first thing to emphasize is the distinction made famous by Rawls (1955), between what he called the “summary conception” and the “practice conception” of a rule. On the “summary conception,” rules result from generalizing over some class of similar cases, serving as a shortcut to deciding each on its own merits. On the “practice conception,” by contrast, rules are adopted as (de facto) sources of authority, precisely because it is only by deferring decision to the rule that certain valuable forms of (inter)action are rendered possible. Rawls’s invocation of the example of promises serves to illustrate the point well: it is only when people see the rule that promises must be kept in accordance with the practice conception, that the voluntary issuing and accepting of reliable assurances appears on the menu of action.²²

The importance of this distinction here is that if the Humean view is to have its day in court (no pun intended), the lawyer should be allowed to see her relation to the rules in terms of how their establishment as rules of a practice makes possible certain valuable forms of (inter)action. It is under that conception that she presumably has regard for them, feeling admiration for anyone else who exhibits a similar attitude.

That is, she admires herself and others not as people who implement the rules of property as handy shortcuts because this predictably redounds to doing good on the whole. Rather, she understands that by regulating themselves in accordance with these rules, in the sense of deferring individual decision to them, people are transformed into honest, reliable, and trustworthy targets for social interaction and cooperation in a world of otherwise fierce competition over scarce material resources. This is essential to understanding the rationale for our system of property, and thus essential to explaining why the lawyer would hate (or lose respect for) anyone who too readily chooses to substitute her own judgment about what ought to happen for what the rules direct.

The second and perhaps deeper point about the example concerns what sort of expectations we should have going in, about the possibilities for ethical conflict, and the conditions of its successful resolution. In the example, the law is evidently clear: the later will should prevail. The lawyer is nevertheless conflicted because, on the one hand, it seems important to her to follow the law. But on the other hand, it also seems important not to be so bloody-minded about it, to the point where one is prepared to bear any cost—in particular, in this sort of case, massive costs to the lives or welfare of others.

One way of approaching this sort of conflict, implied in Korsgaard’s own discussion of the example, is to assume that, to the extent that it is a conflict among genuinely moral concerns, it simply cannot go “all the way down.” Rather, apparent conflict must be resolvable by appeal to some higher-order rule or principle, which tells the agent which of two or more competing considerations really should win out. The force of the cannot and the must here is that of conceptual truth, and the thought can in fact be reformulated in terms of the more familiar way of thinking about what moral obligation “really” is and its attendant role in deliberation, which I discussed in the previous section. We may be tempted, that is, to think of the lawyer as facing two disparate considerations that each have the initial appearance of bona fide obligation. She should therefore see her deliberative task in terms of figuring out which is her actual obligation, so that she can act correctly, morally speaking. A judgment to the effect that she has a genuine moral obligation to disclose the will represents, on this view, the successful solution to the lawyer’s deliberative problem.

But why should a Humean lawyer be forced to see her predicament in terms of determining which of two strictly incompatible courses of action she is, as a matter of fact, unconditionally required to take? One might instead think that certain kinds of conflict are bound to emerge among disparate sources of value and concern, which are in an important sense incommensurable. In particular, a person’s moral obligations may conflict with other forms of value that themselves do not take the form of obligation. When they do, at least sometimes there may be no entirely systematic way of ordering or weighting the various sorts of concerns that she takes to bear on her practical situation. In that case, the perpetual search for some sort of all-in justifying principle may be a hindrance to achieving whatever sort of deliberative success is available to an agent, given her circumstances.²³

A very different characterization of the situation faced by Korsgaard’s lawyer is available along these lines. When I imagine myself in the lawyer’s shoes, I indeed imagine feeling conflicted; but that conflict arises precisely because I already know I have an obligation to bring to light the later will, grounded in my regard for our institutions of property. This sort of obligation is something I take very seriously, and in other instances it would be an entirely sufficient basis for decision. But in this instance, that obligation conflicts with something that is also profoundly important, ethically speaking: the lives and interests of others that would have to be sacrificed, for the sake of following the wishes of a dead man to squander his fortune. My problem, so understood, is not one of determining whether I am “really” obligated, but rather whether to violate my obligation for the sake of something else I regard as important. I feel I must choose, and there is no obvious principle or rule to which I can defer my decision.

