1.1 Wallace's Argument

Let's start by introducing Wallace's solution to the numbers problem and by sketching why it both adheres to the individualist restriction and can overcome two common objections.

Wallace argues that contractualists should consider which principles regarding the behaviour of rescuers individuals would be able to reasonably reject before they know what position they are occupying in the world (Wallace, Reference Wallace2019, pp. 215–19). This is what is generally referred to as reasoning from an ex ante perspective. Without knowing what position they will occupy in the world, Wallace suggests that individuals would assess principles in regard to whether or not a principle maximizes their own chances of being saved, should they ever be in need of aid. Note that the probability that each individual will end up in a larger group of people if ever in need of aid is greater than the probability that they will end up in a smaller group of people. Thus, from an ex ante perspective, individuals would reasonably reject a principle which permits rescuers to save the few. Instead, they would maximize their chances of being saved by agreeing to principles that require rescuers to save the greater number. Due to the fact that individuals only appeal to personal reasons in the form of their own probabilities of being saved, Wallace's argument upholds the individualist restriction.

Wallace's approach may not appear entirely new, since ex ante solutions to deciding how to approach Rock have previously been advanced (Taurek, Reference Taurek1977; Otsuka Reference Otsuka2006). Two standard objections have been pressed against such solutions. First, the claim that contractualists should assess principles from an ex ante perspective in which individuals don't know their position can appear ad hoc. Why should principles be assessed from an ex ante perspective rather than from an individual's actual, fully-informed perspective? Second, contractualists may oppose engaging in ex ante reasoning because it potentially leads back to the kind of unattractive utilitarian conclusions contractualists set out to avoid.

That second worry needs more explanation. Once we posit a certain view about rationality, reasoning from an ex ante perspective can quickly permit the imposition of substantial burdens on the few, in order to secure smaller benefits for the many. Transmitter Room serves to illustrate this: from an ex ante perspective, each individual knows that their chance of being the severely injured individual is minuscule, but that their chance of being a world cup viewer is very large. Thus, from an ex ante perspective, each individual would maximize their expected well-being by agreeing to a principle which would permit the continuation of the broadcast. This is a conclusion that utilitarians accept, but which contractualists universally reject.

But Wallace is alive to these two objections. He offers a single solution to both. He suggests that in the vast majority of cases, contractualists will continue to compare the immediate (rather than ex ante) claims of individuals. However, things are different in cases in which individual claims are equally strong and therefore tied, as in Rock. In those cases, individual claims are ‘inconclusive’: they leave open which course of action the rescuer ought to take. Wallace writes:

This appeal to inconclusive claims can answer both of the foregoing objections. First, the need for more guidance in cases where claims are tied removes the appearance of ad hocness in the move to assessing claims from an ex ante perspective. Second, because principles are only assessed from an ex ante perspective in such tied cases, Wallace's argument doesn't cause contractualism to collapse into utilitarianism. In Transmitter Room, for example, individual claims are not inconclusive about what we should do, and so there is no need for us to move to assessing individual claims from an ex ante perspective.

In establishing a duty to save the greater number without jettisoning the individualist restriction, Wallace's argument seems to throw a lifeline to contractualists who are reluctant to give up the core idea of their moral theory. However, on closer inspection, it appears that Wallace's solution is vulnerable to two objections.

1.2 The Objection from Motivation

In order to offer answers to the two standard objections, Wallace heavily relies on the premise that individual claims are tied and therefore ‘inconclusive’ in Rock. I want to show that this premise cannot motivate Wallace's move to ex ante reasoning.

Note that out of the two possible contractualist responses to Rock outlined in the introduction, Wallace's argument isn't addressing the idea that the rescuer is required to toss a coin in Rock in order to give equal chances of being saved to all individuals. If the rescuer in Rock were required to use this kind of decision-procedure, it would mean that individual claims in Rock would not be inconclusive – they would deliver a single right course of action for the rescuer.

