II. THE CONTENT OF GLOBAL PUBLIC REASON

I begin, to reiterate, with an account of the content of Rawlsian GPR (from now on, note, ‘GPR’ always refers to Rawlsian GPR, unless otherwise specified). It seems helpful to proceed by comparison with Rawls's more familiar idea of LPR. Rawls describes the content of LPR as being ‘given by the principles and values of the family of liberal political conceptions of justice’ (p. 143). ‘To engage in [liberal] public reason’, Rawls adds, ‘is to appeal to one of these political conceptions – to their ideals and principles, standards and values – when debating fundamental political questions.’ Notice that a family of such conceptions provides the content of LPR: when engaging in public justification, liberal citizens are permitted to draw on any conception within that family, whether or not all their fellow citizens accept it. The content of GPR, meanwhile, is significantly more constrained. It is provided not by any family of conceptions of international justice, but by a single conception – the Law of Peoples.

Rawls explicitly contrasts the content of LPR, as provided by a family of conceptions of justice, with that of GPR, as provided by the Law of Peoples alone (p. 57). The reason other conceptions of global justice are not to be introduced into GPR is apparently that the Law of Peoples is unique in being reasonably acceptable to the diverse societies which comprise the international justificatory community.Footnote ⁹ This is in contrast with the domestic case, where Rawls allows that there are ‘many liberalisms’ which citizens can agree are at least reasonable (even if not the most reasonable), among which his own ‘justice as fairness’, ‘whatever its merits, is but one’ (p. 141).

To inquire into the content of GPR, then, is to inquire into the content of the Law of Peoples. The latter is, as noted earlier, based around eight ‘familiar and traditional’ principles of international relations, which Rawls calls its ‘basic charter’. They are:

1. 1. Peoples are free and independent, and their freedom and independence are to be respected by other peoples.

2. 2. Peoples are to observe treaties and undertakings.

3. 3. Peoples are equal and are parties to the agreements that bind them.

4. 4. Peoples are to observe a duty of non-intervention.

5. 5. Peoples have the right of self-defense but no right to instigate war for reasons other than self-defense.

6. 6. Peoples are to honor human rights.

7. 7. Peoples are to observe certain specified restrictions in the conduct of war.

8. 8. Peoples have a duty to assist other peoples living under unfavorable conditions that prevent their having a just or decent political and social regime. (p. 37)

There is more, however, to the Law of Peoples than the basic charter. For as Rawls acknowledges, the charter is not a sufficiently detailed guide to the conduct of peoples without considerable further embellishment. For instance, immediately upon setting out the basic charter, Rawls notes (p. 37) that the fourth principle (and, we might add, the fifth) requires qualification to allow for other-defensive action, including humanitarian intervention in outlaw states. And he says that further principles will be needed to govern, for example, ‘forming and regulating federations (associations) of peoples, and standards of fairness for trade and other cooperative institutions’ (p. 38). Given the multiple ways in which the further development of the basic charter might be accomplished, Rawls tells us that ‘there is no single possible Law of Peoples, but rather a family of reasonable such laws. . .satisfying the representatives of peoples who will be determining the specifics of the law’ (p. 4 n. 4). And he distinguishes between what is the case under ‘a Law of Peoples’ (i.e. on some specific interpretation of that conception of justice) and what is the case under ‘the Law of Peoples’ (i.e. under any of its eligible interpretations). Consequently, the parallel between the respective contents of LPR and GPR is somewhat closer than it initially appears: although participants in GPR may only appeal to one conception of global justice, there nonetheless exists a family of interpretations of that conception on which they may draw. In further laying out the content of GPR, then, I shall focus on identifying the essential features of the Law of Peoples – those that will be preserved in any valid interpretation that might be adduced in the global public forum.

To begin: Rawls stresses that the status of the Law of Peoples as the basis of GPR depends upon its being a political conception of global justice. In virtue of what characteristics, however, does a conception of global justice count as political? In Political Liberalism, Rawls says that a conception of domestic justice is political if and only if it meets three criteria: (a) its principles apply primarily to society's major institutions, or ‘basic structure’; (b) it can be presented in a way that is ‘freestanding’ of the comprehensive doctrines over which reasonable liberal citizens disagree; and relatedly (c) it is derived entirely from fundamental ideas taken from liberal-democratic public culture, which any reasonable citizen can be expected to share.Footnote ¹⁰ In LoP, Rawls does not explicitly return to or amend this definition of the political for purposes of classifying conceptions of global justice. But it seems clear that he regards the Law of Peoples as political in that it satisfies analogues of the foregoing criteria: (a′) it governs what he calls (e.g. at p. 33) the ‘basic structure of the relations between peoples’ – that is, international law, and various cooperative organizations, such as military alliances, free trade areas, federal unions, and so on; (b′) it does not presuppose the truth or validity of any of the comprehensive doctrines or conceptions of justice according to which the reasonable members of the Society of Peoples order their internal affairs; (c′) it is worked up from normative ideas shared by reasonable peoples, as found in their common global public culture, or the ‘history and usages of international law and practice’ (p. 41). The content of GPR, then, is composed of interpretations of the Law of Peoples that preserve these political-making features.

