Jump to ContentJump to Main Navigation
Democracy Beyond BordersJustice and Representation in Global Institutions$

Andrew Kuper

Print publication date: 2004

Print ISBN-13: 9780199274901

Published to Oxford Scholarship Online: November 2004

DOI: 10.1093/0199274908.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2019. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a monograph in OSO for personal use (for details see www.oxfordscholarship.com/page/privacy-policy). Subscriber: null; date: 21 May 2019

Global Justice

Global Justice

Beyond The Law of Peoples to a Cosmopolitan Law of Persons

(p.7) 1 Global Justice
Democracy Beyond Borders

Andrew Kuper

Oxford University Press

Abstract and Keywords

Develops a cosmopolitan theory of global justice, in critical dialogue with John Rawls’s The Law of Peoples. Kuper argues that Rawls has begged some of the central questions of global justice by adopting a ‘thin statist’ conception of legitimate global order. Thus, Rawls effectively supports a system of unitary nation-states with limited sovereignty, while Kuper rejects this idea in favour of multi-level and multi-type political institutions. Similarly, Rawls disavows free speech and democratic rights at the global level, while Kuper establishes that they are fundamental requirements of global justice. Kuper then proposes a new notion of ‘plurarchic sovereignty’ governed by Principles of Democracy and Subsidiarity. Important practical implications are demonstrated in three areas: economic development, the rules of engagement with illiberal states, and the use of force in humanitarian intervention.

Keywords:   capabilities, cosmopolitan, culture, global justice, humanitarian intervention, international law, Law of Peoples, political liberalism, Rawls, realism, rights, sovereignty, theory of justice, toleration, use of force

John Rawls' The Law of Peoples (LP) represents a culmination of his reflections on how we might reasonably and peacefully live together in a just world.6 My aim in this chapter is to show that a theory of global justice can be developed that is more in keeping with the Kantian constructivist procedures Rawls once employed for domestic justice in Political Liberalism and A Theory of Justice.7 This result is important because it helps establish that my alternative conception of global justice better realises some fundamental liberal values, even on Rawls' own terms.

Rawls has a strong hold on the imaginations of political theorists, but that is not the main reason I have adopted the approach of reading and responding to his work so closely. Rather, it seems to me that his later work at once exemplifies the orthodoxy of ‘liberal statism’ that dominates the field and, along with his earlier work, contains powerful conceptual resources for overcoming that inconsistent and pernicious orthodoxy.

My core argument is that Rawls has begged some of the central questions of global justice by adopting at the outset a ‘thin statist’ conception of the legitimate divisions between persons who share a world. Once this ungrounded assumption is removed, the nature and (p.8) boundaries of the basic political units which the principles of global justice coordinate might look quite different, as might the principles themselves. Although my focus is on ideal theory, on formulating a moral vision of justice in a cosmopolitan world order, the closing section does discuss relevant implications for non-ideal theory.

The chapter is structured as follows: The introductory section outlines Rawls' project and constructivist methodology in LP, with a view to characterising his thin statism in particular. It briefly articulates four arguments for his position. Each of the next four sections explicates and then criticises one of those four arguments and in so doing further develops an alternative conception of global justice. Sections 1 and 2 consider how widely the scope of liberal moral and political concern ought to be drawn, arguing for representation of persons through a global rather than a two-stage (domestic and then international) original position. Section 3 explores the cosmopolitan institutional implications of this modified Rawlsian procedure and elaborates a politically liberal conception of ‘plurarchic sovereignty’. Section 4 defends the relevance of that ideal conception of justice for realistic political action in the decidedly non-ideal conditions of the contemporary world. In closing, I provide brief illustrations of appropriate action in two contested issue-areas: the rules of engagement with illiberal states and the use of force in humanitarian intervention.

The most crucial differences between Rawls and me are the following: (1) he effectively supports a system of unitary nation-states with limited sovereignty, while I reject that whole idea in favour of a more multiform institutional configuration; (2) he disavows democratic rights at the global level, while my argument establishes that the rights to full free speech and democracy are fundamental requirements of global justice. We are thus led to support quite different liberal approaches to the aims and methods of world politics. Indeed, it is the need to reconcile my joint commitments to plurarchic sovereignty and to democracy that makes the argument of the rest of the book necessary: I am compelled to provide a normative theory of democracy for a complex world in which political authority is no longer the predominant preserve of the state.

Prelude: Rawls' Constructivism and Thin States

In LP, Rawls attempts to provide ‘a particular political conception of right and justice that applies to principles and norms of international law and practice’ (LP 3). The question to which this conception (p.9) answers is the following: how can the conception of justice as fairness, elaborated in Political Liberalism for a closed and self-sufficient liberal democratic society, be convincingly ‘extended’ to cover relations between societies, including some non-liberal societies (LP 9)? This question emerges since, after the principles of domestic justice have been decided upon, many issues of justice remain to be resolved—namely, those which arise once the assumption of a closed society is dropped. How is a domestically just society to interact with other societies? Rawls thinks the extension of political liberalism to global justice can be achieved by running a second session of the original position.8

At this ‘second level’, the parties in the original position represent peoples, with the result that the constructivist procedure models conditions for arriving at terms of cooperation that are ‘fair to peoples and not to individual persons’ (LP 17, n. 9). Persons are not the relevant ‘(moral) actors’ precisely because persons' basic claims to justice have already been taken into account (LP 10): the principles of domestic justice are established prior to and independently of the principles of global justice (which are either derivative or compatible), and are given lexical priority. This is the methodological heart of LP: Rawls works upwards and ‘outwards’ from sufficiently just societies (peoples) to a just Society of Peoples (LP 3, 23).

A Rawlsian constructivist procedure has three steps. If each step of the procedure can be justified, then the principles chosen will be fair.9 The first step is to say for whom justice is being derived by answering the question ‘what is a people?’; the second is to identify what alterations must be made to the original position if suitable account is to be taken of the change in moral agents at this second level; and (p.10) the third is to determine which principles of global justice would be chosen by representatives of those agents, deliberating under those procedural constraints on argument.

In his 1993 Amnesty Lecture ‘The Law of Peoples’, which served as a prelude to the book, Rawls initially defined a people as ‘persons and their dependants seen as a corporate body and organised by their political institutions, which establish the powers of government’.10 In the book, he provides a more extensive characterisation of peoples as having, in ideal theory, three basic features—institutional, cultural, and moral. Institutionally, each people has a ‘reasonably just … government that serves their [a people's] fundamental interests’: protecting their territory; preserving their political institutions, culture, independence, and self-respect as a corporate body; and guaranteeing the safety, security, and well-being of their citizens (LP 23–9, 34–5). Each people's citizens are also culturally ‘united by what Mill called “common sympathies”’; Rawls clearly means by this ‘an idea of nationality’, generally based on ‘a common language and shared historical memories’ (LP 23–5).11 Finally, each people has ‘a moral nature’, in that each is firmly attached to a moral conception of right and justice which is at least not unreasonable (LP 23–5, 61–8). Each is prepared—in rationally advancing its fundamental interests—to propose as well as abide by fair terms of cooperation, as long as other peoples do so as well.12

This normative idea of a ‘not unreasonable’ and ‘reasonably just’ people is less demanding than the idea of a reasonable and fully just society as specified in Political Liberalism. Politically liberal societies are certainly included, but so are comprehensive liberal societies (such as that specified in A Theory of Justice) and ‘decent peoples’. The latter are also schemes of social cooperation, but they are associationist, in that persons are respected not directly as free and equal individuals (p.11) but rather as reasonable and rational ‘cooperating members of their respective groups’ (LP 64). This minimal criterion of respect, which defines a decent people, is derived from the basic idea of a bona fide system of law (and not from the idea of persons as free and equal), as follows: A law-governed scheme of social cooperation differs from a ‘scheme of commands imposed by force’ precisely because persons are able to recognise, understand, and be moved to act on the law without necessarily being coerced (LP 65).13 Yet without some reassurance that domestic institutions of justice take some account of citizens' important interests—at the very least as members of groups which each cleave to a comprehensive doctrine—a legal system cannot impose such moral duties and obligations for all members of society, since citizens will not be thus (morally) motivated. The pursuit of the common aims of a decent people must thus be constrained by ‘a common good idea of justice’, which at least takes citizens' important interests into account, thus allowing them all to play a responsible role in public life (pp. 66–8).14 Most significant among persons' interests are those in having their basic human rights secured, and in having laws non-arbitrarily administered (LP 78–81).15

The idea of a decent people is a central innovation of LP. As with any construct, it is created to serve a particular analytic purpose; Rawls' main aim is to develop liberal principles of global justice that are also tolerant of peoples with other moral and political traditions. (The idea of toleration in LP is the subject of Section 2, below.) In order to generate these liberal principles and ensure their acceptability ‘from a decent non-liberal point of view’ (LP 10), the second session of (p.12) the original position is run in two stages—once for liberal peoples, and thereafter for decent non-liberal peoples. As in the case of establishing the fair terms of cooperation for a closed society, parties to each stage are situated symmetrically, behind a veil of ignorance which screens out information (this time it is territory size, level of development, particular common good conception of justice, etc.) which might make them less than impartial in the rational pursuit of the good of those they represent. The strong claim that Rawls makes is that, in virtue of sharing the three minimal features described above, delegates in both stages would independently come up with the same law of peoples.

This result may not seem at all intuitively obvious: why should every decent non-liberal people accept a liberal law of peoples? Rawls reminds us that decent peoples are not unreasonable, and so do not engage in aggressive wars or pursue expansionist ends, nor fail to respect the civic order and integrity of other peoples; thus the delegates of decent peoples would accept the symmetrical (equal) situation of the original position as fair (LP 69). He also reminds us of the common good conception of justice, which takes account of persons' important interests, ensuring that decent peoples would accept principles honouring basic human rights (LP 78–81). Finally, a decent people's fundamental interests—in security, independence, the benefits of trade, and so on—would lead it to accept and adopt the laws of peace (non-intervention, war only in self-defence, restrictions on conduct in war) and duties of contract (observing treaties and undertakings, mutual assistance in times of need) (LP 30–43, 89–113).16 According to Rawls, these are nothing less than liberal principles of global justice.

Notably missing from such a law of peoples are principles for respecting persons as free and equal citizens with constitutional democratic rights; if the latter were included, decent peoples would certainly not accept them. But Rawls wants to draw a clear line between basic human rights (‘liberty rights’ to bodily integrity, etc.), on the one hand, and more extensive liberal democratic rights, on the other. The primacy of this aim is repeatedly emphasised in LP and is based on a stipulation that ‘all persons in a decent hierarchical society are not regarded as free and equal citizens, nor as separate individuals deserving equal representation … they are seen as decent and rational and as capable of moral learning as recognised in their society’ (LP 71). He insists that this exclusion of persons in decent societies from treatment as fully free and equal individuals is required by liberal conceptions (p.13) themselves: liberal peoples must express toleration for decent non-liberal ways of ordering society. (In Section 2, below, I evaluate and criticise the basis for this stipulation and conclusion.) Drawing the line in this place does allow Rawls to address a major programmatic concern, by identifying what liberals should not tolerate: ‘We must reformulate the powers of sovereignty in light of a reasonable Law of Peoples and deny to states the traditional rights to war and to unrestricted internal autonomy … included in the (positive) international law for the three centuries after the Thirty Years' War’ (LP 25–7).

