Generic Moral Grounding

Abstract

Moral theories often issue general principles that explain our moral judgments in terms of underlying moral considerations. But it is unclear whether the general principles have an explanatory role beyond the underlying moral considerations. In order to avoid the redundancy of their principles, two-level theories issue principles that appear to generalize beyond the considerations that ground them. In doing so, the principles appear to overgeneralize. The problem is conspicuous in the case of contractualism, which proposes that moral principles are grounded in generic reasons that operate in only a subset of the cases covered by the relevant principle. Arguments that motivate the use of generic reasons on the basis of guidance, institutional necessity, or fairness are unsatisfactory. But the generality of moral principles can be justified in terms of the idea of a generic interest, which is an interest that all occupants of a standpoint have in virtue of what occupants of the standpoint as such typically take an interest in. The notion of a generic interest therefore acknowledges a distinction between having and taking an interest, and its use in moral theory is a way of giving expression to the significance of autonomy. This is reminiscent of the idea of a normative power; but the idea of a generic interest is less expensive and makes better sense of the way in which autonomy features in our moral lives.

1 A Puzzle of Moral Grounding

We turn to moral theory only in part because we want to know the facts about which actions are permissible and impermissible. We also want to know why these particular facts obtain. To be exact, we want to know what morally grounds these facts.¹ We seek to know not simply that our moral judgments can be relied upon to be correct, as they would be if produced by an oracle, but that they do justice to the various considerations that compete for our attention when we are thinking clearly about what to do.

The search for this sort of grounding is especially important for those who think that a moral theory should take our ordinary moral judgments seriously, as do those who find act-consequentialism overly revisionary. Someone who aimed only at tracking commonsense morality might simply provide a list of our ordinary judgments about particular cases. Such a list could not be understood as a theory with the ambition I have in mind. That is only in part because, without more, the list would provide no reason for continuing to accept any of the enumerated judgments. Establishing the coherence of these judgments, or the social convenience or practical necessity of accepting them, or their endorsement by those with better epistemic qualifications than us, may each count toward justifying our acceptance of them, but would not amount to explaining why these judgments hold true. What is needed is to show that our moral judgments are well-grounded. Consider what you might say to a person who is negatively affected by an action that moral judgment seems to require. The affected person’s complaint would not be properly met by facts about the coherence of your view, or its social utility, or even its endorsement by a moral oracle. What is needed instead is an account of how your moral judgment is properly responsive to the features of the situation. Such an account would explain why the action is morally required, and it would do so by grounding the requirement in the moral features of the case.²

This appealing view of the aim of moral theory sits uneasily with the place given to principles by ordinary moral thought. Consider the example of promising. Moral thought does not just involve such judgments as that this promise ought to be kept, and that one too. It also suggests a principle of the following form:

Promising. There is an obligation to keep a promise that is successfully made (i.e. sincerely asserted, not in conflict with antecedent obligations, taken up by the promisee, etc.), barring certain exigencies of limited scope (e.g. unforeseen emergencies).³

Suppose (what is admittedly controversial) that we can successfully set out what counts as a successfully made promise and which exigencies undermine a promissory obligation, without completely stripping the principle of its generality. Now just as we can ask for the grounds of the judgment that a particular promise ought to be kept, we can ask for the grounds of the Promising principle. If the principle is grounded in moral considerations that appear in every case it covers, then the principle is in danger of being superfluous, and will appear to merely abbreviate the fact that these considerations appear in the relevant cases. That is the problem of redundancy (though I set aside whether it is really a problem).⁴ If the principle is grounded in considerations that do not appear in all the cases it covers, then we need an explanation of why the principle extends beyond those cases which do feature the considerations that ground it. That is the problem of overgeneralization,⁵ and it is the problem with which we will be concerned. It is not a problem for every moral theory; it is for those that attempt to give generality a special explanatory role by distinguishing two theoretical stages, one involving the justification of judgments about particular cases in terms of general principles, a second involving the justification of general principles in terms of moral considerations that appear in particular cases.

