Click for: article abstract; all publications by author.
This is an electronic version of an article published in Philosophy & Public Affairs 35.4 (2007): 318-358. Copyright (c) 2007 by Blackwell Publishing. The definitive version is available as a PDF from the publisher here.
ARASH ABIZADEH
It is widely known that in A Theory of Justice Rawls restricted the scope of distributive justice to the domestic context of a single polity.[1] This restriction implies that citizens have responsibilities of distributive justice to each other that they do not have to foreigners. Rawls's readers widely thought that he had motivated this anticosmopolitan restriction, despite his commitment to the equal moral worth of all, on the basis of two assumptions. First, he assumed that principles of distributive justice come into play only within the context of society's "basic structure." Only persons with a shared basic structure have claims upon and responsibilities to each other arising from considerations of distributive justice. Second, Rawls famously assumed that there is no global basic structure, that the societies whose basic structure is the subject of justice are "more or less self-sufficient association[s]" whose boundaries correspond to those of contemporary polities. This was the argument that Rawls's readers thought was implicitly backing the restriction of justice's scope in his Theory of Justice: since "the primary subject of justice is the basic structure of society," and since there is no global basic structure, the scope of justice is domestic.[2]
The force of this Rawlsian basic structure argument [3] concerning the scope of justice hinges on what exactly is meant by "basic structure" and why it is the primary subject of justice. There are at least three distinct ways in which the institutions constituting society's basic structure can be defined, each with some textual support in Rawls: the basic structure might be said to comprise (1) the institutions that determine and regulate the fundamental terms of social cooperation; (2) the institutions that have profound and pervasive impact upon persons’ life chances; or (3) the institutions that subject persons to coercion. These three defining criteria are analytically distinct and, as such, may turn out not to be coterminous in practice. Each criterion also specifies the scope of society in a distinct way: a society (with a single basic structure) might comprise the set of persons (1) engaged in a scheme of social cooperation regulated by the same institutions, or (2) pervasively impacted by the same institutions, or (3) subject to coercion by the same institutions. Consequently, each criterion suggests a distinct way to reconstruct the Rawlsian basic structure argument about the scope of justice, corresponding to three theories concerning the proper subject of justice: what I call a cooperation theory, a pervasive impact theory, and a coercion theory.
This article examines each of these interpretations of a Rawlsian theory of justice and assesses their impact for global justice. By principles of distributive justice I mean, at the very least, comparative principles concerned not merely with how individuals fare or are treated in absolute terms, but also in comparison with others. In other words, they are relational principles making demands of equality. (Rawls's difference principle is one such comparative principle; my argument is not specifically tied to it.)[4] My thesis is that any plausible interpretation of any of the three theories implies that principles of distributive justice are global in scope. Beyond this substantive conclusion, the article makes four further contributions to the burgeoning literature on the relation between Rawlsianism and global justice.
First, it clears up two important confusions in anticosmopolitan appropriations of the cooperation theory. The first confusion is between the site and the scope of justice: the basic structure is the site but does not necessarily limit the scope of justice. The second concerns three different senses in which justice might be said to "require" a basic structure: it may require a basic structure in the sense that it presupposes one before its demands arise, includes (and so constitutively demands) one as a constituent part, or instrumentally demands one as a means to fully realize justice. Anticosmopolitans often assume that justice presupposes a basic structure, i.e., that it is an "existence" condition rather than instrumental condition of justice. I shall demonstrate that, while this assumption is consistent with the pervasive impact and coercion theories, it is not consistent with the cooperation theory: the cooperation theory, properly understood, shows justice to demand a basic structure as one of its instrumental conditions. The result is that, on the cooperation theory, the site and scope of justice do not coincide. It is only when the basic structure is an existence condition of justice that it is both the primary site of justice and that its existing boundaries limit justice's scope.
Second, the article demonstrates that although the coercion theory suggests that the basic structure both is the site and determines the scope of justice, the most sophisticated recent formulations of the theory (by Michael Blake and Thomas Nagel) nonetheless fail to secure the anticosmopolitan conclusions its partisans wish to draw. Blake's anticosmopolitan argument fails to address the interstate system of border coercion, while Nagel's deployment of the coercion theory rests on a perverse normative principle.[5]
Third, the current debate on global justice has to some extent been obscured by its complexity and sheer volume. By systematically mapping out the terrain of one of its most important areas, namely the relation between the basic structure and the scope of justice, the present article clarifies how the failure adequately to distinguish between the cooperation, pervasive impact, and coercion theories has led to confusion. It might be possible, of course, for anticosmopolitans to combine the argumentative strategies and normative premises of the three theories, appealing to the strengths of each in order to compensate for the weaknesses of the others. Indeed, the most sophisticated statement of Blake's coercion argument seeks to meet objections by incorporating elements of the other two theories. This concatenation, however, far from resolving matters, simply suffers from all the difficulties that each of the three theories individually poses for anticosmopolitans.
Finally, though I do not directly address Rawls's argument in The Law of Peoples, my analysis of Rawls's theory of justice poses a significant challenge to Rawlsians who take that work as their point of departure for relating Rawls's theory of justice to the globe. In Law of Peoples, Rawls responded to cosmopolitan appropriations of his theory of justice by explicitly arguing against the view that distributive justice is global in scope, this time on fresh grounds. He now appealed to the moral integrity of distinct "peoples" with "common sympathies" (à la John Stuart Mill) and the value of toleration for nonliberal peoples. However, although the relation between Rawls's Law of Peoples to his theory of justice as fairness remains hotly contested,[6] no one can deny the centrality of the notion of the basic structure to his theory of justice. Indeed, Rawls's thesis that the basic structure is the primary subject of justice is routinely cited as one of his most fundamental and enduring contributions to political philosophy. As such, if my contention is correct, i.e., if any plausible interpretation of this central component of a Rawlsian theory of justice implies that justice is global in scope, then anticosmopolitan "Law of Peoples Rawlsians" must face the fact that a central component of a Rawlsian theory of justice stands against them.
The Rawlsian basic structure argument for the domestic restriction of the scope of justice is deceptively simple:
(1) The primary subject of justice is society's basic structure.
(2) A basic structure global in scope does not exist.
Therefore:
(3) The scope of justice is not global.
Much debate has focused on the second, empirical premise. Rawls's assumption of closed, self-sufficient societies, for example, has been forcefully challenged by cosmopolitans who point to contemporary levels of global interdependence; skeptics, in turn, have challenged the significance of those levels.[7] But for Rawlsians advancing the basic structure argument, the debate concerning levels of globalization is somewhat beside the point: the relevant fact is not whether there is interdependence, but that there ostensibly is no societal basic structure at the global level.[8] This is the relevant fact because the starting point of Rawls's theory is that the basic structure is the primary subject of justice. Thus before the empirical premise can be evaluated, the terms of the normative premise must be clarified. Everything turns on what being the subject of justice means, what exactly the basic structure is, and what the justification for the argument's first premise is.
When Rawlsians say that the basic structure is the primary subject of justice, what they normally mean is that the principles of justice appropriately regulate or apply to the institutions of the basic structure only, i.e., that the basic structure is the primary site of justice. The problem for the basic structure argument, however, is that its second premise and the conclusion assume that "subject" refers to the appropriate scope of justice. The site of justice is not the same as its scope: the site of justice refers to the kinds of objects (individuals’ actions, individuals’ character, rules, or institutions, and so on) appropriately governed by principles of justice, that is, to which the principles of justice rightly apply, whereas the scope refers to the range of persons who have claims upon and responsibilities to each other arising from considerations of justice.[9] For the basic structure argument to be valid, its first premise cannot merely be a claim about site, but must also imply something about scope. In particular, its validity requires showing that the boundaries of the basic structure qua site of justice somehow limit the scope of justice to the range of persons whose lives are regulated by the existing basic structure. This equation of scope with site is what I shall call the site/scope thesis; the point here is that it involves a substantive claim in need of justification.
It might be thought that the site/scope thesis could be secured by arguing that, since the basic structure is the primary site of justice, justice requires a basic structure, which in turn implies that for principles of justice rightly to regulate the relations between persons, their interactions must already be regulated by a basic structure.[10] That thought is misleading, however: it relies on an ambiguity between three senses in which justice might be said to "require" a basic structure. Justice may require a basic structure in the sense that it presupposes the existence of a basic structure before its demands arise; in the sense that it includes (and so constitutively demands) a basic structure as one of its constituent parts; or in the sense that it instrumentally demands a basic structure as a means for justice to be fully realized. Another way to put this is to say that a basic structure may be a condition of justice in one of three different senses: the first sense specifies an existence condition, the second specifies a constitutive condition, and the third specifies an instrumental condition of justice.[11] Only the first sense of "require," corresponding to a necessary existence condition, limits the scope of justice to where a basic structure already exists; the second and third senses limit the scope of justice only to where a basic structure could exist. It is thus crucial to see the fundamental difference between existence conditions, on the one hand, and constitutive and instrumental conditions, on the other. If a necessary existence condition of justice is missing, then demands of justice do not arise and the scope of its application is thereby curtailed; but if a necessary constitutive or instrumental condition is missing, justice precisely demands the realization of those conditions. A necessary constitutive or instrumental condition merely imposes feasibility limits on the scope of justice.
The upshot is that, in order to defend the site/scope thesis, i.e., in order to show that the scope of justice is limited to the range of persons whose relations are regulated by an already existing basic structure, it must be shown that the basic structure is an existence condition of justice. If the basic structure is merely a constitutive or an instrumental condition of justice, then justice may demand that there be a (just) basic structure. In fact, I shall argue, it is only according to the pervasive impact and coercion theories of distributive justice that the basic structure could be an existence condition. By contrast, the cooperative theory rightly understood, contrary to the assumption of many anticosmopolitan Rawlsians who appropriate it, implies that the basic structure is an instrumental condition of justice.
As a result, on any of the three interpretations of a Rawlsian theory of justice, justice is global in scope. On the one hand, the pervasive impact and coercion theories show why the scope of justice extends and only extends to the range of persons whose interactions are already regulated by a basic structure. But both theories’ accounts of the basic structure imply that a global basic structure exists. On the other hand, the cooperation theory renders plausible the claim that a global basic structure does not exist. But its justification for why the basic structure is the primary site of justice also shows why the scope of justice is not limited by an existing basic structure. The cooperation theory rightly understood shows that the existence condition of justice (which for this theory is social interaction) obtains at the global level as well.
