test

Introduction

Almost everybody believes that we are permitted to be partial to our loved ones: that morality allows us to favour our near and dear rather than treating everybody impartially. Sometimes the explanation seems to be that we have special duties: people often feel not only that they are allowed to treat partially those with whom they share close relationships, but that they owe them particular things that they don’t owe to others. There is, of course, plenty of disagreement about what exactly is owed. Are you obliged to try to give your children the best possible start in life, or to do what you can to make your parents’ final years as comfortable as they can be? Are you required to save your drowning wife rather than many strangers? Do you have to spend time with a sad friend, rather than with an even sadder stranger? But there is also a debate at a more fundamental level, about where these duties come from. Here we propose a view that not only explains when we have duties to the people we love, and for what reasons, but also makes sense of why partiality is valuable even when there are no duties in play. We assume that some partiality is permissible in such cases, but take no stand on why or when.

Our suggestion is that all duties towards near and dear can be explained by appealing to general moral principles that do not themselves make any reference to the loving character of the relationship. Once we have in mind the full range of principles from which those duties might be derived, there is no need to posit “sui generis” duties. We agree that people sometimes feel compelled to do things for those they love even when there is no such principle that explains why. But in those cases what they experience is what we will call a “reason of loving attachment” – a reason arising from the value of the relationship – that can register subjectively as no less weighty than a duty.

Most of us need close relationships for a minimally flourishing life - human beings are creatures of attachment (Bowlby 1969-1980) – and our loving attachments always provide some reasons for action: namely, value-based considerations. They do not indicate what morality requires of us but rather concern what is good for us, and for others.[1] As we will explain, they are partly reasons to benefit those we love and partly prudential reasons, i.e. reasons to benefit ourselves, so that when we rescue or visit a close friend or spouse we can be doing something that is good for us both. Regarding people as permitted to act on reasons of loving attachment, rather than being duty-bound to do so, does nothing to trivialize those reasons, which are central to their moral agency and sense of who they are.[2] This centrality may also explain why people are sometimes permitted to be partial towards their loved ones even when they have no duty to favour them.

A moral theory that explains some of the phenomenology of loving relationships as reasons of attachment instead of duties has several attractions: it makes sense of the way in which partiality is commonly experienced, it accommodates a plausible view about the value of close relationships, and it helps to explain the importance of partiality even in the absence of duties. Our answer to our question, then, is that we have duties towards those we love when, only when, and because, those duties can be explained by a general principle that makes no reference to the loving character of the relationship. But we and our beloved can be better off, and partiality towards our loved ones can be more valuable, when there are no duties in play.

The paper proceeds as follows. Part 1 sketches the philosophical debate between views like ours, i.e. ‘reductionist’ accounts of relationship duties, and their ‘non-reductionist’ rivals, and identifies three objections to reductionism. Part 2 responds to two of those objections, focusing on duties to children and showing how principles of procreative liability and parental fiduciarism, neither of which appeal to the loving character of the relationship, can ground many such duties. Part 3 discusses two cases – Friend with Alzheimer’s and Saving Mary – to show how reasons of attachment can make sense of our intuitions about what we ‘must’ do without positing sui generis duties (and, in passing, responds to the remaining objection). In Part 4 we elaborate on two attractive features of our view that what others see as relationship duties are sometimes prudential reasons of attachment. First, the goods of loving relationships are more fully realized when the benefactor is partly motivated by prudential considerations – when it is good for the saver to save or the visitor to visit. Second, the less people are constrained by duties to others, the more they enjoy moral autonomy.

1. Relationship Duties: Reducible or Sui Generis?

Some duties are general: we owe them to all individuals in virtue of their moral status. Others are special: we owe them to particular individuals in virtue of our connection to them – for example, because we have promised them something, or because they have done something for us and perhaps we are bound by reciprocity or gratitude. Some special duties arise in the context of close relationships; we owe them to our near and dear, those we love and care about, such as our friends, lovers, spouses, parents, and children (Scheffler, 2001; Raz, 2011, 2020; Wallace, 2012; Owens, 2012; Keller, 2013; Seglow, 2013; Jeske, 2014; Bazargan-Forward, 2019).[3] Philosophers disagree about how we should understand these “relationship duties”, where they come from or what grounds them.  Some think that they are to be explained in the same way as other special duties, others insist that there is something distinctive about them: that they are sui generis.

On the first view, which we defend, relationship duties are like other kinds of special duties in that they can be explained by appeal to general moral principles that don’t themselves make any reference to close relationships - principles like those that say we have duties to those to whom we have promised something, or who did something for us, or who are particularly vulnerable to us.  Special duties derive from the combination of a moral principle (‘you have a duty to keep your promises’) and a fact about your connection to another person (‘you promised to lend Amy your car’). In so far as near and dear are bound to each other by duty, the duties are not sui generis but, like other special duties, reducible to general principles that make no reference to close relationships: this is a reductionist account.

The second view holds that some relationship duties are irreducible or sui generis – we shall use these terms interchangeably. Sui generis duties cannot be explained as particular applications of moral principles that make no reference to close relationships. Instead, at least some are said to arise from the relationship itself (Kolodny, 2010), or the unique goods generated by the relationship (Seglow, 2013), or the person with whom one stands in the relationship (Keller, 2013) [4]. This is the non-reductionist view, according to which the duties are not only “special” but also sui generis. They are owed to some rather than to all and, because distinctive principles apply to close relationships, the closeness of the relationship is an essential part of their explanation.