The practical appeal of a general principle is that it provides guidance to human beings who are cognitively limited and unable to act on the basis of an enumeration of the correct moral judgments in all possible cases. Some have hoped that we can retain this appeal even while loosening the supposition that we can definitively set out the scope of a principle like Promising by specifying which circumstances undermine a promissory obligation. Key to that aspiration is the thought that when we are uncertain whether a principle covers a case, we can turn to the considerations that ground that principle and compare them to the considerations that feature in the relevant case (Scanlon 1988; Väyrynen 2009). But this idea simply deepens the dilemma for general principles. If each case can be reconsidered in light of the grounds of a moral principle, then the principle does no special moral work. Principles are rendered superfluous in a similar way when Scanlon says, in his defense of two-level moral theory, that commonsense moral principles always stand to be completed by critical thinking about the grounds of those principles (1988: 130). Perhaps then a moral principle adds something morally significant to the particular considerations that ground it, either by applying some force that the considerations do not have on their own, or by extending to cases that do not feature these considerations.⁶ But this appears to render principles overly general.

One way to respond to this apparent dilemma is by jettisoning moral principles. But to do that seems to overlook something important about the structure of ordinary moral thought. That structural possibility is contemplated by David Owens when he suggests that there are cases of bare wronging: cases in which an action is wrong, and indeed wrongs someone, but not because it sets back the wronged person’s interests, or anyone else’s for that matter (Owens 2012, 15–16, 64). We can think of several candidate examples:

3 Am Trespass. I cross your front lawn at 3 am, wearing plimsolls and leaving you to sleep. I leave you and your lawn unharmed, and you are not even aware of the trespass. It is nonetheless a trespass that wrongs you (Owens 2019, 279ff).

Boring Diary. I secretly read the teenage diary you have hidden away. In it you have recorded only each day’s weather. I discover nothing embarrassing or private about you. I don’t even know who you are. Nonetheless I have invaded your privacy, wronging you.

Sewing Machine. A mother promises to give her daughter a sewing machine. The daughter does not want the sewing machine, but doesn’t say so to save the mother’s feelings. (Scanlon 1998, 311).

My aim will be to explain the structure of cases like these, and I will only return to Owens’s competing explanation toward the end. My basic approach involves showing how the cases could be covered by moral principles that are grounded in considerations that do not directly feature in the cases themselves. But the grounding considerations feature indirectly insofar as the relevant patient has an interest in how they would be treated in a case in which the considerations do feature directly. That seems to me to vindicate the intuitive two-level structure of ordinary moral thought, which admits both deliberation about particular cases in terms of general principles, and further deliberation about the basis of those principles in terms of typical cases. I won’t press the case for thinking that such principles escape explanatory redundancy. But I do aim to show that the generality of principles does not simply reflect the shape in which we happen to find our particular moral judgments. Their generality has some moral merit.⁷

2 Contractualism’s Generic Reasons

I turn now to the contractualist theory formulated by Scanlon (1998), since it embraces both principles and their grounding considerations, and so confronts the explanatory challenge I have been articulating. Scanlon’s response involves the notion of a generic reason. I hope to show now that this notion is unsatisfactorily motivated. But I will not completely jettison the idea, and in the sequel I hope to show how the similar idea of a generic interest both solves our puzzle and can be independently motivated.

Contractualism holds that an action is wrong if any set of principles permitting that action could be reasonably rejected by those who seek mutually non-rejectable principles by which to govern their conduct (Scanlon 1998, 189). It has been questioned whether this criterion adds anything of substance to our existing moral views, since the idea of reasonable non-rejectability simply refers us to the considerations which make particular actions wrong.⁸ But the merits of contractualism are primarily structural. One way in which the contractualist criterion illuminates the structure of our moral views is that the idea of reasonable rejection calls attention to the grounds of the principles we accept. That is why Scanlon insists on formulating the test in terms of principles that cannot be reasonably rejected rather than principles that all could accept (1998, 390n8).

There is yet more to the structure of contractualism. In assessing the moral status of an action, the contractualist compares the seriousness of the complaints that could be raised against various alternative principles regulating that action. But Scanlon thinks that complaints may only be formulated in certain ways, and it is these qualifications, such as that complaints may only refer to ‘personal’ reasons, that place substantive constraints on the judgments Scanlon’s theory can endorse. One of these qualifications requires that the complaints to be considered are those that arise from representative standpoints. What this means is that a complaint against a candidate principle must be formulated in terms of generic reasons, which ‘are reasons that we can see that people have in virtue of their situation, characterized in general terms, and such things as their aims and capabilities and the conditions in which they are placed’ (Scanlon 1998, 204–205). These reasons do not exist as a matter of statistics: there may be generic reasons for complaint even if only a few people who occupy the standpoint have grounds to raise such a complaint (205).