Rawls's official definition of the basic structure centers on the notion of social cooperation: after saying that justice as fairness conceives society as "a cooperative venture for mutual advantage," Rawls defines society's basic structure as comprising "the way in which the main political and social institutions of society [a] fit together into one system of social cooperation, and the way they [b] assign basic rights and duties and [c] regulate the division of advantages that arises from social cooperation over time."[12] These three elements are what we might call the fundamental terms of social cooperation. The institutional components of the basic structure that Rawls explicitly mentions all regulate one or more of these three fundamental terms. He mentions:
i. the "political constitution," which determines how major political institutions and branches of government fit together, and which assigns basic legal rights and duties to persons;
ii. the "legally recognized forms of property," which, by determining the nature of property rights, help determine the economic structure and regulate the division of goods;
iii. the "structure" or "organization of the economy" itself, which refers to the nature of markets and ownership of the means of production (as well as, presumably, tax law), and which therefore helps determine how the division of goods is regulated; and, finally,
iv. "the nature of the family," which refers to the legal rights and duties of equal citizens in their roles as spouses, parents, and children. [13]
Rawls's official definition of what the basic structure is lends a certain plausibility to the second, empirical premise of the basic structure argument. Although there may indeed be high levels of global interdependence, on this interpretation the argument's second premise simply claims that there are no global institutions that regulate (in any significant way) how major political and social institutions (a) fit together to form a system of social cooperation, (b) assign basic rights and duties, and (c) regulate the division of advantages arising from social cooperation globally. The major political and social institutions that regulate these three fundamental terms are domestic institutions.
The account also provides a particular interpretation of the first, normative premise of the argument: it suggests that when Rawls says that the primary subject of justice is the basic structure, he means that the principles of justice should primarily regulate only the way in which major political and social institutions fit together in the system of cooperation, assign basic rights and duties to members of society engaged in social cooperation, and regulate the division of advantages arising from social cooperation. This clearly is a cooperation theory of distributive justice: the principles of justice are meant only to govern the fundamental (publicly recognized) terms of social cooperation.[14] This implies that, as Rawls's discussion of the family makes clear, principles of justice are not meant to govern the internal life of institutions or associations, even if they are themselves institutional components of the basic structure.[15] So the division of goods within a family, within a firm, or within the branches of government are not subject to Rawls's difference principle; the difference principle applies to the regulation, by these institutions, of the division of goods in society as a whole. As Rawls puts it, the
difference principle holds, for example, for income and property taxation, for fiscal and economic policy. It applies to the announced system of public law and statutes and not to particular transactions or distributions, nor to the decisions of individuals and associations, but rather to the institutional background against which these transactions and decisions take place.[16]
Rawls's contention also implies that distributive justice is not primarily an allocation of goods problem; as Samuel Freeman has emphasized, it is concerned in the first instance with "fairly designing the system of basic legal institutions and social norms that make production, exchange, distribution, and consumption possible."[17]
The question is how Rawls justifies the premise that the basic structure of society is the primary subject of justice. His first justification—the justification to which the cooperation theory appeals—is a frankly instrumentalist justification. It centers on the indispensable role of society's basic structure in securing the just background conditions that make possible fair transactions and agreements between individuals and associations. Rawls's point here arises in response to the libertarian view of justice defended by Robert Nozick. Rawls argues that, starting with an initially just socioeconomic state of affairs, subsequent states of affairs resulting from fair transactions and agreements will also turn out to be just, as Nozick maintains, only if just background conditions, which make fair transactions and agreements possible in the first place, are preserved over time.[18] The problem, Rawls argues, is that the accumulation of individually fair transactions and agreements will, over time, causally undermine the background conditions that make fair transactions and agreements possible in the first place. It is thus the indispensable role of the institutions of the basic structure to maintain background justice, in order to ensure the fairness of the system of social cooperation. As Rawls puts it,
the conditions necessary for background justice can be undermined, even though nobody acts unfairly or is aware of how the overall result of many separate exchanges affects the opportunities of others. There are no feasible rules that it is practicable to require economic agents to follow in their day-to-day transactions that can prevent these undesirable consequences. These consequences are often so far in the future, or so indirect, that the attempt to forestall them by restrictive rules that apply to individuals would be an excessive if not an impossible burden.[19]
This is what Liam Murphy has called a "division of labor" argument: given that background justice can be fully maintained only via institutions that regulate the fundamental terms of social cooperation, and given that applying the principles of justice exclusively to the institutions of the basic structure is the best means for realizing justice (since applying them elsewhere would place excessive burdens on individuals), the principles of justice only appropriately apply to the institutions of the basic structure. (In particular, they apply to the way in which the institutions of the basic structure determine the fundamental terms of social cooperation by regulating how those institutions fit together, assign basic rights and duties, and regulate the division of advantages.) Thus the thesis that the basic structure is indispensable to justice—that justice requires a basic structure—encompasses two distinct claims, one about necessity, the other about optimality: the basic structure is instrumentally necessary for the full realization of justice; and applying principles of justice exclusively to the basic structure is the best means for realizing any degree of justice. Hence the principles of justice do not appropriately apply directly to the conduct between individuals and associations in their interactions with each other.[20]
Another way to put this is to say that a basic structure is the indispensable means by which a system of social coordination or interaction could become a fair system of social cooperation. Rawls explicitly distinguishes between these two concepts: as he puts it, while mere social coordination is compatible with, "for example, activity coordinated by orders issued by an absolute central authority," social cooperation "includes the idea of fair terms of cooperation." Social cooperation is social coordination or interaction conducted on fair terms and premised on the reciprocity of the system of coordination.[21] The instrumental necessity of a basic structure for realizing the ideal of social cooperation, combined with the putative fact that applying the principles of justice only to the basic structure is the best way to realize a fair system of cooperation, is what justifies the first premise of the basic structure argument (according to which the basic structure is the primary subject of justice).
Regardless of how one puts it, the implication of the division of labor argument, for the cooperation theory, is this: the basic structure is an instrumental condition of justice.[22]
The problem for anticosmopolitans is that, while the second premise and conclusion of the basic structure argument assume that the "subject of justice" refers to the scope of justice, the cooperation theory's justification for the first premise only establishes that the basic structure is the site of justice. For the basic structure argument to be valid, its first premise must furnish a claim about scope. In other words, its validity requires a defense, consistent with the cooperation theory, of the site/scope thesis. As I hope to demonstrate, since for the cooperation theory the basic structure is an instrumental and not existence condition of justice, there is no defense to be had.
The question is what the cooperation theory can plausibly say about the scope of justice. Only a crude interpretation of the theory would have it say that claims and responsibilities of justice only arise between participants in a shared scheme of social cooperation for mutual advantage, regulated by a shared basic structure.[23] Recall that a system of social cooperation is not a mere system of social coordination or interaction: it is a fair or just system of social interaction. Cooperation, in other words, incorporates a moralized ideal that is a constituent of Rawlsian justice. The interpretation is crude because by restricting the scope of distributive justice to those persons who participate in what is in fact a scheme of social cooperation, it mistakes a constituent of the ideal of justice, namely the fairness of social interaction, for one of its existence conditions (which must be satisfied before the demands of justice as fairness arise). This conflation of the existence and constitutive conditions of justice incurs a perverse and arbitrary bias in favor of the status quo. By using the ideal of social cooperation to specify (and restrict) the scope of application of the ideal itself, the crude interpretation perversely implies that demands of distributive justice arise only between persons whose social interactions are already conducted on fair terms, i.e., that demands of justice would not arise for persons whose social interactions are unjust.[24] It is true, of course, that justice "requires" that social interaction be fair. But justice does not require fairness in the sense of presupposing the fairness of social interaction before its demands arise; rather, it demands the fairness of social interactions as a constituent part of justice.[25]
Nor could the first premise of the basic structure argument be understood to say, on any plausible reading of the cooperation theory, that since social cooperation "requires" a basic structure, then a shared basic structure is a precondition for relations of justice to arise. This would also be a crude interpretation of the cooperation theory, incurring an arbitrary status quo bias, because it conflates the existence and instrumental conditions of justice. Consider three scenarios: (i) where there exists no socially coordinated interaction at all, (ii) where there exists social coordination and a basic structure but no background justice, and (iii) where there exists social coordination but no basic structure. (i) According to the cooperation theory, where there is no social interaction, the demands of justice do not arise. Because on the cooperation theory the point of justice is to regulate social interaction, social interaction is a necessary and sufficient existence condition of justice: the cooperation theory does not demand that social interaction exist, in order to create occasions for justice to do its work; rather, justice only needs to do its work when social interaction already exists. (ii) But the cooperation theory cannot say the same about the basic structure. Where there is unfair social interaction and a basic structure exists, the cooperation theory demands that the basic structure be effective and just in order to realize justice in the system of social interaction and to make it a fair system of social cooperation. (iii) The same instrumental relation holds in the third case, where there exists social interaction without a basic structure: the cooperation theory demands the creation of a just basic structure, when a basic structure is feasible, in order to realize justice in the system of social interaction.[26]
The conflation of the existence with the constitutive or instrumental conditions of justice, and in particular the failure to see the instrumental role of the basic structure, is the central difficulty with many anticosmopolitans’ appropriation of the cooperation theory. Samuel Freeman, for example, argues that Rawlsian "distributive justice presupposes social cooperation," which in turn requires "political cooperation," all of which requires institutions of a basic structure; and he concludes that since there is no global basic structure, global principles of justice do not arise.[27] But on the cooperation theory, distributive justice does not presuppose social cooperation: as Freeman himself acknowledges, "for Rawls social cooperation also involves an idea of reciprocity and fair terms of cooperation" and, as such, "incorporates a distinctly moral component."[28] According to any plausible cooperation theory, justice presupposes only social interaction, not cooperation. Nor does justice require a basic structure in the sense of presupposing one before the demands of justice arise. Rather, since the cooperation theory's justification for taking the basic structure as the primary site of justice is instrumental, for it the institutions of a basic structure, including institutions of political cooperation that may be necessary to realizing justice, are instrumental conditions of Rawlsian justice, not existence conditions.[29]
Thus the fact that a basic structure is indispensable to background justice means that, where social interaction exists, there ought to be a just basic structure to secure background justice. It does not mean that if a basic structure does not exist then there is no need for background justice. It is true that, on the cooperation theory, principles of justice require a basic structure because such principles "define the appropriate distribution of the benefits and burdens of social cooperation."[30] But this requirement is a demand, not a presupposition. The cooperation theory therefore cannot sustain the anticosmopolitan basic structure argument about the scope of justice. According to the cooperation theory rightly understood, either the basic structure argument's first premise (about the primary subject of justice) is true but the argument itself invalid, or the first premise is false. The first premise is true if it is a claim about the site of justice, but it is false if it is about the scope of justice, which it must be for the conclusion to follow validly from the premises. According to the cooperation theory, the basic structure is the primary site of justice because it is the necessary means for fully realizing justice. However, because on the cooperation theory the basic structure is merely an instrumental condition of justice, the site/scope thesis turns out to be false: the scope of justice is given by the range of persons engaged in a system of social interaction (the justice of which demands a just basic structure) that is or could become a scheme of social cooperation. If a system of social interaction does exist, and if a just basic structure could exist, then there ought to be a basic structure to which the principles of justice can be applied.