Wallace must therefore be assuming that contractualism does not yield an obligation to toss a coin and instead yields the permissibility of saving either group. On this view, the rescuer may decide which group to save by following an inkling, flipping a coin, throwing a die, running a lottery, or choosing on a whim. We can understand Wallace as arguing that the move to ex ante reasoning is warranted by the need to provide further direction to rescuers.

Here is a worry about Wallace's proposal: It seems to follow that the possibility that tied claims could render several moral options permissible is regarded as unsatisfactory, and as implying a need for further moral input. But this doesn't seem quite right. Consider a two-person case in which a rescuer can save either Jill or Joan but cannot save both of them. Let's call this Choice. Given that Jill and Joan's claims are tied, it seems permissible for the rescuer to save either of them.Footnote ⁴ Thus, Jill and Joan's claims are, according to Wallace, inconclusive. But if the existence of inconclusive claims is considered to license a move to ex ante reasoning, then this equally applies to Choice. Just like Rock, we should consider Choice to be morally unresolved, and likewise strive to gain additional moral guidance through ex ante reasoning. Yet this seems false. In Choice, the existence of tied claims, and the resulting permissibility of multiple options, seems unobjectionable. It is perfectly permissible for the rescuer to save either Jill or Joan. In Choice, no further moral guidance is required. Intuitively, then, Wallace's argument delivers the wrong conclusion about Choice.

A defender of Wallace may insist that Jill and Joan's claims are, indeed, inconclusive in Choice, and require us to move to ex ante reasoning in order to gain further moral input. They may highlight that from an ex ante perspective, many options would be considered permissible in Choice, such as picking whoever is nearest, plumping, or flipping a coin. Thus, the fact that the permissibility of multiple options is intuitively unobjectionable in Choice can be explained even when engaging in ex ante reasoning and doesn't indicate that Jill and Joan's claims aren't inconclusive.

Alternatively, it's possible that Choice is a special case in which the fact that several options are permissible doesn't deeply trouble us and doesn't push us to seek additional moral guidance. Even then, there might still be a general reason to seek out additional moral guidance through ex ante reasoning in cases in which claims are tied. If it's possible to obtain any more guidance on which action to perform in cases in which claims are tied, surely rescuers will want to access it. For instance, even though the permissibility of several options isn't troubling in Choice, rescuers might still prefer to know whether they should rather flip a coin to decide between Jill and Joan, or directly plump for one of them.

But the reasoning of both replies seems to me mistaken. The fact that tied claims render several options morally permissible doesn't by itself justify a move to ex ante reasoning. Choice illustrates that we don't generally expect morality to be uniquely action-guiding. Once several courses of action are rendered permissible by a moral principle, it's irrelevant that morality doesn't direct us towards one single option. In fact, we normally don't think that there is anything inconclusive about views which tells us that multiple options are morally permissible. Thus, what matters is that agents adhere to the correct moral principle by pursuing one of the permissible courses of action, irrespective of which action they choose.

Put differently, given that there is no requirement for morality to isolate a single right action, the fact that individual claims in Rock render several options permissible cannot justify a move to ex ante reasoning. Morality has already spoken, as it were, and no further moral guidance is required.Footnote ⁵

This is not to dismiss the rescuer's own desperate wish for accessing further moral guidance when being faced with several permissible options in Rock. Far from it. Instead, the rescuer's plea for more guidance is best understood as a reflection of the grave tragedy of the situation. After all, although the rescuer is choosing between morally permissible options, all options are terrible ones, and the choice between them is a terrible one to have to make.

1.3 The Objection from Counterintuitive Implications

Even if the previous objection could be answered, there is a second objection to Wallace's argument. Wallace's argument leads to counter-intuitive, maximizing conclusions which contractualists are likely to find objectionable. This can be seen in the following case. In Charity, an individual is planning to donate a third of her relatively low yearly income to charity and, without having any specific preference,Footnote ⁶ is choosing between three charities which tackle poverty: Oxfam, the Hunger Project, or Local Poverty Action. Contractualists like Wallace and Scanlon tend to believe that donating such a large portion of one's income is morally optional in light of the significant burden it imposes on the individual, even if the individual might be required to donate a smaller portion (Scanlon, Reference Scanlon1998, p. 152, p. 224). Let's stipulate that Oxfam is the most efficient charity in this case because it saves the greatest number of lives per pound donated. Many contractualists, Wallace included, may will still want to claim that the individual in Charity isn't required to donate to Oxfam but can permissibly choose to donate her money to any of the three charities (Wallace, Reference Wallace2019, pp. 206–207). Indeed, this conclusion reflects an intuitive and popular way of thinking about charitable giving.