Now, let us look further at criterion (c′), on which the Law of Peoples must be based on ideas shared by reasonable peoples.Footnote ¹¹ Since any publicly admissible interpretation of the Law of Peoples must conform to (c′), we can go deeper in specifying GPR's content by isolating the common normative commitments of reasonable peoples, as Rawls defines them.

In order to qualify as reasonable, a people must, Rawls says, accept and honour certain moral commitments regarding both their fellow peoples and their own members. He suggests that societies meeting these conditions ought to be admitted as full members of the Society of Peoples, and given justifications they can accept for the shape of the global order, both on grounds that they have ‘certain institutional features that deserve respect’, and on grounds that doing so is required for achieving durable peace in a pluralistic world (p. 84).Footnote ¹²

Consider first how reasonable peoples are defined as viewing each other: as free and equal participants in a scheme of global cooperation, entitled to cooperative terms that are both fair and mutually justifiable (see e.g. pp. 34–5). Reasonable peoples, in other words, are committed to political values of freedom, equality, fairness and public reason among peoples – analogues of the values of interpersonal freedom, equality, etc. to which reasonable citizens in a political liberal state subscribe. These global values are depicted by Rawls as moral cornerstones of the Law of Peoples, and thus of GPR.

Regarding their own members, meanwhile, reasonable peoples hold moral beliefs that qualify them as at least ‘decent’, if not necessarily liberal. A decent people accepts, first, that all its members have at least basic human rights against their government, where human rights constitute ‘a special class of urgent rights, such as freedom from slavery and serfdom, liberty (but not equal liberty) of conscience, and security of ethnic groups from mass murder and genocide’ (p. 79). A decent people also accepts that its members should all be included somehow in the scheme of social cooperation, that they should be governed according to a ‘common good’ conception of justice that takes everyone's interests into account to some degree (though not necessarily an equal degree), and that their views be given a measure of political representation – if not directly then through a ‘decent consultation hierarchy’ (see pp. 65–75). The content of GPR also includes, then, a shared idea of human rights, and of decency, understood according to the foregoing standards.

Decent peoples need not accept, however, that their members should have the extensive equal rights and liberties that liberalism prescribes, such as full liberty of conscience and association, sexual and reproductive freedom, and rights to vote and seek public office. Nor need they accept, more foundationally, that their members have the moral status of free and equal persons (p. 68). Accordingly, a people is not unreasonable by Rawlsian lights in adhering to a conception of domestic justice under which a wide range of non-basic rights are accorded only to certain favoured groups – males, say, or members of the official religion. And it is also fully consistent with Rawlsian decency for a people to hold that infringements of the basic human rights and associated interests of some of its members are more morally grave or tragic, and to be condemned, punished and guarded against more strenuously, than those of others. For as we have just seen, the criteria of decency specify only that peoples must honour their members’ human rights, and give weight to their interests and perspectives, not that they must do so equally. By that token, then, it is reasonable not merely for a people, say, to deny women the vote, or gay people the liberty to engage in sexual relations (to mention two policies that involve abridging liberal rights but not basic Rawlsian human rights), but also for it to impose, say, harsher legal penalties for the murder or enslavement of those to whom the official doctrine accords privileged status (at least assuming the human rights of each are protected to an adequate minimum degree).Footnote ¹³

In short, as Samuel Freeman puts it, ‘[i]t is not a condition of a decent society that it affirm the equality of its members or give them equal political rights. . .or even that it provide for equality of all basic human rights’.Footnote ¹⁴ The upshot of this, for the content of GPR, is that doctrines and arguments affirming the fundamental freedom and equality of persons, and the equal importance of their basic rights and interests, have non-public status, and may not be invoked in international deliberation and justification. Liberal societies may try to persuade their non-liberal counterparts that greater equality would be in their interests as peoples. But they may not claim, consonant with the rules of GPR, that decent peoples are mistaken about the moral worth of their citizens.

The fact that the liberal and non-liberal members of the global justificatory community share an understanding of the moral status of their fellow peoples, but not of the person, also accounts for two further distinctive features of the Law of Peoples – its contractualism, and composition in terms of principles specifying duties for peoples, rather than duties which persons across borders owe directly to each other. Consider first its contractualist nature. According to Rawls, the allegiance of reasonable peoples to the abstract political values of freedom, equality and fair cooperation among peoples will also lead them, more concretely, to agree that the appropriate perspective from which to endorse, interpret and refine the principles of the Law of Peoples is that of a global original position, wherein agreement is brokered between representatives of peoples who have an equal say and veto, irrespective of their geographical sizes, populations, conceptions of internal justice, and levels of wealth and power (which differences between them are obscured behind a veil of ignorance).Footnote ¹⁵ In the global original position, as Rawls depicts it, the representatives of peoples are not modelled as facing a choice between the Law of Peoples and alternative conceptions of global justice. Rather, they ‘simply reflect on the advantages of these principles of equality among peoples and see no reason to depart from them’ (p. 41). They also select, however, between more fine-grained interpretations of the Law of Peoples. In Rawls's words, ‘[i]n the Law of Peoples the many difficulties of interpreting the eight principles. . . take the place of the arguments for first principles in the domestic case. The problem of how to interpret these principles can always be raised and is to be debated from the point of view of the second-level [i.e. global] original position’ (p. 42).