The cooperation theory rightly understood thus makes the following claims about the site and scope of justice. First, it claims that the site of justice is those institutions capable of regulating the fundamental terms of social cooperation in such a way that a system of social interaction can be a fair system of social cooperation over time. Second, it claims that the scope of justice comprises participants and only participants in a joint scheme of social interaction that is or could become a fair scheme of social cooperation. This suggests three limits on the scope of justice. The first limit arises from justice's existence condition, i.e., the putative fact that justice presupposes the existence of social interaction before its demands arise. This limits the scope of justice to the range of persons actually engaged in social interaction. The second limit arises from its constitutive condition, i.e., the putative fact that the justice of a society is constituted (at least in part) by its being a fair system of social cooperation. This limits the scope of justice to the range of persons whose participation in a system of social interaction could, at least in principle, become participation in a fair system of social cooperation for mutual advantage. The third limit arises from an instrumental condition, i.e., the putative fact that a (just) basic structure is a necessary means for fully realizing distributive justice, and that exclusively applying the principles of justice to the basic structure is the best means for realizing it. This suggests that the scope of justice is limited to the range of persons whose social interaction could be regulated by a (just) basic structure, i.e., for which a (just) basic structure is possible. The second and third conditions, unlike the first, merely impose feasibility limits on the scope of justice.
In light of these ostensible limits, one might revise the anticosmopolitan basic structure argument in one of two ways. The no globalization argument, which abandons any reference to the basic structure, appeals to the first limit:
1.1. The scope of justice is restricted to participants in a system of social coordination that could be a (just) system of social cooperation in which these persons participate.
2.1. A system of social coordination global in scope does not exist.
Therefore:
3. The scope of justice is not global.
The two feasibility limits, by contrast, suggest a second (revised) basic structure argument:
1.1. The scope of justice is restricted to participants in a system of social coordination that could be a (just) system of social cooperation in which these persons participate.
2.2. A (just) basic structure is a necessary means for realizing a fully just system of social cooperation; and
2.3. A basic structure global in scope does not exist; nor is a global basic structure, or at least a just global basic structure, feasible.
Therefore:
3. The scope of justice is not global.
I take up each argument in turn.
The no globalization argument limits the scope of justice by focusing on the claim that a system of social interaction is an existence condition for demands of justice. Charles Beitz, who previously had been the primary exponent of the view that economic interdependence is a necessary and sufficient existence condition of justice, has more recently rejected the imposition of such an existence condition, arguing that the scope of justice is subject only to feasibility limits. What motivates Beitz's revised view is the worry that imposing an existence condition on justice "would arbitrarily favor the status quo."[31] To this one may respond that the arbitrary status quo bias does not arise from imposing an existence condition on justice per se, but from conflating its existence conditions with its constitutive or instrumental conditions.[32] I do not claim to settle this debate here. My point simply is that Beitz's worry exposes a dilemma facing anyone who wishes to ground the no globalization argument in the cooperation theory: either the existence condition on which the argument relies is extremely weak and so is easily met by contemporary levels of globalization (rendering premise 2.1 false), or the existence condition is stronger than mere social coordination, which shores up premise 2.1, but falls prey to the status quo bias by conflating the existence conditions with constitutive or instrumental conditions of justice.
The second horn of the dilemma is illustrated by two attempts to strengthen the existence condition of justice. Consider first Brian Barry's suggestion (recently echoed by Andrea Sangiovanni) that the demands of Rawlsian distributive justice do not arise unless there is a scheme of socially coordinated action that actually provides public goods and meets "conditions for reciprocity." Barry's argument for this stronger existence condition is that social coordination consisting in mere transactions or exchanges gives rise only to considerations of "justice as requital," i.e., the only demand of justice arising from trade relations is the demand to give each party a "fair return" in exchange.[33] According to Barry, it may be collectively irrational if no public goods scheme is developed in such contexts, but without such a scheme, demands of distributive justice do not arise.[34] The problem with this suggestion is that it ignores Rawls's argument against Nozick. It ignores the background conditions that Rawls considers necessary for individually fair transactions and exchanges to be possible in the first place. As we have seen, according to Rawls, it is not possible to specify and to ensure a fair return unless the conditions of background justice have already been secured. And one of the roles of applying principles of distributive justice is precisely to ensure that it is possible, in the long run, for individual transactions and exchanges to be fair.[35] Indeed, it may turn out that it is impossible to maintain the fairness of transactions unless some public goods are provided. This would make such provision not just a demand of rationality, but of justice. That is why embedding the provision of public goods into the existence condition of justice misconstrues a potential demand of justice for one of its existence conditions.[36] Recall again that Rawls does not claim that every society in fact constitutes a scheme of social cooperation for mutual advantage; rather, he assumes that it must be conceived as a scheme of social cooperation for the normative purpose of determining the principles of justice. His claim specifies a normative element of the ideal of justice. The provision of public goods may be a constituent element of such an ideal, but it is not an existence condition.
The same point holds with respect to the interpretation according to which Rawlsian principles of justice arise only among persons whose system of social coordination occurs within the context of a public culture already recognizing, as the subject of an overlapping consensus, the equality and freedom of persons. Again, this is to smuggle a constituent of the Rawlsian ideal of justice—an overlapping consensus on the ideals of freedom and equality—into its existence condition.[37] So it is not surprising to find this interpretation falling prey to the status quo bias: the implication is that if my equality or freedom is not recognized by the ideals embedded in my public culture, then I have no standing to demand that, as a matter of justice, my equality or freedom be publicly recognized.[38] This interpretation conflates a feasibility limit with an existence condition.[39]
The first horn of the dilemma arises from trying to avoid conflating the existence and constitutive conditions of justice by imposing a very weak existence condition. That social coordination is an existence condition of justice follows straightforwardly from the cooperation theory's claim that the purpose of justice is to regulate the terms of social cooperation in order to ensure that the system of social coordination is, and remains over time, a system of fair social cooperation for mutual advantage. Thus a cooperation theorist might reply, to the worry about an arbitrary status quo bias, that an existence condition of mere social coordination is weak enough to avoid it. This rejoinder may or may not stand to reason. The problem for the anticosmopolitan is that, even if the rejoinder were to stand, the requirement of social coordination is too weak to sustain the no globalization argument: significant amounts of social interaction of any kind qualify, including the strategic kinds of interaction involved in commerce or trade.[40] The upshot is that, regardless of how controversies about the nature of contemporary globalization are resolved, premise 2.1 of the no globalization argument is manifestly false, since agents do coordinate their action globally in this minimal sense (whether through trade and commerce, or through social, cultural, and political relations). The only way to shore up premise 2.1 is to strengthen the relevant existence condition, which, as we have seen, gives rise to the status quo bias.
Moreover, the rejoinder may in any case not stand to reason. Consider the case of two societies who really do meet the fantastic Rawlsian assumption of being closed, i.e., two societies who have no flows of goods, capital, people, and the like across their borders and whose members consequently do not socially coordinate their action with each other. Even though there is no social interaction, the social activities of the members of one society may nonetheless have pervasive negative impact on the life chances of members of the other society, through "nonsocial" causal relations involved in polluting the atmosphere or poisoning downstream lakes and the food chain, for example, precisely because one society refuses to develop a scheme for coordinating action. Some may thus argue that the relevant existence condition for justice is simply the existence of pervasive impact on life chances, regardless of whether social coordination actually exists.[41] In other words, some may argue that social coordination can indeed be a demand, not an existence condition, of justice.[42] Again, I do not claim to settle the issue here; the point is that such arguments move us away from the cooperation theory and towards a pervasive impact theory.
There are at least two problems with the revised basic structure argument. One way to undermine it would be to argue that a global basic structure already exists in the relevant sense.[43] However, let us grant, for the sake of argument, that no global basic structure exists in the sense relevant to the cooperation theory. Once it is recognized that a basic structure is not an existence condition, but an instrumental condition of justice, a critic need only demonstrate the feasibility of a (just) basic structure. A proponent of the basic structure argument, in turn, must rely on one of two feasibility claims: that no global basic structure of any kind is feasible, or that no just global basic structure is feasible. Since the first is a rather heroic claim, I assume that proponents of the revised basic structure argument rely on the second claim: not that a global basic structure is unfeasible per se, but that it is undesirable, presumably because any full-fledged global structure, capable of regulating the fundamental terms of social cooperation, would be intolerably unjust.[44]
Let us grant this empirical assumption (which in fact I believe is false) as well. The point is this: even if it were true that a just global basic structure is not feasible, the scope of justice would still remain global, if some other means were available for at least improving justice at the global level. The reason, again, is that on the cooperation theory a (just) basic structure is a necessary instrumental condition for fully realizing justice. If justice can be realized to lesser and to greater degrees, and if the necessary instrument for its full realization is not available, all that follows is that other, second-best means must be deployed for at least a partial realization of justice. To deny this is to make the possibility of fully realizing an ideal an existence condition of the ideal itself, which is absurd. As far as the scope of justice is concerned, the feasibility limit that ultimately matters is the feasibility of justice itself, not the feasibility of any of its particular means of realization (such as the basic structure). Without a full-fledged basic structure, more or less justice could still be realized depending on how global social coordination were regulated by major social and political institutions, even if these institutions did not necessarily form a single, global basic structure.
