3.1. Promissory obligations & rectificatory obligations

Before addressing the challenge to the idea that groups bear these two kinds of obligation, we should say more about how we understand each of them. We will take promising first. Promissory obligations are a species of a genus that includes both promissory and contractual obligations (the latter involves consideration, the former doesn’t). Within this genus, we focus on promising for two reasons. First, it has a well-developed philosophical literature. Second, promissory obligations loom large for groups. We hold political parties to account for their election pledges; corporations to account for their advertisements; and (to the extent we can) international organisations to account for their expressed aspirations. Our question is whether this practice is justified.

We’ll use the popular ‘normative powers’ view of promising (Owens Reference Owens2006; Shiffrin Reference Shiffrin2008). On this view, to promise is to exercise a normative power: we generate specific obligations for ourselves, and rights for others, by making promissory utterances. Recent advocates of this view justify it by reference to the interests that are served in the exercise of this power. According to David Owens, this is ‘an authority interest: I often want it to be the case that I, rather than you, have the authority to determine what you do’ (Reference Owens2006, 51). According to Seana Shiffrin, this is an interest in having the ‘ability to engage in special [i.e. close] relationships in a morally good way, under conditions of equal respect’ (Reference Shiffrin2008, 485).

On this rationale, if groups are to be involved to the give-and-take practice of promising, then groups must have interests in having authority over others or in having close relationships with others. And for particular promissory obligations to bear, groups must be able to make utterances. There are no problems here. On the functionalist model of agency, entities with decision-making procedures can produce decisions about what the entity desires. The entity can be understood as having an interest in those desires being fulfilled. And it looks plausible that many groups have a desire – an interest – in both (a) holding authority over others and (b) having close relationships with others. On (a), simply consider the archetype preferences of states, churches and corporations, which is to hold authority over their members qua individuals, and often over non-members too. These preferences generate groups’ interests in receiving promises, which in turn generating interests in giving promises, as a way of indicating a commitment to the practice. On (b), consider the close relations that group agents often have with other group agents, such as allied countries, political coalition partners and sister organisations, and the ways in which group agents try to forge such close relationships by making promises, often through formal mechanisms such as contract law.

Additionally, it’s plausible that groups can make utterances. Groups do this in two steps. First, the decision-making procedure produces decisions of the form ‘the group commits to C’ and ‘role bearer R will publicise the group’s commitment to C’. Second, role bearer R publicises the group’s commitment to C, and does so (i) without violating any other aspects of her role in the group and (ii) partly because of the group’s decision that she should do this. It looks like collectives can both make promises and be obliged to keep them.

Rectificatory obligations are much more straightforward. We understand them in the following way: if one violates an obligation, then one acquires a duty to compensate for, undo or otherwise ‘rectify’ the violation. This kind of obligation is justified by the fact that one has not followed through on some other obligation.

As we mentioned above, both promissory and rectificatory obligations face a common challenge when it comes to their application to collectives. The challenge arises because of issues surrounding groups’ identity across time. Individuals are rarely thought to undergo changes that block their duties to keep promises, or rectify harms done, before the change. For groups, though, such changes might be systematically caused by features particular to groups. This would create a disanalogy in the way groups and individuals should be treated with regard to obligations to keep their promises and to rectify wrongs. Think about a promise that a corporation makes to its shareholders, soon before being subsumed by a larger corporation. Or consider the obligation of West Germany to pay reparations to Israel after Second World War: one might think that West Germany was not descended in the right way from Nazi Germany, because that was West Germany together with East Germany, and furthermore under radically different leadership. So the duty here, it might seem, is not rectificatory. (This is not to say that there was no moral obligation on West Germany, just that it was not an obligation to rectify a violation committed by West Germany.)