Wallace's argument leads to the opposite conclusion. Because the individuals aided by all three charities have tied claims, and all three options are regarded as morally permissible, Charity would be assessed from an ex ante perspective. Yet from this standpoint, individuals would agree to a principle which requires agents to donate to Oxfam in Charity, as this will maximize each individual's chances of benefiting, should they ever find themselves in need of aid.

There are two reasons for why Wallace should resist this conclusion. First, notice that a duty to donate to the most efficient charity has severely counter-intuitive implications. It suggests that even though it's morally permissible for the individual in Charity to spend most of her income on frivolous knick-knacks, it's morally wrong for her to donate this money to the Hunger Project or Local Poverty Action. But suggesting that individuals who ignore the sufferings of others and spend their money on their own amusement are doing the right thing, while individuals who choose to use their money to help others act wrongly seems like a troubling verdict. Intuitively, or so it seems, we make the opposite moral judgement. Second, while philosophers sympathetic to the effective altruism movement conclude, similarly, that one has a duty to donate to the most efficient charity even in cases in which one could permissibly refrain from donating (Pummer Reference Pummer2016; Horton Reference Horton2016),Footnote ⁷ Wallace himself rejects their conclusion, arguing that contractualists have no reason to affirm that individuals are under a ‘standing requirement to maximize the impartial good’ (Wallace, Reference Wallace2019, p. 206).

Indeed, the fact that Wallace's argument delivers such maximizing conclusions in Charity also undermines the anti-maximizing thought that motivates the individualist restriction and further conflicts with the contractualist idea that reasonable principles are principles which foster relations of mutual recognition, rather than principles which best advance each individual's interest (Scanlon, Reference Scanlon1998, p. 194).

All in all, looking closer at Wallace's argument makes plain that an acceptable solution to the numbers problem needs to fulfil desiderata beyond the individualist restriction. A satisfactory solution to the numbers problem must flow naturally from the core ideas of contractualism and exemplify the anti-maximizing spirit behind the individualist restriction rather than just adhering to the letter of that restriction.

2.1 Charity

I want to suggest that assuming a duty to save the greater number in Rock is inconsistent with supporting another, similarly common-sense intuition: that one can permissibly donate to any cause in Charity.

There is a widely shared intuition that, when costs are equal, one has a duty to secure more rather than less good and thus perform the better act. This common-sense intuition is exemplified in a duty to save the greater number in Rock and would seem to threaten contractualism. But this common-sense intuition to perform the better act may itself conflict with similarly compelling intuitions, leading us to doubt its initial plausibility. Perhaps, then, the contractualist should not worry about failing to accommodate the idea that when costs are equal, one ought to secure more good rather than less good.

If we posit a duty to perform the better act when costs are equal, then this duty would deliver the wrong conclusion in Charity. As a reminder, in Charity, an agent chooses between donating a third of her income to Oxfam, the Hunger Project, or Local Poverty Action. Here, because the agent can save a greater number of individuals at no additional cost by donating to Oxfam, donating to Oxfam constitutes the better act from an impartial perspective. Therefore, from the assumption that there is a duty to perform the better act when costs are equal, it would follow that the agent in Charity has a duty to donate to Oxfam.

But, as I emphasised earlier, most people, at least outside the philosophical movement of effective altruism, find this conclusion counterintuitive: it seems permissible for the agent in Charity to give her own money to whichever cause she chooses. What's more, a duty to donate to Oxfam in Charity has the highly implausible implication that it's morally wrong for the agent to help people in need by donating to the Hunger Project or Local Poverty Action, but morally permissible for the agent to spend her money on knick-knacks.