The reductionist view is charged with three important shortcomings. First, some find it difficult to see how duties grounded in principles that make no reference to close relationships can explain the fact that we owe our near and dear loving treatment. For them, only sui generis considerations can yield that implication. One aspect of this is that some of the considerations that can justify special duties, such as promises or contracts generate discrete duties, unlike many relationships duties which are open-ended (Hurka 2016). Second, some putative relationship duties are too stringent – their demands are too weighty or forceful - to be plausibly explained by appeal to the usual grounds for reduction such as promises, vulnerability or gratitude (Bazargan-Forward, 2019). How, for instance, can such considerations explain the intuition that, in rescue cases where one cannot save everybody, one has a duty to rescue one’s child rather than use a fair procedure to decide whom to rescue? Third, many find reductionism implausible because it fails to account for the way people think and speak (Stocker 1976). For instance, we typically explain why we ought to help a friend by saying simply “because she is my friend”, rather than appealing to a relationship-independent principle. Some philosophers make sense of this phenomenon by appealing to the thought that relationship duties are the appropriate response to valuable relationships or individuals (Keller 2013, Seglow 2013, pp.13-14).

Section 2 shows that appeal to vulnerability can explain many weighty relationship duties, and then considers cases involving children, where the first two objections are particularly relevant. Debates about relationship duties have overlooked principles of liability and fiduciarism that, while not themselves invoking close relationships, can ground duties to treat children lovingly and explain their weight and open-endedness. Invoking them expands the range of reducible duties by pointing to additional ways in which duties to our near and dear may be reduced. Section 3 suggests that some of the phenomena we experience as duties to our loved ones, and whose apparent irreducibility leads some philosophers to regard them as sui generis, are not actually duties but reasons of loving attachment. This view is in line with the phenomenology of close relationships. If we are right, some alleged counterexamples to reductionism fail because the phenomena in question are neither reducible nor sui generis duties. They are not duties at all.

2. Duties to our near and dear

Many of the special duties that we owe to our near and dear are easily explained. Our relationships with them typically involve more interactions, and more types of interaction, than with other people, and there are correspondingly many ways in which familiar moral principles governing those interactions may give rise to duties. In cases like these, the closeness of the relationship on its own does no normative work and the fact that we owe special duties arises only when and because our relationships with the people we love involve these other kinds of interaction.[5] The challenge to a view like ours is that some relationship duties seem not to be reducible in this way. In some cases, the grounds that are usually invoked to explain those duties - promises, legitimate expectations, reciprocity, gratitude, reliance, trust, and vulnerability – seem unable to account for their existence, or stringency, of open-endedness.

In this section we make two moves that between them go a long way towards meeting the challenge. The first concerns the implications of the principle that we have special duties to those who are vulnerable to us. Since those to whom we are close are vulnerable to us in special ways, we often owe them special duties. The principle grounding the duty makes no reference to the closeness of the relationship. Rather, the closeness of the relationship creates the context in which certain vulnerabilities to another person emerge, and in this sense is part of the explanation of the special duty. The second shows that two further principles explain some of the special duties, namely the ones arising in relationships where some are liable to, or fiduciaries for, others. The case for regarding some relationship duties as sui generis derives much of its intuitive appeal from cases involving children. Recognising that procreators are liable to their procreatees, and that parents are fiduciaries for their children, allows us to explain many of those cases without invoking a principle that appeals to the closeness of the relationship.

Vulnerability

It is widely acknowledged that we owe special duties to those who are vulnerable to us, where vulnerability is understood as dependence on a particular agent for the provision of goods important to one’s wellbeing (Goodin 1985). Since close relationships involve a great deal of vulnerability, this explains many relationship duties. There are three ways in which close relationships trigger a vulnerability principle. Our near and dear often need company and emotional support from us rather than from just anybody: their attachment to us means that others cannot act as substitutes. Moreover, we usually know a lot more about them, and so are better situated to help (Jackson 1991). Close relationships are neither necessary nor sufficient for superior knowledge but they greatly facilitate it.[6] But even if a stranger were equally well-situated, and leaving our non-substitutability aside, the existing relationships can bear on the meaning of our behaviour: for instance, the indifference revealed by our leaving it to someone else to help may be hurtful. And our hostility, or indiscretion, can be particularly distressing, especially if they involve information that we have acquired through our closeness, and so involve betrayal. The discharging of duties grounded in vulnerability often is what we call loving treatment, namely treatment that is particularly attentive, affectionate and generous. Moreover, the duties in question tend to be open-ended, since relationships unfold over time in unpredictable ways, as do people’s vulnerabilities to each other. Vulnerability can also explain why relationship duties are so stringent: when others depend on you for their emotional wellbeing, your duties to them are not easily overridden.

If so, the thought that we owe loving treatment to our near and dear, which might seem to support a case for sui generis duties, actually fits the reductionist picture. We may owe special duties to others in virtue of having a close relationship with them, but the principle grounding those duties makes no reference to the closeness of the relationship. The duties derive from the general principle that we have special duties to those who are vulnerable to us, combined with the claim that those with whom we have close relationships are indeed vulnerable to us.

Of course, general vulnerability and the existence of close relationships can come apart: a person can be highly vulnerable to another in the absence of a special relationship, for instance when the second has a great deal of knowledge about the first person – say, about sensitive medical information. Here, too, vulnerability can ground duties, but, absent emotional vulnerability there is no duty to provide loving treatment. Conversely, it is possible to be in a loving relationship with someone who is not particularly emotionally vulnerable, for instance, with a friend who is unusually emotionally resilient, and whose feelings are difficult to hurt. In such cases, as we explain in section three, we have reasons of attachment to give priority to this friend over strangers, but fewer duties of (emotional) vulnerability – and hence it may be more often permissible, or even required, to give priority to others over this friend.