Consider next what may be described as the ‘statist’ or ‘corporatist’ character of the Law of Peoples. As we have seen, the principles of the Law of Peoples are framed in terms of what politically organized peoples, not individuals, may and must do. Rawls says variously of the Law of Peoples that ‘its principles are addressed to peoples as peoples’ (p. 55), that it ‘applies to how peoples treat each other as peoples’ (p. 81), and that ‘it conceives of liberal democratic peoples (and decent peoples) as the actors in the Society of Peoples, just as citizens are the actors in domestic society’ (p. 23). He recognizes, moreover, that we will naturally want to know why the Law of Peoples is so structured: ‘What is it about peoples’, Rawls asks (p. 17 n. 9) on our behalf, ‘that gives them the status of the (moral) actors in the Law of Peoples?’

Unfortunately, Rawls's answer to this question is somewhat oblique. He advises us (p. 17 n. 9) that the answer is contained – along with an answer to the related question, ‘Why does the Law of Peoples use an original position. . .that is fair to peoples and not to individual persons?’ – in §11 of LoP. Yet there only the latter question is explicitly addressed. The substance of §11, however is a rejection of cosmopolitanism, on grounds that it presupposes an egalitarian conception of the person that is unacceptable to decent peoples. And this suggests an implicit answer to the question why the Law of Peoples consists of principles enjoining action from peoples rather than persons, as follows. First, any conception of justice specifying duties which individuals owe to each other globally must perforce take a stand on whether persons matter equally from the moral point of view, since otherwise we will be unable to determine how much each is required to sacrifice for the sake of the interests of others. Yet, in the absence of agreement on that issue, no conception of global justice that extends to global interpersonal relations can be based entirely on moral commitments shared by liberal and decent peoples. Those peoples do, however, agree that their fellow peoples are free and equal, and that their interests merit equal consideration. Hence, in the name of mutual acceptability to decent and liberal peoples, the Law of Peoples is restricted to specifying the terms of relations between peoples, leaving peoples to treat their own citizens in accordance with their own reasonable conceptions of domestic justice.

To sum up, the content of GPR is provided by a family of interpretations of the Law of Peoples – a contractualist conception of global justice that Rawls takes to embody political values of (inter alia) freedom and equality between societies that reasonable peoples share. The freedom and equality of persons, however, are values only for liberal peoples. And GPR therefore requires neutrality over whether the liberal view of the moral status of the person, or some decent non-liberal view, is correct.

III. THE JUST WAR DOCTRINE OF THE LAW OF PEOPLES

With the foregoing general account of the content of GPR in hand, I now turn to how GPR guides deliberation specifically about war. In Part III of LoP, Rawls aims, inter alia, to ‘work out the content of the principles of the Law of Peoples for the conduct of war’ (p. 91). Notice that his concern is with the principles of war for the (not merely a) Law of Peoples. The elements of just war theory presented are, in other words, purportedly commitments of any reasonable interpretation of the Law of Peoples that might be raised in GPR. Rawls depicts his account of war as largely faithful to ‘traditional thought on the subject’ (p. 94), and in particular to conventional just war theory, as surveyed and distilled in Michael Walzer's seminal Just and Unjust Wars (p. 95 n. 8). As we shall see, however, it is highly unconventional in certain respects. In this section I set out the main outlines of the Rawlsian theory of war, draw out its implications for the practice of GPR, and advance an initial objection that I believe to be damaging, but concede not everyone would regard as fatal.

Rawls gives us only a fragment of a complete theory of the just war, leaving considerable scope for further debate within the international community over the best interpretation of the relevant principles. He touches on aspects of both jus ad bellum and jus in bello, thereby elaborating on the fifth and seventh principles of the Law of Peoples. Regarding jus ad bellum, he focuses on the circumstances under which, so reasonable peoples would agree, there exists just cause for war. He does not, however, address the way in which peoples might interpret the other traditional restrictions on the right to instigate war, such as the legitimate authority or necessity requirements. Even Rawls's discussion of just cause leaves considerable room for further refinement. He says that reasonable peoples abjure expansionist war, and accept as just causes only self-defence, other-defence and humanitarian intervention (pp. 91–2). But should the Law of Peoples include, as part of the right to engage in defensive war, a right to go to war pre-emptively or preventatively, say, or to extract by force resources that are unjustly withheld by another regime?Footnote ¹⁶ These questions, among a host of others relating to the interpretation of jus ad bellum, remain to be resolved through collective deliberation within the terms of GPR.