This claim needs careful interpretation, lest it appear more radical than it is. Rawls endorses the existence of sovereign states and of an international state system, with the important caveat that such sovereignty is not absolute. When he writes that peoples are not ‘states as traditionally conceived’, he means only to ‘emphasise’ that his conception of states is very far from the traditional Realist conception of states as predominantly concerned with power (LP 25–7). Realist states pursue their ‘rational prudential interests’ in power, unconstrained by ‘the reasonable’, and are thus unmoved by the criterion of reciprocity; Rawlsian peoples have moral conceptions of justice and regimes which ‘limit their basic interests as required by the reasonable’, but they are still states (LP 28–9). Indeed, as we have seen, they are nation-states, each with a single independently derived system of law, and a ‘so-called monopoly of power’ on the enforcement of that law, and on the pursuit of persons' politically important interests, in a particular territory (LP 23–6, esp. n. 20 and n. 22). The difference is that in Realist theory the shell of state sovereignty may not be pierced or removed if and when the regime acts unjustly—this exemplifies what I shall call ‘thick statism’—whereas in Rawls' theory, the law of peoples reasonably constrains what a state may rightly do to its own people and other states—this exemplifies what I shall call ‘thin statism’ or ‘liberal statism’.

The crucial methodological question, which Rawls himself asks, is why this issue of extension—from justice within a closed society to international justice—is what a Rawlsian theory of global justice ought to address. Why are peoples assumed to be the politically relevant subjects with which to start? Rawls himself once pointed to much the same question, as follows: ‘Wouldn't it be better to start with the world as a whole, with a global original position, so to speak, and discuss the question whether, and in what form, there should be states or peoples at all?’.17 At the time, he had ‘no clear initial answer to this (p.14) question’; indeed, he saw no reason why such a starting-point would not result in the adoption of exactly the same principles.18 In the book, however, his reasons for preferring a thin statist procedure can now be discerned. I term them the arguments from incorporation, toleration, cohesion, and realism:

  1. 1. Incorporation—if peoples are stipulated to take members' interests into account, and all persons are members of peoples, then all persons' interests are fully accounted for and given due consideration.

  2. 2. Toleration—liberal principles require respect for other cultures and ways of ordering society, and so imposing on them a conception of global justice based on the idea of persons as free and equal would be wrong.

  3. 3. Cohesion—the alternative to a Society of Peoples is an illiberal, strife-torn world state; thus, even if the former involves some injustice, it is preferable.

  4. 4. Realism—as a practical matter, to best secure the great goods of world peace and respect for human rights, liberal regimes should fully engage decent non-liberal peoples rather than excluding them from international forums and law.

In the ensuing four sections, I explicate and rebut these arguments for liberal statism in turn; in doing so, I show how an alternative conception of global justice might be developed from less unsatisfactory basic assumptions. To put it another way, Rawls has not gone far enough in distancing himself from the Realists; he still tolerates too much.19 I sketch a theory of global justice that is not statist at the outset, and is, I argue, more in keeping with political liberalism.

1. Incorporation: Different Interests of Persons and States

‘In laying out the Law of Peoples, we begin with principles of political justice for the basic structure of a closed and self-contained liberal democratic society’ (LP 86).

(p.15) Familiarity with Rawls' theory of justice should not mask just how odd it is to take ‘society … as a closed system,’ ‘self-contained and … having no relations with other societies’,20 as the founding assumption of a theory of international or, better, global justice. As Onora O'Neill points out, this assumption is not the mere ‘considerable abstraction’ that Rawls claims it is, ‘since abstractions (taken strictly) omit or bracket certain predicates true of the matter from which they abstract. Rather the idea of a closed society is an idealisation, that assumes predicates which are false of all existing human societies’.21

Now, like the idea of a frictionless surface used in natural science, this idealisation is not necessarily objectionable, so long as there are very strong arguments for why the false construction can, by analogy or resemblance (for no strict inference to a true conclusion is possible), show something useful about cases that are not idealised.22 Rawls would maintain that the idea of a closed society is a useful device for representing persons' fundamental interest in having basic human rights respected in their own society; the same interest will lead each society to endorse a global legal framework which supports societies' respect for rights, and in this way, all persons of the world can have their rights respected. The underlying idea here is this: if both peoples and persons are stipulated to have a fundamental interest in basic human rights, then their interests coincide. If this were true, persons would in no way be disadvantaged by starting from societies and not persons, and societies could form the basis for a stable global human rights order.23

But I now argue that there is a strong presumption against Rawls' idealisation: the assumption of a closed society obscures the fact that (p.16) the interests of persons and of peoples do not necessarily coincide. So even if a confederation of peoples secures urgent rights, it may well do so in a less than optimal way; other institutional configurations may better secure persons' basic rights as well as other rights and fundamental interests.

Do peoples' and persons' interests necessarily coincide? There is good reason to think not: depending on how subjects are divided into sets at the outset, the outcomes of reasonable and rational deliberation—about what their interests are and how best to pursue those interests—will differ. Consider the following example.24 In a world of two states, U and D (Underdeveloped and Developed), the government of each intends to act rationally so as to secure the interests of persons in their territories to the maximal extent possible. It might be rational for D to restrict immigration because the cost of supporting new residents would result in a slight reduction in standards of living for its current citizens; and it might be rational for U to restrict emigration, because it would deplete the skills base for securing current citizens' rights and well-being.25 If two parties representing these states, though they did not know which, had to establish a law governing their relations, it would be one that allows for only highly restricted movement of persons between the two from U to D.

Yet it is not true in principle that this law best secures the rights and well-being of all the persons in both countries. It may be the case that allowing some more movement of people between the two would result in a gain for those who are worst off or even in a more extensive scheme of basic liberties for all: a minor worsening of the well-being situation of those who were citizens of D and for those left behind in U might make immigrants from U significantly better off, sufficiently to justify the movement. This is not, however, a consideration which could count for parties representing U and D's respective citizenries separately, but only for parties representing all the persons in U and D at once, as individual persons.

This example evidences a more general point about social choice: what is rational to agree upon at the level of two parties representing (p.17) two sets of persons' interests (that together exhaust the set of existing interests) is not the same as what is rational if it is the interests of each and every person that are being considered. Thus there is no reason to think that what proves—as Rawls put it26—‘more or less sound’ for one domain (justice for persons in a closed society) is appropriate to another (global justice for persons), any more than there is reason to think that the principles for packing eggs into padded boxes are extendable to the principles for packing egg-boxes into a crate. Nor is it apparent that the sequence should be to design egg-boxes first and only later ask questions about how to design the crate. Therefore Rawls' theory of domestic justice might provide tools for the independent construction of global justice, but it cannot simply be incorporated as the first step in that construction. Since the idea of decent peoples as a starting-point embodies two layers of distortion (ascriptive associations, thin states) in representing individuals' interests, liberals—for whom individual persons are the ultimate locus of concern—should be deeply wary.27

We have seen that the interests of all human individuals and those of the same persons assumed to be grouped as members of states do not necessarily coincide, and that we may come to have good reason to jettison thin statism in favour of a global original position which represents all the persons of the world. But Rawls might object that my example concerns interests that are not ‘fundamental’, and so risks impugning a possible global human rights order by raising less urgent (socioeconomic) claims. This objection is telling only if one accepts the implausible stipulation that the important interests of persons can be narrowly confined to barely adequate domestic justice only. But it is profoundly counter-intuitive to assume that parties should take no interest at all in the well-being or standard of living of persons, ‘beyond the minimum necessary for [minimally] just institutions’.28 (p.18) Rawls himself acknowledged that a just regime cannot be a final and circumscribed end in itself, rather it is ‘something we ought to realise for the sake of individual human persons, who are the ultimate units of moral concern … Their well-being is the point of social institutions’.29

Of course, some interests are more important than others, and it might be thought that a thin statism secures the most important interests of all persons, in world peace and respect for minimal human rights. But this begs the question: it cannot be assumed that thin states best secure persons' important interests; if states do so, then that is something that will count for parties representing individual persons, thus parties will endorse thin states (as the basic institutions of global justice). There is the greatest difference between a liberal constructivism which takes thin states as a possible outcome of the procedure, on the one hand, and thin statism—which assumes states as foundational to global justice—on the other. The former leaves two possibilities open: (1) thin states may not best secure those important interests; and (2) there may be an alternative which secures those interests and more, such as added security and increased well-being. (I offer such an alternative in Sections 3 and 4, below; here I have simply established that these are live issues.) In sum, because of its potentially suboptimal results for persons, any initial demarcation of groups must be justified. Rawls' first main argument—that persons' fundamental interests would be addressed already by peoples since peoples take members' interests into account—fails to justify his basic assumption of thin statism.

2. Toleration: The Universal Scope of Global Justice

We have seen that the law of peoples may represent the fair terms of cooperation for peoples, but it certainly does not necessarily represent the fair terms of cooperation for all the persons of the world; this is a serious concern for the liberal. But Rawls' most powerful and explicit argument might be thought to provide reasons to override this concern, since it stresses the overwhelming importance of recognising—by starting with the idea of peoples—the value of (p.19) national–cultural affiliation:

Leaving aside the deep question of whether some forms of culture and ways of life are good in themselves (as I believe they are), it is surely, ceteris paribus, a good for individuals and associations to be attached to their particular culture and to take part in its common public and civic life. In this way political society is expressed and fulfilled. This is no small thing. It argues for preserving significant room for the idea of a people's self-determination and for some kind of loose or confederative form of a Society of Peoples (LP 61).

The common sympathies arising out of a shared history and tradition are profoundly valuable to individuals, and an adequate theory of global justice must recognise and respect that fact—rather than insensitively and destructively ignoring it. I am in full agreement with Rawls that it would be foolish and wrong not to recognise the value of culture to individual persons; but the question is not whether to tolerate cultures, rather it is how to do so. In this section I argue—from politically liberal premises—that the Rawls of LP seeks toleration of the wrong kind. Only an original position that includes all the persons of the world as free and equal persons can express toleration in the right way.

Rawls' argument for toleration of decent non-liberal peoples seeks to establish that this kind of toleration is required because of features internal to liberal justice theory. His argument proceeds by analogy to domestic justice:

If all societies were required to be liberal, then the idea of political liberalism would fail to express due toleration for other acceptable ways (if such there are, as I assume) of ordering society. We recognise that a liberal society is to respect its citizens' comprehensive doctrines—religious, philosophical, and moral—provided that these doctrines are pursued in ways compatible with a reasonable political conception of justice and its public reason. Similarly, we say that, provided a nonliberal society's basic institutions meet certain specified conditions of political right and justice and lead its people to honor a reasonable and just law for the Society of Peoples, a liberal people is to tolerate and accept that society (LP 59–60).