Still, someone may argue that there are circumstances in which the agent does have a duty to perform the better act in Charity. What if the individual in Charity has the explicit aim of alleviating poverty? Having done research and knowing very well that donating to Oxfam would better fulfil this aim, surely it would be irrational for the individual to nevertheless donate to the Hunger Project or Local Poverty Action.

Even if someone who explicitly aims to do as much good as possible should give to Oxfam in Charity, most people who give to charity do not have this explicit aim. People generally don't aim to do ‘as much good as possible’ when choosing to give to charity. Rather, people care about and empathize with individuals in need and want to advance their claims. And, since the Hunger Project and Local Poverty Action both fight poverty by advancing the claims of individuals in need, just like Oxfam, it seems permissible for people to donate to any of those three causes. What's more, the fact that it might be irrational for this individual to donate to the Hunger Project or Local Poverty Action in light of her aims does not suffice to show that it would also be morally wrong for her to do so.

A critic might insist that Rock and Charity are morally disanalogous and hence reject the claim that a duty to perform the better act in Rock is incompatible with the permissibility of donating to any cause in Charity. This critic might argue as follows: While the rescuer in Rock is morally required to save someone, acting is optional for the agent in Charity, given that acting imposes a substantial burden on the agent. But since any course of action in Charity is morally optional, a duty to perform the better act cannot apply. What is more, the respective burdens which each of the three possible donations impose on the individual is precisely what exempts the individual from a requirement to donate. Even if the individual is prepared to donate to the Hunger Project or Local Poverty Action, the individual can still appeal to the financial burden which donating to Oxfam would impose on her in order to explain why she is not required to donate to Oxfam.

I want to deny that this disanalogy between Rock and Charity can explain why there is a duty to perform the better action in Rock but not in Charity. To see this, it's helpful to assimilate both cases by stipulating that individuals are morally required to donate a substantial part of their income - let's say a third- to charity, given the very strong claims which individuals who are dying from deprivation around the world have on those who are better-off. Let's call this Required Charity. If moral optionality is the only distinction between Rock and Charity, it would follow from this that individuals ought to perform the better action in Required Charity and donate to Oxfam.

But this seems controversial. One may maintain that it's counter-intuitive to think that the individual in Required Charity acts wrongly if they choose to support the Hunger Project or Local Poverty Action. The individuals in need who would be helped by both of these charities merit and require our support just like individuals who would be aided by Oxfam. They, too, have strong claims to our assistance. And, as Wallace has emphasized, there is no reason for contractualists to share a commitment to maximizing the impartial good, as commonly asserted by proponents of effective altruism. Thus, contractualists are apt to deny that the individual would be required to donate to Oxfam in Required Charity.

Note that the only thing which currently resembles morally required donations is taxation. Yet it's clear that tax revenue isn't allocated according to what does the most good: Even when leaving aside pluralist causes such as culture, taxes are still used to fund treatment for ailments such as hair loss or nail infections although many more lives could be saved if funds were allocated differently (Voorhoeve, Reference Voorhoeve, Eyal, Hurst, Murray, Andrew Schroeder and Wikler2020).Footnote ⁹ Now of course there may be other justifications for this, but it's important to keep in mind that this use of tax revenue is generally considered unobjectionable.

In summary, the appeal to moral optionality cannot explain why agents intuitively don't have a duty to donate to the most efficient charity in Charity. Thus, despite the fact that both claims are rooted in common-sense morality, it's not possible to consistently affirm both a duty to perform the better action in Rock as well as the permissibility of performing any action in Charity.