We won’t take a stand on whether these cases involve the ‘descent’ relation appropriate for promissory and rectificatory obligations. All we need, for the parity claim to hold, is that later groups are often descended from earlier groups in the way relevant to these obligations and that there is nothing particular to group agency that systematically undermines the descent relation. Two substantive proposals for the appropriate ‘descent’ relation for individuals are as follows (Olson Reference Olson and Zalta2010). There is a psychological approach, according to which B’s mental states must be caused in the right way by A’s mental states. There is also a somatic approach, on which physical continuity connects B to A. It is obviously beyond our scope to discuss all the issues here. The key point is this: whichever descent relation one prefers, it can be readily applied to groups.

Psychological descent will amount to connection (of some appropriate form – here this is dispute) between the group’s earlier and later beliefs, desires and decisions. Such descent often holds: group agents have the beliefs, desires and decisions formed by past exercises of the group’s decision-making procedures by past members of the group; unless those beliefs, desires and decisions have been superseded by more recent exercises of the group’s decision-making procedures. Superseding of this kind is analogous to an individual changing her mind, and is consistent with psychological descent.

Physical continuity of a group amounts to substantial continuity in its membership. Requiring perfect continuity would lead to the implausible view that, for example, a new country comes into existence every time a new citizen is born. There is minimal continuity when there are no abrupt and complete changes in membership, such that not one individual is a member at both t ₁ and t ₂. Substantial continuity exists when a substantial number of members remain members between t ₁ and t ₂, even while new members join (or are born) and old members leave (or die). That means a group can entirely change its membership over time, without becoming a different group. (It is impossible to say in the abstract exactly how many (or which) members must remain for substantial continuity to hold – this will vary with the size and structure of the group.) This is the same reason for thinking that Theseus’s ship is the same ship even when all the original parts have eventually been replaced, and for thinking it would be a different ship if we were to simply demolish the original and build a new one where the original used to stand. It’s also the same reason for thinking that there’s not a new human individual every time a finger is sliced off or a beard grows – and that there would be if the human was destroyed and Davidson’s swamp person appeared in their place (see e.g. Lewis Reference Lewis and Rorty1976).

However, these descent relations do not always apply cleanly to actual groups. Consider fusion and fission. Fusion occurs when a group is formed by (partially or totally) combining two or more previously existing groups, which then cease to exist. Fission occurs when two or more groups arise out of one previously existing group, which then ceases to exist. Nazi Germany became East Germany and West Germany (fission), and later reunified (fusion). In such cases, we have change that is the result of an earlier decision procedure, with substantial continuity in membership, but we have a different number of entities than we had previously (Schwarz 2014a, 2014b).Footnote ⁷

Does this mean that the scope of groups’ duties does not extend to keeping promises, or rectifying past violations? No, on both counts. Fusion and fission are simply problems that occur more often for groups than for individuals. They’re also easier to recognise in groups, because more visible. When a country divides, we see the effects everywhere: the Berlin Wall divided East from West Berlin and largely prevented mixing of the two groups’ members. By contrast, when an individual suffers dissociative personality disorder, we still see only one physical form, and when two individuals become so interdependent that they almost wholly merge their identities, we still see two physical forms. But frequency and ease of recognition of fission and fusion do not undermine the parity claim. Individual identity can cease, just as group identity can.

As with individuals, there are two ways to go on fission or fusion cases when it comes to groups’ obligations to rectify violations. The first is to say that the later and earlier groups are associated in a morally relevant way, whether or not the relation is one of identity. Continuity across time establishes a relationship between the earlier and later entities – a kind of associative taint – that gives the later one obligations to rectify the violations of the earlier one. The second route is to say the later entity or entities do not have obligations regarding the earlier entity’s or entities’ violations. Whatever we would say about individual fusion or fission cases can extend, mutatis mutandis, to groups.

3.2. Beneficiary obligations

The idea has recently been gaining currency that we have obligations when we benefit from the wrongs of others. The precise nature of these obligations is under dispute. Some argue that if one intentionally benefits or welcomes the benefits, the obligation is stronger (Pasternak Reference Pasternak2014). Others suggest these obligations arise when the benefit is received non-voluntarily, but that there must be a victim of the wrong, who benefits from the discharge of the beneficiary obligation (Butt 2007, 2014). Still others argue that one can have a duty simply to disgorge the benefits of injustice, regardless of whether there is (or ever was) an identifiable victim (Goodin and Barry Reference Goodin and Barry2014). Whichever specification one uses, the fact remains that groups regularly benefit from injustices. States – particularly those who already have high material wealth – benefit from unjust trade agreements. Firms benefit from governments’ unjust corporate tax policies. And so on. Do beneficiary obligations throw up particular problems for groups?