A law of peoples, then, embodies ‘principles of the foreign policy’ of a liberal people, where the requirement of toleration of other societies' comprehensive doctrines is met by ensuring that such policy norms could also be acceptable ‘from a decent non-liberal point of view. The need for such an assurance is a feature inherent in the liberal conception’ (LP 10).

This argument is unconvincing because it trades on a partial analogy between peoples and persons—organising their respective ‘lives’ around reasonable comprehensive doctrines—that Rawlsian constructivism (p.20) cannot sustain and liberals should not endorse. States, even thin states, institutionalise political coercion, and any coercive institution raises questions about its legitimacy. Rawls recognises this element of disanalogy and tries to deal with it by stipulating that each people simply is legitimate, in virtue of having decent institutional features. To tolerate decent societies is, then, to tolerate what is sufficiently tolerant of persons already, all with a view to achieving broad agreement on common principles of justice. But here Rawls is mistaken.

‘Sufficient tolerance’ is not simply a pale approximation of full liberal tolerance; rather the two are deeply contradictory. Liberal tolerance expresses ethical neutrality, by remaining impartial between particular moral conceptions of the good; for this very reason, liberalism must reject any political neutrality, that is, neutrality in respect of justifications for coercion: ‘a commitment to ethical neutrality necessarily entails a commitment to a particular type of political arrangement, one which, for one, allows for the pursuit of different private conceptions of the good’.30

As Thomas Pogge put it, while a society or world can contain numerous associations and conceptions of the good, its basic political structure ‘can be structured or organised in only one way … There is no room for accommodation here’ since it is precisely the characteristic of a fundamental law backed by coercive force that it must apply to and be justifiable to all.31

The idea of tolerance in LP is, then, fundamentally different from and opposed to—and not simply a less demanding version of—the idea of a liberal regulatory framework presented in Political Liberalism. There Rawls argues repeatedly that it would be intolerant, oppressive, and unjust for a state to be organised around any one comprehensive doctrine—precisely because such endorsement fails to respect other such doctrines, and the persons that hold them.32 And there he is correct. It is the essence of a politically liberal regulatory framework that it expresses toleration by not incorporating any comprehensive doctrine in the principles of justice; to fail to do so is not to extend but rather to eliminate liberal tolerance. In LP, on the other hand, he is mistaken. Decent peoples are not ethically neutral, nor is a Law of Peoples which recognises their comprehensive doctrine ethically (p.21) neutral; thus at neither stage is there any basis for saying that what is being expressed counts as liberal toleration.33

One can imagine Rawls making the following reply: ‘The idea of ethical toleration in a liberal society is all very well, but it is much too stringent for global toleration. For one thing, it is simplistic to regard decent peoples as ethically intolerant; they are better characterised as ethically “not intolerant” [my term, though Rawlsian in spirit]. Unlike outlaw states, decent peoples show significant respect for persons not only by honouring basic human rights but also by allowing associations in civil society that hold a range of comprehensive moral doctrines. Further, the political structure of a decent society, although organised around a comprehensive common good conception of justice, does not entirely reject citizens' comprehensive doctrines. One condition stipulated for a decent people is, as we saw, that citizens' interests must be taken into account; but for this to be the case, “the basic structure of the society must include a family of representative bodies whose role in the hierarchy is to take part in an established procedure of consultation and to look after what the people's common good idea of justice regards as the important interests of all members of society” (LP 71). This is a strong demand for what might be called a “decent consultation hierarchy” (LP 71–8). It cannot be said that where peoples have these features, there is no bona fide system of law: extensive consultation ensures an “institutional basis for protecting the rights and duties of the members of the people”, thus persons have their interests taken into account and can be morally motivated to obey the law (LP 71). These protective, expressive, and deliberative features “deserve respect, even if their institutions as a whole are not sufficiently reasonable from the point of view of political liberalism” (LP 84). Ethical toleration is not an on–off affair: a decent people is structured around one comprehensive doctrine only in a very limited way and does pass an adequate threshold of respect for persons; we must, in turn, respect the basic institutions of such a people.’

This argument appears to be appealing in that it seems to take cultural pluralism seriously, but—I now want to argue—it does so by not taking seriously the reasonable pluralism of individual persons. Consider, for one thing, some profoundly anti-liberal implications of (p.22) ‘toleration’ as specified in LP:

The fact is that none of Rawls's ‘well-ordered’ hierarchies will be free of natives who are themselves inspired by liberal ideas of liberty and equality. There is no Islamic nation without a woman who insists on equal rights; no Confucian society without a man who denies the need for deference. Sometimes these liberals will be in a minority in their native lands; but given the way Rawls defines a ‘well-ordered’ hierarchy, it is even possible that they might be a majority … [Why] should we choose to betray our own principles and side with the oppressors rather than the oppressed?34

When a liberal regulatory framework recognises a decent hierarchical regime as sufficiently just, it participates in the denial of freedom and equality to such individuals. Dissenting individuals with liberal views would surely, it seems, dispute the idea that accommodation of reasonable pluralism requires that their individual moral claims be taken less seriously. But then one could not really know what they would think, since their views could well be sealed off from view by the decent consultation hierarchy.

To take only one significant instance, a regime with a decent consultation hierarchy does not allow free speech: persons may only dissent as members of associations, and only with reference to the common good conception of justice (LP 72–5). It follows that citizens must argue within the conceptual terms of the regime, and only through representatives of a group; this closes off large domains and numerous types of discussion. The most serious exclusion is that it prevents proper critical discussion of how the rules of discussion might be altered—the latter are determined by some interpretation of the dominant cultural tradition only (though it must make some room for the survival of others). A claim that this represents ‘freedom’ of speech but ‘not equal’ freedom of speech is thus farcical, as is Rawls' claim that barring persons of certain religions from occupying public offices represents ‘liberty of conscience, but not equal liberty’ (LP 65, n. 2).

These so-called inequalities are in fact serious restrictions on liberty which would rightly horrify a liberal at home, and it is not apparent that they should be any less rightly horrifying when perpetrated against people that are not part of one's liberal society. This would certainly be apparent to parties in a single global original position who, when the veil lifts, might find themselves dissenters in a non-liberal society. Rawls might say that other societies do not share and cannot be expected to share our conception of the person as free and (p.23) equal, and so no such original position can be constructed, but this misses the point. From a liberal perspective—for reasons Rawls himself has done most to elaborate—the geographical location and group membership-status of a woman born into an inegalitarian Islamic state, or indeed anyone else, is morally arbitrary. She does not suddenly come to be a free and equal person for us when she crosses the border into a liberal society. On the contrary, as Charles Beitz points out, ‘Although the basis of the [liberal] conception of the person may be parochial, the conception itself … is not … One might say that we are compelled to take a global view in matters of social justice by features internal to our conception of moral personality, however parochial it may be.’35

It is this most basic internal feature—the respect for persons captured by the idea of ethical toleration—that must be the cornerstone of a consistent liberal global regulatory framework. Decent hierarchical peoples could not agree with it, but that is precisely the ethical problem with them, and not grounds to seek their agreement on some lesser mixture of respect in some parts and disrespect in others. As Rawls argues on the second page of A Theory of Justice, ‘Being first virtues of human activities, truth and justice are uncompromising.’36

The foregoing points are no less true of non-metaphysical liberalism than they are of a metaphysical liberalism. Political liberalism starts from ideas implicit in a liberal democratic culture, but none of the major intellectual figures who are taken to be progenitors of that culture—Locke, Rousseau, Kant, Mill, to name but a few—does not begin with some idea of all persons as free and equal. Each thinker then goes on to justify the state or something like it, on the very grounds that such an institutional formation is in some way rational to will for persons thus construed.37 Indeed, even such a vociferous opponent of liberalism as Carl Schmitt is clear that liberal community rests on the idea of a ‘democracy of mankind’—which, though maybe not practically achievable, is philosophically universalist at its core: ‘Every adult person should eo ipso be politically equal to every other (p.24) person’.38 To the extent that the moral claims of states have any normative force in liberalism, that force is derivative—states must be justified. In political liberalism, we do not close off the possibility that parties representing free and equal persons in a global original position would decide in favour of thin states, or even in favour of an inferior position for that woman within a particular state (though I doubt they would); rather we say that thin states, and her occupying this position, must be justified.39

What then about the good of community? It would be a mistake to interpret my cosmopolitan position as a form of abstract individualism. A global original position does not rule out people banding together in communities with special bonds of sentiment and obligation between them; all it demands is that such a form of organisation—as Rawls himself once wrote—cannot be assumed as foundational or not subject to justification: ‘we want to account for the social values, for the intrinsic good of institutional, community and associative activities, by a conception of justice that in its theoretical basis is individualistic’.40

If loyalties and sentiments of affiliation to particular cultural and national groups have value for the members of those communities (and I believe they do), then ‘on a cosmopolitan point of view, this fact should matter for practical reasoning. The important question is not whether it should matter but how’.41 This is the question that—I have argued—LP does not properly consider. The fact that community and solidarity enrich and partly constitute a valuable human life should not block consideration of the implications of such arrangements for (p.25) non-members and dissenters. A liberal background culture implies universalist justification. That is to say, the importance of cultural differences does not obviate the requirement to refer in the last instance to individual lives and not to a social formation as ‘an organic whole with a life of its own distinct from and superior to that of all its members in their relation to one another’.42 These are Rawls' words. To say that a social milieu or institutional formation is not automatically sealed away from critical scrutiny—by minimal gestures towards human rights and consultation—is not to abstract from real individuals; rather it is to treat their claims to moral consideration, including their cultural claims, entirely seriously.

3. Cohesion: Towards Non-statist Principles of Global Justice

Rawls might be taken to impugn my conclusions above, in one fell swoop, with his third argument, as follows: On the second page of A Theory of Justice, another statement can be found: ‘an injustice is tolerable only when it is necessary to avoid an even greater injustice’.43 Now, although ‘the social world of liberal and decent peoples is not one that, by liberal principles, is fully just’, there are ‘strong reasons’ for ‘permitting this injustice’ (LP 62). One primary reason is that a world state—which Rawls might also think is the outcome of a global original position—would have even greater drawbacks (i.e. cause even more injustice) than a law-governed Society of Peoples:

These principles … will not affirm a world-state. Here I follow Kant's lead in Perpetual Peace (1795) in thinking that a world government—by which (p.26) I mean a unified political regime with the legal powers normally exercised by central governments—would either be a global despotism or else would rule over a fragile empire torn by frequent civil strife as various regions and peoples tried to gain their political freedom and autonomy (LP 36).

He also approvingly cites Kant's dictum that ‘laws always lose in vigour what government gains in extent’.44

Rawls seems then be making one or both of the following psychological claims: persons as they are would not, on an ongoing basis, morally affirm a world state; and the centralisation and cumulation of power in a world state would encourage extreme administrative abuse, laxity, or ineffectuality. It follows that although a confederation of peoples may have some illiberal consequences, these are far less severe—given persons as they are—than the consequences of concentrating power excessively in a world state.