The second problem with the revised basic structure argument lies directly in the normative premise 1.1. By construing justice as being constituted (at least in part) by a fair system of social cooperation, the cooperation theory effectively limits the scope of justice to those persons who could, at least in principle, become participants in a fair system of social cooperation for mutual advantage. Yet as Robert Goodin has pointed out, this effectively excludes the unproductive and infirm, who are incapable of making advantageous contributions to society, from the scope of social justice.[45] Rawls has conceded that an adequate theory cannot exclude the congenitally handicapped from the scope of distributive justice; and he is himself unsure whether his theory of justice as fairness is equipped to handle this problem.[46] Regardless of whether it can handle the problem or not, what is relevant for our purposes is the problem's source: the fact that the set of participants in a system of social coordination that is, or could become, a scheme of social cooperation is not coterminous with the set of persons whose life chances are profoundly and pervasively affected by such a scheme. This points directly to the second justification that Rawls gives for why the basic structure is the primary subject of justice.
Rawls offers two distinct justifications for his claim that society's basic structure is the primary subject of justice. The first justification, whose implications we have just examined, is the division of labor argument. His second justification is that the basic structure of the society to which a person belongs has a "profound and pervasive" impact on his or her life chances and, indeed, on his or her "aims, aspirations, and character."[47] This second justification has rather different implications than the first. If the basic structure is the primary subject of justice because of the second justification, then it would seem that the principles of justice appropriately apply to all institutions that have pervasive impact on persons’ life chances. Which is to say that if the primary subject of justice is the basic structure, then, as G. A. Cohen has pointed out, the basic structure must comprise, by definition, all major social and political institutions that have pervasive impact on persons’ life chances (regardless of whether they regulate the fundamental terms of social cooperation). In contrast to the first justification, the second further suggests that if the primary site of justice is a society's basic structure, then the scope of justice is the range of persons whose life chances are pervasively impacted by it.[48] Which is to say that the pervasive impact theory, unlike the cooperation theory, is compatible with the site/scope thesis.
Although both the cooperation and the pervasive impact theories have textual support in Rawls, once they are distinguished, it becomes clear that they differ, and it is important to see how. According to the cooperation theory, (a) the basic structure comprises the institutions that regulate the fundamental terms of social cooperation; (b) the primary site of justice is the institutions that regulate these fundamental terms; and (c) the scope of justice is those persons who participate in a potentially cooperative scheme of social coordination. The problem with the cooperation theory, from the perspective of the pervasive impact theory, is that the range of persons pervasively impacted by the institutions of the basic structure is not equivalent to the range of persons who participate in a potentially cooperative scheme of social coordination. This is true because of a rather simple fact: every such scheme has externalities, i.e., those who (can) participate are not necessarily the same as those who are affected.[49] So if the basic structure has normative import for determining the scope of distributive justice only because it is serving as a proxy for the scope of pervasive impact, then the cooperation theory flounders when the proxy fails, as in the case of the infirm or of foreigners. According to the pervasive impact theory, then, (a) the basic structure comprises the institutions that have pervasive impact on persons’ life chances; (b) the primary site of justice is the institutions that have pervasive impact; and (c) the scope of justice is those persons whose life chances are pervasively impacted by these institutions.
On such a view, the existence condition of justice is met, and considerations of distributive justice arise, if and only if there exist major social institutions (i.e., society's basic structure) with pervasive impact on the life chances of persons. The pervasive impact theory thus yields a third version of the basic structure argument:
1.2. The scope of justice consists of those persons whose life chances are pervasively impacted by a society's basic structure.
2.4. The range of persons whose lives are pervasively impacted by any given existing basic structure is not global in scope.
Therefore:
3. The scope of justice is not global.
The advantage of the pervasive impact theory for the anticosmopolitan is that, unlike the cooperation theory, it allows the basic structure to be viewed as an existence condition of justice: if and only if there are major social institutions with pervasive impact on people's life chances do the demands of justice arise. Hence the prima facie plausibility of premise 1.2. There are, however, two major disadvantages that the pervasive impact theory has for the basic structure argument, one for each of the argument's supporting premises.
The problem with premise 1.2 is that its specification of the scope of justice threatens to undermine the pervasive impact theory's specification of the site of justice, viz., the institutions that have pervasive impact on persons’ life chances. G. A. Cohen, who finds that Rawls equivocates between a pervasive impact theory and a coercion theory, has argued that the pervasive impact theory undermines Rawls's special focus on institutions: if the principles of justice apply to the institutions of the basic structure because these institutions have pervasive impact on life chances, then why, Cohen asks, should they not also apply to other objects, such as individuals’ actions, which also have such impact? Rawls would have answered this question, of course, with his division of labor argument. This is also, in part, how Thomas Pogge defends Rawls: that the basic structure ought to promote some goal D does not mean that other institutions or individuals ought also to promote D via their own actions. This is because D may only be a constituent part of the total set T of morally significant goals, and the fact that the basic structure ought to promote the D component of T may simply be because exclusively assigning the promotion of D to the basic structure is the best way of ultimately realizing T.[50] The problem with invoking the division of labor argument in the present context, however, is that it depends on the cooperation theory, not the pervasive impact theory in question.
However that may be, the argument's fatal premise is the empirical premise 2.4. Although the pervasive impact theory allows the anticosmopolitan to include the basic structure as an existence condition of justice, the theory's interpretation of what the basic structure is renders premise 2.4 wildly implausible. Recall that the cooperation theorist was able to deny, with at least some degree of plausibility, the existence of a global basic structure because what she meant by basic structure involved institutions systematically regulating the fundamental terms of social cooperation. But on the theory now in question, the basic structure is simply defined in terms of the profound or pervasive impact of social institutions on persons’ life chances. It is under the influence of the pervasive impact theory that Rawls, in order to defend the anticosmopolitan thesis that distributive justice is bounded within a single polity, is compelled in later work to resort to rather dodgy empirical claims about the impact of foreign polities and nondomestic institutions on the life chances of human beings across the globe. And note that premise 2.4 can be defeated in one of two ways: by showing either that there is indeed a global basic structure or that there exist some domestic basic structures with profound impact on the lives of outsiders. Since it seems to me that whatever plausibility the no-global-basic-structure premise has depends on construing the basic structure in terms of cooperation or coercion, I will not pursue the pervasive impact version of the anticosmopolitan basic structure argument any further: it seems to me to provide the weakest possible basic structure argument against the cosmopolitan scope of principles of distributive justice.[51]
Much more promising for anticosmopolitan Rawlsians, in my view, is the coercion theory of distributive justice, the most important recent defenders of which have been Michael Blake and Thomas Nagel. For Blake and Nagel, what grounds the requirement and scope of distributive justice is not the fact of cooperation, nor the fact of pervasive impact per se, but the fact of state coercion.[52] Rather than tying the scope of distributive justice to persons under shared institutions regulating the fundamental terms of social cooperation, they tie it to state coercion governed by the shared political-legal apparatus of a single state. Responsibilities of distributive justice attach to state practices and institutions because of the need to justify the state's use of coercion. The basic structure within which requirements of distributive justice arise, on this view, comprises the major coercive institutions of a society. This focus on coercion is supposed to explain why the scope of Rawlsian distributive justice is restricted to the domestic context. I focus here on Blake's formulation of the coercion theory, while briefly considering Nagel's distinctive contribution.
The power of Blake's argument derives from his appeal to a prominent liberal value: the freedom of individuals from state coercion. Blake frames the issue in terms of the value of autonomy, by which he understands (following Joseph Raz) a "vision of people controlling, to some degree, their own destiny," such that they can see themselves as "part creators of their own moral world."[53] Acts of coercion directly encroach upon autonomy by eliminating options or courses of action otherwise available to an agent, thereby subjecting the agent to the will of another.[54] Blake thus adopts what he calls the "autonomy principle," according to which state coercion, which "is necessarily an affront to autonomy," must "be either justified or eliminated."[55] (The autonomy principle, then, provides Blake's answer to G. A. Cohen's question about what is so normatively significant about coercive institutions, as opposed to institutions with pervasive impact.)
For Blake, the challenge is that without "some sort of state coercion, the very ability to autonomously pursue our projects and plans seems impossible; settled rules of coercive adjudication seem necessary for the settled expectations without which autonomy is denied."[56] So if, for the sake of autonomy itself, state coercion cannot be eliminated, then state coercion must be justified. More formally, the autonomy principle can be formulated as a demand for justification:
A1: If x is ongoing state coercion against an individual, then x requires a justification consistent with the autonomy of that individual.
The ensuing question is what such a justification would require. Beyond showing that the particular practices and institutions of state coercion are necessary for establishing the general conditions of autonomy, Blake's view is that, to justify ongoing state coercion, the autonomy principle demands that the state concern itself with the relative material deprivation of those whom it coerces; it demands, in other words, the application of comparative principles of distributive justice to state action.[57] This provides Blake's reconstruction of Rawlsian theory. More formally, the principle of distributive justice complementing the autonomy principle states:
A2: If x is ongoing state coercion against an individual that requires justification consistent with the autonomy of that individual, then x requires concern by the state for the relative deprivation of that individual.
The combination of the autonomy principle A1 and the principle of distributive justice A2 yields:
A3: If x is ongoing state coercion against an individual, then x requires concern by the state for the relative deprivation of that individual.
I will not here call into question the coercion theory of distributive justice, encapsulated by theses A1–A3. I will take its truth for granted. I focus instead on the anticosmopolitan conclusion that Blake derives from applying the coercion theory to questions of global justice.
Blake's anticosmopolitan thesis is that the scope of Rawlsian principles of distributive justice is restricted to the citizens of a single state: "we owe distinct things to fellow nationals" because "an impartial principle will give rise to distinct burdens of justification between individuals who share liability to the coercive power of the state." Hence, "a concern for specifically economic egalitarianism is only morally required within the context of a domestic legal system."[58] More formally, Blake's anticosmopolitan thesis holds:
ACT: If someone is a foreigner of a state, then nothing requires concern by the state for her relative deprivation.
