The dimensions of global poverty and inequalities are striking, and they prompt questions about justice. Do we have a duty of justice to contribute to eradicating global poverty and to reducing inequalities? If so, what are we required to do? Assuming we agree on a conception of justice, can we apply the principles it contains globally? One may think that these questions can be answered automatically by taking a stance on the debate between cosmopolitans and statists. Cosmopolitans hold that the scope of distributive justice is global. Statists, by contrast, argue that the scope of distributive justice is limited to a narrower range of persons such as citizens of the same state.Footnote 1 I argue in this article that answers in this controversy do not automatically settle questions of justice about global distributive issues even when there is agreement about the principles of justice. Even if we start out with a cosmopolitan conception of justice, there are reasons why its distributive principles might not be applicable to certain global distributive questions, if we suppose that their application requires that they should be capable of evaluating distributions and guiding individual action and institutional design. In the course of the argument, I explore the role of economic and political institutions in determining the content of principles of justice, and the way this constrains the applicability of principles of justice. This article is structured as follows. First, I outline a version of the cosmopolitan position which is based on a non-relational view of justice. I then show that it does not follow from the non-relational view that principles of distributive justice can be applied to adjudicate distributive questions globally. I argue that, in the absence of political and economic institutions, under certain conditions, principles of justice are not determinate enough to define fair individual distributive shares, and even if they could do so, they are not determinate enough to guide individual conduct with regard to justice. Therefore, I argue that existing institutions may delimit the scope of application of principles of justice. They do not represent the outer bounds of justice, however, since the global institutional setup can be reformed so as to become more sensitive to the demands of global justice. The article's argument illuminates one sense in which the institutions and policies Rawls covers by the term ‘basic structure’ are special.Footnote 2
I. RELATIONAL AND NON-RELATIONAL THEORIES OF JUSTICE
To situate my argument, I briefly outline two positions from the recent literature on global justice, representing the statist and the cosmopolitan view respectively. They give contrary answers to the question whether or not political or economic institutions delimit the scope of principles of justice because they are based on two conflicting outlooks about the role of these institutions: relational and non-relational conceptions.Footnote 3
Relational conceptions of justice hold that individuals’ standing in a specific practice-mediated relation is a necessary condition for requirements of distributive justice to exist among them.Footnote 4 These relations are often taken to correspond to institutions and policies that regulate social and economic inequalities, most importantly ‘the political constitution and principal economic and social arrangements’ covering domains such as the legally recognized forms of property and the organization of the economy.Footnote 5 These institutions have been regarded as necessary for the existence of justice-relations under three descriptions: as having a profound and pervasive effect on the lives of their subjects; as employing coercive force against their subjects; or as representing cooperative schemes among their participants.Footnote 6 As I will refer to it, relational theories regard these relations as playing a foundational role in grounding requirements of justice.
Non-relational views deny this. They claim that at least some demands of justice can emerge even in the absence of practice-mediated relations. Political or economic institutions do not ground all requirements of justice.
The disagreement between relational and non-relational conceptions of justice can motivate different views about the scope of justice. If we are adherents of the relational view, the scope of justice will be contingent on the kinds of relation we stand in with others. Absent the requisite relation at the global level, there will be no global requirements of justice. If, however, we hold a non-relational view of justice, then we have a strong case for the global scope of justice. Non-relational theories are necessarily cosmopolitan. They allow for demands of justice to exist even in the absence of practice-mediated relationships. At first glance this seems to imply that principles of justice contained by non-relational conceptions should automatically be applicable to adjudicating globally arising distributive issues and that non-relational theories are committed to a requirement of global redistribution on the basis of justice.Footnote 7
This article concerns the inference drawn from non-relational theory to the global applicability of the principles of justice. I will show that it does not follow from the non-relational view, or indeed from any other cosmopolitan egalitarian position, that principles of distributive justice can be applied with a global scope to all arising distributive issues. Institutions can affect the scope of application of principles of justice even when they do not give rise to them.
II. THE APPLICATION OF PRINCIPLES OF JUSTICE
For the sake of argument, suppose we hold a non-relational egalitarian view of distributive justice. To keep the argument applicable to a fairly broad range of egalitarian theories, let us suppose we accept a principle of justice that prescribes that resources must be distributed in a way that limits ‘the range of permissible social inequalities among individuals’.Footnote 8 Does it follow from the non-relational outlook that this principle can be applied to adjudicate distributive questions globally?
We should clarify what it means to say that principles of distributive justice have a global application. It seems plausible to hold as a necessary condition for principles of justice to be applicable at a certain time that they should be capable of assessing states of affairs, guiding and evaluating individual action and/or institutional design with regard to a just distribution at that time.Footnote 9 In other words, in order to be applicable, principles of justice must enable judgments about the justice or injustice of distributions, actions or institutions. Thus, they can have global application only if they are able to do so with regard to a globally just distribution for the agents and institutions concerned. If we understand the application of principles in this sense, non-relational cosmopolitan theories prescribing a global egalitarian distribution do not imply that egalitarian distributive principles can be applied to all globally arising distributive issues. It does not follow from them, for example, that we individually should redistribute our surplus resources to those who are worse-off.
To see why, let us distinguish between three different functions institutions can play in a theory of justice. They can be regarded as having a role in grounding requirements of justice; in implementing the pre-institutional content of distributive principles; and in constitutively determining the content of principles. The first function is emphasized by relational theories, which attribute a foundational role to institutions. The standard view of institutions focuses on the second function. I will defend the third one. We will shortly see that institutions partly constitute the content of principles of justice. These principles do not define a fair distribution or a set of requirements for individuals antecedently to and independently of the rules making up the institution. Since these principles do not yield a sufficiently determinate standard for assessing distributions and for guiding and evaluating actions and institutional design, they may not be applied to adjudicate distributive matters outside existing institutions.