But, leaving aside whether this is a plausible view of moral psychology (and of institutions), we must ask: are these the only two options? We need not debate the relative merits of two illiberal conceptions of the outcomes of the global original position if there is another alternative that does not fail to be liberal. In this section I present an alternative, albeit a modest sketch. As is appropriate at this phase of Rawlsian theory, my conception is an ideal; its relevance and feasibility in non-ideal conditions for justice are considered at length in the next section.

We want principles of justice to regulate a global institutional scheme, principles which are not statist in their assumptions. It will be a cosmopolitan conception because it requires institutions to meet three criteria: taking individual human persons as the ultimate units of concern (individualism); attaching that status to every human being equally (universality); and regarding persons as the ultimate unit of concern for everyone (generality).45 It will be a Rawlsian conception in that it uses the original position as a device to represent conditions for agreeing on fair terms of cooperation for all. But ‘all’ will not be defined as a closed community, involved in a scheme for mutual advantage, that needs to agree on rules of engagement with other peoples. Rather, the ‘parochial’ assumptions of liberalism require that parties in the original position act as if they represent all human persons who share this world and affect one another. These parties will be far more concerned with individuals' abilities to pursue their reasonable (p.27) conceptions of the good, and with individuals' capabilities and well-being, than would be delegates of peoples (thin states). It is not possible in this limited space to consider the many issues on which parties would decide, therefore I concentrate on only one central issue: the nature and limits of sovereignty—its appropriate moral bases and political extent.

In constructing a Law of Persons to address the question of sovereignty, I begin by recalling what the original position does: it embodies constraints on substantive argument for principles of justice; it tells us what kinds of reasons cannot count. By making some substantive points about why the notion of a people cannot be assumed to ground even thin state sovereignty, I want to show what reasons cannot count for parties in a global original position. Thereafter, I argue, by considering what kinds of non-arbitrary reasons are left—as legitimate bases for principles of global justice—we can get a surprisingly long way toward an outline of just cosmopolitan institutions.

The first thing to note is that there is a veritable catalogue of empirical difficulties in identifying any people that is not deeply contested in practice or that clearly coincides with a particular political boundary.46 Each difficulty can be seen to underpin a reason why the idea of a people cannot justify legitimate divisions between sovereign political entities.47 Now, it might be thought that such a catalogue, while (p.28) sufficient to reject closed and organicist views of the nation, does not constitute an argument against Rawls' liberal account of peoplehood. For one thing, Rawls fully acknowledges that

if those [common] sympathies were entirely dependent upon a common language, history, and political culture, with a shared historical consciousness, this feature would rarely, if ever, be fully satisfied … Notwithstanding, … [there is a] need for common sympathies, whatever their source may be. My hope is that, if we begin in this simplified way, we can work out political principles that … enable us to deal with more difficult cases where all the citizens are not united (LP 24–5).

For another,

It does not follow from the fact that boundaries [of thin states] are historically arbitrary that their role in the Law of Peoples cannot be justified. On the contrary, to fix on their arbitrariness is to fix on the wrong thing. In the absence of a world state, there must be boundaries of some kind, which when viewed in isolation will seem arbitrary, and depend to some degree on historical circumstances (LP 39).

Such arguments do not, however, answer the empirical catalogue of critique: it may be necessary to have simplifying assumptions and historically contingent boundaries, but this does not show that these particular assumptions (those of Rawls in LP) are not bad ones. For instance, it is not apparent that bonds of sympathy need be primarily between citizens of thin states. It is true that many people have cultural allegiances, but persons have many other legitimate allegiances too, and the idealisation of a homogenous nation removes the possibility of any basic political consideration of how and to what extent claims arising from these allegiances ought to be prioritised. Those other claims are only accommodated once the basic regulatory framework has been determined, in favour of an international thin state system—to which (p.29) those with other allegiances must adjust (gaining as much respect as is possible within that system). Rawls has confused the putative value of common national sympathies with their moral primacy for establishing political institutions. The effect is to give peoples or nations a veto on what identities and bonds persons may take to be of predominant political significance. Yet this cannot be assumed to be just: ‘A central objective of politics may be the reconstrual of political identities, the separation or the merging of destinies rather than the working out of principles of justice to be shared within a closed society.’48

It is not apparent that sovereignty has to be located at one level, nor is it evident that there are not better bases for borders—bases which take different historical contingencies into account, and for good reasons. One of the most interesting attempts to reformulate the notion of sovereignty so as to encompass these complexities is that of Thomas Pogge in ‘Cosmopolitanism and Sovereignty’. He proposes a multilayered institutional scheme in which the powers of sovereignty are ‘vertically dispersed’ rather than concentrated almost entirely at the level of states:

What we need is both centralization and decentralization … Thus, persons should be citizens of, and govern themselves through, a number of political units of various sizes, without any one unit being dominant and thus occupying the traditional role of the state. And their political allegiance and loyalties should be widely dispersed over these units: neighbourhood, town, county, province, state, region, and world at large. People should be politically at home in all of them, without converging upon any one of them as the lodestar of their political identity.49

(p.30) He argues that dispersing governmental authority over such ‘nested territorial units’ would have significant benefits, such as reducing the stakes and hence the intensity of the ‘struggle for power and wealth within and among states, thereby reducing the incidence of war, poverty, and oppression’ and environmental degradation.50

An obvious objection to this idea is the Hobbesian claim (which makes its way into much of the later social contract literature) that there must be a final decision mechanism that uniquely resolves any dispute—thereby preventing formal, ongoing, destructive conflict—and this can only be a supreme agency of last resort. As Pogge points out, though, the history of the last two hundred years—particularly the success of division of powers within states—attests to the possibility of law-governed coexistence even when ultimate conflicts between legitimate powers are theoretically possible. The three traditional branches of government within states rarely engage in all-out power struggles, and when more minor constitutional crises do arise they tend to be rapidly resolved.51 Similarly, the vertical division of sovereignty in federal regimes—in ways that leave open some conflict over constitutional allocation of powers and hence no authoritative path of resolution—has proved a remarkably effective underpinning of robust and enduring institutions in many cases (including the United States, the United Kingdom, Switzerland, Germany; the list is long …). It therefore seems mistaken rigidly to insist that sovereignty must be located in one place in the last instance.

Dispersed sovereignty of some kind may well be effective and highly beneficial—though I have not yet said which kind, if any, is most promising (see Chapter 3). For now, note simply that Rawls thinks we must rid ourselves of the cult of unqualified state sovereignty, but he does not see that the cult of unitary sovereignty may (p.31) similarly deserve to be consigned to the flames. In any case, Pogge's suggestion stands as at least one broad, plausible, non-statist (in its assumptions and outcomes) alternative to both a world state and a state-dominated system. Rawls cannot assume that the problem of a world state provides, by process of elimination of alternatives, a justification for representing peoples in the original position, nor for endorsing them as the primary political configurations in the principles of global justice.

However, it seems to me that even Pogge does not go far enough, in that he has imported an unjustified assumption that sovereignty can and should only be exercised over territorially defined units. It seems that we may have escaped the cults of unqualified and unitary state sovereignty only to fall prey to the lesser cult of territorial sovereignty. Why should the parties in the original position accept this arbitrary restriction on the domain of government? Why should sovereignty not be dispersed horizontally as well as vertically? Considering the numerous issues that territorial demarcations of governmental functions could not (best) resolve—for example, crime on the Internet, prosecution of violators of human rights, and environmental protection—there seems good reason to divide the tasks of governments on functional rather than territorial lines.52 Some governmental functions may be best exercised within territorial demarcations, and some groups of functions may coincide at various levels of Pogge's vertical scheme; but it is unlikely that they would exhaust all governmental functions.

But how are we to understand these ‘functions’? The rapid pace of globalisation and technological innovation in several areas suggests that non-territorial spaces of interaction (from web pages to financial markets) will have an increasingly significant role in human affairs. It may thus be more appropriate to think of functional sovereignty as (p.32) legitimate power over kinds of human practice and resources.53 In considering the political authorities that would best fulfil various government functions, parties to the original would position consider territory—restricted or global—a derivative basis for inclusion (along with syndicalism, to mention at this point only one other). What is needed, then, is a conception of political agencies that appropriately regulate different spheres of human action, and since not all spheres of action are primarily territorially based, neither need those authorities always be.54 Persons would still band together over various functions, many of them territorially and communally based, but there would by no means be the overweening concentration of legitimate power in states or a state system.55

One initial model for such a world order is to be found in the literature on polyarchy, which looks to smaller, diverse metropolitan areas such as New York or Los Angeles as a guide to a more cosmopolitan vision:

These megalopolises comprise numerous municipal and special-purpose jurisdictions, many of which overlap, but they lack a strong central government for the whole metropolitan area. The polyarchic ‘global city’ could exhibit an analogous structure … [T]ransnational institutions, with memberships and spans of control congruent with the interdependent relationships, could be accorded with responsibility for co-ordination, rule making, and even rule enforcement.56

(p.33) In such a world order, the dispersal of jurisdictional authority over a plurality of agencies would of course be limited, on functional grounds, by a need for effective coordination. Each will operate ‘on the lowest possible level’, but there will generally be advantages to assigning jurisdiction over several functions to each agency, and benefits to creating agency-clusters—not least that it is then possible (as I have shown in Chapters 3 and 4) to exercise greater democratic control over governance. Grounds of this kind for allocating authority are already recognised in Article 5 of the Treaty Establishing the European Community, which endorses both a Principle of Democracy and a Principle of Subsidiarity (this is the principle that powers and tasks should be vested in subunits, ‘as close as possible to citizens’, unless a more encompassing unit can better achieve the specified goals).57

The Treaty that enshrines these principles is a pioneering practical document for organising non-unitary sovereignty; as is well known, however, it retains a bias in favour of allocating powers, including powers to shape decision rules, to states. Both subsidiarity and democracy are understood as delegative principles: the authority of localities, regions, and Union is conferred by states (as a legal, not a merely historical matter), and the authority of states to delegate in this way is officially derived, in good part, from the purported fact that those states are internally democratic. These restrictions, which are applied to the detriment of other levels and kinds of political formations, are unlikely to find a place in a non-statist, non-territorialist order. Instead, that order would be governed by principles of distributive subsidiarity and democracy—principles that do not presume at the outset the (superior or original) legitimacy of states, or the transitivity of intra-state democratic mandates.58

There are, then, non-arbitrary reasons for drawing boundaries of sovereignty—boundaries that are historically contingent only in the sense that they take account of the best current means to reach the ends (p.34) of free and equal persons.59 The result would be what I am calling a system of functionally plural sovereignty—or ‘plurarchy’ for short. This is merely to sketch the broad outlines of an alternative global institutional configuration, one that cannot be assumed out of parties' purview arbitrarily by stipulating peoples—thin, and often illiberal, states—either as the represented subjects or as the necessary institutional outcomes of appropriately constrained practical reasoning.