The fact that many duties might flow from a principle of vulnerability could plausibly explain the possibility that we are under a duty to maintain emotionally close relationships, once they have been established. Bazargan-Forward (2019) believes that we have a duty, owed to close friends and some family members, to adopt and enact an identity which itself requires us to give them special weight in our practical deliberation. For example, one may have a duty to think of oneself as somebody’s friend, which involves caring about them more than about a stranger, and to act on that basis, for example by spending time with them and being responsive to their needs. On this view, discharging the duty requires living up to what non-reductionists usually take to be sui generis duties. This account is promising but incomplete: it offers no justification for the duty to adopt and enact the relevant identity. Bazargan-Forward observes that the usual grounds invoked by reductionists to explain the derivative nature of relationship duties – vulnerability, gratitude, reciprocity, legitimate expectations and promises – are implausible explanations of the priority that many feel that they owe to their near and dear. They don’t seem weighty enough to defeat duties of impartial justice. But if such considerations are too weak to explain a duty to, say, rescue one’s child in rescue cases, how can they be strong enough to generate a much more demanding duty; namely to adopt, and then to enact, a particular practical identity which itself requires one to rescue one’s child?

An appeal to vulnerability might supply the missing part the above account, just in case becoming vulnerable to someone for the provision of a particular close relationship can ground a duty on that person to continue to provide it. We are agnostic about the possibility that an adult can owe another adult a close relationship – which is what Bazargan-Forward’s view, in effect, says. But when it comes to relationships between parents and children, the view that one can have a duty to maintain - and indeed to establish - a close relationship is very plausible. Parents owe their children a range of loving attitudes and behaviours and procreators are duty-bound to ensure that their procreatees have parents. This duty to provide a close relationship can easily be explained by principles that are rarely invoked in the discussion of relationship duties, namely liability and fiduciarism.

(Procreative) Liability

It is clear that relationships of liability involve special duties that can be explained in a reductionist manner. For instance, negligently injuring another creates special duties to compensate them. Here we are particularly concerned with procreative liability, where we take procreators to include at least the people who (intentionally) engage in genetic procreation, but possibly more. Almost everybody agrees that procreators owe their procreatees lives worth living, at least when they can provide that. That very modest standard will not account for all the examples of parental duties in the literature on relationship duties, but more plausible views set the bar higher. Many hold that procreators owe their procreatees the necessary conditions for reasonably flourishing lives (Kant, 1996, 64); Mill, 1989, 108; O’Neill, 1979; Archard, 2010), while some identify a standard that is in some respects more demanding (Shiffrin 1999).

Plausible interpretations of both the modest and the demanding view hold that at least one of the things owed by procreators is the willingness to provide the kind of caring, stable relationship that the child needs to enjoy reasonable flourishing (Olsaretti, 2017). If so, this makes sense of many of the duties that we typically think parents owe their children, such as treating them in a protective, loving, caring, and nurturing way. A loving parent-child relationship, on this approach, is something a procreator must be willing to provide to their procreatee, at least in cases where nobody would be able and willing to provide an even better relationship (Gheaus 2021). That relationship is not itself the ground of duties owed by parents to their children. To see this, consider an analogous, but hypothetical, case of liability incurred because of bringing a person into existence: assume the procreatee is a non-human person – an Artificial Superintelligence perhaps – whose wellbeing does not require relationships. In this case, the same principle applied to a different empirical context might yield a duty to ensure its wellbeing, but not the duty to provide it with a close relationship. The fact that human procreators owe their procreatees a loving relationship can also explain the content of duties in tension with impartial moral demands – such as spending time with one’s child rather than with people in greater need of company.[7]

The debate on relationship duties has so far neglected procreative liability as a ground for reducible duties – although of course it is often appealed to in debates on permissible procreation and procreative duties. This neglect may be a symptom of a more general tendency to fail to distinguish procreation from parenthood; in this case, normative demands are pinned on the parental role that properly attach rather to the fact of having brought the child into existence. Parenting is distinct from procreating and need not involve any biological connection between parent and child. (This raises interesting questions about special duties to children in cases where the procreator is not also the parent. It is possible, for example, that adoptive parents owe less to their children than procreative parents but, even so, the transfer of procreative liabilities to adoptive parents is likely to leave intact, as reducible duties, much of the content of relationship duties regarded as sui generis by non-reductionists.) Another possible explanation for the neglect of procreative liability as a ground for the special duties of parents is that there are no other relationships in which someone brings another person into existence and therefore acquires liabilities towards them, although such situations are not unimaginable. Our Artificial Superintelligence is one example, Frankenstein’s monster is another.

The fact that a close relationship is the content – rather than the ground – of some special duties should be clear in the case of procreative liability. When, for example, we think about a procreator’s liability to pay for the upbringing of their own procreatee rather than another person’s, we are not likely to be misled into taking closeness in the relevant sense to be an essential feature of the relationship between procreator and procreatee. They are, obviously, biologically close, but that is not the kind of emotional or psychological closeness that is relevant here. Often, indeed, insisting on – and trying to enforce - a procreator’s financial liability is most important in cases where there is no close relationship between the two. The suggestion here is that procreators owe their children not only financial support but also the kind of relationship they need for their healthy emotional and cognitive development. In that case, the close – and loving – relationship is the content of the duty but, like the duty to pay for the procreatee’s upbringing, it derives from a principle that makes no reference to procreator and procreatee having a close relationship.

(Parental) Fiduciarism

The right to parent is conditional on the parent discharging certain duties. Children are vulnerable and lack the capacity to protect and further their own interests. The parent’s role is thus to act as a fiduciary: to exercise authority over the child in the child’s interests. Children lack the power to give consent to, or to exit from, the parent-child relationship. Unlike adults, who sometimes enter fiduciary relationships with doctors, lawyers, or others to whom they give power of attorney, they cannot exercise their own judgment about whom, if anybody, to entrust with their current and future wellbeing. Therefore, on a range of views, the allocation of parental rights must be justified, at least in part, by appeal to the benefits that children will enjoy in virtue of that allocation. Parents have duties of care to their children in virtue of occupying the parental role (Brighouse and Swift, 2014).