The invocation of generic reasons suggests that the trouble for contractualism is not redundancy so much as overgeneralization. The proposal seems to be that the contractualist procedure will generate a principle that covers some set C of cases but is fully morally grounded in the generic reasons that appear only in some (possibly small) subset C_S ⊂ C of cases. But we typically think that some fact F fully explains G only if G follows from F,⁹ and that the inference from F to G goes from a stronger claim to a weaker claim. For example, it is bad to infer from the fact that most days are sunny in Philadelphia that all days are sunny in Philadelphia. Likewise it seems bad to infer from the fact that the reasons for not ϕing appear in some cases C_S ⊂ C that one should not ϕ in cases C.

The appeal and limitation of generic reasons is well illustrated by Scanlon’s account of promising, which rests upon the justification of a principle that can be abbreviated as follows:

Principle F: If (1) A assures B that A will ϕ, (2) A and B have the right kind of mutual knowledge and intentional attitudes toward this assurance, and (3) B does not consent to A not ϕing; then in the absence of special justification A must ϕ (Scanlon 1998, 304).

This principle is based on reasoning about the sorts of interests at stake in a representative case of promising. While a promise is often made because the promisee wishes to rely on the promisor’s performance, reliance is not always at issue. More generally, promisees seek assurance that the promisor will act as promised. The value of such assurance lies not just in the satisfaction and peace of mind experienced by the person assured, but rather in the fact that the promisee is thereby able to determine the way things will go (Scanlon 1998, 303). Principle F is a principle of promising because it is triggered by assurance of the sort conveyed by promise-making. It is not enough that one person wants assurance though. There must be something approaching common knowledge between promisor and promisee, as specified by clause (2) of F. Finally, the promisee’s interest in assurance is not the only interest at stake, and the promisor will have an interest in being free not to perform if circumstances change. But the promisor will also have an interest in being able to provide assurance, since this allows coordination of the type required by all sorts of joint activity. So while a promisor may have reason to reject a principle like F, his own reason to endorse the principle may undermine his complaint. But clause (3) allows that the promisor’s interest in non-performance can be met by gaining release from the promisee. In this way the principle strikes a balance between the interests of the relevant representative standpoints in such a way that it could not be reasonably rejected from either standpoint.

Grounding Principle F in assurance rather than reliance ensures that in many cases of promising a promisee will indeed have the interest in assurance that F protects. But it is still possible to imagine cases in which we judge that a promise is binding even though the promisee lacks the interest in assurance, or has a weak interest that is outweighed by the promisee’s interest in non-performance. Sewing Machine is one such case. But while the daughter lacks an interest in being assured that she will receive the sewing machine, and the mother may easily develop an interest in non-performance, we nonetheless think that the mother has made a valid promise and is bound unless released. Scanlon defends F by denying that this is the right judgment, asserting that if the mother learned that the daughter had no interest in receiving the offending object, the daughter could not then insist on performance (Scanlon 1998, 313–14). But we should distinguish this scenario from one in which the daughter tells the mother that she doesn’t want the sewing machine, effectively releasing her mother. It would be no release at all if a third party told the mother that the daughter didn’t want the sewing machine, because the third party might be wrong, or because the daughter might since have changed her mind, or because the daughter might have some other reason for wanting performance. And even if the daughter in fact lacked any interest in performance, these possibilities make it unreasonable for the mother to set aside her obligation without something amounting to consent on the daughter’s part.

It is in cases like these that the contractualist may rely on generic reasons in order to derive a principle of promising that covers a wider range of cases than F does, but is nonetheless restricted to cases that bear an important relation to those of which F accurately describes the balance of interests.¹⁰ Notice that Principle F is not formulated as a principle of promising, but instead describes the standpoint of promisee only insofar as that standpoint is one from which assurance is generally sought, and the interest in such assurance generally outweighs the interest of a promisor in non-performance. Framed in this way, F describes the generic reasons which a promisee has to reject non-performance. But it is the promisee as such who may raise these generic reasons in complaining about the principles that govern promising. So the contractualist procedure allows us to reformulate F in terms of the standpoints of promisor and promisee, justifying it in terms of the generic reasons given for F:

Principle F^*: If (1) A promises B that A will ϕ, (2) A and B have the right kind of mutual knowledge and intentional attitudes toward the promise, and (3) B does not consent to A not ϕing; then in the absence of special justification A must ϕ.