4. Realism: Practical Application in a Non-ideal World

‘But’, an imagined critic might say, ‘your liberal principles of global justice will put us at odds with hierarchical regimes: they will not accept the privileging of the individual as free and equal, and so will not endorse the same principles, and we will surely come into conflict. Whereas Rawls at least recognises pragmatic limitations and strategic difficulties, the cosmopolitan scheme you have provided is uncompromisingly idealistic and of little help as a guide to political action. What are you going to suggest that liberal regimes do to decent non-liberal regimes? Are they to be subject to military attack, colonisation, or intervention in their domestic affairs? Are they to be ostracised from cooperative international schemes altogether? Rawls gives us good reasons why liberal regimes must—by their own principles—rule out these kinds of action. Not only that, he gives us good reasons to engage such societies, so as at least to ensure minimal adherence to international legal rules of non-aggression and minimal respect for human rights. And he even gives us reasons to think that non-liberal regimes would accept engagement on these terms. In any case, setting the ideal too high may lead to frustration and disillusionment for liberals, and that would adversely affect progress on peace and human rights issues. Finally, tolerance of this injustice is not as distasteful as you claim. Liberals are still able to criticise non-liberal regimes, since acceptance of decent peoples in international law by no means implies endorsement of their principles by liberals more generally nor does it require that non-liberal regimes are viewed as beyond reproach. Rawls has provided us with a “realistic utopia” (LP 7, 11–23); you have given us an impracticable pipe-dream’.

This forceful critic would surely be right in one important respect: the distance between an imagined cosmopolitan world and current (p.35) grim realities ‘is so great that it would be madness to use the conclusions of ideal theory as the unmediated basis for a practical application program’.60 However, in this section I want to argue that the cosmopolitan scheme I have presented is a more useful ideal than that presented in LP, on two grounds: Rawls is neither sufficiently utopian nor sufficiently realistic. This may seem paradoxical, but careful consideration of the relationship between ideal and non-ideal theory reveals that Rawls has created a unitary term—‘realistic utopia’—by watering down both its elements; further, a perspective which maintains the distinct role of each element is far more practical. In what follows, I discuss each element in turn.

Rawls' argument for his (in my view, limited) utopianism is in part an argument from stability: the situation of ideal justice must generate ongoing support and not be subject to ‘assurance problems’ arising out of shifts in power. It must be a sustainable ideal. But if this is to be achieved at the global level, he insists in LP, the standards of the reasonable need to be ‘relaxed’, since treating all persons as free and equal—as having the two moral powers regardless of culture or location—‘makes the basis of the law of peoples too narrow’.61 Actual peoples do cleave so strongly (and not unreasonably) to their ‘different cultures and traditions of thought’, and a liberal world order will constantly be faced with civil strife if it expects them to sacrifice what is most dear to them in order to endorse global legal rules and institutions (LP 11): ‘Historically speaking, all principles and standards proposed for the law of peoples must, to be feasible, prove acceptable to the considered and reflective public opinion of peoples and their governments.’62

This is not simply a matter of actual recognition (fortunately for Rawls, since in Section 2 we saw that this argument must fail), rather it is a matter of ‘speculation’ on what is ‘feasible and might actually exist’, given that persons do in general organise themselves into peoples (LP 12–13). Liberals should thus adopt ‘a minimum standard of realism which requires that the law of peoples not call into question the existence of the international state system’, at least where that refers to a confederation of what I have called thin states.63

(p.36) The fatal flaws in this argument are immediately apparent if we recall the meaning of stability in Political Liberalism: a regime is stable when ‘members will tend increasingly over time to accept its principles and judgements as they come to understand the ideas of justice expressed in the law among them and appreciate its benefits’.64 Breadth of agreement, Rawls writes there, can establish a wide modus vivendi—an agreement based on prudential considerations, which is therefore unstable—but not an overlapping consensus, which is the only kind of agreement that is stable. The former is a question in non-ideal theory of how assent could be won from within current societies, given that so many are organised into states that cleave to comprehensive doctrines; the latter is a question of ideal theory, and involves a moral affirmation by the politically relevant subjects of the social framework regulated by principles of justice.65 Who those subjects are is not settled by the ease of achieving assent from peoples rather than individual persons (once the veil is dropped). Rather it is settled by (1) who is an authentic source of moral claims in this domain; and (2) whether those agents would continue to support the resultant conception of global justice, even if shifts occurred in their conceptions of the good.

Regarding (1), I have tried to show not only that peoples (thin states) are not self-authenticating sources of valid moral claims but also that their having taken each person in each territory's minimal interests (in liberty rights and peace) into account is insufficient to secure that authentic status. It is in any case deeply questionable whether an historical analysis would guide us to an acceptance of peoples as the politically relevant subjects. The horrors of nationalistic wars, xenophobia, and unnecessary starvation might motivate instead a greater focus on human individuals regardless of their geographical location, and—as Pogge argues—on lowering the stakes and incentives to abuse that attach to each level or locus of authority.66 If history suggests anything, it is that we should scrupulously interrogate and dismiss assumptions which might be destructively ‘trapping us in the buildings and boundaries’ of the past or present.67

(p.37) The goal of any Kant-inspired utopian political theory of global relations thus cannot be to show which principles are likely to be accepted at present (by the powers that be or by persons simpliciter, with all our distorting prejudices). Rather, it is to specify which principles ought to be accepted by those subjects in this domain, considered in terms of the morally relevant features of their persons and situations. In this sense, the Rawls of LP fails to heed Kant's injunction in ‘Perpetual Peace’ not to end up tailoring a political morality to the concerns of those currently in power: ‘I can easily conceive of a moral politician, i.e. one who so chooses political principles that they are consistent with those of morality; but I cannot conceive of a political moralist, one who forges a morality in such a way that it conforms to the statesman's advantage.’68

A vital task of the liberal political theorist is to subject the status quo to (sometimes speculative) critique: ‘One finds no great concern to stabilise every existing order, nor should one. There is no reason to mourn the destruction of unjust social and political orders.’69

As for the question (2) above, as to whether the ideal presented in LP is sustainable, it seems quite evident that decent peoples are profoundly unstable in the modern world—and thus so too is a Law of Peoples—since they are organised around comprehensive conceptions of the good.70 There may be demographic shifts such that a majority does not cleave to that conception, or (perhaps due to justified liberal support for those demanding greater freedom and equality within these societies) particular individuals may strenuously oppose the existing organisation of such a society, leading to civil strife. Because persons (even the majority) can reasonably claim that they should be treated equally, such fundamental strife can occur even within an ideal society where everybody acts reasonably (as we have seen, the decent consultation hierarchy does not provide any negotiating mechanisms for such (p.38) fundamental conflicts).71 Worse still, illiberal reactions to, say, the institutionalisation of one religion might be aroused when such historically contingent shifts occur. Civil strife is hardly the basis for consistent performance of one's duties in respect of a global scheme of cooperation.72

Plurarchy, the functionally differentiated global scheme for which I have argued, does not face the same problems: if persons change their comprehensive doctrines, or there are shifts in demographics or power, an ethically neutral scheme of political structures is equally and always accommodating. As Ackerman points out, liberals must insistently not accommodate the exigencies of current power relations in ideal theory:

Of course government officials [representing a dominant religion or comprehensive view] will not accept a fundamental critique of existing boundaries—their political power presupposes their legitimacy. Giving them a veto on the question of boundaries is like giving the rich a veto on the distribution of wealth … Rawls proposes a disastrous political compromise … [Even if illiberal regimes] satisfy these very minimal minima … I fail to see why it justifies anything more than a modus vivendi with oppressor states.73

Why is it important to keep in view the fact that this supposed ‘“overlapping consensus” is really just a modus vivendi among quite different models of society’?74 One reason is to be found in Kant: to limit a conception of global justice to a ‘hybrid solution such as pragmatically conditioned right halfway between right and utility’ is to ‘eternalise the violation of right’.75 A practical foreign policy is one thing, sacrificing the proper regulative ideal is quite another. It was in (p.39) this spirit that Rawls wrote in ‘The Idea of an Overlapping Consensus’ that: ‘The politician looks to the next election, the statesman to the next generation, and philosophy to the indefinite future.’76 This is what is meant when I say that the Rawls of LP is insufficiently utopian: his conception is neither robust nor aspirant enough.

But there is an even greater pragmatic danger than limiting our distant future—a danger of not being appropriately realistic at present, if we do not keep the correct ideal in view. We need, as Rawls put it in A Theory of Justice, ‘a standard for appraising institutions and for guiding the overall direction of social change’77; and if we do not get the standard right, we will misjudge how to pursue justice in the present non-ideal conditions: ‘Non-ideal theory is … more immediately relevant to practical problems, but ideal theory is more fundamental, establishing the ultimate goal of social reform and a basis for judging the relative importance of departures from the ideal.’78

A modus vivendi with hierarchical regimes may be the best that is achievable right now, and it will have its own special prudential rules; but let us be aware of what we are compromising, and let us be able to judge which are the least offensive such rules. It is well, then, to constantly bear in mind an injunction from Thomas Pogge, which is very much in the spirit of Kant and the early Rawls:

Realism hardly requires that the principles of justice conform themselves to the prevailing sordid realities. We don't feel justified to give up our ideals of domestic justice or personal honesty just because we despair of achieving them fully. We cannot reasonably demand of moral principles that they vindicate the status quo. All we may ask is that a conception of justice provide a criterion for assessing our global order that allows us to [identify and] choose from among the feasible … avenues of institutional change and thus specifies our moral task gradually to improve the justice of this order.79

I now want to complete this reply to my imagined critic by demonstrating that Rawls' conception in LP leads him to prioritise the wrong things, and that my ‘relevant utopianism’ serves as a better guide to political action.80 I will focus on the omission of free speech and democracy from Rawls' list of basic human rights. Consider the (p.40) following two statements from LP about political strategy:

With confidence in the ideals of constitutional liberal democratic thought, it [the Law of Peoples] respects decent peoples by allowing them to find their own way to honor those ideals (LP 122).

The Law of Peoples considers this wider background basic structure and the merits of its political climate in encouraging reforms in a liberal direction as overriding the lack of liberal justice in a decent society (LP 62).

These points suggest that peoples are more likely to come round to liberal views if they are engaged as full participants and are included as members in good standing of the Society of Peoples under international law.81 But, insofar as we can make informed judgements about these matters, is any part of this claim plausible? Why should the liberal expect that deliberative rationality—leading to a liberal outcome—is a characteristic of the decent consultation hierarchy? The most that dissent can expect to achieve, according to Rawls, is that the government ‘spells out how the government thinks it can both reasonably interpret its policies in line with its common good idea of justice and impose duties and obligations on all members of society’ (LP 78). There is no reason that constrained internal critique within a comprehensive notion should have liberal results; if anything, it is more likely to result in a spiral of doctrinal self-confirmation. So perhaps Rawls means only that engagement will encourage decent peoples to respect human rights and the laws of peace—the limited ‘liberal’ aims of a Law of Peoples—on an ongoing basis.