How demanding are these duties? Commonsense morality and the law set the standard of care at the modest level of avoiding abuse and neglect, but many philosophers believe that the moral right to parent is in fact conditional on parents’ meeting a higher standard, which they call “adequate” or “good enough” parenting (Brighouse and Swift, 2014). Still, even the modest view usually regards parents as duty-bound to provide their children with the kind of relationship that is needed for their healthy development. On our analysis, again, a loving relationship is not the ground of parental duties but concerns the content of what parents owe their children. The parental duties that non-reductionists typically appeal to are not “sui generis”; they are indeed owed to their children, but they are a kind of fiduciary duty that derive ultimately from justifications of the right to parent.

As with the principle of procreative liability, the debate about relationship duties has so far neglected parental fiduciarism as a ground from which those duties can be derived – although such accounts are often discussed in the literature on family ethics. Parental fiduciarism can explain stringent duties to rescue, and more generally to favour one’s child in various – and open-ended - ways, where appeal to principles such as vulnerability or promises would be unsuccessful. The thought here is that, qua fiduciaries, parents owe their children specific benefits which they don’t owe others, because the parental role itself is justified, at least in part, by the way that it entrusts the fate of specific children to specific adults (Brighouse and Swift 2014). Parents qua fiduciaries owe their children whatever special treatment is needed to maintain the kind of parent-child relationship they have a duty to provide.

The case of parental fiduciarism can be more confusing than that of procreative liability because it is easy to build at least some degree of closeness into one’s understanding of what it means to parent a child, or to confound being a parent with being a good parent. For us, to be a child’s parent is simply to exercise some degree of authority over them; to exercise those rights that constitute the role of parent-as-custodian. This is, of course, a relationship. But it is not one that is necessarily close in the relevant sense – however much one may wish that it were. The claim is that those who exercise parental authority should serve as their children’s fiduciaries: it is from a general principle concerning the fiduciary character of parental authority, grounded ultimately in justifications of the right to parent, that parents’ duties may be derived. 

Given children’s developmental needs, it is plausible that close, and, indeed, loving, relationships are among the things that children are owed by those who are related to them by principles of (procreative) liability or (parental) fiduciarism. If so, even duties to one’s children, where the intuitive case for sui generis duties seems especially strong, can be explained without invoking the closeness of the relationship as their ultimate ground. That these duties are owed by procreators and parents independently of whether they love their children brings out starkly the fact that the duties do not derive from the loving character of the relationship.

By adding these two to the list of general duty-grounding principles – and explaining the wide application of the principle of vulnerability that is already on it – we can account for the vast majority of relationship duties, including many that are often claimed to be sui generis. Our analysis does not seek to show that all putative relationship duties are reducible but, as we now argue, a theory that explains away some such duties has its attractions.[8]

3. Beyond duty: reasons of loving attachment

We have claimed that many intuitively compelling cases of relationship duties can be explained by principles that make no reference to close relationships. But some cases seem to resist this kind of analysis. Here we discuss two examples from the philosophical literature that have been taken to support the view that some duties are sui generis. Friend with Alzheimer’s was devised precisely as a case in which no general principle could serve as the ground of what looks like a relationship duty. Saving Mary has a different target but is sometimes read as showing that relationship duties are not reducible to general principles but rather are generated by close relationships themselves, or the people in them. We use these examples to introduce our suggestion, developed in section 4, that sometimes what look or feel like sui generis duties are not duties at all, but reasons of loving attachment. With reasons of loving attachment in the picture, reductionism can explain why people feel they owe things to their loved ones, and can identify powerful reasons for them to act on that feeling, even in cases where there are no duties in play.

The case we call Friend with Alzheimer’s is formulated by Jay Wallace (2012, p.186):

“Consider our relationships to friends and life partners of many years, where the person we love has become sick and infirm, and in need of constant care and attention (perhaps in the way of an Alzheimer’s patient). Suppose that the beloved in this situation has made it clear to us long ago that they do not expect us to be there for them should things become especially grim in these ways. There are financial resources that have been set aside, and institutional arrangements that have been made; so it seems that there is no moral obligation to attend to the beloved on grounds of vulnerability. And we can stipulate, further, that we have already done at least as much for our friend or partner as they have done for us over the course of our active lives together, so that there are no outstanding debts of gratitude or reciprocity. We might nevertheless feel under these circumstances that we owe it to our beloved to be there for them in this phase of our lives, and that it would somehow be disloyal of us to turn our backs on them in the present situation. This would seem to be a sui generis duty of love, which cannot plausibly be reduced to moral principles of the kind that potentially apply to other cases, since by hypothesis we would not be in violation of such principles if we were to fail to live up to the duty that is in question.” 

To preserve the spirit of Wallace’s example, we assume that the friend with Alzheimer’s is no longer capable of enjoying any of the goods of friendship. They are not aware, at any level, of who is visiting them. If they were, a duty to continue the relationship might be explained by their vulnerability to their loved one for the provision of those goods. Cases like Friend with Alzheimer’s, where one feels that one must do something for one’s near or dear even though all considerations from which that feeling might be derived have been stipulated away, offer the best evidence for non-reductionism.