This principle of promising applies squarely to Sewing Machine, vindicating our pre-theoretical judgment that the mother is obliged to perform unless she obtains release. But it is much less clear how we are supposed to derive a principle like F^* from the idea of a generic reason. The mother’s obligation to perform is not grounded in any interest that the daughter actually has, but rather in the interest that others who occupy the same standpoint typically have. Yet it seems a bad inference to go from the fact that a principle like F is required to protect those promisees having an interest in assurance to the fact that a principle like F^* is required to protect promisees in general. So the idea of a generic reason appears not only poorly motivated, but incoherent.

3 Guidance and Fairness

Moral theory seeks to provide guidance to human beings rather than angels. Human rationality is bounded in a number of ways. We have limited means to ascertain information, including information about how the interests lie; and in order that we can cope efficiently with our environment, our reasoning makes pervasive use of heuristics and biases that fall short of rational inference. Given these limitations, perhaps it is simply practical—because efficient, or because it’s the best we can do—that we make use of principles, even when the considerations from which those principles draw plausibility are not accurately captured by them.¹¹ This is the basic idea behind the rule utilitarian thought that we do better by following rules that produce suboptimal outcomes in some cases than by attempting to calculate what optimality requires in each particular case. At the limit, it is not just difficult to find out how the interests lie, but impossible, given that interests will often depend upon private facts about mental attitudes. So when Scanlon resorts to a similar approach in defending generic reasons, we should note that he is concerned not just with the efficiency or optimality of following rules, but with their justifiability in the light of our epistemic boundedness. Thus:

… an assessment of the rejectability of a principle must take into account the consequences of its acceptance in general, not merely in a particular case that we may be concerned with. Since we cannot know, when we are making this assessment, which particular individuals will be affected by it in which ways … our assessment cannot be based on the particular aims, preferences, and other characteristics of specific individuals. (Scanlon 1998, 204)

Given that these are the limits on our deliberation about which principles to take as guidance, the idea of a generic reason, which adverts to what is paradigmatically at stake in a representative case, is more appealing than the rule utilitarian’s statistical notion of expected consequences. It is nonetheless difficult to understand why we should think that principles grounded in generic reasons are any more than helpful heuristics for bounded agents. Most importantly, our need for guidance presents no moral reason for thinking that the correct action in a particular case is one that conforms to what generic reasons would require instead of what the moral features of the particular case require. One might worry that it is pointless to be concerned with morality if it is not a practice that human beings can engage in, and that may lead one to think that morality accommodates basic facts of human nature, such as our bounded rationality. But morality needn’t accommodate us in this way by consisting in principles that do not accurately reflect the moral considerations at stake in particular cases. Instead, morality can accommodate us by excusing us in cases in which we act on the basis of what we reasonably believe to be the case, even if it turns out that some other action would have been better given what actually is the case.¹²

The idea that rules are presented for our guidance is a version of what Rawls famously called the ‘summary view’ of rules, contrasting it with a practice view which treats rules not as generalizations from decisions about particular cases, but rather as constitutive of a practice which renders certain types of cases intelligible (1955). This view is easiest to make sense of in the case of institutions, which really do require rules ‘defining offices, moves, and offenses’ (Rawls 1955, 25). There too there is reason to think that ‘[o]nly by reference to the practice can one say what one is doing’ (27). But in the case of promising, as opposed to contract, it is not as clear that we need to assign to the promisor the office of promisor in order to make sense of the fact that his failure to perform counts as a promise-breaking, and to evaluate whether moral reason is on his side or the promisee’s. Conventionalism of this sort seems implausible to me. Even if one thinks that a social practice is required in order to make sense of how a promissory obligation arises, it is contrived to deny that we can make sense of what is at stake in a case of promise-breaking in advance of the practice being established. Moreover, even if the conventionalist were correct, we could always ask whether we should adopt instead of F^* a principle like it that makes an exception for some particular case; or for that matter, whether it would be better to adopt Principle F. So it is not clear why the practice view would not end up recommending rules of the redundant kind.