One might think that firms, in particular, are resistant to this obligation. Imagine a coffee shop that is grossly vandalised. During repairs, its regular customers head to the coffee shop down the road. The second shop benefits from an injustice that is done to the first one.Footnote ⁸ Yet we would strongly resist the idea that the second shop thereby owes anything to the first shop. Indeed, we could make the case stronger: imagine the second shop puts up an advertising sign outside the first one, intending to benefit by attracting the first shop’s regular customers into its doors. It’s still unclear whether the second shop derives any obligations from these benefits. Companies enter the market in full knowledge of the possible misfortunes that might befall them, and in awareness of the vagaries of their customers’ tastes and loyalties. So it seems companies do not have beneficiary obligations.

This is all plausible. But it does not speak to a disanalogy between group and individual agents. Replace each coffee shop – which implies a team of owners, investors, managers, etc. – with a single person selling flowers. Each is under the radar of the tax system, such that they are not legally incorporated. Suppose one seller has all their flowers stolen, leading to a boom in sales for the other seller. Again, facts about markets, known risks and competition lead us away from thinking the second seller should disgorge or pay back these benefits, even if they are welcomed by her – even if they are intended by her. So there is no special problem for groups when it comes to beneficiary duties in the market. There may well be special problems for beneficiary duties in market contexts as compared with non-market contexts. But groups are a red herring for these problems. This means we don’t have to solve them in order to view groups as falling under beneficiary obligations.

There are, however, other contexts in which groups seem to avoid beneficiary duties. Imagine a government has a policy that unjustly favours Christian organisations over Muslim ones when it comes to funding, taxation, government promotion of the Creed and so on (make the government’s policy as extreme as needed for it to be clearly unjust). Suppose this policy leads many individuals to reject Muslim organisations and join Christian ones. This can be construed as a benefit to the Christian organisations and a harm to the Muslim organisations, insofar as ‘gaining members’ plays the role of a desire in the decision-making procedures of proselytising religions. The harm is, ex hypothesi, unjust. Do the Christian organisations owe anything to the Muslim ones?

Before answering that question, consider another case. Take the state as employer of civil servants. Imagine the state unjustly favours men over women when it comes to employment, promotion, paid leave, mentorship and so on (make the policy as extreme as needed for it to be clearly unjust). Suppose this policy leads to Rawiri being employed (or promoted, or given leave, or mentored) over Ngaio. Ngaio loses out, and Rawiri benefits, because of the employer’s discriminatory policy. Does Rawiri owe anything to Ngaio?

Our contention is that there is no good reason to answer differently in the Christian/Muslim case than in the Ngaio/Rawiri case. Both are cases where an unjust government policy – something over which the protagonists have no control – leads one group or individual to receive benefits that should have gone to some other group or individual. Either obligations are owed in both cases, or in neither case. We cannot imagine what reason there could be to give different judgements on these cases.

These two pairs of thought experiments (the coffee shops vs. flower sellers; the religious organisations vs. civil servants) do not prove that groups have beneficiary obligations in just the same way as individuals. But the debate about the preconditions of beneficiary obligations is nascent. So to prove the parity here, we would have to go through a huge number of paired thought experiments, each sensitive to different preconditions, tweaking the set-up in each case so that the only variable was whether the (putatively) duty-bearing agent was individual or collective. This would be a cumbersome task. The two pairs of thought experiments we have outlined what we see as the main problem cases for parity in beneficiary duties, giving us reason (though defeasible) to believe that parity holds here.