Consider next Rawls's treatment of jus in bello, or what is sometimes called ‘the war convention’. It has been claimed that Rawls endorses the doctrine of the moral equality of combatants, whereby soldiers, on both sides of a conflict, are morally permitted to kill each other, and lack a right not to be killed by the enemy, irrespective of whether their cause is just.Footnote ¹⁷ In fact, however, he does not. For where targeting the enemy is concerned, he only explicitly addresses the question of what policies reasonable peoples, as collective agents, are required to adopt, on the assumption that they are fighting to resist the unjust expansionism of an outlaw state. This is to leave aside both what targeting policies are required or permitted of the outlaw state in prosecution of its ex hypothesi unjust war, and the question of what individual soldiers in the field may or must do.Footnote ¹⁸

Rawls suggests that reasonable peoples would accept non-combatant immunity as a limitation on their conduct, but could not agree to divest themselves of the option of killing enemy soldiers, even though they are often unwilling instruments of their unjust regime. ‘The reason why they may be attacked directly’, Rawls writes of an outlaw state's combatants, ‘is not that they are responsible for the war, but that well-ordered peoples have no other choice. They cannot defend themselves in any other way, and defend themselves they must’ (pp. 95–6). This does not entail, however, that reasonable peoples would also endorse an interpretation of the Law of Peoples under which, once war is in progress, outlaw states are likewise permitted to pursue a policy of targeting enemy combatants. To be sure, it may well be that reasonable peoples would agree to obey, and enshrine in international law, such a symmetrical in bello code, rather than one in which an outlaw state is not permitted to authorize any attacks on the enemy, even if they respect non-combatant immunity. Whether or not reasonable peoples would do so may depend, inter alia, on whether they, or their representatives in the original position, judge that, if outlaw states were denied permission to attack even enemy combatants, this would lead them to abandon restraint, increasing the destructiveness of war to everyone's detriment. Contractualist war ethicist Yitzhak Benbaji has argued that peoples that are at least decent in Rawlsian terms would accept a symmetrical war convention on precisely these grounds.Footnote ¹⁹ Whether or not Benbaji is right, the question of whether the constraints of jus in bello ought to be symmetrical between reasonable peoples and outlaw states is one which Rawls leaves to the Society of Peoples.

A further noteworthy aspect of Rawls's account of the principles governing reasonable peoples’ conduct in war is its inclusion of a supreme emergency exemption from the principle of discrimination, whereby a people is permitted intentionally to attack civilians where necessary to save itself from an imminent threat to its survival, or that of the Society of Peoples (see pp. 98–101). While the idea of a supreme emergency exemption is itself familiar, Rawls's treatment of it is atypical in at least two respects. First, whereas on the standard view targeting civilians represents an infringement of jus in bello – albeit sometimes a defensible one overall – under the Law of Peoples the exemption is a prerogative granted under the war convention itself. Second, the supreme emergency exemption is standardly given a lesser evil justification, under which the deontological constraint on targeting civilians is overridden if and only if the consequences of respecting it would be dramatically worse than those of infringing it. As part of the Law of Peoples, however, the justification behind the Rawlsian supreme emergency exemption must instead be that it would be adopted as part of a fair contractual agreement between reasonable peoples.

The lesser evil and contractualist justifications carry different implications for the range of circumstances in which the exemption may be invoked, with the contractualist justification being, in an important way, more permissive. Consider the case of a people that cannot save itself from annihilation or enslavement except by targeting some number of civilians of the enemy regime that vastly exceeds its own population. On the lesser evil view, this people cannot claim the exemption, since it would thereby cause far more evil than it prevents. On the contractualist view, however, it seems that this people must be allowed to save itself, even at the cost of a much greater evil. This is because, as we saw above, reasonable peoples accept that a fair agreement over the principles of the Law of Peoples treats them as equal parties irrespective of size. Thus, when they adopt the perspective of the original position, Rawlsian deliberators are to imagine themselves as unaware of the size of the people they represent. I take it that, under those informational constraints, each party would rationally insist on an equal right for peoples to avail themselves of the exemption, whether they are populous peoples aggressed against by smaller ones, or vice versa (presumably subject to the different limitation that they not cause the destruction of more peoples than they save). For otherwise, once the veil of ignorance was lifted, smaller peoples might find that, notwithstanding their equal fundamental interest in survival as a people, they are prohibited from saving themselves under conditions in which a more populous people would be permitted to proceed. If this is right, GPR constrains deliberators to endorse this more generous prerogative.