Our view about this kind of case is that there is indeed no duty to visit. It is permissible not to do what one feels one must.  But one’s reasons to visit – reasons of loving attachment – are very weighty. Not to visit is to miss out on something of great value. One’s friend with Alzheimer’s would not be wronged, since there would be no breach of duty, but to abandon them would be to deprive oneself of the important good of acting on one’s loving attachment. It would also deprive the friend of the benefit of a visit, albeit one that might as well be a visit from a stranger. As we elaborate in section 4, these considerations explain why we experience those reasons as compelling in a way that leads us to mistake them for sui generis duties. It matters to us so much that we can visit the people we love, and we care so much about benefitting them with a visit, that we feel that we have no choice. These considerations also vindicate our intuition that visiting would qualify as permissible partiality.

Saving Mary is our name for Bernard Williams’ famous case of a man’s choice between saving a stranger and his wife.[9] Williams not only thinks it obvious that he should rescue his wife but also that something would have gone wrong if he were to justify that choice by appeal to a moral principle; to do that would be to have “one thought too many”. Some take this to show that he owes a duty to his wife, and that the duty is generated by the relationship itself, or perhaps by the beloved wife herself (Keller, 2013). By contrast, we suggest that the only duties in play here are the duty to rescue someone plus any special duties that are reducible to, e.g. vulnerability, promises, reliance, legitimate expectations, gratitude or reciprocity. For instance, Mary might be vulnerable not only to drowning but also to a sense of feeling unloved should he fail to try and save her, and the latter might be a significant additional harm. Alternatively, implicit understandings and conventions are particularly important sources of special duties; the history of interactions between Mary and her husband, perhaps in combination with conventional understandings of what spouses owe one another, may well have created legitimate expectations and trust about what each will do for the other. This, in combination with general normative principles, might supply a reductionist explanation of the husband’s duty to save his wife.[10]

But, whatever general moral principle, or principles, might explain the duty to rescue her, and whether or not those duties are stringent enough to outweigh impartial reasons, Williams’ husband also has a weighty reason to rescue Mary that cannot be explained in this way. This is a reason of loving attachment. In cases like this, one of the reasons to rescue his spouse is to act on his desire to protect her. Her well-being, and not that of the stranger, matters intrinsically for his own well-being. It matters, that is, in ways that have nothing to do with any instrumental benefits that might accrue to him from her survival. In this sense, he has an intrinsic prudential reason to save her. Loving relationships often generate weighty reasons of this kind, and such reasons may strike us with the force of duties. We assume, moreover, that people have some personal prerogatives to advance their own good and that these imply a permission both to form close relationships and to act on at least some of the reasons they generate.[11]

We are now in a position to respond to the third objection to reductionism that we identified in section 1, that the phenomenology of close relationships seems to be shaped by non-reductionist convictions. We feel, and talk, as if we must do certain things for particular people, and we explain that feeling and talking by saying not “because she is vulnerable to me”, or “I have a debt of gratitude”, but simply “because she is my friend”.[12] It is not difficult to see why we might refer only to the relationship in cases where there is a relationship duty. That duty arises from a combination of a general moral principle and facts about the relationship in question. Pointing to, or thinking of, the latter as a reason for action can be a useful shorthand and is not strictly speaking incorrect, even though it is incomplete.[13] In cases of loving attachment, it is even more natural that we would think about, and point to, our near and dear themselves, or perhaps to our relationship with them, as our reason for action. We do not think, or say, “because my wellbeing depends on it”, but “because she is my wife”, or “because she is Mary” or, perhaps, simply “because I love her”.

4. The appeal of reductionism: two thoughts about the value of love and attachment

We accommodate the intuition that something goes wrong when one fails to visit in cases like Friend with Alzheimer’s – even though what goes wrong does not amount to a wronging. Not to visit is to decline to act on a permission to realise important goods for oneself, partly by benefitting the friend. Here we elaborate on our view, by discussing two value considerations at stake. First, one kind of value – in different parlance, a special good – can only be realised in close relationships: the good of loving attachment.[14]  By its very nature, that good is fully realised, for both parties, only if the motivation behind the visit is of the intrinsic kind mentioned above, and if some of it is prudential; i.e. for the benefit of the visitor herself. Friendship is more valuable if a person seeks her friend’s company because she feels it is good for her, and her friend knows this.[15] Second, and independently, the visitor enjoys a kind of autonomy when she visits without being duty-bound to do so. In elaborating them, we hope both to present the attractions of our view and to show how they explain some moral phenomenology better than does invoking sui generis duties.

On the first consideration: We distinguish between attachment and love. The central element of love is disinterested: a concern with the beloved’s wellbeing for their own sake. However, the full sense of love – at least the kind that connects human beings – typically also includes an element of attachment, which is selfish.[16] To be attached to someone means that one’s relationship with them, and implicitly the person themselves, are necessary for one’s emotional wellbeing. This gives us a prudential interest in preserving them and our relationships with them. There could be cases of disinterested, i.e. unattached, love (perhaps accomplished Buddhists are capable of it), but they are highly unusual. Moreover, in one sense, unattached love would also be less valuable than attached love. This is because it is also good for a person to be the object of another person’s loving attachment, since this is a powerful way of having one’s value affirmed.

Saving Mary can nicely illustrate the difference between attachment and love. Commenting on it, Derek Parfit disputed the content of the “one thought too many” (Murphy, 2000, 140). Might Mary not be disappointed to find out that her rescuer’s reason for saving her was that she was his wife? That itself might count as one thought too many. Surely the loving motivation for saving her should be that she was Mary. One interpretation of Parfit’s point is that, in thinking of her as his wife, the rescuer is valuing his marriage – his relationship with, or his attachment, to Mary – rather than the woman herself.[17] If so, then in our terms this would be an attachment, but not a loving one. The rescuer cares about Mary’s safety only in so far as this is necessary for the preservation of their relationship. By contrast, if the rescuer loves his wife he will be motivated directly by a concern for her safety.