Giving up the intuition behind Charity might seem to some people to be the more attractive option. After all, affirming the duty to donate to the more efficient charity in Charity would nicely reflect the very strong intuition behind Rock, namely that individuals should be moved by the fact that there are more lives to be saved in one group than in the other. In addition, it seems possible to explain why the intuition in Charity appears appealing. In many real-life cases we simply don't know which charity is better than others, so even if we accept the intuition about effectiveness in the hypothetical case, this is clearly disanalogous from real-world cases in which our intuition of wide discretion would still persist. But it is far from obvious that we should give up the intuition in Charity rather than Rock. So, giving up the claim that there is a general duty to save the greater number cannot be as contrary to common-sense morality as opponents of contractualism often assume.

In the remaining sections, I argue that the option of giving up the intuition in Rock, and hence rejecting a general duty to save the greater number, is preferable to giving up the intuition in Charity. I also gesture towards a deeper explanation of why individuals can permissibly choose to save either group in both Rock and Charity.

2.2 Mother

I want to look at another widely shared common-sense intuition that is inconsistent with the pervasive idea that the rescuer in Rock has a duty to save the greater number. In Mother, Jane is the mother of three children, and has to make a decision about which of her children to save from death. She can either save Anny, or she can save both Betty and Cady, but she cannot save all three. What should Jane do? Jane loves all her children. The loss of any of them would be devastating for her. Indeed, having to choose between her children in this way is one of the worst decisions Jane could ever be faced with, and she would gladly give her own life if that would allow her to save them all. Jane desperately looks for any considerations which could help her in making this harrowing choice. She considers in vain whether she could save all three of them, whether one of them may have a higher chance of survival, or whether she could find help elsewhere. At this point, those who support a duty to save the greater number in Rock will tell Jane that she ought to save as many of her children as she possibly can.

Yet suppose that in this tragic case, after having exhausted all other options, Jane chooses to save Anny. It might be reasonable to ask Jane why she saved Anny rather than Betty and Cady. But would we consider Jane to have acted wrongly for having saved Anny rather than Betty and Cady? I suggest not. We just don't think Jane is obligated to choose the option which maximizes the number of lives saved. We respect her choice as a mother who is trapped between two tragic options, and who chooses to save Anny.

Plausibly, one may suggest that the special features of this case render Mother disanalogous to Rock and therefore explain why Jane, unlike the rescuer, doesn't have a duty to save the greater number. I will consider four such explanations and show why each of them is unsuccessful.

First, it appears that Mother is set apart from Rock by special psychological features, as Jane's situation is extremely psychologically demanding. Deciding between the lives of her children is inconceivably painful for Jane, understandably impairing her ability to rationally and level-headedly assess her options.Footnote ¹⁰ Perhaps Jane cannot reasonably be expected to make the right choice and save the greater number.

However, this explanation can be contested. Say Jane has nominated a proxy to choose in her place. Given that this individual is representing Jane, they may act in whichever way they believe to be reflective of Jane's interest. Here, it still seems that Jane's proxy could permissibly save Anny. But this can no longer be justified by reference to the psychological burden imposed by this choice, meaning that such a burden cannot furnish an explanation for why Jane doesn't have a duty to save the greater number.

A second explanation might draw attention to the special duties a mother has towards her children. One may suggest that a mother's duty to save her children is much stronger than a rescuer's duty to save strangers. For instance, a mother might have a duty to save her children even if this means risking her own life. Thus, Jane's duty towards each of her children may be uniquely stringent and powerful in a way which doesn't aggregate, meaning that Jane's duty to save Anny cannot be outweighed by her respective duties to save Betty and Cady.

This explanation falls short. Although a mother's duty to aid her children differs in strength to a rescuer's duty to save strangers, we have no reason to think that it also differs in kind. But a mere difference in strength cannot plausibly by itself make a difference to whether duties aggregate or not.

Third, one might appeal to the special duties that a mother has towards her children in a different way. One might propose that mothers have a very strong duty to treat each of their children equally. Thus, Jane might be required to give equal chances to be saved to each child, for instance by flipping a coin.

This, too, seems mistaken. It seems rather cruel to expect Jane to fetch and flip a coin before being able to permissibly save her children. Indeed, should Jane fail to do so and opt to save right away, it seems counter-intuitive to suggest that Jane acted wrongly. Just like it seems permissible for Jane to save the many or the few, it seems permissible for her to save either party without first flipping a coin.