A later Rawlsian idea is that participants are required to act in accordance with certain practices not because of logic but because of fairness (Rawls 1999, 301–308). Fairness is in fact tentatively suggested as a basis for the generality of principles by Scanlon (1998, 205). The idea, familiar from discussions of moral responsibility (Wallace 1994), is that moral requirements must accommodate our cognitive limitations not because of our need for guidance, but because it would be unfairly burdensome to require of an agent that he act in a certain way if he is unable to easily determine that he is so required. More specifically, the concern is that the contingent burden of discerning the actual incidence of considerations in a particular case will vary arbitrarily with the nuances of the situation and the epistemic standpoint of the agent. This might strike a particularly epistemically burdened agent as unfair compared to a system of principles that requires only that each agent take into consideration the generic reasons pertaining to the relevant representative situation. Such a system need not abstract from the cognitive capacity of agents, which is also a matter of luck, but would aim at mitigating the burdens that arise for agents because they have no control over the comparative epistemic accessibility of the particular instances of moral scenarios they find themselves in. Perhaps that is what the contractualist appeal to generic reasons achieves.

While this concern may be compelling in some settings (perhaps those involving permissible competition), it enters at the wrong stage if taken as a justification for generic reasons, since bringing fairness to bear at this meta-deliberative stage gives it an unvarying force that is not properly sensitive to the other interests at stake in a particular situation. The problem is not just that an agent’s interest in being treated fairly according to his epistemic position might have been outweighed by a competing consideration favoring a less general rule, had the interest in fairness been introduced at the same deliberative moment as the competing consideration. In some cases, the generality of a principle does not even favor the agent whose interest in fairness supposedly motivates it. The proposal, in the context of Sewing Machine, is that it would be fairer, given that promisors face situations of varying moral opacity, were the mother to take up only as much of a deliberative burden as any other promisor does; and that this would be the case if each promisor only consider what would be required in a representative case of promising. The first problem is that this fairness interest may well be unimportant when compared to the other interests at stake, and that it gathers undue force under the current proposal because it is taken into account at the different methodological moment of considering how generally the other interests are to be formulated. The second problem arises when we consider that the mother has a serious interest in being at liberty to perform or not as she wishes, a liberty interest that may well tip the balance in her favor were it put in competition with whatever interest in assurance the daughter in fact has—given that the daughter in fact has none. It would be odd to conclude that it is the mother’s interest in fairness that weighs in favor of treating her as a generic promisor, and so bound to her daughter despite the balance of their actual interests.

4 Generic Interests

My proposal is that we have an autonomy interest in generalizing to some extent from the considerations that operate in paradigmatic cases. The proposal rests on the claim that protecting the autonomy of an individual S who occupies some representative standpoint involves treating S in accordance with the normative expectation that she would be treated in a way that protects the interests that someone occupying that standpoint would typically have qua occupant of the standpoint in paradigm cases that give rise to the standpoint. I will call the interests typically possessed by an occupant of the representative standpoint generic interests, the term indicating both that they are basic considerations of individual wellbeing like other interests, and that they gain their normative force from their internal connection to the generic features of the representative standpoint. As an example, a particular promisee S has a generic interest in assurance, even if in their particular case they do not in fact take an interest in being assured that the promisor will perform. Why then does S have any sort of interest at all in the performance? Not because her life goes better if she is assured of performance, or if the promisor performs, but rather because S is treated badly by not being given as much room to take an interest in assurance as the typical promisee would have. That is, S may expect (in the normative sense), in virtue of her position as promisee, to be treated just as if she had an interest in assurance.

In sketching the basic idea of a generic interest, I talked of both having an interest and taking an interest, and it is worth noticing that these are slightly different notions. The difference between them is much like that between judging that some entity is valuable and valuing it (Scheffler 2010). Judging that a practice like capoeira is valuable involves being disposed to acknowledge that it is a source of reasons, making statements to that effect and affirming the statements of others, such as that its value gives us reason not to prohibit its public performance. But valuing capoeira would involve a much richer set of affective and attentional dispositions, and an entanglement of these dispositions with features of one’s psychology. One who values capoeira may also be disposed to spend more time watching performances, to mourn the departure of a master of the tradition, or even to spend her evening hours raising funds for a local academy. There are many things in the world that are valuable because they are worth valuing in this way; but we cannot possibly give every valuable thing this kind of space in our lives. Valuing therefore involves a choice to live a life of a certain kind, at least in one small respect. That we have the choice to value different things is therefore an important source of autonomy, and autonomy is enabled by circumstances which give one a wide range of things worth valuing.