To be sure, and importantly for our view, rescuing Mary for love still involves some intrinsic prudential motivation. To love someone means that your good is not independent of theirs; rather, their wellbeing is constitutive – at least in some respects – of your own life going well. For example, the mere fact that they are safe makes your life go better for you. Saving them benefits you as well, even though, and partly because, it is motivated by their wellbeing. This is how we understand the distinction between attachments in general and loving attachments in particular. Where one is merely attached, the other’s wellbeing has only instrumental prudential significance – there is no special concern for the other’s wellbeing as such. One’s loving attachments are distinguished precisely by the way in which the other’s wellbeing is nested within one’s own. This is not the view that love is entirely selfish. Reasons to favour one’s loved ones are not exhausted by the desire to benefit one’s beloved just insofar as their well-being is part and parcel of one’s own. On the contrary, loving someone necessarily involves being motivated by their wellbeing independently of any benefit to oneself. But in a loving attachment part of one’s motivation to benefit is prudential; the rescuer also benefits in one way (even if not all things considered, so it is possible for us to sacrifice ourselves for the people we love).

Reductionism can make sense of the moral phenomenology of close relationships, and in particular of the strength of people’s reasons to act partially in favour of those they love. We may well feel that we are breaching a duty when we fail to benefit others to whom we are close. Yet, on our analysis, what goes wrong in Friend with Alzheimer’s is rather a failure to act on reasons of loving attachment. The values at stake are weighty enough to make it a mistake not to act on those reasons. Not visiting is objectionable because it fails to respond to the value of our attachment, not because it violates a claim.

A little speculation about our tendency to experience some reasons of loving attachment as duties helps to explain why the full value of loving relationships is realised only when the motivation to do things to, with or for one another is of the right kind. Sometimes our attachment to the people we love makes us feel that we must do certain things for them. In this respect, reasons of loving attachment, unlike other kinds of reason on which we are free to act or not to act, are perceived as stringent demands. But this is not mysterious. On its own, our desire to do something for the people we love, and more specifically our desire not to deprive them of the goods they would enjoy if we benefitted them, can be so intense that not acting on it may feel like a betrayal. We can feel that we are not only disappointing them but letting them down – and in certain situations this would be the case, for instance when vulnerability grounds special duties. That does not apply in Friend with Alzheimer’s, who, by assumption, is no longer vulnerable in the way that friends typically are to each other; they would not, for instance, interpret a failure to visit as neglect or indifference. But even in this case, a visit can be beneficial, just as a visit from a stranger would be. The powerful disposition to benefit one’s beloved explains why, in such a case, failing to visit can trigger feelings of guilt even though such feelings are not apt, since staying away is not in fact blameworthy. At the same time, our own wellbeing depends heavily on the beloved’s presence in our life, and on how well their life goes. It depends, moreover, on the fact that at least some of the beloved’s wellbeing is the result of what we do for them. So, our own lives would be diminished by a failure to act on reasons of loving attachment. At least sometimes, when our lives would be greatly diminished – as in the case of failing to rescue a friend – the prudential reasons are correspondingly weighty. Together, these considerations constitute a debunking explanation of why reasons of loving attachment appear to us as duties, making full sense of the “musts” of loving relationships: Duties owed to others are very weighty reasons which are other-directed. We have weighty reasons of attachment to benefit loved ones. Since they are reasons to benefit, they are other directed. Phenomenologically, they feel like duties because they are weighty and other-directed, but this is a confusion, since the weightiness comes from the self-directed, prudential, aspect.

The same considerations also explain the intuition that it is especially good for a person when her near and dear benefit her because they have an intrinsic desire to do so. It is worse for her if they are doing it because they think they have a duty to, since this would introduce an extrinsic motivation. The intrinsic desire to benefit the other is constitutive of the attachment; it also indicates the presence of the prudential value that we derive from being the object of another person’s loving attachment. It matters not only that our loved ones want to benefit us, but also that we know that they want to. The suspicion that another is doing something for us because they feel duty-bound to do so itself diminishes the value of the relationship.

Further reflection on the phenomenology of loving relationships offers some evidence for this claim (Gheaus 2017). If asked why we do so much for someone we love – rather than for ourselves, or for someone else who may need us more – we often say: “because she means so much to me” and even, sometimes, “because I feel cannot do otherwise”. It would be hard for us not to do those things, and we would perceive it as harmful to ourselves not to do them. The deep motivation can be affection: “I don't want your life to go badly and it makes me unhappy to know that it goes badly”; or inclination: “I take a lot of pleasure in helping you”; or fear: “If I don't help you I feel disconnected from you and you are too precious to me to risk losing you, so this feels threatening”. Conversely, we perceive our relationships with others as fully displaying love – rather than, say, some kind of universal benevolence – precisely when we believe that our beloved’s behaviour towards us is in part moved by such prudential reasons. Similarly, when someone we love dies, we can feel sorry for her (assuming death was bad for her), but we also feel sorry for ourselves. In such cases, we think “I suffer because I lost that person” at least as much as “I suffer because I am sorry she missed out on this or that by dying.” All this is easily explained if the full value of loving attachments is realized when benefitting a beloved is good for both parties, and both of them know that.

If we all have weighty interests in the goods of loving relationships, it is desirable that people, and their relationships, provide some prudential reasons for others. It is good for an individual that another would suffer if her life goes badly, that another takes genuine pleasure in helping her, and that another’s wellbeing depends on not losing her. This is in part instrumentally good, because it secures the relationship. But being the object of another person’s loving attachment is also valuable as a way of having one’s value recognised, or perhaps of having value bestowed on oneself.[18] Think of someone who is valued as a human being, respected by all, but loved by none. Suppose that when she died those who knew her would be sorry for her, but nobody would be sorry for themselves because they had lost that person. Her life probably went a lot worse than it would have done if some people loved her and so had prudential reasons to regret losing her.