There are extremely strong empirical reasons to doubt this latter claim. The single most extensive analysis of the effects of non-democratic political structures on the rights and well-being of persons is to be found in the work of Amartya Sen. In wide-ranging diachronic and synchronic studies, Sen and his collaborators have demonstrated repeatedly that non-democratic regimes are in fact almost unfailingly detrimental to human rights and well-being.82 The most significant (p.41) reason for this—though there are many—is that there are insufficient political incentives for the regime to secure decent social, economic, and legal conditions for persons. Rulers owe their legitimacy to a tradition, a way of life, and not so much to their efficacy in achieving the present important interests of individual persons; states are thus insufficiently attentive and unresponsive to persons' plights:

[There are] extensive interconnections between political freedoms and the understanding and fulfilment of economic needs. The connections are not only instrumental (political freedoms can have a major role in providing incentives and information …), but also constructive. Our conceptualization of economic needs [as well as, Sen goes on to say, our interpretation and application of rights-claims] depends crucially on open public debates and discussions … Furthermore, to express publicly what we value and to [effectively] demand that attention be paid to it, we need free speech and democratic choice.83

The highly likely and disturbing results of not having democratic freedoms are dramatic: social disintegration, famine, and abuse of rights. This is not good for cultures or ways of life either. Sen adds that respect for a number of fundamental human freedoms and for a range of democratic ideas is by no means a function only of ‘Western’ values or impositions, but that elements are to be found in all major cultures and traditions, despite the claims of non-democratic rulers (who have an interest in persons thinking otherwise).84 He also makes the crucial epistemological and eminently practical point that without (p.42) democracy it is impossible to tell which interpretations of a culture are the impositions of semi-autocratic rulers, and which are widely held and justifiable.85

So we can accept, with Rawls, the notion that we should respect cultures, and even that ‘the crucial element in how a country fares is its political culture’ (LP 117), even while we strenuously insist that political cultures ought to become liberal and democratic. This seriously undermines Rawls' project, since he writes that, ‘Should the facts of history, supported by the reasoning of social and political thought, show that hierarchical regimes are always, or nearly always, oppressive and deny human rights, the case for liberal democracy is made. The Law of Peoples assumes, however, that decent hierarchical peoples exist, or could exist …’ (LP 79).

On these terms, the case for liberal democracy has been made: there are no secure minimal human rights without (a governing principle of) democracy. And it has been argued—though I shall not do so much here—that without the political incentives that liberal democratic political institutions provide, (thin) states are likely to become aggressive.86 It is the less powerful sections of the population—and not the leaders of associations or states, unless they are subject to popular judgement—that suffer the full ravages of war; so giving every person a say is likely to be more conducive to the perpetual peace we seek. In sum, my arguments together with the best empirical evidence available establish that the notion of an ongoing scheme of cooperation that has the features of decency is unstable, unrealistic, and undesirable from the point of view of justice.

At this point, my critic throws up his or her hands: ‘Fine, our liberal principles do not lead us to accept decent peoples as sufficiently just. But what do you propose to do?!’ My answer is that decent regimes must be engaged in a global legal structure but only to a limited extent. First, the conditions for entry must require reforms in a democratic direction rather than avoiding this issue. Far from being impracticable, that is increasingly the practice of transnational bodies such as the Commonwealth (as with its suspension of Zimbabwe), the European Union (which makes democratic reforms a condition for entry), and the World Bank Group (which—whatever else it does (p.43) wrong—puts some key democratic conditions on aid and loans). Rather than being a pipe dream, this kind of perspective has won the support of, and been implemented by, numerous hardened pragmatists.

Moreover, the theory and practice of international law increasingly, and rightly, invokes such democratising imperatives. Since the initial United Nations endorsement of those imperatives in Article 19 of the 1966 ‘International Covenant on Civil and Political Rights’, full free speech and democracy requirements have been increasingly crisply and demandingly formulated in broad African, American, and European conventions on human rights, and now are to be found in most such charters.87 In Rawls' scheme there is, on the contrary, no ‘political case for intervention [of any kind] based on the public reason of the Law of Peoples’ (LP 84). While liberals can express their private views about the injustices in non-liberal societies, ‘Rawls's international law principles do not even authorise representatives of liberal societies to publicly (i.e. in an international forum such as the United Nations) criticise the non-liberal practices (e.g. suppression of speech) in hierarchical societies, when such practices are consistent with hierarchical conceptions of the good.’88

Liberal complaints have the same political status in Rawls' idea of international law that comprehensive doctrines have in his idea of domestic society; that is, they carry no political weight whatsoever. While minimal human rights remain a national duty—with potentially more efficacious institutions than the state having only default obligations—rights such as free speech and democracy are removed from global view entirely. It seems to me deeply regrettable that Rawls has taken a step backwards from recent hard-won advances in international law,89 and from the possibility of using global forums to (p.44) bring about change in a liberal democratic direction—especially since political practice has shown this compromise to be unnecessary.90

This leads to my second illustration of the practical application of ideals: humanitarian intervention. I simply want to point out that Rawls and my imagined critic have confused two things. It is one thing to treat illiberal regimes as outlaws—to a greater or lesser extent, depending on the extent of violation, that is, on where they fall on the decent–tyrannical continuum—but it is quite another to think that it is morally permissible to colonise or eliminate them by force. The legitimacy of a regime is only one among many reasons that preclude war and the use of force:

War [may be] excluded because it is grossly disproportionate to the goal sought. Even in cases where the regime is overtly tyrannical (such as the present Chinese regime) waging war would be wrong because of the impossibility or prohibitive cost of victory, that is, for purely prudential reasons. So humanitarian intervention (that is, wars to liberate oppressed populations) is subject to a number of moral constraints that counsel moderation …91

It is a simple and unfortunate category error to confuse illegitimate interventions, on the one hand, with judgements of regime illegitimacy, on the other.

The use of force must be reserved for cases where force is the only realistic way to encourage sustainable democracy or to prevent egregious abuses of human rights. Clearly if force will do more harm than good, then force is not to be adopted, and clearly there are other methods that should be preferred. Every effort should be made first to use moral suasion and bring diplomatic pressure to bear on illiberal regimes. It follows that there are independent grounds for rejecting heavy-handed intervention. Further, the efficacy of less drastic and morally vexing means than force would be increased by the existence of a global law that recognises the underlying value of the ends that cosmopolitans promote.92

(p.45) It is absolutely critical that these shifts in conceptions of sovereignty be encouraged along with a focus on how institutions can be made democratically responsive, otherwise we are likely to be left with either non-democratic dispersed sovereignty or a state-dominated order. I have suggested that neither sort of global basic structure is an appealing prospect. There are good reasons to develop a relevantly liberal and realistic utopian democratic alternative. As Edvard Hambro put it, when he was President of the UN General Assembly, ‘We ought not to be satisfied when people tell us that politics is the art of the possible. Politics should be the art to make possible tomorrow what seems impossible today.’93

Commitment to such an ideal vision is entirely consistent with and even requires realism about practical obstacles, constraints, and opportunities. Our practical task—a task explored in the coming chapters—is gradually to pluralise the current global order by creating a variety of forms of democratically responsive, semi-autonomous legal authority; they could in turn develop a texture of relationships that is sufficiently complex and that meets an important range of interests, so that the entire scheme is widely accepted and stable. In the coming chapters I shall explore one scheme that could meet these requirements. It is time to end the dominance of what David Luban has called ‘the romance of the nation-state’94 and to discern principles for a more complex and more promising global institutional configuration. Those principles will take individuals to be the normative epicentre of a system of plurarchic sovereignty. In this chapter, I have critically engaged the work of John Rawls to begin constructing such a cosmopolitan Law of Persons. (p.46)


(6) John Rawls, The Law of Peoples (Cambridge, MA: Harvard University Press, 1999a).

(7) John Rawls, Political Liberalism (New York: Columbia University Press, 1993a) and A Theory of Justice (Oxford: Oxford University Press, 1971). Also see his ‘Kantian Constructivism in Moral Theory’, Collected Papers (Cambridge, MA: Harvard University Press, 1999b), 303–59.

(8) The original position is a ‘device of representation’. What it ultimately represents is the contractarian conviction that we can derive principles of justice that are fair to all members of a society by asking what all members would agree to if they were deciding in a way which took everyone else into account equally. Since we are all partial to our own interests and allegiances, however, how do we identify this impartial perspective? In A Theory of Justice, Rawls argues for the adoption of a hypothetical situation or thought experiment in which each decision-maker represents a person or group in the society, but also has no knowledge of the race, gender, social class, and the like of those represented. In this original position, a veil of ignorance screens out any knowledge that could bias decision-makers one way or another. Because the parties in the original position could turn out to represent anybody in the society, they are driven to take everybody into account equally. The principles that the parties would then agree to necessarily aim to be fair to all.

(9) This constructivist procedure relies on the idea of pure procedural justice developed by Rawls (see esp. Theory, 83–90).

(10) John Rawls, ‘The Law of Peoples’, in S. Shute and S. Hurley, eds., On Human Rights: The Oxford Amnesty Lectures, 1993 (New York: Basic Books, 1993b).

(11) ‘[J. S. Mill] uses an idea of nationality to describe a people's culture’ (LP 25, n. 20), and ‘I think of the idea of nation as distinct from the idea of government or state, and I interpret it as referring to a pattern of cultural values of the kind described by Mill …’ (LP 23, n. 17). Rawls approvingly quotes Considerations on Representative Government, where Mill writes of ‘common sympathies, which do not exist between them and any others—which make them cooperate with each other more willingly than with other people, desire to be under the same government, and desire that it should be a government by themselves, or a portion of themselves, exclusively …’ (ibid.).

(12) For Rawls' ideas of the reasonable and rational, see Political Liberalism, 48–54.

(13) Here Rawls follows P. Soper's A Theory of Law (Cambridge, MA: Harvard University Press, 1984), esp. 125–47. On several problems with Soper's theory, see J. Raz, ‘The Morality of Obedience’, Michigan Law Review, Vol. 83, No. 4 (1985), 732–49.

(14) ‘Well-ordered societies with liberal conceptions of political justice also have a common good conception in this sense: namely, the common good of achieving political justice for all its citizens over time and preserving the free culture that justice allows’ (LP 71, n. 10). The idea of a comprehensive ‘common good conception’, which includes a ‘decent consultation hierarchy’, is discussed in Section 2, below.

(15) These claims about interests are examined in Sections 1 and 2, below. In addition to liberal and decent societies, Rawls discusses ‘outlaw states’ (those which fail even to be decent), ‘burdened societies’ or ‘states suffering from unfavourable conditions’ (to which decent and liberal peoples have duties of assistance), and ‘benevolent absolutisms’ (which honour human rights but in which citizens play no major role in public life) (LP 4, 90–112). The principles for dealing with each type are different and important, but my focus is the liberal–decent divide.

(16) These are unpacked as eight principles of international justice (Rawls provides a summary at LP 37).

(17) Rawls, ‘Law of Peoples’, 42. In the book, he writes: ‘Why does the Law of Peoples use an original position at the second level that is fair to peoples and not to individual persons? What is it about peoples that gives them the status of the (moral) actors in the Law of Peoples?’ (LP 17, n. 9).

(18) ‘Offhand it is not clear why proceeding in this way should lead to different results than proceeding, as I have done, from separate societies outwards. All things considered, one might reach the same law of peoples in either case’ (Rawls, ‘Law of Peoples’, 54–5).