Those who favour the lesser evil position on targeting civilians (as, I observe, most contemporary war ethicists do) will find the wider exemption that GPR produces unacceptable.Footnote ²⁰ This gives us a significant initial objection to GPR's handling of war. The objection, however, may not be decisive. For not everyone, I acknowledge, finds implausible the supposition that societies have prerogatives to save themselves even at much greater cost to innocent life. Notably, Walzer writes that, on the lesser evil view, ‘large nations and small ones would have different entitlements in [supreme emergency] cases, and I doubt very much that that is true’.Footnote ²¹ I shall, then, leave the objection that GPR takes too lax a view of supreme emergencies in reserve, as an extra consideration for those who are persuaded by it. The objections on which I concentrate now, meanwhile, should be of concern to all.

IV. PROPORTIONALITY

To recapitulate, the just war doctrine developed by Rawls for the Law of Peoples contains a number of significant omissions, which are left for the Society of Peoples to fill in through deliberation within the constraints of GPR. One matter which Rawls leaves entirely aside is that of how to understand and apply the idea of proportionality. In the just war tradition, proportionality is a condition of both jus ad bellum and jus in bello. And clearly, members of the Society of Peoples will need to know when to censure parties who initiate disproportionate wars, or employ disproportionate means within war. As I now argue, however, the restrictions of GPR leave them entirely unable to do so.

By ‘proportionality’, note, I shall mean the issue of whether the harm inflicted upon innocent civilians by some (act of) war is, or would be, excessive in relation to the good thereby brought about.Footnote ²² That broad-brush characterization masks complexities which I cannot address within the confines of this article, concerning which harms and benefits count with the proportionality calculus, and how heavily they should be weighted.Footnote ²³ For our purposes, it suffices to say that one crucially relevant variable in calculations of proportionality is the moral status of war's victims and beneficiaries – that is, how much they, their rights, lives and interests count for from the moral point of view. To be sure, just war theorists have not tended to examine how proportionality assessments are to be made when the good and bad effects of war accrue to individuals with varying moral statuses. For the orthodox theory factors only harms and benefits to human persons into assessments of proportionality, and has a cosmopolitan moral complexion, to the extent that it takes all persons, irrespective of group membership, to merit equal concern and respect. Nonetheless, just war thinking implicitly acknowledges that moral status makes a difference to proportionality, precisely in so far as it completely discounts the effects of war on non-human animals, on the apparent assumption that they lack significant moral standing.Footnote ²⁴

Unfortunately, however, as we have seen, GPR requires a high degree of neutrality over the moral status of the person, owing to the fact of disagreement over that question between liberal and decent peoples. And precisely because they are disbarred from introducing into justificatory dialogue the claim that persons have equal moral standing, participants in GPR will be unable to determine whether, in particular cases, the harms caused to some by war are morally outweighed by the goods thereby realized.

To illustrate, consider a simple Collateral Damage Case, in which liberal people L is fighting a just war of self-defence against outlaw society O. L, suppose, faces the choice whether to launch an aerial raid that will cripple O's offensive capabilities, saving the lives of some large number of civilians_L, while collaterally killing some smaller number of civilians_O. Suppose that if the lives of each civilian_L and civilian_O were given equal weight (as L, in light of its liberal convictions, takes to be the case), L's bombing would be proportionate. Suppose also, however, that according to the prevailing religious doctrine of some decent member of the Society of Peoples, D, the lives of civilians_L are worth less than those of civilians_O, such that the raid would fall foul of proportionality (perhaps, for instance, D holds that civilians_L are of lesser value than civilians_O on grounds that they do not follow the true religion). The representatives of D cannot, of course, press for the Society of Peoples to censure L for violating proportionality without transgressing the terms of GPR, since their view presupposes the truth of their unshared doctrine. Yet by the same token, L's egalitarian view also relies on what are, under the rules of GPR, non-public considerations. Neither D nor L can claim that their conclusions about proportionality are or could be derived entirely out of moral reasons native to GPR. Thus, GPR is indeterminate on the question of whether L's act violates proportionality – it cannot provide the route to any answer to it. It is simply not possible to do the moral accounting without taking a stand on whether those who will be killed if the bombing does and does not go ahead are, one for one, of equal worth.

It is strange that Rawls would apparently not have noticed, in the context of war, that one cannot evaluate whether a given allocation of harms and benefits is morally permissible (or required) without first determining whether the individuals to whom they will accrue count equally. For this is a point that he himself emphasizes, in the cognate context of distributive justice. In explaining why the Law of Peoples cannot require decent societies to allocate burdens and benefits within their own schemes of domestic cooperation according to the principles generated in an original position populated by representatives of individuals, Rawls says (p. 70) that the original position is only appropriate for determining the claims of equal parties. For what would constitute a just distribution of burdens and benefits between equal persons will be unjust if their interests do not matter equally. This point applies just as strongly to the distribution of harms and benefits in war, however.