This thesis is the conclusion of an argument with one normative and one empirical premise. The normative premise is that only ongoing acts of state coercion require concern for relative deprivation, and only for the individuals coerced: "concern with relative economic shares" is required "only when [liberal] principles are applied to individuals who share liability to the coercive network of state governance"; "only between people who share the coercive mechanisms of a state does a concern for specifically economic egalitarian justice become appropriate."[59] More formally, Blake's normative premise is:
N1: If x requires concern by the state for the relative deprivation of an individual, then x is ongoing state coercion against that individual (requiring justification).[60]
Blake's empirical premise is that no ongoing state coercion exists at the international level: that people are only subject to ongoing state coercion by their own state:
In the international arena . . . no institution comparable to the state exists. No matter how substantive the links of trade, diplomacy, or international agreement, the institutions present at the international level do not engage in the same sort of coercive practices against individual moral agents.[61]
In other words, there is no global coercive basic structure, and domestic coercive basic structures do not coerce foreigners. More formally (and granting Blake's simplifying assumption that all and only the residents of a given state are its citizens), to obtain ACT from N1 Blake requires the following empirical premise:
E1: There is no ongoing state coercion against any individual such that she is a foreigner of the state.
Combining N1 and E1 yields Blake's anticosmopolitan thesis ACT. In Blake's succinct formulation, "To insiders, the state says: Yes, we coerce you, but we do so in accordance with principles you could not reasonably reject. To outsiders, it says: We do not coerce you, and therefore do not apply our principles of liberal justice to you."[62]
Of course, the state may say "we do not coerce you" to foreigners until it is blue in the face, but its say-so does not make it so. The empirical premise of this formulation of Blake's argument is straightforwardly false: it is one of the most salient features of the contemporary Westphalian interstate system that individuals are subject to a vast network of ongoing coercion by foreign states that restrict their movement across state borders. States today, including self-proclaimed liberal states, use coercion against foreigners on a massive and ongoing basis to prevent them from entering their territory at will.[63]
Blake explicitly considers this objection; his response is that interstate border coercion against noncitizens comprises a different form of coercion. "[E]ach form of coercion," he says, "requires a distinct form of justification. . . . The mere fact that exclusion is coercive does not erase the distinction between prospective and current membership."[64] The obvious question to ask is, What is distinctive about the form that ongoing state coercion takes domestically?
One answer is that ongoing domestic state coercion is different from coercion against foreigners because only the former "profoundly and pervasively" affects a person's life chances. This, of course, is to supplement Blake's coercion theory with elements from the pervasive impact theory. More formally, the argument here is:
N2: If x requires concern by the state for the relative deprivation of an individual, then x is ongoing state coercion against that individual that also has pervasive impact on her life chances.
E2: If someone is a foreigner, then ongoing state coercion against that individual does not have pervasive impact on her life chances.
The most immediate problem here, once again, is that the empirical premise is manifestly false: the reason why a tremendous number of human beings sell every possession they have and risk life and limb to cross state boundaries is because the Westphalian system of state borders and the policing of those borders do "profoundly and pervasively" affect human beings’ life chances in our world.
Another potential answer for distinguishing between types of coercion draws on a well-worn distinction between law-governed state coercion, i.e., coercion regulated by a legal system, and lawless state coercion, i.e., coercion carried out by executive prerogative or decree and not tempered by the rule of law. Blake's point, it might be argued, is not that there is no ongoing state coercion at the international level—he explicitly concedes that "Coercion can, after all, occur both between nations as well as within them"[65]—but that there is no law-governed state coercion at the international level. A coercive basic structure is one that is legally articulated. On this revised, third version of the argument for ACT, Blake is concerned with "the circumstances of all those facing legal coercion."[66] His empirical premise is simply that there is no international coercive legal system.[67] More formally:
E3: If someone is a foreigner, then there is no ongoing coercion against her by the state (to which she is foreign) that is regulated by a system of law.
I take it that E3 is more or less true: there is no coercively enforced international legal system, and state policy towards foreigners and other states is, by and large, conducted according to executive prerogative. There are qualifications to this, of course, and increasingly so, but I shall grant the premise without qualification for the sake of argument. The real problem with this revised empirical premise is that N1 and E3 do not yield ACT. The normative premise that E3 requires to yield ACT is the premise that concern for relative deprivation is required only by ongoing state coercion and only if that coercion is conducted according to a system of law. More formally, it requires:
N3: If x requires concern by the state for the relative deprivation of an individual, then x is ongoing state coercion against that individual regulated by a system of law.
This is, rather surprisingly, in fact what Blake seems to say: "the existence of a coercive network of law is a precondition of a concern with relative deprivation," i.e., it is a necessary existence condition of distributive justice.[68] That Blake actually says this is surprising because making N3 explicit exposes the perversity of the third version of the coercion theory's argument for ACT. The state no longer says to foreigners, as Blake had suggested, "We do not coerce you, and therefore do not apply our principles of liberal justice to you"; it says, rather, "We not only coerce you, but we coerce you without subjecting our ongoing coercion to the constraints of a legal system and the rule of law, and therefore we have no responsibilities of comparative distributive justice to you." The perversity of the argument is clear: on this interpretation, Blake is suggesting that the fact that states mistreat foreigners by coercing them lawlessly, without tempering coercion with the rule of law, is what justifies denying them concern for relative deprivation and constricts the scope of distributive justice.
Essentially the same problem, the problem of perversity, plagues Thomas Nagel's formulation of the coercion theory and, in particular, his attempt to specify what is distinct about domestic state coercion. According to Nagel's appropriation of Rawls, demands of justice arise if and only if two existence conditions are met: persons are (1) subject to ongoing coercion (2) that is carried out in their name. The first condition refers to being subject to the norms of a coercively imposed basic structure; the second condition holds only if those subject to coercion are the "putative joint authors of the coercively imposed system" that "actively engages" their will.[69] Domestic state coercion meets this condition but, according to Nagel, interstate forms of coercion such as border coercion do not: "Immigration policies are simply enforced against the nationals of other states; the laws are not imposed in their name, nor are they asked to accept and uphold those laws. Since no acceptance is demanded of them, no justification is required." Nagel's premises are thus:
N3': If x requires concern by the state for the relative deprivation of an individual, then x is ongoing state coercion against that individual regulated by a system of law carried out in her name, i.e., actively engaging her will.
E3': If someone is a foreigner, then there is no ongoing coercion against her by the state (to which she is foreign) that is carried out in her name (i.e., that engages her will).
The problem is that Nagel's normative premise, like the premise N3 that it parallels, is perverse: it implies that a state can exempt itself from the demands of justice simply by ensuring that the coercion to which it subjects persons is pure coercion without any pretense of accountability, i.e., by denying to those whom it coerces any standing as putative authors of the system of coercion. This seems to suggest, for example, that colonial rule is exempt from the demands of justice precisely because it does not claim to act in the name of the colonized. Nagel disputes this by interpreting the criterion of "engaging the will" expansively: "there is a sense," he claims, "in which it [colonial rule] is being imposed in their [the colonized] name." In particular, colonial rule (a) requires the "normative engagement" of its subjects since they are "expected to uphold as participants" the enforced system of colonial laws, and (b) "is intended to serve their interests."[70] The veracity of Nagel's empirical claim, that colonial rule was intended to serve the interests of the colonized, can be disputed, but that would be beside the point. The point is that no matter how expansively the "engaging the will" existence condition is interpreted, insofar as N3' is supposed to be distinct from N1, there must be some cases of pure coercion that do not meet the condition. However it is defined, the perverse implication of this condition is that a tyrant can exempt himself from the demands of justice by relying solely on pure coercion (whatever it means). The closer a tyrant's rule approaches pure slavery, the less it can be criticized for being unjust.[71]
Let us take stock of what I have argued so far. I have considered three possible versions of Blake's argument for his anticosmopolitan thesis ACT. The first version of the argument fails because its empirical premise E1 is false: there is ongoing state coercion against foreigners. The next two versions each rely on normative premises that are more demanding than N1: they are more demanding because, in each case, they add a further necessary existence condition beyond state coercion before responsibilities of distributive justice come into play. The normative premise N2 of the second version, for example, requires not only ongoing state coercion but also pervasive impact. However, this second version fails because, like the first, its empirical premise is false: pace E2, the ongoing coercion to which states subject foreigners has pervasive impact on their life chances. And although the third version's revised empirical premise E3 is more or less true, the new stipulation added by its normative premise N3 is perverse: by denying responsibilities of distributive justice to foreigners on the grounds that state coercion against them is not constrained by the rule of law (or, as in Nagel's argument, "in their name"), it simply adds insult to injury.
Before turning to perhaps the most sophisticated formulation of Blake's argument, we must first clarify a potential ambiguity, namely, the difference between necessary and sufficient conditions. It will be recalled that Blake's coercion theory merely establishes a sufficient existence condition of justice:
A3: If x is ongoing state coercion against an individual, then x requires concern by the state for the relative deprivation of that individual.
Yet Blake's anticosmopolitan argument requires a necessary condition. Recall the argument's normative premise in its most basic version:
N1: If x requires concern by the state for the relative deprivation of an individual, then x is ongoing state coercion against that individual.
Blake's coercion theory says that ongoing state coercion is a sufficient condition for requiring concern for relative deprivation, but the premise his anticosmopolitan argument needs (in all its versions) is that ongoing state coercion is a necessary existence condition of concern for relative deprivation. So when Blake says that the "legal system is coercive, and thus stands in prima facie conflict with the liberal principle of autonomy,"[72] the conclusion follows straightforwardly from his autonomy principle: a legal system requires justification because it is coercive.[73] Yet when he appeals to his coercion theory to establish his anticosmopolitan thesis, there is a difficulty:
It is only because [A1] they [impartial liberal principles] are required for the justification of coercive force to all those who face it, I argue, that [A2] a moral concern with relative deprivation is implied by a liberalism committed to autonomy, and therefore [ACT] a concern for specifically economic egalitarianism is only morally required within the context of a domestic legal system.[74]
The difficulty is that A1 and A2 establish A3, not ACT. The autonomy principle simply tells us what is the case when there is ongoing state coercion; it does not tell us what is the case when there is none. Blake could potentially avoid this difficulty: he could argue that, although strictly speaking the coercion theory only demonstrates that state coercion is a sufficient existence condition of distributive justice, as a matter of fact no other conditions can be plausibly shown to give rise to demands of distributive justice. In other words, the autonomy principle explains why ongoing state coercion is inherently related and gives rise to demands of distributive justice; but no plausible explanation can be given for why any other phenomenon is inherently related and gives rise to such demands. Effectively, therefore, state coercion is a necessary existence condition of justice as well.[75]
This retort may be plausible for the most basic version of Blake's anticosmopolitan argument, according to the first premise (N1) of which only ongoing state coercion gives rise to demands of justice.[76] Its plausibility vanishes, however, once N1 is strengthened with additional conditions beyond state coercion, because each additional condition must be construed as another necessary existence condition of justice as well. We can see this with the most complete and promising explanation for the distinctiveness of domestic coercion that one might find in Blake. He ultimately seems to argue that there is something distinctive about ongoing coercion conducted through private or civil law, which, through laws of property, tort, and contract, regulates the fundamental terms of economic production, exchange, and distribution. Blake wants to argue that the coercive imposition of private law requires a kind of justification distinct from other forms of ongoing state coercion (such as border coercion, or the coercive enforcement of criminal law). In particular, it requires the application of comparative principles of distributive justice.[77]
The problem with this account of what is distinct about domestic state coercion arises as soon as we ask why private law should be inherently related in this way to comparative principles of distributive justice. Indeed, since this kind of coercion must effectively be a necessary existence condition of justice, we must also ask why the private law should be uniquely related in this way. There are two responses here. One is that private law is related to comparative principles of distributive justice because, by regulating the fundamental terms of economic production, exchange, and distribution, it has "profound and pervasive" impact on the distribution of resources and, so, on the life chances of those subject to it. As we have seen, however, such an answer fails to distinguish private law from border coercion, much less other kinds of domestic state coercion. The coercive policing of state borders has profound distributive consequences; indeed, it is in large part because of these consequences that states police their borders in the first place. At best, then, this first response merely shows that imposing private law is a sufficient, but not necessary, existence condition of justice.