The claim that there is no single, decisive principle for settling such conflict, need not imply that any choice is thereby rendered arbitrary. One can hold that a decision may be made for better or worse reasons, or in a better or worse way, without assuming that there is some way of fully justifying that decision against any alternative—in particular, justifying it to all parties concerned or affected. I suspect that I, in the lawyer’s shoes, would eventually come down on the side of exposing the later will, despite the consequences.²⁴ I also imagine I would feel less than fully justified, in at least one familiar sense of that term. While I might offer grounds for the choice in the seminar room, I do not know that I could face the families of those whose lives were lost as a result of choosing to uphold the rules in this instance and expect them to accept my decision. They might resent or blame me for refusing to make an exception, and I do not know that I could defend myself to my own satisfaction, to say nothing of theirs, in the face of their criticism. What I could manage would no doubt take the form of appealing to the immense importance of not arrogating to oneself the choice to violate the rules of an immeasurably valuable practice, even when the stakes are this high. But I do not know that this should be “enough” from their point of view, even if it proved enough for me to decide as I did.²⁵

When I imagine, on the other hand, someone else making the opposite choice, I want to know more about how she deliberated. If someone showed no regard for the rules because she thought them obviously “outweighed” by the good that could be done by setting them to the side, I will have little esteem (or patience) for her. But a different possibility, for which I have more respect, might involve coming to reflect critically on the rationale of our institution of property. Suppose the lawyer came to wonder how any system that gives some the power to control so enormous a fortune that its disposition might literally determine whether countless others live or die could possibly be worth upholding? Such reflection might weaken her resolve, and lead her ultimately to violate the practice. This would amount to a radical critique not of the application of the rules in this instance, but of the ethical desirability of the very way of life the establishment of those rules makes possible. This is another instance of the sort of critique toward which I gestured at the end of the previous section in the discussion of marriage—a critique that lessens the sense of bondage, through erosion of the agent’s sense of our practice’s rationale.

6. 

I have been discussing a view of our sense of moral obligation and how it moves us, which runs counter to one embraced by many philosophers. On the more “Humean” picture, to feel oneself morally bound is to be moved by a sense of devotion to, or reverence for, certain social practices or the rules or norms that constitute them. I have argued that this sort of motive, properly understood, neither wants nor admits of ultimate legitimation. It does, however, invite asking after what I called the rationale of our obligation-engendering practices. A better understanding of our practices and their rationales may serve either to reinforce or to undermine the operation of the motives in question, depending on what sort of answer we are able to provide. I next argued that, when obligation is seen this way, it makes little sense to conceive of deliberation in terms of determining whether one is “really” morally obliged. Moral obligation is, on this more Humean view, better thought of as among deliberation’s inputs. Ordinarily, that I have an obligation is an entirely sufficient basis for decision, and serves as an ample justification for acting as obliged. But the question can still arise, sometimes, whether to do what I am morally obligated to do. I have tried to show how this sort of view is not subject to familiar worries about the reflective stability of our motives to follow moral rules, while simultaneously demonstrating how serious ethical criticism can nevertheless have traction. I take these all to be selling points of the view: considerations that favor of adopting it, in the sense of thinking through ethical questions and challenges concerning our obligations in its recommended terms.

There are, nevertheless, ways the view will remain unappealing to at least some people’s ethical sensibilities. Raising questions about our relation to our obligation-engendering practices in any serious way requires a willingness to face the uncomfortable possibility that we may come to regard our own cherished moral motivations as, at best, grounds for ambivalence. I more than hinted at this possibility with respect to two of the practices I considered in this paper: the marriage promise, and institutions of private property. I do not know why we should expect anything else concerning any of the other various practices that give rise to obligations for us, to which we are also devoted. When that sort of ambivalence does result, such a view promises no resolution. It may be our lot in life to feel deeply invested in certain practices that, when considered from another view available to us, want for a sufficiently strong rationale.

Whether one regards the possibility of this kind of deep and intractable conflict within the soul as cause for alarm, relief, or both, is perhaps a matter of one’s ultimate temperament. Korsgaard goes on to say, in discussion of her lawyer, that she

. . . may find that she cannot destroy a valid will without intense feelings of humility or self-hatred. These may or may not be strong enough to cause her to desist. But even if they are there will have been a normative failure. The lawyer does not believe that the claims her moral feelings make on her in this case are well-grounded. If she could cure herself of them then that is what she would do. (1996: 88)

I confess that I do not see why this is, as such, the description of a normative failure. Indeed, it seems to me that, thought differently, it may serve as a striking acknowledgment of the sort of thing we all have to face up to, as complicated creatures who live in a messy ethical world that is never fully to our liking. We are all of us moved, at least sometimes, by feelings of bondage that we do not get to choose for ourselves. The ideal of the person according to which some subset of her feelings—in particular, the properly moral ones—can be autonomous in the special, Kantian sense of existing beyond the influence of sources outside her genuine self, is itself the product of a particular moral culture, the rationale for which is shaky at best.

For those of us made uneasy by mainstream thinking in moral philosophy, perhaps one of its more objectionable features is that it persists in telling us we are not allowed to feel ambivalent about our own relationship to morality in this way.