Let us now consider what responses might be available to the problem I am posing for GPR. First, a Rawlsian might argue that, although liberal and decent peoples do not agree in the first instance on the moral status of the person, a determinate, mutually justifiable position on proportionality might still be available to them by consulting the global original position. Representatives of peoples in that choice situation would know that the Society of Peoples will need to make collective judgements about proportionality, and, one might think, would therefore be inclined to agree to some convention for weighing lives. It might even be suggested that, for the parties in the original position – deprived as they are of knowledge about the conception of justice around which their internal affairs are organized, but aware that they have an interest in the security of their members – the default agreement point would be on a convention of weighing lives equally for proportionality purposes.Footnote ²⁵

I do not believe that this suggestion can be correct. Note first that there are, strictly speaking, two global original positions described in LoP. In the first, liberal peoples endorse the Law of Peoples, and, as needed, debate and agree to its further development. In the second, decent peoples do the same. Now, if denizens of the first global original position know they are liberal, and are presented with an egalitarian proportionality principle, they would seem to have no reason not to adopt it, since they are aware that they all accept the equality of the person. But denizens of the second global original position do not know how, at the bar of their socially regulative doctrine, the moral status of the person is understood. What they do know, however, according to Rawls, is that they have a fundamental interest in preserving the integrity and stability of that doctrine (see e.g. pp. 29 and 33). If they know that their interests lie in the preservation of their doctrine, do not know what it pronounces on the moral equality of persons, and are asked to endorse a principle of equal valuation of civilian lives and interests in war, what would they do? It seems to me that they would be unable to reach a decision, owing to the risk that they will have seriously undermined their doctrine once the veil is lifted. If this is right, we cannot appeal to the global original position to resolve the question of how to weigh harms in proportionality calculations, since the problem of indeterminacy infects the original position itself.

A Rawlsian might reply that failure to agree rules of war would render all peoples less secure, and that it is rational for the representatives of decent peoples in the original position to subordinate their interests in the preservation and pursuit of their doctrines for the sake of security. But this response seems inconsistent with the way Rawls depicts the representatives of peoples as conceiving of and reasoning about their interests. For those representatives appear to understand their peoples’ interests in the preservation of their respective members and territories as derived from and conditioned by the commitments of their doctrines. Thus, regarding liberal peoples, Rawls writes (p. 34):

And later, regarding decent peoples, Rawls suggests the parallel – that their interests are specified by their ‘common good’ conceptions of justice (p. 69). This indicates, then, that decent peoples will be viewed by their representatives in the global original position not as having an interest in protecting their members that is independent of, and capable of being traded off against, their interest in maintaining and implementing their doctrines, but as having an interest in protecting their members to the degree required under their doctrines, as part of their general interest in pursuing those doctrines. Yet, if the representatives of decent peoples see their interests in this way, while being deprived of knowledge of the value that their doctrines place on the lives and interests of their members, it again appears that they will be unable to come to any judgements regarding whether some proposed proportionality rule is compatible with their interests or not.

GPR's problem of indeterminacy regarding proportionality cannot, then, be resolved by recourse to the global original position. This is not yet to say, however, that there are no other means (short of resorting to the use of non-public reason) by which the Society of Peoples can cope with this indeterminacy, and reach satisfactory practical decisions regarding whether to authorize or prosecute wars or acts of war. Because the literature on GPR is relatively small, and the incompleteness objection has not (to my knowledge) previously been pressed against it, the question of how global deliberators might handle cases in which GPR fails to issue sufficient guidance on the resolution of political problems has not been considered. However, in the more extensive literature on LPR, Rawlsians have argued that, when public reasoning ‘runs out’, there are a number of decision-making strategies available that enable a political community to come to a conclusion concerning how, practically, to proceed, without the need for citizens to invoke their comprehensive doctrines. Indeed, Micah Schwartzman has identified five such strategies: (1) deferring the decision until later, in the hope that additional public reasons come to light; (2) deferring to someone else who is thought better able to come up with an answer using the available public reasons; (3) finding a mutually acceptable compromise between the opposing perspectives; (4) giving up on deliberation, and putting the matter to a democratic vote; (5) employing a random decision procedure, like a lottery.Footnote ²⁶ I believe, however, that none of these coping mechanisms would be of assistance to the Society of Peoples in the matter at hand.

We can, I think, safely discount options (1), (2) and (3). There are no additional relevant public reasons that will come to light by waiting, or that some parties have privileged access to, to whom others should therefore defer. Nor is this a matter that admits of a mutually acceptable compromise: splitting the difference between rival views of the moral worth of the person, even if possible, would only lead to conclusions that everyone would find morally repugnant (where by ‘everyone’ I mean not only the members of the Society of Peoples themselves, but those who, like you and I, are appraising Rawlsian GPR from the perspective of the theorist, and to whom its details and implications must be acceptable in reflective equilibrium).