A fourth explanation points us towards the fact that permissible partiality is a factor in Mother, but not in Rock. Permissible partiality generally exempts individuals from a duty to bring about the greater good in cases in which individuals would be required to sacrifice their own significant interests or loved ones. For instance, an individual can permissibly save their loved one's life rather than the lives of five strangers (Williams, Reference Williams1981). Thus, we may think that Jane is exempt from a duty to save the greater number because she loves all of her children.

However, the explanation from permissible partiality doesn't actually justify why Jane can permissibly save Anny. The concept of permissible partiality generally exempts individuals from a duty to bring about the greater good in cases in which bringing about the greater good would impose substantial costs or sacrifices on them, such as the loss of a loved one. But this isn't what's going on in Mother. Given that Jane will lose at least one of her children either way, a duty to save the greater number doesn't impose a substantial sacrifice on Jane. Hence, Jane's exemption from a duty to save the greater number cannot be justified by appealing to permissible partiality.Footnote ¹¹

Common-sense morality, then, tells us that in this tragic case, it's intuitively permissible for Jane to save either Anny or Betty and Cady, yet this cannot be explained by appealing to the special features of this case. Thus, Mother offers a key counterexample to the claim that there is a general duty to save the greater number. Again, we find that denying a duty to save the greater number is less in conflict with common-sense morality than opponents of contractualism typically assume.

2.2 Valuing Individuals and Permissibly Saving the Few

The last two sections considered the contractualist conclusion that the rescuer in Rock doesn't have a duty to save the greater number and argued that this conclusion is in fact much more in line with common-sense morality than is often suggested. Still, some questions for the contractualist remain: In particular, a reader might worry that while a contractualist bar on aggregation has been shown to be compatible with common-sense morality in several cases, we have yet to offer independent argumentation in favour of the conclusion that it's permissible to save either group in Mother and Rock. In other words, aside from a concern for adhering to the contractualist framework provided by the individualist restriction and a concern for individualist justification, do we have any other reason for thinking that both Jane and the rescuer can permissibly save either group?

Answering this question comprehensively and providing an argument for why the permissibility of saving either group coheres with our value judgements more generally is beyond the scope of this paper. However, let me indicate a possible direction in which such an analysis might go.

Whenever we assess and compare individuals and individual claims, we do so in light of a ‘covering value’. A covering value is a value in virtue of which meaningful comparisons are made between two bearers of value. In regard to a covering value, a bearer of value may generally be equal, better, or worse than another. For instance, when comparing cheese and chalk in regard to the covering value ‘goodness as a housewarming gift’, cheese is better than chalk, but when comparing both in regard to the covering value ‘goodness as a toy’, chalk is better than cheese (Chang, Reference Chang and Chang1997, pp. 5–7). Individuals can be compared in regard to a vast number of different covering values: in regard to their ability to play chess, their quality as a conversation partner, their coffee-making skills, their generosity, their amount of well-being etc.

So how do we assess Rock? Contractualism tells us to compare and rank individual claims with regard to their strength. However, this doesn't help us decide what to do in Rock, given that all claims in Rock are of equal strength and thus tied. Thus, we need to apply a different covering value when assessing Rock. We may, for instance, compare the individuals in Rock in regard to morally relevant covering values such as intrinsic value, moral status, capacities, rationality, or well-being. Indeed, we might stipulate that the individuals in Rock are equal in regard to these covering values (even if this is unlikely to be the case in regard to all of them). Because the value of individuals is equal in regard to each of these covering values, we can aggregate the value of individuals in regard to these values and secure more value by saving the many than by saving the few. Thus, so the thought goes, we ought to save the greater number. In other words, if our aim is to secure as much value as possible in regard to such covering values as intrinsic value or well-being, then a duty to save the greater number delivers on this goal.