3.3. Associative obligations

Following Samuel Scheffler, we understand associative obligations as ‘responsibilities that the members of significant social groups and the participants in close personal relationships have to each other’ (Reference Scheffler2001, 49). For individuals, these are usually held towards friends, family or members of one’s cultural group.

Again, there is dispute over the justification of associative obligations. There are three rough camps. The first sees associative obligations as grounded in a voluntary act by their bearer. These will likely collapse into a tacit promissory obligation. Having already discussed promissory obligations, we put this view aside. The second sees associative obligations as grounded in the fact that our personal relatives (friends, family, perhaps co-nationals) are especially vulnerable to our actions. On this view, associative obligations transmute into assistance obligations. Again, we put this view aside.

The third view, on which we will focus, sees associative obligations as irreducible to obligations of other kinds. Scheffler’s account is paradigmatic. He says that ‘one’s relationships to other people give rise to special responsibilities to those people when they are relationships that one has reason to [non-instrumentally] value’ (Reference Scheffler1997, 197–198). The precise content of the obligations generated by such relationships varies depending on the ‘nature of the relationships’ (Reference Scheffler1997, 199). For example, the nature and norms of friendship vary across cultures, meaning the specific associative obligations between friends vary across cultures. The generally described justification, based on relationships with objective non-instrumental value, applies universally.

For groups to have associative obligations, it must be possible for relations between groups (or between groups and individuals) to be non-instrumentally valuable: valuable not as a means to anything else. To make this concrete, take two states – say, Australia and New Zealand. These two states are connected by a closely shared history, in which they were each colonised by the UK, before gradually gaining independence. They fought together in two world wars, are perceived by members to have a similar culture and outlook on life, have high levels of mutual immigration, and are committed to (more or less) similar values. This connection gives rise at least to perceived obligations between the states: obligations, for example, to assist one another in the case of a natural disaster. These obligations are perceived to hold even in the absence of express promises of such assistance, and would hold even if some other agent (say, the USA) was better-placed to provide assistance. So these obligations are not subject to the promissory or assistance-based interpretations of associative obligations. Instead, we can explain them by acknowledging the complex value – the value of shared experiences, shared perceptions, and shared circumstances – that is realised by the past, present and projected future interactions between these countries.

Scheffler’s account can also be applied to two other kinds of group-group relationships. First, consider group–group relationships that are only instrumentally valuable: the relationship between two banks competing in the market, or between two civil society organisations with fundamentally different aims. It is hard to make sense of associative obligations between such pairings – and it is also hard to make sense of the possibility that such relationships are non-instrumentally valuable. So when we resist positing associative obligations between groups, we can make sense of this in terms of the absence of the non-instrumentally valuable relationship that exists in the Australia/New Zealand case. As for beneficiary obligations, notice that the same is true of individuals in competitive contexts: the lack of a non-instrumentally valuable association implies a lack of associative obligations.

Second, there are group-group relationships that are not dyadic, but are constituted by membership in a social group. Scheffler points out that associative obligations are held between individuals who are co-members of non-instrumentally valuable social groups, even if the two members are unaware of one another’s existence. (Scheffler gives the example of two members of a fan club (Reference Scheffler1997, 198)) For groups, the relevant ‘social groups’ have groups as members: social groups like ‘international civil society’ or ‘ethical companies’ or ‘democratic states’. As long as such social groups are non-instrumentally valuable, they can give rise to obligations for members – say, obligations to help one another maintain the ethos or achieve the characteristic aims of the social group’s members.

Finally, Scheffler’s account can be applied to group–individual relationships. Most obviously, groups might have non-instrumentally valuable relationships with their members. It might be valuable that I am a member of this club or nation or church, not as a means to anything else, but simply in itself. This could generate obligations for me to pursue my club’s interests over the interests of other clubs, and for the club to pursue my interests over the interests of non-members. This might appear to be a disparity: groups have a kind of obligation – an associative obligation to their members – that individuals do not have (and do not have because proper parts are agents in one case and not the other). However, while there surely are associative obligations between groups and members, these are not a further kind of obligation. They fall under the umbrella of associative obligations, which are premised on a relation of non-instrumental value between agents. Because some agents have other agents as proper parts, it stands to reason that some associative obligations hold ‘inwardly’, that is, to those parts (members). But it’s still the intrinsically valuable relationship that’s doing the duty-generating and kind-defining work, not the inward-directedness. So this is not a disparity for the scope of collective’s and individual’s obligations.