(19) I use the term ‘Realism’ to denote schools of power politics in international relations theory; ‘realistic’ and ‘realism’ denote practical workability.

(20) Rawls, Theory, 8 and Political Liberalism, 12.

(21) O. O'Neill, ‘Political Liberalism and Public Reason: A Critical Notice of John Rawls Political Liberalism’, Philosophical Review, Vol. 106 (1997), 411–28. O'Neill develops the implications of this distinction in Towards Justice and Virtue (Cambridge: Cambridge University Press, 1996).

(22) O'Neill thinks, however, that there are still ‘considerable disanalogies between uses of idealisation in practical and theoretical reasoning, because the direction of fit is reversed. In theoretical reasoning idealisations that are wide of the mark will reveal their failure, or are likely to. In practical reasoning we may conclude that we ought to live up to the idealisations’ (personal correspondence, but see ibid.).

(23) Rawls could also reply that the idealisation of a closed society is justified because it recognises and represents the existence and value of common sympathies or nationhood, while at the same time it at least forms a constructivist basis to secure persons' important interests (especially in human rights). In Section 2, below, I assess this argument and show that Rawls seeks to tolerate common sympathies in the wrong way.

(24) This example raises various issues about the status of immigration in Rawls' work, and about the social embeddedness of persons' identities—issues discussed later in the chapter. For the moment I use this example to illustrate a more general point about (grouping for the purposes of) social choice, and the relevance of that point for establishing political boundaries.

(25) Many such scenarios are imaginable; indeed is arguable that US–Mexico and South Africa–Mozambique relations, among others, fit this model (my point is, rather, conceptual).

(26) Rawls, ‘Law of Peoples’, 43.

(27) There is a historical story to be told which impugns the convergence claim too; for this, see H. Arendt, The Origins of Totalitarianism (New York: Harcourt Brace & Company, 1973), 290–302. She writes that after the French Revolution, humankind ‘was conceived in the image of a family of nations, [and] it gradually became self-evident that the people, and not the individual, was the image of man. The full implication of this identification of the rights of man with the rights of peoples [was “severe”]’, especially for marginalised (not to mention stateless) individuals and minorities (LP 291–3). Rawls' vision would avoid many but arguably not all of these adverse consequences.

(28) This point has been developed eloquently by several theorists of economic justice who criticise Rawls' refusal to extend the difference principle to global distributive justice (my focus lies elsewhere). The quotation is from T. Pogge, ‘An Egalitarian Law of Peoples’, Philosophy and Public Affairs, Vol. 23, No. 3 (1994a), 195–224, at 209–10. Also see B. Ackerman, ‘Political Liberalisms’, Journal of Philosophy, Vol. 91, No. 7 (1994), 364–86, esp. 381–2; C. Beitz, ‘Cosmopolitan Ideals and National Sentiment’, Journal of Philosophy, Vol. 80, No. 10 (1983), 591–600, esp. 594.

(29) Rawls, Theory, 115.

(30) Kok-Chor Tan makes this point in respect of Rawls' earlier ‘Law of Peoples’ article, in ‘Liberal Toleration in Rawls's Law of Peoples’, Ethics, Vol. 108, No. 2 (1998), 276–95, at 283.

(31) Pogge, ‘An Egalitarian Law’, 217.

(32) Rawls, Political Liberalism, 10, 37, 60, 137, and 154. It is especially evident from 190–200 that politically liberal toleration is of an ethically neutral sort.

(33) Which is why, contra Rawls' assertion in LP at 82–3, we can know that decent regimes are unacceptable from a liberal point of view. Respect for persons' comprehensive conceptions of the good precisely does not translate into respect for comprehensive regimes.

(34) Ackerman, ‘Political Liberalisms’, 382–3.

(35) Beitz, ‘Cosmopolitan Ideals’, 596.

(36) Rawls, Theory, 4.

(37) The assumption that persons are free and equal is more complex and varied than this makes it sound; and there are of course notorious blind spots, such as Mill's refusal of political equality to those in a ‘primitive’ cultural state. But leaving aside such badly justified and unjustifiable exceptions, the broad idea can be found alike in Locke, Rousseau, and Mill—and also comes out clearly, it seems to me, in Immanuel Kant, ‘Theory and Practice’ and ‘Groundwork of the Metaphysics of Morals’, in Practical Philosophy, M. Gregor, ed. and trans. (Cambridge: Cambridge University Press, 1996).

(38) C. Schmitt, The Crisis of Parliamentary Democracy (Cambridge, MA: MIT Press, 1986), 11. See also C. Schmitt, The Concept of the Political (Chicago and London: Chicago University Press, 1996).

(39) There is some question as to the nature of the cooperative activity which gives rise to the global original position. Brian Barry and Charles Beitz initially disagreed over whether international society ‘constitutes a scheme of cooperation in Rawls' sense’; Onora O'Neill and Beitz now both think that in any case, since ‘human beings possess these essential [moral] powers regardless of whether, at present, they belong to a common cooperative scheme’, there is no need to ‘depend on any claim about the existence or intensity of international social cooperation’ (Beitz, ‘Cosmopolitan Ideals’, 595; and O'Neill, Towards Justice, 91–121). I will not discuss this debate here; my arguments concerning political liberalism's universalism strongly endorse Beitz and O'Neill's position.

(40) Rawls, Theory, 264.

(41) C. Beitz, ‘Cosmopolitan Liberalism and the State System’, in C. Brown, ed., Political Restructuring in Europe: Ethical Perspectives (London: Routledge, 1994), at 129.

(42) Rawls, Theory, 264. It is not the case that the cosmopolitan position necessarily assumes a self that is not embedded or that is prior to its ends, as is claimed by Michael Sandel, in Liberalism and the Limits of Justice (Cambridge: Cambridge University Press, 1982), and several so-called communitarian thinkers. Rather, the cosmopolitan position demands that a ‘critical moment’ be possible for persons living in different societies, that the value of social practices and a way of life for persons should not be fixed and determined. Just because culture is significant, in that it provides us with a context for becoming who we are, does not mean that it necessarily has normative significance in determining what it is about people that we respect. No political community has a priori legitimacy. Yet it seems quite evident that for there to be a critical moment, persons must have the freedoms that allow for such reflection and debate—that is what communitarian positions often ignore and wrongly sign away, in their enthusiasm to have culture respected. There is no need to pay such a heavy price.

(43) Rawls, Theory, 4.

(44) I. Kant, ‘Perpetual Peace’ (Ak: VIII:367), quoted in LP, at 36, n. 40.

(45) T. Pogge, ‘Cosmopolitanism and Sovereignty’, in C. Brown, ed., Political Restructuring in Europe: Ethical Perspectives (London: Routledge, 1994b), at 89.

(46) The catalogue includes: (1) real persons are often unsure about their (political) identity or have multiple such identities (they may ‘belong’ with no people or many); (2) persons ‘may find those whom they live with in a particular society are not identical with those whom they regard as of their own culture or people (O'Neill, ‘Political Liberalism’, 16); (3) there is no clear cut distinction between peoples, cultures, and other kinds of groupings (Pogge, ‘Cosmopolitanism’, 197); (4) virtually no national territory contains only the members of a single people (ibid.); (5) official borders do not coincide with ‘the main characteristics that are normally held to identify a people … such as a common ethnicity, language, culture, history, tradition’ (ibid.); (6) ‘whether some group does or does not constitute a people would seem … to be a matter of more-or-less rather than either-or’ (ibid.); (7) appeals to ‘a mythical past’ or a ‘desired future’ as constituting a people beg the question of why that particular conception of national identity ought to be constructed (O. O'Neill, ‘Justice and Boundaries’, in C. Brown, ed., Political Restructuring, at 76); and (8) ‘national and community identity is always framed in terms of’ concepts that have no ‘sharp boundaries, and hence cannot provide a basis for sharp demarcations such as political boundaries between states’ (ibid.). This list is hardly exhaustive.

(47) For defences of the nation-state premised on the ‘rights’ of peoples to self-determination, see: D. Miller, ‘The Nation-State: A Modest Defence’, in C. Brown, ed., Political Restructuring in Europe: Ethical Perspectives (London: Routledge, 1994); A. Margalit and J. Raz, ‘National Self-Determination’, Journal of Philosophy, Vol. 87, No. 9 (1990), 439–61; and several articles usefully collected in both W. Kymlicka, ed., The Rights of Minority Cultures (Oxford: Oxford University Press, 1995), and G. Balakrishnan, ed., Mapping the Nation (London and New York: Verso, 1996). I do not think that any of these arguments defeat all the criticisms in my catalogue above. The collections by Kymlicka and Balakrishnan also contain articles that dispute—to my mind, convincingly—the validity of notions of nation and nationalism and the normative claims that purportedly follow. Especially worthwhile is J. Waldron's ‘Minority Cultures and the Cosmopolitan Alternative’, in Kymlicka, ed., Minority Cultures. A pioneering, radical cosmopolitan argument to refute exclusivist statism—especially when the latter is based on culturalist and nationalist claims—is J. Carens, ‘Aliens and Citizens: The Case for Open Borders’, Review of Politics, Vol. 49, No. 2 (1987), 251–73.

(48) O'Neill, ‘Political Liberalism’, 16. Rawls does say that the ‘psychological principle [of limits on affinity] sets limits to what can sensibly be proposed as the content of the Law of Peoples’ (LP 112, n. 44), and that ‘the moral learning of political concepts and principles works most effectively in the context of society-wide political and social institutions’ (LP 112). But this mere assertion rests on an implausibly narrow moral psychology to which we have very little reason to subscribe. For one thing, ‘society’ and ‘nation’ come in so many sizes that stipulating numeric limits to affinity is dubious. For another, identity—and its motivational force—is not essentially unipolar. That unipolarity should even seem plausible to us is due largely to the contingent (and now changing) configuration of Europe after 1648g. Those who remain in the grip of a picture which allows only a state system or a world state would do well to recall that the nation-state is such a historically recent phenomenon, and thus evidently not a necessary, eternal unit of political organisation and psychological affiliation, let alone the fundamental or exclusive unit. Citizenship in overlapping institutions and institutions of great scope is a central topic of Chapter 3.

(49) Pogge, ‘Cosmopolitanism’, 99–100.

(50) Pogge, ‘Cosmopolitanism’, 89, 102–5.

(51) Ibid., 100–1. Contra Weber, it is not the case that a monopoly on the legitimate use of force is foundational to sovereignty, as numerous empirical examples attest: who has a monopoly on legitimate violence over Bavaria—the State itself? Germany? The European Union? NATO? The United Nations Security Council? Each is legally entitled to exercise force only to achieve certain ends and only under certain conditions. True, some of these powers are a function of intergovernmental and treaty arrangements that are rescindable by the German state, but others are not (e.g. there exists no entitlement to strip Bavaria of its autonomous powers). It should be added that, contra Schmitt, it is not the case that the absence of a unitary sovereign makes decisive positive action impossible. Indeed, as I shall show in subsequent chapters, quite the reverse is likely to be true: properly understood and implemented, pluralised sovereignty would improve the overall quality of political judgement and action.