This analysis is meant to provide a more attractive account than our non-reductionist rivals’, but it cannot constitute a defeating argument against them. It doesn’t aim to, and could not, show that there are no sui generis relationship duties. Suppose that there are in fact some duties of that kind. In that case, we show that there is an important value not being realized if people are motivated by – or are even believed to be motivated by - those duties rather than by reasons of loving attachment. This point, indeed, applies to all relationship duties, whether derived or sui generis. Consider parental duties to treat one’s child lovingly: should a child be loved, or believe that they are loved, out of duty, this would significantly detract from the value of that loving relationship.

The second value consideration supporting our view – the connection between close relationships and autonomy – can be set out more briefly. To the extent that duty regulates loving relationships, participants in them lack a valuable kind of freedom. Here we appeal to Seana Shiffrin’s (1991) account of “moral autonomy”, understood as the possibility of exercising one’s autonomy in a morally acceptable fashion. This sort of autonomy depends on people having a range of morally permissible options available to them. If so, friends and family lack moral autonomy when they are duty-bound to act in a particular way.

It is, of course, plausible that loving relationships diminish this type of autonomy. We often, and correctly, say of people that they are not as free as they were before they had a family, for instance (Alstott, 2006). But in these cases, the duties can easily be understood as reducible. Note also that friendships and families in which members release each other from such duties are often regarded as particularly loving. It is usually a sign of genuine love when people liberate one another from promises, expectations, and any duties rooted in, e.g., gratitude, reciprocity or vulnerability. It would, then, be unfortunate if relationships themselves created sui generis duties. To endorse them would be to aggravate the extent to which loving attachments are, in one way, regrettable. Love would lead not only to reducible duties but also to sui generis duties that would restrict people’s autonomy still further. Regarding as reasons of loving attachment what others take to be duties, and taking ourselves to be sometimes permitted to act on those reasons, makes more room for moral autonomy.

To be sure, some philosophers regard duties – both derivative and sui generis – as integral to the value of relationships. For them, it is better that near and dear owe one another duties than that they are free of such requirements (e.g. Raz, 1989). David Owens (2012, 2021), for instance, believes that we have a normative interest in being duty-bound to those with whom we are in close relationships. He doesn’t provide an explanation of this interest beyond pointing to the importance of being bound to others. Our view accounts for this interest while denying that the bonds in question are those of duty.

Finally, it should be clear how the two considerations bear on each other.  If I have no duty to visit my friend, and both of us know this, then my decision to visit indicates how much they mean to me. Indeed, the mere suspicion that others are motivated by duty in loving relationships can undermine the relationships (Stocker 1976). The full value of relationships is realized precisely when people know that their loved ones have, and are partly motivated by, prudential reasons; that doing something for us is good for them too. If so, the widespread belief in sui generis relationship duties is not only mistaken, but corrosive of close relationships. It invites the worry that people might be doing things for one another for the wrong reason, and thus not recognizing each other in the way that they would hope to be recognized.

Conclusion

Our answer to our question is that we owe duties to those we love when, only when, and because, they can be explained by a general principle that makes no reference to the loving character of the relationship. We thereby disagree with those philosophers who argue that some of the duties generated by close relationships are sui generis.

Our argument for that answer came in two stages. In stage one we showed that some of the duties that have been taken to support non-reductionism are in fact reducible, once we have in view the full range of principles that might explain them. When others are vulnerable to us, or when we are connected to them via principles of procreative liability or parental fiduciarism, the principles that account for our duties make no reference to the loving character of the relationship. This move does not account for all the phenomena we experience as duties to our loved ones; there are cases that look or feel like relationship duties but cannot be reduced in this way. Stage two argues that in such cases the way things look or feel is deceptive. We are bound not by duty at all but by reasons of loving attachment. We supported this avowedly revisionist account of the ethics and morality of close relationships by explaining how it provides a more attractive view of the role that such relationships play in our lives.

Our overall argument is an appeal to the best explanation: one need not think of close relationships as imposing sui generis duties in order to explain their phenomenology. The resulting view has several virtues. By showing how to account for all genuine relationship duties as reducible, and how reasons of loving attachment explain some of what we experience as the “oughts” of loving relationships, it accommodates much of the phenomenology of loving relationships in a more parsiminious way than alternative views. And in showing how acting on duties is inimical both to some goods of loving attachment and to moral autonomy, it vindicates widespread intuitions about the disvalue of people being motivated by duties in this domain.

Acknowledgements

Over the eight years during which we have been working on this paper we have accumulated numerous debts of gratitude. We thank Matthew Clayton, Rebecca Brown, Harry Brighouse, Roger Crisp, Tom Douglas, Serena Olsaretti, David Owens, Christian Seidel, Adam Slavny, Kartik Upadhyaya, Andrew Williams, and two anonymous reviewers for this journal.

Funding

Anca Gheaus' research was funded in whole or in part by the Austrian Science Fund (FWF) [COE 03].

References

Alstott, Anne. 2006. No Exit: What Parents Owe Their Children and What Society Owes Parents. Oxford: Oxford University Press.

Archard, David. 2010. “The Obligations and Responsibilities of Parenthood,” in David Archard and David Benatar (eds.) Procreation and Parenthood, Oxford: Oxford University Press, pp. 103–27.

Baron, Marcia. 2008. “Virtue Ethics, Kantian Ethics, and the “One Thought Too Many” Objection”, in Monika Betzler (ed.) Kant's Ethics of Virtue, Berlin and New York: Walter de Gruyter.