4.1. Demandingness

The demandingness problem is this: holding fixed the kind of duty and the agent’s capacity to exercise the duty, is the threshold of ‘overdemandingness’ higher for collectives than for individuals? Does morality legitimately infringe more upon groups’ pursuit of their ends than on individuals’ pursuit of theirs? Interestingly, there is good reason to think less may be demanded of groups than individuals and good reason to think more may be. In order to defend our parity claim, we need to defuse each possible disanalogy.

First, there is reason to think less may be demanded of groups than individuals. This reason is that collectives are created for the specific purpose of pursuing certain ends. These are its ‘constitutive ends’: ends that are deeply important to the collective and serve to underpin many of the collective’s more specific ends and plans. This suggests it’s overdemanding for them to do much to sacrifice those ends. Thus, even if all kinds of obligations apply to groups, each obligation applies with less demandingness when it applies to groups. Corporations and churches might be the clearest examples here: while there is a pro tanto moral demand for corporations and churches to have environmentally friendly practices, we tend to think this obligation is overridden if it is incompatible with the full pursuit of their (permissible) constitutive ends. Thus, it seems an oil company cannot have an obligation to stop searching for oil and pursue green energy instead – even if the company is perfectly able to pursue green energy, and would cause harm if it searches for oil, has caused rectification-requiring harm through oil drilling, is able to help others through pursuing green energy, and so on. The oil company cannot have an obligation to pursue green energy, one might think, simply because this is not its raison d’être. In the Introduction, we saw that Singer holds a similar view about art galleries’ obligations to give to poverty relief. The same looks true for churches: any obligation churches have to, say, alleviate global warming had better be consistent with the ardent pursuit of their aims of worshipping, proselytising and so on. Anything else is too demanding.

Second, though, there is reason to think more can be demanded of groups than individuals. This reason is that groups do not have phenomenal experience. They don’t feel demands in the way individuals do. Imagine for a moment that the oil company has a duty to pursue green energy. If this requires the company to wholly revise its constitutive ends, we would not regret the effects on its conscious experience and unique subjective feel of the world, because it does not have these things. It would not feel despair or loss. The group would not – in an experiential, phenomenal sense – even be aware it was revising its ends (though obviously it would believe this in the functional sense outlined in Section 2). This suggests we can ask more of collectives than individuals, since they don’t really feel the demands anyway.

The upshot is this. The overdemandingness threshold for any given agent will vary with: (i) whether the demand conflicts with the agent’s (permissible) constitutive ends, and (ii) whether the agent can feel the demand. However, we don’t think there’s any entailment between having one of these properties and being a group.

Take constitutive ends. While it’s true that well-functioning groups have constitutive ends, it’s also true that well-functioning humans have constitutive ends. An individual’s constitutive ends might include being an excellent teacher and researcher, being a dedicated parent, maintaining lifelong friendships, or seeking new cultural experiences. It is typically thought that morality may not thwart such ends (Cullity Reference Cullity2004). Some individuals do not have any constitutive ends, but then some groups are also in the midst of identity crises and so lack constitutive ends. Well-functioning agents – whether individuals or groups – have constitutive ends that morality cannot override. There’s no disanalogy here.