Even if the strategy of compromise were not here morally distasteful, moreover, there is a further reason why it cannot help in this particular case – a reason which also serves to rule out strategy (4). This is that (3) and (4) are not really eligible solutions to indeterminacy in public reason, but only to a somewhat different obstacle to decision-making in public reason, known as inconclusiveness. To explain this distinction (which is originally due to Gerald Gaus),Footnote ²⁷ indeterminacy refers to a situation in which, as in the example of proportionality, public reason supports no answers to the question asked of it (that is, there are no or insufficient public reasons to justify giving an answer). Inconclusiveness, meanwhile, occurs when a number of answers can be reached on the basis of the available public reasons, but different views obtain regarding which is best, between which public reason fails to adjudicate.

It is consistent with the Rawlsian ideal of public reason, in both the domestic and global contexts, to resolve instances of inconclusiveness by proceeding to a vote, or by compromising on a third option that is regarded by the deliberators as better than nothing, albeit not optimal. For in such cases, the resultant policy is justified by a reasonable balance of public reasons, notwithstanding that some or all parties believe another policy to be preferable. Matters are very different, however, in cases of indeterminacy. Here there are ex hypothesi no policy options supported by public reason alone. To make a decision by majority vote, in these cases, means simply for the majority to impose its non-public doctrine on the minority, while for opposing parties to compromise involves striking a balance between their non-public doctrines, in a way that Rawls contrasts with the practice of public reasoning, as ‘political in the wrong way’.Footnote ²⁸ In short, then, where proportionality is concerned, strategies (3) and (4) fail to uphold the ideal of GPR.

This leaves only strategy (5). And if this is what Rawlsian deliberators are forced to fall back on, all seems lost for GPR. Although there are plausibly some political and moral issues that call for the use of random decision procedures, it seems impossible not to regard it as a reductio of the Rawlsian view if it requires the random resolution of political questions as fundamental as whether the Society of Peoples will prosecute, aid or issue statements of support for wars, in preference to resolutions arrived at via the use of ordinary moral reasoning. When what is at stake is whether the members of the Society of Peoples will materially contribute to an armed conflict, deciding the matter randomly rather than by judgement would be intolerably cavalier. And when what is at stake is whether to offer some public statement condemning or condoning a war, or the means being employed within it, no declaration whose content had been determined randomly would enjoy any modicum of moral authority. No belligerent power could feel chastened, for instance, to be censured over its conduct in war, if the decision to do so was made through the drawing of straws. According to Rawls (pp. 92–3), an important purpose of the Law of Peoples, and GPR, is to provide a shared language within which reasonable peoples can influence global agents into changing their unethical ways, through moral persuasion. The example of proportionality indicates that the theory is a failure in those terms.

To conclude, then, GPR is indeterminate with respect to the important issue of proportionality in war. And none of the mechanisms suggested by defenders of LPR can be imported to alleviate this problem adequately. That GPR should struggle with an issue of such centrality to the ethics of war seems fatal in itself. As I argue next, however, GPR suffers from another war-related defect that is at least as serious.

V. INDIVIDUAL PARTICIPATION

As we have seen throughout this article, the Law of Peoples is a conception of international justice under which peoples, not persons, are the primary loci of concern and respect, and the agents to whom its principles issue injunctions to act. Because the Law of Peoples provides the content of GPR, these features of the conception produce a framework for deliberation that is highly inhospitable towards the pursuit of cosmopolitan moral concerns. For they ensure that arguments regarding fairness in the meeting of individual interests across borders, and how far persons may owe duties directly to each other to promote and protect each other's well-being, cannot even be raised in the global public forum, much less acted upon.

Even as a defender of Rawls, Leif Wenar acknowledges that the people-centric nature of the Law of Peoples ‘comes at a price’.Footnote ²⁹ ‘Because Rawls's global theory works exclusively in terms of peoples’, he writes, ‘it cannot show any direct concern for individuals.’Footnote ³⁰ Even when the Law of Peoples prescribes humanitarian intervention in an outlaw regime, or material aid in burdened societies,

According to Wenar, however, while the Law of Peoples may seem, from a cosmopolitan perspective, ‘disappointingly conservative’Footnote ³² – especially in its opposition to egalitarian global redistribution – it is vindicated overall by providing what Wenar sees as an attractive – indeed indispensable – practical account of war and security.