Bazargan-Forward. Saba. 2019. “The Identity-Enactment Account of Associative Duties”, Philosophical Studies 176: 2351–2370.

Bowlby, John. 1969-1980. Attachment and Loss, vols. 1-3, New York: Basic Books.

Brighouse, Harry, and Adam Swift. 2014. Family Values: The Ethics of Parent-Child Relationships, Princeton NJ: Princeton University Press.

Brink, David. 2001. “Impartiality and Associative Duties,” Utilitas, 13: 152–172.

Cohen, G. A. 2009. Rescuing Justice and Equality, Cambridge MA: Harvard University Press.

Estlund, David. 1998. “Liberalism, Equality and Fraternity in Cohen’s Critique of Rawls”, Journal of Political Philosophy 6(1): 99-112.

Frankfurt, Harry. 1988. “The Importance of What We Care About,” in The Importance of What We Care About: Philosophical Essays, New York: Cambridge University Press.

Frowe, Helen. 2023. “Assisting the Assisters: The Comparative Claims of Afghan Refugees”, Philosophy & Public Affairs 51: 294-326.

Gheaus, Anca. 2017. “Love and justice: A paradox?”, Canadian Journal of Philosophy 47 (6): 739-759.

Gheaus, Anca. 2021. “The Best Available Parent”, Ethics 31 (3):431-459.

Goodin, Robert. 1985. Protecting the Vulnerable: A Reanalysis of Our Social Responsibilities, University of Chicago Press, Chicago and London.

Goodin, Robert. 1988. “What is so Special about our Fellow Countrymen?”, Ethics 98(4): 663-686.

Hurka, Thomas and Esther Shubert. 2013. “Permissions To Do Less Than the Best: A Moving Band”, in Mark Timmons (ed.) Oxford Studies in Normative Ethics, Vol. 3, Oxford University Press.

Hurka, Thomas. 2016. “Love and reasons: The many relationships”, in Katrien Schaubroeck and Esther Kroeker (eds.), Love, reason and morality, London Routledge.

Jackson, Frank. 1991. “Decision-theoretic Consequentialism and the Nearest and Dearest Objection”, Ethics 101(3): 461-482.

Jeske, Diane 2014. “Special Obligations”, Stanford Encyclopedia of Philosophy.

Kant, Immanuel. 1996. The Metaphysics of Morals, Cambridge: Cambridge University Press.

Keller, Simon. 2013. Partiality, Princeton, NJ: Princeton University Press.

Koch, Felix. unpublished. “Skepticism About Special Obligations”.

Kolodny, Niko. 2010. “Which Relationships Justify Partiality? The Case of Parents and Children”, Philosophy and Public Affairs 38(1): 37-75.

Lange, Benjamin. 2022. “The Ethics of Partiality”, Philosophy Compass 17(8).

Mill, John Stuart. 1989. On Liberty and Other Writings, Cambridge: Cambridge University Press.

Murphy, Liam. 2000. Moral Demands in Nonideal Theory, New York: Oxford University Press.

O’Neill, Onora. 1979. “Begetting, Bearing and Rearing”, in Onora O’Neill and William Ruddick (Eds.), Having children: Philosophical and Legal Reflections on Parenthood, Oxford: Oxford University Press, 25–38.

Olsaretti, Serena. 2017. “Liberal Equality and the Moral Status of Parent-Child Relationships”, in David Sobel, Peter Vallentyne, and Steven Wall (eds.) Oxford Studies in Political Philosophy Vol. 3, Oxford University Press.

Owens, David. 2012. “The Value of Duty”, Proceedings of the Aristotelian Society Supplementary Volume lxxxvi: 199-215.

Owens, David. 2021. Bound by Convention. Unpublished manuscript.

Railton, Peter. 1984. “Alienation, Consequentialism, and the Demands of Morality”, Philosophy & Public Affairs 13: 134–171.

Raz, Joseph. 1989. “Liberating Duties”. Law and Philosophy 8: 3–21.

Raz, Joseph. 2011. “Attachments and Associated Reasons,” Columbia Public Law Research Paper No. 11-287. Available at SSRN: https://ssrn.com/abstract=1956026.

Seglow, Jonathan. 2013. Defending Associative Duties, New York and London: Routledge.

Setiya, Kieran. 2014. “Love and the Value of a Life”, Philosophical Review 123, pp.251-280.

Scheffler, Samuel. 1982. The Rejection of Consequentialism. Oxford: Clarendon Press.

Scheffler, Samuel. 2001. Boundaries and Allegiances, Oxford: Oxford University Press.

Shiffrin, Seana. 1991. “Moral Autonomy and Agent-Centred Options,” Analysis 51: 244–54.8.

Shiffrin, Seana. 1999. “Wrongful Life, Procreative Responsibility, and the Significance of Harm”, Legal Theory 5 (2): 117-148.

Stocker, Michael. 1976. “The Schizophrenia of Modern Ethical Theories”, The Journal of Philosophy 73:453-466.

Singer, Irving. 1994. The Pursuit of Love. Baltimore, MD: Johns Hopkins University Press.

Velleman, J. David. 1999. “Love as a Moral Emotion.” Ethics 109: 338–374.

Wallace Jay. 2012. “Duties of Love”, Proceedings of the Aristotelian Society Supplementary Volume lxxxvi: 175-198.

Williams, Bernard.1981. “Persons, Character and Morality” in Moral Luck: Philosophical Papers, 1973-1980. Cambridge: Cambridge University Press: 1-19.

Wonderly, Monique. 2016. “On Being Attached”, Philosophical Studies 173:223–242.