It might seem that individuals have more control over the content of their constitutive ends than collectives have over the content of theirs. If so, then perhaps morality can require individuals to choose particular constitutive ends rather than others, whereas it cannot require this of collectives. If so, then morality would be more demanding on individuals than on collectives: a disanalogy.Footnote ⁹ We resist this on two fronts. First, individuals have less choice over their constitutive ends than liberal visions of the self-suggest. Even in diverse and socially liberal societies, many forces other than autonomous choice account for shifts in people’s personalities, attitudes and self-conceptions, Footnote ¹⁰ not to mention that these commonly become ‘locked in’ over time, all of which plausibly limit one’s ability to genuinely choose one’s constitutive ends. Second, many collectives are able to change their constitutive ends. Consider firms that decide to branch out into, and then solely focus on, new products, or decide to take the ‘green bottom line’ seriously. Or consider how states can shift from central planning to free marketism. These are genuine choices of the collective itself, just as long as they’re made via the collective’s distinct internal decision-making procedure.Footnote ¹¹ So there is not sufficient difference between collectives’ and individuals’ freedom to determine constitutive ends to make for disparity in obligations.

What about feeling demands? Again, poorly functioning individuals may lack the relevant affective response to demands that undermine their constitutive ends. And, again, well-functioning groups have constituent parts – members – that would feel despair or loss upon the dissolution of the group or the undermining of its constitutive ends. (Imagine how deeply disaffected many people would feel if their football team changed its fundamental aims, let alone their church or state.) So while constitutive ends and phenomenology may well make a difference to demandingness, they do not drive a wedge between individuals and groups.

4.2. Secondariness

The second strength-related problem arises from double counting: for every obligation, there is the possibility that group members will be ‘hit twice’ by that obligation. This will happen if both the individual and a group of which she is a member bear the obligation. Suppose, for example, that for me the assistance obligation triggers a duty to give money to charity. Suppose I fully discharge that duty. But then my state or my church has a duty to give international aid as a result of the fact that it, too, has the obligation (because it has the capacity to meet the need). And suppose that, in discharging its obligation, my group must pass at least some costs to me (through tithing or taxation). As a result, I have now been hit twice by assistance obligations.

The objection, then, is this. In order to avoid double-counting, we must hold that groups’ obligations are secondary to individuals’ obligations. To see the secondariness, suppose all members have done everything they individually should, up to the limit of what can be reasonably demanded of them, and there are no costs that can get absorbed entirely at the group level. In this case, the group cannot be reasonably demanded to discharge the obligation. All the shares or parts of its obligation are excluded as distributable to members, because members are already at the limits of what can be demanded of them. So groups’ obligations can be precluded by individuals having fully discharged their individual obligations, when doing so takes them up to their limit of demandingness. In this way, groups’ obligations might be secondary to – because conditional upon facts about – the discharging of duties that members have as individuals considered independently of the group. But if groups’ obligations are secondary to the obligations of their members, then there is not parity between the two. So our parity claim is wrong.

An initial reply to this objection is to simply accept double-counting. Perhaps being a member of a group entails that more costs can be imposed upon me when my group has a duty, even if I’ve already fully discharged my individual obligations. Perhaps this is because the group can do much more than me on my own. To make our opponent’s view as strong as possible, though, let’s assume it’s impermissible for the group to distribute any costs to me if I’ve hit my demandingness threshold.

The correct reply is to deny that individuals’ obligations always take priority. In this way, we deny that groups’ obligations are inherently secondary. An individual will have obligations as an individual, and obligations in virtue of her membership in the various collectives to which she belongs. Sometimes it will not be possible for her to fulfil all obligations without exceeding her demandingness threshold. But that doesn’t tell us which she should prioritise. It seems rather obvious that what she should prioritise depends on the case. Sometimes, collectives’ obligations will be extremely important, i.e. their fulfilment (or not) will involve very high stakes. In such cases, members should surely prioritise their member-distributed obligations over less important individual obligations (such as keeping a promise to meet a friend for lunch). Sometimes the reverse will be true. If and when it happens that all members are in this position, the collective made of those members will not have an obligation (or it will have the obligation, but be excused from non-performance on grounds of over-demandingness). Likewise, individuals will be excused from the non-fulfilment of their qua-individual obligations when the discharging of their qua-member obligations took them up to the relevant threshold on demandingness. It works both ways. There is no systematic relegation of either obligation, and so no disparity.