The distinction between having an interest and taking an interest is analogous with, and perhaps ultimately based on, the distinction between judging valuable and valuing. One has an interest in some state of affairs if one’s life goes better when that state of affairs obtains. One takes an interest in that state of affairs if one has the affective and attentional dispositions that go together with caring about whether that state of affairs obtains and giving the hope of its obtaining a significant place in one’s life. For example, S may have an interest in whether there is a swimming pool on campus because swimming would be a particularly good form of exercise for her health. But S takes an interest in there being a swimming pool on campus only if she has some of the dispositions involved in hoping that there is, such as the disposition to be disappointed on finding out that there isn’t, or lobbying the health and recreation committee to open one.¹³ Similarly, S may have an interest in being assured of a promisor’s performance whether she cares about performance or not. And even if she believes it doesn’t matter whether she knows whether the promisor will perform or not, her planning and peace of mind may in fact improve given the assurance that the promisor will perform as a backdrop. Taking an interest in assurance is different, and if she does S will pay particular attention to whether she can count on performance and be particularly disposed to be frustrated if she does not know. These dispositions may arise automatically— S may have an anxious streak—but in general someone like S will be put to a choice as to whether to take an interest in assurance or not. Even where the matter is not entirely voluntary, the fact that taking an interest involves attentional and affective entanglement means that it matters to an agent to what degree she is able to craft or tailor her interest-takings.

There is a twofold normative connection between the idea of having an interest and that of taking an interest. First of all, the fact that people are ordinarily likely to have an interest in x in a certain sort of case adds weight to the argument that it is reasonable to take an interest in x in that sort of case. The fact that one ordinarily has an interest in assurance in cases in which one is a promisee makes it reasonable to take an interest in assurance in any particular case of promising, even if assurance does not in fact make one’s life go better in the particular case. This aspect of the connection should not be understood as a straightforwardly statistical relation, since whether people are ordinarily likely to have an interest in assurance when occupying the position of promisee is not just a matter of how often they have the interest, but of how often they have the interest in ordinary cases. That leaves open the possibility of a logical conception of the relation, such as that ordinarily having an interest in x is sufficient for taking an interest in x. But consider that ordinarily people have an interest in being esteemed by others in their community, but that it is arguably overly self-absorbed and so unreasonable to take an active interest in this sort of esteem, though we ordinarily do so. This suggests that the claim that it is reasonable for someone to take an interest in x requires a moral argument, perhaps one that invokes our reasons for thinking that one ordinarily has an interest in x and not just our bare judgment that one does.

Second, the fact that people ordinarily have an interest in x makes it reasonably foreseeable that one will take an interest in x. The connection is again not obviously a statistical one. The reasonableness of foreseeing that one will take an interest in x rests not on our (no doubt unreliable) capacity to judge the probability that people take such an interest, but rather on our ability to judge the ordinariness of people doing so. To the degree that such judgments of ordinariness are non-normative, they are likely calibrated not so much on frequency as on social acceptance.)

The reasonableness of taking an interest in x, and the reasonable foreseeability that one might do so, argue for morality leaving one some room to take an interest in x. Why? Consider that the fact that it is reasonable to take an interest in x in a certain case means that it serves one’s autonomy that one be able to take such an interest (since it adds to one’s choices worth having). That does not on its own imply that one has an interest in being able to take an interest in x. Such an inference threatens to go too far, since there may be ways of expanding one’s autonomy that do not deserve protection, even before being weighed against competing moral interests.¹⁴ But if taking an interest in x is reasonable, then there is reason for morality to make room for doing so since it is already responsive to the fact that people ordinarily have such an interest. Moreover, if people ordinarily have an interest in x, then it is possible for morality to make room for one’s taking such an interest, since those with whom one interacts will be able to respond to the fact that one might take such an interest, given that it is reasonably foreseeable that one will do so.

Return to the case of promising. That promisees ordinarily have an interest in assurance argues for the reasonableness of a promisee taking an interest in assurance, even though they would otherwise lack the interest. Daughter may take no interest in getting the sewing machine, and she may in fact have no interest in it; but it would be reasonable for her to take an interest in it, and it would have been reasonably foreseeable that she had done so. Taken together, the fact that it is reasonable for a promisee to take an interest in assurance in such a case, and that it is reasonably foreseeable that she does so, makes it reasonable for the promisee to expect to be able to take such an interest. That is, she may demand that the promisor leave her room to do so, even though she in fact does not end up taking an interest. And to ask that the promisor leave her that space is to ask that morality do so, in the form of a principle that protects her ability to take an interest. Having the choice whether to take such an interest or not is a way in which the promisee shapes her life, and so she has an autonomy interest in being able to take such an interest. Since this is an interest she has, as an actual promisee rather than an abstract viewpoint, it is a consideration to which morality is quite properly responsive.