The second response is that private law is inherently related to distributive justice precisely because private law regulates the fundamental terms of social cooperation (through enforcing an economic regime and creating entitlements to material holdings). If the first response draws on the pervasive impact theory, the second draws on the cooperation theory. But this response now faces the dilemma that the cooperation theory itself poses for anticosmopolitans. Either social cooperation is understood in idealized terms (so that a fair system of genuine social cooperation characterized by reciprocity is thought to be an existence condition for principles of distributive justice), or it is not (so that mere social coordination jointly with coercion is an existence condition of justice). If it is, then this defense of Blake's theory illicitly conflates the existence and constitutive conditions of justice (which, as we have seen, incurs an arbitrary status quo bias). If it is not, then practices of border coercion (not to mention coercively enforced interstate economic institutions and regimes) also satisfy the relevant condition: border coercion, like the coercive imposition of private law, is a significant tool for regulating the terms of production, exchange, and distribution. So it will not do to point out that the state imposes its rules of property, contract, and so forth, only on its own citizens, because, even if this were true (and it is not always so), it is simply false that the terms of social coordination and access to material resources are coercively regulated only by domestic rules of property ownership or contract. Coordination and access are also regulated by border coercion: as is obvious to any migrant worker, the coercive regulation of membership and cross-border movement are important components of the state-imposed regime regulating the terms of economic production, exchange, and distribution.[78] Again, it will not do to retort that border coercion does not amount to a system of social cooperation: cooperation in the idealized sense is not an existence condition but a constituent of Rawlsian justice.
More formally, then, the premises of Blake's most complete argument are:
N4: If x requires concern by the state for the relative deprivation of an individual, then x is (i) ongoing state coercion against that individual (ii) through a legal system to which she is subject that (iii) regulates the fundamental terms of social cooperation (economic production, exchange, and distribution) and/or that (iv) has pervasive impact on her life chances.
E4: If someone is a foreigner, then ongoing state coercion against that individual does not subject her to a legal system regulating the fundamental terms of economic production, exchange, and distribution and/or having pervasive impact on her life chances.
We can summarize the problems with this argument as follows: condition (i) does not distinguish the coercive imposition of private law from other forms of state coercion; condition (ii) is perverse and, in any case, does not distinguish private law from other forms of law; if "social cooperation" in condition (iii) is understood to mean (nonidealized) social coordination, then it is satisfied empirically at the interstate level; and condition (iv) is satisfied empirically at the interstate level as well.
One of the central claims of Rawls's theory of justice is that the primary subject of justice is society's basic structure. I have argued that there are three different ways plausibly to understand this idea, corresponding to the cooperation, pervasive impact, and coercion theories of distributive justice. Each theory (a) defines the basic structure in a distinct way; (b) appeals to a distinct justification for why the basic structure, so defined, is the appropriate (and only appropriate) site for applying principles of justice; and, consequent to this definition and justification, (c) draws conclusions about the scope of distributive justice. The cooperation theory defines the basic structure as those major social and political institutions that regulate the fundamental terms of social cooperation; justifies the claim that such institutions are the primary site of justice with an instrumental argument according to which applying the principles of justice to such institutions is the best way to realize justice; and equates the scope of justice with the scope of social interaction. The pervasive impact theory defines the basic structure as those major social and political institutions that have pervasive impact on persons’ life chances; justifies the claim that such institutions are the primary site of justice by referring to their pervasive impact; and equates the scope of justice with the scope of that impact. Finally, the coercion theory defines the basic structure as those institutions subjecting persons to ongoing state coercion; justifies the claim that such institutions are the primary site of justice by reference to the autonomy principle's demand that coercion be justified; and equates the scope of justice with the range of persons subjected to ongoing state coercion.
Regardless of which theory is used to interpret it, I have argued, the Rawlsian idea that the basic structure is the primary subject of justice implies that justice is today global in scope. The cooperation theory has cosmopolitan implications because the scope of social interaction, which the theory construes to be the necessary and sufficient existence condition of justice, is global. The pervasive impact theory, in turn, supports cosmopolitanism because the pervasive impact of currently existing basic structures is global in scope. Finally, the coercion theory supports cosmopolitanism because of the global scope of the interstate system of border coercion and the economic regime it imposes.
This finding may be surprising to anticosmopolitans, because many have turned to this Rawlsian idea—that the basic structure is the primary subject of justice—in the hopes of restricting the scope of distributive justice, without having to appeal to nonindividualist or partialist moral premises (concerning, for example, the intrinsic significance of national ties) that many see as external to a Rawlsian theory of justice. Yet these hopes have flourished, it seems to me, thanks to a failure in the literature to distinguish between the cooperation, pervasive impact, and coercion theories. Distinguishing them helps to demonstrate that their concatenation does not solve the problems that each individually poses for anticosmopolitans. The more sophisticated construal of Michael Blake's anticosmopolitan argument is a good illustration of this: by shifting between these different components, it obscures rather than resolves the defects of the anticosmopolitan basic structure argument. Once the relevant distinctions are made, anticosmopolitan Rawlsians' turn towards the basic structure is seen to backfire: not only does the Rawlsian basic structure idea fail to support anticosmopolitans, but on any plausible interpretation it also positively stands against them.
This finding also means that "Law of Peoples Rawlsians" who follow Rawls himself in denying the global scope of distributive justice face a considerable challenge: although the relation between Rawls's Law of Peoples and his theory of justice is a matter of considerable controversy, the centrality of the notion of the basic structure to his theory is incontrovertible. If I am right, Rawlsians who choose his Law of Peoples as the appropriate model for thinking about global justice must abandon—unless they are able to develop a novel theory of the basic structure and its relation to justice—an inseparable feature of Rawls's celebrated theory of justice. They must, in other words, abandon his theory of justice altogether.
For valuable comments, I would like to thank Michael Blake, Pablo Gilabert, Catherine Lu, Don Moon, Alan Patten, Veronica Ponce, Sarah Stroud, Daniel Weinstock, the Editors of Philosophy & Public Affairs, and participants at a colloquium of the Centre de Recherche en Éthique de l’Université de Montréal, April 16, 2005. This research was supported by a team grant from the Fonds québécois de la recherche sur la société et la culture.
[1] In this article, unless otherwise noted, the term justice refers to distributive justice, and not to what Rawls called natural justice; by cosmopolitan I simply mean the view that the scope of distributive justice is global, by anticosmopolitan the view that it is not.
[2] John Rawls, A Theory of Justice (Cambridge, Mass.: Harvard University Press, 1971), pp. 4, 96, 7; cf. p. 457: the "boundaries" of "the basic structures of well-ordered societies . . . are [assumed to be] given by the notion of a self-contained national community."
[3] I say "Rawlsian" rather than "Rawls's" because (as an Editor of Philosophy & Public Affairs has pointed out) whether Rawls himself was ultimately committed to the basic structure argument is a matter of some interpretive ambiguity. On the one hand, in Theory of Justice and Political Liberalism, paperback ed. (Cambridge, Mass.: Harvard University Press, 1996), Rawls presented his anticosmopolitanism merely as a provisional but methodologically useful assumption. He did not provide an explicit argument for his anticosmopolitanism and, in The Law of Peoples (Cambridge, Mass.: Harvard University Press, 1999), when he did provide one, he did so without explicitly appealing to the basic structure argument. On the other hand, the basic structure argument is arguably implicit to Theory of Justice. When in that work Rawls relaxed the assumption of closed, self-sufficient societies to consider the issue of international justice, he did not conclude that "justice between states" requires distributive justice but, rather, that it requires principles familiar from traditional international law for "adjudicat[ing] conflicting claims among states" (pp. 377–79). The implicit assumption backing this conclusion seems to have been that while empirical reality requires relaxing the assumption of self-sufficiency, it does not require relaxing the assumption that there is no global basic structure: even when he relaxed the assumption of self-sufficiency to consider issues in international justice, he continued to assume that "we have already derived the principles of justice as these apply to societies as units and to the basic structure" (p. 377, my emphasis). Moreover, as Allen Buchanan, " Rawls's Law of Peoples: Rules for a Vanished Westphalian World," Ethics 110 ( 2000): 697– 721, suggests, the basic structure argument is even implicit to Law of Peoples. Finally, in Justice as Fairness: A Restatement, ed. Erin Kelly (Cambridge, Mass.: Harvard University Press, 2001), p. 11, Rawls appeared to endorse the argument when he justified the distinction between principles of domestic justice and principles of global justice (the latter of which are not principles of distributive justice) by exclusively equating the former to "the principles applying to the basic structure of society" and the latter to "principles applying to international law," which again implies that the assumption that there is no global basic structure (in the sense relevant to distributive justice) is not relaxed. However the interpretive ambiguity is resolved, what is clear is that the basic structure argument has been enormously influential for contemporary Rawlsians. See, e.g., Charles R. Beitz, Political Theory and International Relations, 2nd ed. (Princeton, N.J.: Princeton University Press, 1999), p. 132, and, for the attribution of the basic structure argument to Rawls himself, J. Donald Moon, " Rawls and Habermas on Public Reason," Annual Review of Political Science 6 ( 2003): 257– 74, at p. 265, and Samuel Freeman, "Distributive Justice and The Law of Peoples," in Rawls's Law of Peoples: A Realistic Utopia?, ed. Rex Martin and David A. Reidy (Oxford: Blackwell, 2006).