Worryingly, this approach seems to treat individuals as mere vessels for these covering values. In particular, this approach reduces our concern for saving individuals in Rock to a concern for securing and maximizing amounts of value and thus fails to recognise and value each individual for their own sake. Or, as Samuel Scheffler puts it, sacrificing one bearer of value in order to save a greater number of bearers of value seems to ‘giv[e] the value that things have priority over the things themselves’ (Scheffler, Reference Scheffler2018, p. 110).

Properly valuing individuals when assessing Rock, or so we might think, should involve valuing individuals in a way which isn't derivative from the amount of value which each individual adds to this world.Footnote ¹² That is, we should assess individuals in regard to the covering value ‘value for one's own sake’. Indeed, valuing individuals for their own sake is a familiar and common-sense way of valuing individuals, as it generally captures how we value our loved ones.Footnote ¹³ This can be seen in Mother, where Jane loves Anny for being Anny, Betty for being Betty, and Cady for being Cady.

I want to draw attention to a significant implication of valuing individuals in this way. I suggest that when valuing individuals in regard to the covering value ‘value for one's own sake’, the value of individuals is incomparable.Footnote ¹⁴ It is often suggested that two bearers of value are incomparable if one bearer of value is not better than the other, worse than the other, or equally as good as the other in regard to a covering value (Chang, Reference Chang and Chang1997). Upon reflection, any time we compare the value of two individuals in regard to the covering value ‘value for one's own sake’, we will find that two individuals S₁ and S₂ are incomparable. This is because any individual S₁ is neither better than, worse than, nor equal to any other individual S₂ in regard to the covering value ‘value for one's own sake’ given that the value of each individual is grounded in, and individuated by each person: S₁'s value is grounded in S₁, S₂'s value is grounded in S₂, etc. A crucial implication of the claim that the value of individuals is incomparable in regard to the covering value ‘value for one's own sake’, is that the value of individuals cannot be aggregated in respect to this covering value. As a result, saving a larger number of individuals doesn't secure more value than saving the few does. Thus, if we care about valuing individuals for their own sake rather than as vessels for value, then we arrive at the conclusion that it may be permissible to save either group in Rock.

One challenge for presenting an argument for why the permissibility of saving either group in Rock coheres not only with the contractualist framework but also with our theoretical value judgements more generally, is to specify the value which individuals have for their own sake. This includes responding to two key objections. First, one may argue that the value of individuals is in fact non-aggregable not because individuals have incomparable value, but because individuals have infinite value. Second, one may propose that the value of individuals in regard to the covering value ‘value for one's own sake’ isn't incomparable, but roughly equal or on a par.Footnote ¹⁵ However, more work must be done on carving out such an argument for the permissibility of saving either group, with this section offering no more than a rough outline.

2.3 Practical Implications: Charitable Giving and Policy Making

So far, I've suggested that permitting the rescuer to save either group in Rock is compatible with common-sense morality and may even accurately reflect our concern for valuing individuals for their own sake. It now remains to return to Charity and similar concerns which arise in the context of public policymaking.

It's now possible to explain why one should reject a duty to save the greater number in Rock rather than the claim that it's permissible to donate to any cause in Charity. The Hunger Project and Local Poverty Action help different people than Oxfam. Different people need our help and different people are aided through different charities. We care about donating to charity because we care about each individual for their own sake and care about aiding them. In other words, while it's possible to take a stance similar to the position of a public health official and aim to maximize well-being by donating to the most efficient charity, this may fail to value those individuals who are in need of support for their own sake.

In fact, public officials may have to make decisions slightly differently than individuals. When making public policy decisions for the benefit of a society, they may have to compare options in regard to the amount of well-being each option brings about, or the amount of lives it secures and therefore use aggregative reasoning. This is because it may be the specific role and task of a public official to maximally benefit the well-being of the people, just like it is the task of a bodyguard to protect their client. Thus, they may allocate foreign aid to the charity which will save the most lives or redirect a terror threat from a crowded city to a small village. But, again, this is a feature of their public role. If, outside of their official capacities, an individual regards other people as vessels for value rather than as being valuable for their own sake, they are no longer properly valuing individuals, and possibly have abandoned the task of assessing what they owe to others.