Abstracting from the case of promising, and summarizing, we can say that S has a generic interest in x if it is reasonable and reasonably foreseeable that S take an interest in x; and conjecture that it will be reasonable and reasonably foreseeable that S take an interest in x in circumstances C if someone in S’s position in circumstances like C would ordinarily have an interest in x. Note that having a generic interest in x is not a matter of having an interest in x, but of having an interest in being able to take an interest in x. A generic interest is really an interest in autonomy; but it is appropriately indexed to x insofar as x is a dimension along which the interest-holder may exercise her autonomy. This structure means that a generic interest can play a role in the explanation of general principles that can overcome the problem of overgeneralization, and that the role is morally motivated rather than ad hoc.

The problem of overgeneralization arises for the contractualist account of promising because the interest it identifies as the grounds of promissory obligation, namely the promisee’s interest in assurance, is not present, or not as strong as the promisor’s interest in non-performance, in all the cases which intuitively require that the promisor keep his promise. That is because the interest in assurance at most grounds a principle like F, which does not cover a case such as Sewing Machine. But assurance can be invoked in the explanation of F^* without making it a complete characterization of the interest protected. The fact that a promisee typically has an interest in assurance, even though she won’t always have such an interest in every case, makes it reasonable for a promisee to expect to be treated as if she had such an interest in every case in which she is a promisee and which is similar enough to the paradigm case. Such treatment gives the promisee space to take an interest in assurance if she wishes; thus requiring that the promisee be treated in this way protects her autonomy. This means that every promisee (in cases similar enough to the paradigm) has an interest in autonomy that is set back when she is not treated as if she has an interest in assurance. In other words, every promisee has a generic interest in assurance. But that fact is enough to ground a principle like F^* that covers every promisee (in cases similar enough to the paradigm), rather than every such promisee who actually has an interest in assurance. And the daughter in Sewing Machine is just such a promisee, for if she had an interest in assurance she would fall under F. The fact that she does not is, to put it crudely, nobody’s business (not even her mother’s) from the standpoint of autonomy, given that promisees typically have such an interest.

Some may be concerned that this account will generalize without limit, seeming to prove that promisors are on the promissory hook even when performance becomes impossible or is superseded by a moral emergency. More generally, there is a question as to how much generalization the idea of a generic interest permits. Another way of putting the question is by asking just how narrowly we are to understand the idea of a representative standpoint, which is the foundation of our judgments about the ordinariness of having an interest. That is not a question I have set out to answer here, because it seems to me a substantive moral question that may be answered differently in different areas of life. Since the idea of a generic interest is so closely tied to that of autonomy, its relative importance and exact contours will depend on how important we think autonomy is in various parts of our lives. In the case of promising, the fact that promises are so fundamental to our ability to make arrangements with others that expand our ability to plan and control the future, means that autonomy plays a significant role. Something similar will probably be said for moral principles demarcating spheres of privacy and ownership, but probably not for fiduciary relations that serve to protect our vulnerability rather than expand our autonomy.

It is worth pointing out that the current proposal does not throw out entirely the insights of the arguments examined in the previous section. A representative standpoint is in some ways like an office, in the sense that it is the abstract notion of the representative standpoint, rather than the circumstances of an individual occupant, that determines the normative status of the individual occupant. But unlike Rawls’s practice conception, the idea of generic interests does away with the contrivance that certain actions can only be regarded as promise-breakings, or trespasses, or privacy invasions, once a social practice has instituted the relevant office.

The idea of a generic interest also has a connection with the idea of guidance, since part of what makes it reasonable for a promisee to expect to be treated in the same way as a typical promisee is the fact that it is reasonably foreseeable that someone in her position might have (or develop) the same interest as the typical promisee. This fact about reasonableness depends to some extent upon the privacy of the promisee’s evaluative attitudes and the relative inaccessibility of their affective and attentional dispositions, and so it rests upon the same data about human boundedness that suggested guidance as an explanation for the generality of principles. But it is not merely a matter of convenience that we treat the promisee according to their generic interest, but a matter of respect for their autonomy. In this way the idea of a generic interest, like the idea of fairness, provides a moral reason for the generality of moral principles.