Wenar's defence of Rawls first emphasizes his commitment to public justification, arguing that the Law of Peoples centres on relations between peoples, not persons, because the global public culture from which it must, as a political conception of justice, be derived, is itself ‘primarily international, not interpersonal’.Footnote ³³ ‘There simply is no robust global public political culture’, Wenar claims, ‘which emphasizes that citizens of different countries ought to relate fairly to one another.’Footnote ³⁴ Wenar then contends that, even if we regret that Rawls's approach to the justification of global principles of justice is not more conducive to a cosmopolitan economic agenda, we should nonetheless accept that (a) the goals of international peace and stability are prior to distributive justice, and (b) absent a world state, the principles of war needed to achieve these goals will be, as in the Law of Peoples, ‘inescapably statist’.Footnote ³⁵ By this, Wenar means that these principles will grant particular sovereign peoples responsibility over identifiable territories, and rights to defend them militarily. Statism of that kind, Wenar avers, underpins the existing laws of war and traditional just war doctrine, and its adoption within the Law of Peoples is warranted because it represents ‘the only realized approach to global political morality that we have’.Footnote ³⁶

There is a serious problem, however, with Wenar's invocation of the just war tradition in defence of the Law of Peoples.Footnote ³⁷ For the Law of Peoples is statist in a far more radical sense than traditional just war doctrine. To be sure, Walzer's influential presentation of the just war position employs the so-called ‘domestic analogy’, whereby nations are seen as having rights to defend their borders against aggressive regimes that mirror the rights of citizens to defend their homes against intruders.Footnote ³⁸ But the tradition is not committed to thinking of war as if states were the only actors involved, nor are its principles addressed exclusively to nations or peoples as collective actors. On the contrary, the theory holds that persons, even if they are not in a position to influence whether their group goes to war, or what overall strategy it pursues, are still individually subject to the rules of jus in bello, and morally answerable for their acts and omissions. The Law of Peoples, by contrast, addresses all its principles, including its in bello requirements, to peoples, acting collectively through their institutions, and thus speaks to individuals at best only indirectly, in so far as they are responsible, as office holders of one sort or another, for enacting their people's will, or carrying out its duties. It does not, then, speak to the question of what individuals, acting on their own initiative, or in an uncoordinated way, may or must do in war. Far from showing fidelity to traditional just war theory, then, the structure of the Law of Peoples results in the excision of the traditional theory's commitment to the moral appraisal of individual wartime conduct.

This means that claims on the world stage to the effect that particular individual participants in war (whether members of a people that is a party to the war, of a people outside the conflict, or indeed of no people at all) acted wrongly by engaging in such-and-such acts of unjust harming or refusals to save cannot be genuine expressions of GPR. To be sure, given that peoples are subject to in bello duties under the Law of Peoples, they must exercise control over their soldiers (and members more generally), and can be condemned within the terms of GPR for allowing them to go off the rails. Yet suppose that a people, through its government, pursues a military strategy consonant with jus in bello, and puts in place all due precautions against violations of civilian immunity etc. by its members. If, nonetheless, some of them, acting spontaneously, carry out wrongful intentional attacks on civilians, GPR does not provide the language within which to condemn their behaviour, since the agents to whom its principles are addressed acted in full compliance with their obligations.

Because claims about individual wrongdoing fall outside the remit of GPR, Rawls's theory withholds a special status from them: that of having normative authority cross-culturally, such that they cannot be reasonably dismissed as an artefact of some particular parochial doctrine or outlook. Yet, whilst there is, of course, room for reasonable disagreement (in a non-technical sense) over the precise content of the duties that apply to individuals in war, it is unacceptable to deny that any such duties apply to them. Of course, GPR does not, strictly speaking, deny that individuals in war are under their own moral duties – it merely refuses to register or pursue that question. In so doing, however, it unfortunately implies that the view that there are no such duties is reasonable, and fails to provide members of the Society of Peoples with the conceptual and argumentative resources needed publicly to challenge or repudiate that position.

It is not sufficient consolation that, whilst individual activity in war cannot be directly morally criticized in GPR, governments remain free to educate their citizens morally about their wartime obligations under their home-grown moral or religious doctrines, and indeed sanction them for violations. For even if liberal and decent peoples do so, this leaves untouched persons living in badly ordered societies, who are at greater risk of seeing war, and whose leadership and culture may well reject the idea that there are wartime moral duties for individuals, or take a repugnant view of what those duties are (‘kill as many unbelievers as you can’, say).

Nor is it sufficient consolation that, despite the gap I have exhibited in its content, GPR is nonetheless capable of justifying the prosecution, under international criminal law, of individuals who commit war-related crimes. To be clear, GPR can make the case for such prosecutions. What it cannot do, however, is justify them in the most intuitively powerful or common-sense fashion: as the enforcement of some of the most stringent duties of justice to which persons are individually subject – not to kill, maim, rape, and so forth – and which they owe to each other. In GPR, the justification for this legal practice must instead be that it is necessary to the protection of the interests of peoples, who would insist, in an appropriate choice situation, on its being part of the global basic structure. That is a rather morally anaemic defence of such an important feature of international law. Worse, it may even undermine the basis of respect for that law. For it seems misleadingly to imply that the law instrumentalizes those whom it targets, for the sake of the interests of peoples as collective entities, by punishing them for acts from which, as far as could be gleaned from the pronouncements of the international community, they were under no personal duty to refrain.

In sum, GPR is silent over what justice demands of individuals in war. But that silence implicitly says something, and that something is unacceptable: that it is not unreasonable to deny that persons are under any individual duties of justice when they take part in armed conflict.