(52) Rawls cleaves to an entirely territorial ideal on the following grounds: ‘Unless a definite agent is given responsibility for maintaining an asset and bears the responsibility and loss for not doing so, that asset tends to deteriorate. On my account the role of the institution of property is to prevent this deterioration from occurring. In the present case, the asset is the people's territory and its potential capacity to support them in perpetuity; and the agent is the people itself as politically organised’ (LP 8). Not only does Rawls here beg the question as to why peoples should necessarily be that responsible agent—as has become apparent, they should not—it also shifts spuriously from the idea of property to that of closed territories. Yet it is not the case that all property is or need be underpinned by territorial regimes, nor—as will also become evident—that the kinds of property we take to be important are best conserved by the political regimes of peoples.

(53) The deeper aspiration here is to have a universalist account of global justice which avoids assuming territorial boundaries, while still taking the particularity of human practices into account. (On the apparent differences between particularists and universalists more generally, and why these differences are widely misunderstood and largely bridgeable, see O'Neill, Towards Justice.) I cannot provide an account of this kind here. What has been provided is a range of illustrations of this line of thought. Most notably, my discussion of how to expand membership in the United Nations General Assembly (in Chapter 4) proposes some ways to individuate relevant practices and actors in that context.

(54) For example, a global or regional environmental agency for preserving wildlife habitats, wherever they may be, might better achieve that end than nation-states—or certainly nation-states on their own. Chapters 3 and 4 provide extensive examples and analysis of such non-state agencies and their place in a pluralised global order.

(55) My argument that this kind of dispersal would be beneficial is even stronger if Carl Schmitt is at all correct that bounded territoriality necessarily leads to conflict (Schmitt, Crisis, 53 and 69–71). In the system I propose, authorities to which sovereignty is dispersed are not individuated solely on the basis of their capacity to exercise legitimate violence; rather, in Chapters 3 and 4, I bring out the relevance and legitimacy of other forms of effective, circumscribed, and identifiable compulsion.

(56) S. Brown, ‘The World Polity and the Nation-State System’, in R. Little and M. Smith, eds., Perspectives on World Politics (London: Routledge, 1991), 263–72, at 271. See also S. Krasner, ‘Power Politics, Institutions and Transnational Relations’, in T. Risse-Kappen, ed., Bringing Transnational Relations Back In Non-State Actors, Domestic Structures and International Institutions (Cambridge: Cambridge University Press, 1995).

(57) Article 5 of the Consolidated Version (post-Amsterdam) of the Treaty Establishing the European Community, available at www.europa.eu.int. On one interpretation, the principle of subsidiarity aims to locate administrative power where it is most effective; on another (and this is probably closer to the letter of the text) the principle aims to locate that power ‘as close as possible’ to citizens—that is, to citizen control, largely through (local, etc.) elections.

(58) I shall return to and develop these distributive principles in later chapters.

(59) Rawls is quite clear that these kinds of ‘general facts’, including economic theory, are allowed behind the veil (Theory, 136–42).

(60) Ackerman, ‘Political Liberalisms’, 377–8.

(61) This is his most succinct, explicit statement of the stability argument (Rawls, ‘Law of Peoples’, 55).

(62) Ibid., 43.

(63) D. Moellendorf, ‘Constructing the Law of Peoples’, Pacific Philosophical Quarterly, Vol. 77 (1996), 132–54, at 135. Moellendorf characterises Rawls' position only in order to dispute its force.

(64) Rawls, Political Liberalism, 48. See also: ibid., 133–72; Rawls, Theory, 336; and Moellendorf, ‘Constructing the Law’, 147–8.

(65) Rawls, Political Liberalism, 64.

(66) Indeed, Rawls has answered one set of Realists, but not all. Structural Realists insist that it is the interaction between states and not features internal to states per se that generate conflict.

(67) This felicitous constructivist phrase is from O'Neill, Towards Justice, 212.

(68) Beck's 1957 translation, quoted by Moellendorf, ‘Constructing the Law’, 139.

(69) Moellendorf, ‘Constructing the Law’, 147.

(70) Again, stability in this Rawlsian sense implies not merely the survival of the society (illiberal regimes from ancient Egypt to today's China have endured for long periods) but that they obtain adequate moral affirmation and hence political legitimacy from the populace, including but not limited to powerful role-players, as conditions change over time. By ‘unstable in the modern world’ I do not mean primarily that so-called decent states will face overwhelming exogenous pressures from powerful liberal states; liberal and democratic notions of legitimacy (as requiring the assent of the populace) prove compelling to a significant number of dissenters whether or not other powerful states endorse this view of legitimacy or indeed pay any extensive attention whatsoever to such regimes and dissidents.

(71) Rawls surely cannot brand politically liberal persons in comprehensive societies unreasonable. They may not agree with this hierarchical scheme of cooperation, but wanting benefits and burdens to be shared more equally is wrongly characterised as unreasonably uncooperative. Those persons would hold a liberally justifiable interpretation of cooperation, and it does not cease to be so because they happen to be unlucky enough to be born into a less than liberal society. To say otherwise would be to repudiate the value of even the minimally free exercise of reason, to characterise such liberal dissenters as obstructionist misfits.

(72) Here again we note the deep disanalogy between tolerating persons as opposed to regimes with comprehensive doctrines: internal strife in most individuals tends not to impact on the broader system of cooperation, whereas internal strife in hierarchical societies has the potential to impact far more adversely.

(73) Ackerman, ‘Political Liberalisms’, 381–3.

(74) S. Hoffman, ‘Dreams of Just World’, New York Review of Books, Vol. 42 (1995), 52–7, at 54.

(75) I. Kant, ‘Perpetual Peace’, in Hans Reiss, trans. and ed., Kant's Political Writings (Cambridge: Cambridge University Press, 1991); see 119–25.

(76) See Pogge's discussion of this statement in ‘An Egalitarian Law’, 224. Rawls' two essays on the overlapping consensus are reprinted in his Collected Papers, 421–48 and 473–96.

(77) Rawls, Theory, 263.

(78) Carens, ‘Citizens’, 225.

(79) T. Pogge, Realizing Rawls (Ithaca, NY: Cornell University Press, 1989), 260.

(80) The term ‘relevant utopia’ was suggested to me by Stanley Hoffmann. Hoffmann's historically informed and prescient works include Duties Beyond Borders: On the Limits and Possibilities of International Politics (Syracuse, NY: Syracuse University Press, 1981).

(81) In particular, international law ought to recognise and support peoples' domestic institutions of public reasoning, including their decent consultation hierarchies. The idea seems to be that cultures have their own modes of argument and processes of change, and where these are sufficiently deliberative we not only have grounds to believe that they will lead to more liberal principles for regulating a people's common life, but grounds to respect those modes themselves.

(82) There are numerous related and confirming studies, and there is simply no more impressive body of theory and evidence for the empirical judgements I discuss. The most relevant of Sen's monographs in this respect are Development as Freedom (New York: Alfred A. Knopf, 1999), Inequality Re-examined (Oxford: Oxford University Press, 1992), Poverty and Famines: An Essay on Entitlement and Deprivation (Oxford: Clarendon Press, 1981), and—together with J. Dreze—Hunger and Public Action (Oxford: Clarendon Press, 1989). Edited volumes confirming the accuracy and fecundity of this approach to development and democracy include: J. Dreze and A. Sen, eds., The Political Economy of Hunger (Oxford: Clarendon Press, 1995); M. C. Nussbaum and A. Sen, eds., The Quality of Life (Oxford: Clarendon Press, 1989); and M. C. Nussbaum and J. Glover, eds., Women, Culture and Development: A Study of Human Capabilities (Oxford: Clarendon Press, 1995).

(83) Sen, Development as Freedom, 147–8 and 154–5. I am in full agreement with Sen's arguments for democratic rights as intrinsically important too. Rawls mentions Sen's work repeatedly (see esp. LP 108–11), but fails to see that Sen's ‘insistence on human rights’ includes an insistence on democratic rights. Sen's perspective on this is borne out by other commentators on global justice, including T. M. Franck, ‘The Emerging Right to Democratic Governance’, American Journal of International Law, Vol. 86, No. 1 (1992), 46–91.

(84) Sen, Development as Freedom, 227–48.

(85) Sen, Development as Freedom, 227–48; also A. Sen, ‘Human Rights and Asian Values’, The New Republic, 14 and 21 July 1997; and A. Sen, ‘Our Culture, Their Culture’, The New Republic, 1 April 1996.

(86) See O. Otunnu and M. Doyle, eds., Peacemaking and Peacekeeping for the New Century (Oxford: Rowman & Littlefield, 1998).

(87) See F. R. Teson, ‘The Rawlsian Theory of International Law’, Ethics and International Affairs, Vol. 9 (1995), 79–99, at 95.

(88) Teson, ‘Rawlsian Theory’, 88–9. This marks an increasingly conservative turn in Rawls, since—in debarring all but intrastatal efforts to change cultures (excepting minimal rights)—he leaves liberals wringing their hands about socially oppressed minorities beyond their current society's borders. Cultures are difficult to change and must be treated with sensitivity, but that does not imply a counsel of despair so much as a careful and inclusive (proto-democratic) approach to the process of change. For a wide-ranging survey of just how many abhorrent behaviours are cloaked and defended with the exculpatory phrase ‘in our culture’, see Roger Sandall's The Culture Cult: Designer Tribalism and Other Essays (Oxford: Westview, 2002).

(89) The telling contrast between Rawls' conception of international law and current international law is illuminated in J. Tasioulas, ‘From Utopia to Kazanistan: John Rawls and the Law of Peoples’, Oxford Journal of Legal Studies, Vol. 22 (2002a), 367–96.

(90) See Chapter 4; pioneering arguments to similar effect can be found in P. Allot, Eunomia: New Order for a New World (Oxford: Oxford University Press, 1990) and G. Fox, ‘The Right to Political Participation in International Law’, Yale Journal of International Law, Vol. 17, No. 2 (1992), 539–607.

(91) Teson, ‘Rawlsian Theory’, 97.

(92) The implications of cosmopolitan theory for development practice are explored in the ‘Debate on Global Poverty Relief’ between Peter Singer and me. See A. Kuper, ‘More Than Charity: Cosmopolitan Alternatives to the “Singer Solution”’; P. Singer, ‘Poverty, Facts, and Political Philosophies: Response to “More Than Charity”’; A. Kuper, ‘Facts, Theories and Hard Choices: A Reply to Peter Singer’; and P. Singer, ‘Achieving the Best Outcome: Final Rejoinder’—all in Ethics and International Affairs, Vol. 16, No. 1 (2002), 107–128.

(93) Speech to the American Society of International Law (1971), cited in K. Suter, ‘Reforming the United Nations’, in R. Thakur, ed., Past Imperfect, Future Uncertain: The United Nations at Fifty (Basingstoke: Macmillan, 1998), at 203.

(94) D. Luban, ‘The Romance of the Nation-State’, in C. Beitz, M. Cohen, T. Scanlon, and J. Simmons, eds., International Ethics (Princeton, NJ: Princeton University Press, 1985).