[4] I borrow the expression "comparative principles" from David Miller, "The Limits of Cosmopolitan Justice," International Society, ed. David R. Mapel and Terry Nardin (Princeton, N.J.: Princeton University Press, 1998), pp. 164–81, at p. 169. As I define the terms, Andrea Sangiovanni's recent article denies that principles of distributive justice (which, in my sense, are concerned with equality) are global in scope. " Global Justice, Reciprocity, and the State," Philosophy & Public Affairs 35 ( 2007): 3– 39.
[5] Michael Blake, " Distributive Justice, State Coercion, and Autonomy," Philosophy & Public Affairs 30 ( 2002): 257– 96; Thomas Nagel, " The Problem of Global Justice," Philosophy & Public Affairs 33 ( 2005): 113– 47.
[6] See Thomas W. Pogge, "Do Rawls's Two Theories of Justice Fit Together?" in Rawls's Law of Peoples: A Realistic Utopia?, ed. Rex Martin and David A. Reidy (Oxford: Blackwell, 2006), for questions about whether Rawls's assumptions about the moral integrity of peoples is compatible with the moral individualism he defends in Theory of Justice. Although I do not make the case here, I believe that Rawls's toleration of peoples argument in Law of Peoples is detachable from his theory of justice as fairness as a whole.
[7] See, e.g., Beitz, Political Theory and International Relations, pp. 169–70. For the debate concerning contemporary levels of globalization, see David Held et al., Global Transformations (Cambridge: Polity Press, 1999) and the articles collected in The Global Transformations Reader, ed. David Held and McGrew Anthony (Cambridge: Polity Press, 2000).
[8] See Moon, "Rawls and Habermas on Public Reason," p. 265. As noted above (n. 3), in Theory of Justice Rawls was willing to relax the assumption of self-sufficiency but not that the basic structure is domestic in scope.
[9] This important distinction, between scope and site, has not been adequately appreciated in the literature. Andrew Williams, for example, uses the term "scope" to refer to what I call site. See his " Incentives, Inequality, and Publicity," Philosophy & Public Affairs 27 ( 1998): 225– 47, at pp. 229, 231.
[10] As we shall see, this is Samuel Freeman's interpretation, according to whom an already existing shared basic structure is a necessary condition for relations of justice to arise in the first place ("Distributive Justice and The Law of Peoples").
[11] I borrow the term "existence condition" from Charles R. Beitz, " Cosmopolitan Ideals and National Sentiment," The Journal of Philosophy 80 ( 1983): 591– 600, at p. 595.
[12] I quote Rawls's most mature formulation, from Justice as Fairness, p. 10, which follows very closely the wording of Rawls, Theory of Justice, p. 7. See also Thomas W. Pogge, Realizing Rawls (Ithaca, N.Y.: Cornell University Press, 1989), pp. 20–23.
[13] John Rawls, Political Liberalism, p. 258, and Justice as Fairness, p. 10. Cf. Theory of Justice, p. 7.
[14] See Thomas Scanlon, "Rawls's Theory of Justice," Reading Rawls, ed. Norman Daniels (New York: Basic Books, 1974), especially pp. 192–93.
[15] The principles of justice as fairness "do not apply directly to or regulate internally the institutions and associations within society. Firms and labor unions, churches, universities, and the family are bound by constraints arising from the principles of justice, but these constraints arise indirectly from just background institutions within which associations and groups exist, and by which the conduct of their members is restricted." Justice as Fairness, p. 10; see also pp. 7, 162–68.
[16] Rawls, Political Liberalism, p. 283.
[17] Freeman, "Distributive Justice and The Law of Peoples," p. 245. The assumption that Rawlsian distributive justice is in the first instance addressed to the allocation of goods is one of the shortcomings of Joseph Heath's recent defense of anticosmopolitan Rawlsianism in "Rawls on Global Distributive Justice: A Defence," in Global Justice, Global Institutions: Canadian Journal of Philosophy Supplementary Volume 31, ed. Daniel Weinstock (Calgary: University of Calgary Press, 2005).
[18] Political Liberalism, pp. 265–69; Justice as Fairness, pp. 52–55.
[19] Political Liberalism, p. 266.
[20] Justice as Fairness, p. 54. The instrumental character of Rawls's division of labor argument for taking the basic structure as the primary site of justice assumes that a normative aim of both institutions of the basic structure and the individuals who inhabit them ought to be distributive justice. The division of labor argument simply asserts that as a matter of empirical fact the best way to realize this aim is to apply the principles directly to the basic structure. This means that, given the way that Liam B. Murphy, " Institutions and the Demands of Justice," Philosophy & Public Affairs 27 ( 1999): 252– 91, p. 280, defines the distinction between monist and dualist normative theories, he is not entitled to attribute dualism to Rawls. As Murphy acknowledges, the division of labor argument is indifferent between monism and dualism (p. 262). In principle, however, he is entitled to challenge—as he does—the division of labor argument on the grounds that applying principles of justice directly to individual conduct (as opposed to the basic structure's institutional design) need not inherently be more burdensome.
[21] Justice as Fairness, p. 6.
[22] It might be argued that, for Rawls, a (just) basic structure is a constituent part of the ideal of justice, not merely an instrument for realizing it. This would not undermine the argument that follows, because (unlike existence conditions) constitutive and instrumental conditions both merely impose feasibility limits on the scope of justice.
[23] Something like this view is advanced by Brian Barry, "Humanity and Justice in Global Perspective," in Ethics, Economics, and the Law: NOMOS XXIV, ed. J. Roland Pennock and John W. Chapman (New York: New York University Press, 1982), pp. 232–33. Barry argues that duties of "justice as fair play" arise only from social coordination that meets "conditions for reciprocity."
[24] Charles Beitz makes a similar point in his example of slaves forced to contribute to a scheme of social coordination against their will without any advantage to themselves (Political Theory and International Relations, p. 131).
[25] Others, including Barry himself, have worried that a status quo bias plagues Rawls's theory of justice. See Brian Barry, Justice as Impartiality (Oxford: Clarendon, 1995). Cf. Aaron James, " Constructing Justice for Existing Practice: Rawls and the Status Quo," Philosophy & Public Affairs 33 ( 2005): 281– 316. See also David A. J. Richards's critique of Barry, in which Richards argues that existing reciprocity is not an existence condition for duties of justice but, rather, that some claims of justice arise because of moral demands for reciprocity, regardless of whether reciprocity already exists ("International Distributive Justice," in Ethics, Economics, and the Law, ed. J. Roland Pennock and John W. Chapman [New York: New York University Press, 1982], pp. 288–89). Sangiovanni misconstrues the nature of the objection here when he takes such objections to deny all "relational" views of justice ("Global Justice," p. 36). (A relational view of justice, as Sangiovanni describes it, takes some shared practices, values, or institutions to be existence conditions of justice.) The objection indicts views that conflate constitutive and existence conditions of justice, and not relational scope restrictions per se. The pervasive impact and coercion theories I consider in the following sections, for example, are relational views that do not necessarily make that conflation.
[26] James's reconstruction of Rawls ("Constructing Justice") identifies an existence condition for Rawlsian justice as well, but James fails to distinguish social interaction as such from a basic structure (or social interaction regulated by it). As a result, his discussion tends to slide from invoking mere social interaction (p. 295: any "principle of social justice has as a condition of its application the existence of some social practice") to invoking basic structure (p. 298: "among the many and various existing social practices . . . we should attend first to the most ‘basic’ of existing social structures"). But this slide from social practices in general to the basic structure in particular, which leads James to conclude that the "very existence" of a basic structure "cannot be socially just or unjust" (p. 314), rests on a confusion. The cooperation theory's justification for focusing on social practices in general, and, even more generally, on social interaction, establishes mere social interaction as an existence condition of justice. The justification for focusing on the basic structure, by contrast, does not establish it as an existence condition: the cooperation theory's appeal to the basic structure is instrumental. It is hence no surprise that when James himself justifies the focus on the basic structure, he appeals to its pervasive impact (pp. 298–99).
[27] Freeman, "Distributive Justice and The Law of Peoples," p. 245–47. Despite his claim that the absence of a world state makes claims for global principles of justice "absurd," Freeman adopts a cooperation and not a coercion theory of justice, as he makes clear when distinguishing his position from Nagel's, on p. 259, n. 4.
[28] Samuel Freeman, " The Law of Peoples, Social Cooperation, Human Rights, and Distributive Justice," Social Philosophy & Policy 23 ( 2006): 29– 68, at p. 37.
[29] Joseph Heath provides another example of the failure to see the instrumental role of the basic structure in the cooperation theory: "far from instrumentalizing social institutions" of the basic structure, he claims, Rawls "takes them as his point of departure." The putative "absence of a basic structure at the international level" is thus not "merely an implementation problem"; the "rule of law" enforced by effective institutions of the basic structure is, rather, an existence condition of justice: "the absence of the rule of law at the international level is not simply . . . an implementation problem" (Heath, "Rawls on Global Distributive Justice," pp. 200, 201, 205). Heath's thesis is problematic on several fronts. First, his primary argument consists in pointing out that the basic structure is the "subject" of justice (p. 200), hence equivocating between site and scope. Second, he argues that certain "types of free-rider problems," whose resolution may be necessary for realizing social cooperation in Rawls's normative sense, "can only be resolved through" coercive enforcement via institutions of a basic structure (p. 204). Heath's point, then, is that justice requires social cooperation (in Rawls's sense) and social cooperation requires terms of cooperation coercively enforced by a basic structure. This may be true, but on the cooperation theory that Heath apparently favors, coercive enforcement is "required" only in the instrumental sense; enforcement is not an existence condition of justice. Finally, Heath argues that Rawlsian justice is constituted by considerations of both equality and efficiency, meaning that "Efficiency questions, for Rawls, are normative, not instrumental." As such, "the loss of efficiency" that would result from "global redistribution of national savings" would "trump the gain in equality" (p. 211). The problems with this final argument are at least two. First, Heath's argument against global redistribution of national savings proceeds by applying his interpretation of the Rawlsian principle of distributive justice to the globe, not by showing that it is inapplicable. Second, his distinction between "instrumental" and "normative" considerations fails to distinguish between two kinds of normative considerations. He merely shows that efficiency considerations specify constitutive conditions of justice, but not, as his anticosmopolitan argument requires, existence conditions.