5 Normative Powers

Generic interests do not mediate in an ad hoc way between the interests present in typical cases and general principles. They represent the interest in autonomy that all who occupy a given standpoint have in virtue of what is typically true of those who occupy the standpoint. As such, a generic interest serves to preserve the authorial control that it is reasonable to expect in virtue of occupying the standpoint.

Understood in this way, the role of generic interests is reminiscent of the thought that bare wrongings are to be explained by attributing to their victims a normative power to determine the wrongness of such actions (Owens 2012; Raz 1972). On Owens’s account, the existence of a normative power on the part of a promisee is explained by a special kind of interest that they have:

[W]e are wronged by bare wrongings because we have an interest in being able to ensure that certain deeds constitute wrongings should we so declare. What grounds these bare wrongings is a normative interest, an interest in being able to control the normative situation by declaration. It underlies our need to be able to accept promises, consent to sex, and likely many other things. … These wrongings are grounded in an interest which need not be affected by the wronging, namely an interest in its counting as a wronging, a normative interest. (Owens 2012, 64–65)

Both generic interests and normative powers appeal to the idea that individuals have an interest in being able to control their interactions with others. What differs is where each account locates control. Owens thinks that it is the agent, e.g. the promisor, who has an interest in being able to control the normative situation, and that this interest is satisfied by the agent’s power to alter normative statuses. It is slightly puzzling why the promisor’s interest in control should underwrite such a power, but not a coextensive power on the part of a promisor to rescind their promise. It is also puzzling how this proposal is to account for the vital role that the promisee plays in the creation of a promissory obligation by accepting or taking up the promise in the right way. In contrast, the generic interests account locates the relevant interest on the part of the promisee, and this is in harmony with the fact that it is the promisee who has the power to demand performance or release the promisor from performing. More generally, the account admits that control is potentially of importance to every standpoint that can be characterized in terms of generic interests.

Another difference is that the normative powers account attributes control to an individual at the level of normative principles, whereas the generic interests account identifies control as having an important role at the level of what grounds those principles. The latter approach leaves open the possibility that an individual may have a significant interest in autonomy that does not underwrite the conferral upon them of a power. That is suggestive of the kind of explanation that might be given for one’s inalienable right against being enslaved, which protects one’s autonomy so zealously that it does not grant the rightholder the power to give it up.

There are two further advantages that generic interests have over normative powers as explanatory device. First, generic interests provide a more parsimonious explanation of principles. Owens insists that normative powers are operative only once social practices are up and running, and this seems to be a requirement in order to explain the bounds of those powers; whereas the existence of a social practice might influence what interests we ordinarily have, and so what generic interests we have, yet it seems plain that the generality of our moral principles need not always hang on the existence of a social practice. That is one way in which normative powers invoke more explanatory apparatus than we should be willing to permit. What should make us even more reluctant is the kind of choice-dependence invoked by the idea of a normative power. It is one thing to think that the moral facts change in response to one’s choice because one’s choice alters the considerations that ground those facts. It is altogether more puzzling, but required by the idea of a normative power, that the normative landscape changes merely in response to one’s choice and not because the content of one’s choice or the fact of exercising it adduces any moral considerations that could be independently articulated.

Second, the normative powers account seems to misunderstand the moral significance of our interest in autonomy. For the explanation of why an individual has a particular normative power is just that she has a normative interest in determining whether someone’s action wrongs her. Owens takes care to say that such an interest need not be grounded in any further non-normative interest. But this can make the normative interest in control seem like a peculiar fetish, or at best an interest in having control for its own sake. Humans may be distinctive in their ability to exercise control over all aspects of our world; but why should we attribute any moral significance to the protection of that capacity? The capacity is at least as capable of destruction as of anything morally worthy. The idea of a generic interest points out that the legitimate interest in control is always keyed to some further interest that one may reasonably take.¹⁵ The morality of promising does not protect a promisee for the bare reason that the promisee therefore exercises control over the promisor. Rather, what is at stake is the promisee’s legitimate interest in having the ability to take an interest in assurance, where assurance is the kind of thing that could make her life go better. In this way, the idea of a generic interest acknowledges that autonomy has moral significance because it enables us to pursue projects that potentially make our lives better.