[30] Rawls, Theory of Justice, p. 4.
[31] Beitz, "Cosmopolitan Ideals and National Sentiment," p. 595. Compare with Beitz's afterword to the revised edition of Political Theory and International Relations, p. 203. See also Richards, "International Distributive Justice," p. 292.
[32] As such, one could argue that Beitz's own original reformulation of the Rawlsian account in Political Theory and International Relations, which avoids including cooperation in the idealized sense as an existence condition of justice, already avoids the conflation.
[33] Barry, "Humanity and Justice," p. 227.
[34] "Justice as fair play arises not from simple exchange but from either the provision of public goods that are collectively enjoyed . . . or from quasi-insurance schemes for mutual aid" (Barry, "Humanity and Justice," p. 233). Sangiovanni's proposed existence condition of reciprocity is even more restricting: he argues that comparative principles of distributive justice arise only "among those who support and maintain the state's capacity to provide the basic collective goods necessary to protect us from physical attack and to maintain and reproduce a stable system of property rights and entitlements" ("Global Justice," p. 20).
[35] Discussing the principle of fairness or fair play, Rawls says: "We are not to gain from the cooperative labors of others without doing our fair share. The two principles of justice define what is a fair share in the case of institutions belonging to the basic structure. So if these arrangements are just, each person receives a fair share when all (himself included) do their part" (Theory of Justice, p. 112, emphasis added).
[36] Richards, "International Distributive Justice," has made the same point about Barry's argument.
[37] Rather than a constituent, this overlapping consensus may be a necessary means for realizing justice (or, at least, for maintaining its stability over time).
[38] See Jürgen Habermas's critique of Rorty's interpretation of Rawls, in Between Facts and Norms (Cambridge, Mass.: MIT Press, 1996), chap. 2, particularly pp. 58–63. See also Beitz, "Cosmopolitan Ideals and National Sentiment," p. 596.
[39] The same point can be made regarding Andrew Williams's claim that the public recognition of rules is a constitutive feature of the institutions of the basic structure ("Incentives, Inequality, and Publicity"). The public recognition of rules is a constitutive condition of justice, not an existence condition.
[40] See Scanlon, "Rawls's Theory of Justice," p. 202.
[41] Kok-Chor Tan makes this point in Justice without Borders (Cambridge: Cambridge University Press, 2004), pp. 33–34, 169–70, 174.
[42] Or, more weakly, social coordination may be a normative demand of broader normative considerations (not necessarily of justice), and demands of distributive justice may arise not only when a scheme of social coordination exists, but also when it ought to exist. In either case, the demands of justice would arise, as Rawls puts it when specifying the circumstances of justice, when "human cooperation is both possible and necessary" (Theory of Justice, p. 126). For discussion of potential duties to create institutions of social coordination and cooperation, see Pablo Gilabert, " Contractualism and Poverty Relief," Social Theory and Practice 33 ( 2007): 277– 310.
[43] Buchanan, "Rawls's Law of Peoples."
[44] Intolerably unjust because the question is not whether the ideal is fully realizable in practice, but whether attempts to realize the ideal through a basic structure would undermine or contribute to the realization of the ideal. One can extrapolate the claim that it would be intolerably unjust from Rawls's assertion that "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" (Law of Peoples, p. 36). This kind of argument, it seems to me, rests on rather old-fashioned Hobbesian views about the indivisible and centralized nature of sovereignty. A more plausible Rawlsian argument for why a (just) global basic structure is unfeasible appeals to the claim that a just basic structure requires an overlapping consensus, and that the overlapping consensus required by Rawlsian justice in particular—about the freedom and equality of persons—is not a part of whatever global public political culture there is. (A "well-ordered society," Rawls argues, is one whose basic structure is effectively regulated by a public conception of justice accepted by all its members [Theory of Justice, pp. 453–54; cf. Justice as Fairness, p. 9].) But, if an overlapping consensus is an instrumental or constitutive condition rather than an existence condition for a just basic structure, then the existence of such a consensus is not the issue here: the issue is whether such a consensus is feasible. For if an overlapping consensus about the equality and freedom of persons is instrumentally or constitutively necessary for realizing or stably maintaining Rawlsian justice over time, then Rawlsian justice would demand the recognition of the equality and freedom of persons. It is not at all clear why such recognition is not feasible at the global level in the long run; the secular trend since World War II arguably points in such a direction.
[45] Robert E. Goodin, " What Is So Special about Our Fellow Countrymen?" Ethics 98 ( 1988): 663– 86, at p. 683. See also Buchanan, "Rawls's Law of Peoples," p. 230.
[46] See footnote 59 in Justice as Fairness, p. 176.
[47] Justice as Fairness, p. 10. Cf. Rawls, Theory of Justice, p. 96: "The primary subject of justice . . . is the basic structure of society. The reason for this is that its effects are so profound and pervasive, and present from birth." In Political Liberalism, p. 269, Rawls also says that "the basic structure shapes the way the social system produces and reproduces over time a certain form of culture shared by persons with certain conceptions of their good." See also Justice as Fairness, pp. 52–57.
[48] G. A. Cohen, " Where the Action Is: On the Site of Distributive Justice," Philosophy & Public Affairs 26 ( 1997): 3– 30, at pp. 20–21. Cohen overlooks Rawls's division of labor justification for why the basic structure is the primary subject of justice.
[49] See Samuel Scheffler's discussion of the "distributive objection" in Boundaries and Allegiances (Oxford: Oxford University Press, 2001).
[50] Thomas W. Pogge, " On the Site of Distributive Justice," Philosophy & Public Affairs 29 ( 2000): 137– 69, at p. 157.
[51] See Buchanan, "Rawls's Law of Peoples," for the argument that there is indeed a global basic structure (which he construes in terms of institutions having profound and pervasive impact). Rawls claims that the life chances of persons in impoverished, politically weak states are mostly determined by the country's own "political culture—its members’ political and civic virtues" (Law of Peoples, p. 117). Domestic factors surely matter as well, but Rawls flatly ignores how international factors help shape precisely those domestic factors he thinks decisive in the first place. For a critique of Rawls's "explanatory nationalism," and the pervasive role of international financial institutions, see Thomas W. Pogge, "Human Rights and Human Responsibilities," in Global Justice and Transnational Politics, ed. Pablo De Greiff and Ciaran Cronin (Cambridge, Mass.: MIT Press, 2002), pp. 151–95, and World Poverty and Human Rights (Cambridge: Polity, 2002).
[52] "Coercion, not cooperation, is the sine qua non of distributive justice, making relevant principles of relative deprivation" (Blake, "Distributive Justice," p. 289). Cf. Nagel, "Problem of Global Justice," p. 128.
[53] Joseph Raz, The Morality of Freedom (Oxford: Clarendon Press, 1986), pp. 369, 154.
[54] See ibid., pp. 377–78; Blake, "Distributive Justice," p. 270.
[55] Blake, "Distributive Justice," pp. 272, 265.
[56] Ibid., p. 280.
[57] Ibid., pp. 260, 289.
[58] Ibid., pp. 258, 264, 265.
[59] Ibid., pp. 258, 276.
[60] Since according to the autonomy principle A1 all ongoing acts of state coercion require justification, I shall henceforth drop the qualifying parenthetical phrase "requiring justification."
[61] Blake, "Distributive Justice," p. 265.
[62] Ibid., p. 287.
[63] For analysis of the nature of border coercion, see Arash Abizadeh, "Democratic Theory and Border Coercion: No Right to Unilaterally Control Your Own Borders," Political Theory, forthcoming. Blake's recognition that coercion "in the international arena" may "exist in forms other than state coercion" ("Distributive Justice," p. 265) ignores the fact that states are themselves institutions who act in the international arena. Indeed, states are arguably the most important actors in the international arena, and they regularly act in that arena coercively. It is true that there is no world state that acts coercively against the world's population, but there certainly are states that do. On different types of state coercion, see A. J. Julius, " Nagel's Atlas," Philosophy & Public Affairs 34 ( 2006): 176– 92, at p. 184.
[64] Blake, "Distributive Justice," p. 280, n. 30, emphasis mine. Cf.: "There is no ongoing coercion of the sort observed in the domestic arena in the international legal arena. It is . . . only this form of coercion that makes a concern for relative deprivation relevant for a liberal political theory. . . . Material equality becomes relevant only in the context of certain forms of coercion, forms not found outside the domestic arena" (ibid., pp. 280, 284, emphasis mine).
[65] Ibid., p. 280.
[66] Ibid., p. 282, emphasis mine.
[67] "International legal institutions . . . do not engage in coercive practices against individual human agents" (Ibid., p. 280; cf. p. 279).
[68] Ibid., p. 284.
[69] Nagel, "Problem of Global Justice," pp. 128–29.
[70] Ibid., p. 129.
[71] On this point, see Julius, "Nagel's Atlas," pp. 179–83.
[72] Blake, "Distributive Justice," pp. 281–82, emphasis mine.
[73] Though not because it is regulated by laws: its coerciveness is sufficient to require justification. The justification, to be successful, may in turn require the rule of law.
[74] Blake, "Distributive Justice," p. 265, emphasis mine.
[75] Blake has in fact argued this in a personal communication to me.
[76] For some doubts, however, see Sangiovanni, "Global Justice," pp. 10–13.
[77] See Mathias Risse, " What We Owe to the Global Poor," The Journal of Ethics 9 ( 2004): 81– 117, at pp. 100–105, for this interpretation of Blake.
[78] For ethnographic evidence of the tremendous significance of this point, see Charles V. Carengie, Postnationalism Prefigured: Caribbean Borderlands (New Brunswick, N.J.: Rutgers University Press, 2002). On the historical origins of the enforcement of membership-boundaries as an economic tool, see Rogers Brubaker, Citizenship and Nationhood in France and Germany (Cambridge, Mass.: Harvard University Press, 1992), especially pp. 64–72.