III. INSTITUTIONS AS EFFECTIVE INSTRUMENTS TO ACHIEVE JUSTICE
Before I defend in detail the constitutive view of institutions, let me briefly outline what I take to be the standard view. This view regards institutions as merely the most effective instruments to carry out justice-based requirements individuals have, defined independently of the rules of existing institutions. They are instrumental in overcoming the shortcomings of moral agents in performing their duties. This view would have the following justification.
Suppose the content of principles of justice can be fully determined in abstraction from the rules of actual institutions. Individuals are to contribute to a fair distribution of resources by reducing socio-economic inequalities, where fair shares are fully specifiable without reference to institutional rules. However, in complex societies it would be very difficult for an individual, left to her own devices, to know exactly how to act in order contribute her fair share. In order to know what to do, we have to have access to all sorts of relevant information, such as that about the current distribution and the expected effects of all possible courses of action available to us, possibly leading into the indefinite future. Because of insuperable informational and cognitive limitations, individual actions aiming at justice could approximate it only imperfectly. Justice can be promoted more effectively by creating and maintaining just institutions, i.e. systems of rules following which is more likely to result in a just distribution than could independent individual actions. For instance, a social division of labor, involving experts with special skills or knowledge, and the assignment of special rights and responsibilities, might make attempts to pursue justice more effective.
A version of this argument has been advanced by Saladin Meckled-Garcia on the basis of some of Rawls's remarks about the function of the basic structure. Meckled-Garcia argues that because of the impact individual actions have on the lives of a large number of individuals through a long-term horizon, individuals cannot realistically be required to foresee all the ramifications of their actions, and so they cannot be charged with the performance of principles of distributive justice.Footnote 10 The institutions of the basic structure are capable of making the necessary ‘background adjustment’ which individuals cannot. For this reason, he argues, the proper site of distributive justice is the basic structure of society in the sense that primary principles of justice apply to the authoritative institutions of the basic structure only, and not to individual actions, or even to institutions taken individually. The epistemological case Meckled-Garcia presents is certainly one reason why institutions are important for distributive justice.Footnote 11 I explain below, however, that the institutions of the basic structure are special for justice in a further sense. They would be required even if individuals were blessed with perfect foresight and spotless cognitive capacities. There are two general reasons for this. One is rooted in the indeterminacy of fair distributive shares, the other is due to the strategic character of human interaction. Under any plausible egalitarian conception, fundamental principles of justice underdetermine both what fair individual shares are and what individuals ought to do with regard to justice.
In what follows, I will present an alternative conception of institutions, in which they are not merely seen as instruments of the performance of justice-based requirements individuals have, defined independently of and prior to the rules of institutions. The institutions of the basic structure are central for distributive justice by performing two further functions: first, economic markets and political institutions are necessary for defining a just distribution; second, political institutions are also necessary for determining the content of requirements of justice by regulating strategic interaction between individuals, thus determining the underdetermined content of requirements. I will now look at economic markets and political institutions in turn.
IV. MARKET INSTITUTIONS IN RESOURCE EGALITARIANISM
One reason why institutions may be indispensable for the applicability of principles of justice to actual distributive issues is that markets may be essential for determining the value of resources to be distributed in accordance with principles of justice, and thus for defining fair distributive shares. This is the case, for instance, in those versions of resourcist theories of distributive justice, such as Ronald Dworkin's theory of equality of resources, that rely on market mechanisms in determining the value of resources which are to be allocated.
In Dworkin's theory, the content of requirements of justice is partially defined by markets that enable transactions between individuals. In order to be able to arrive at a just distribution of resources – equality of resources, as Dworkin terms his conception – we need the operation of markets to determine the value of resource bundles belonging to individuals. In the absence of markets, there is simply no way to tell what an individual's fair share of resources is. The problem is the following. Resources which distributive justice is in the business of distributing are very heterogeneous, and individuals assign differing value to them. In order to overcome the problem of determining an interpersonally comparable value for these resources, Dworkin proposes a thought-experiment, in which people, using a token currency distributed equally at the outset, participate in a series of auctions until everybody has such holdings that nobody envies anybody else's bundle of goods. At this point, we can say that everyone has an equal share of resources, Dworkin argues.Footnote 12 For this reason, in devising redistributive institutions governments should aim to mimic fair and efficient market mechanisms that define fair distributive shares. Market institutions are seen here not only as effective instruments in securing a just outcome that can be defined independently of them. They define what a just distribution is.
Even though Dworkin presents the series of auctions only as an analytical device and not as an actual mechanism, it seems likely that the theory requires the working of actual markets as constituents of a just institutional scheme, and that hypothetical markets are not sufficient. The reason for this is that the goods distributive justice is concerned with should be valued in a way that takes account of the differing ideas of a good life people have, and takes account of them equally. In Dworkin's theory the only way to measure the value of some resource allocated to one person in a manner that takes an equal account of everyone's interests is to ask what the actual costs of his having that particular resource are for others. This question can be answered only by letting some market mechanism work it out.Footnote 13 The market price will reflect the true cost of the asset held by someone to the other market actors, providing a standard for an interpersonal comparison of resource holdings that is not biased towards any particular conception of the good life.Footnote 14
If markets have a role to play in equality of resources, not merely as an analytical device to model hypothetical decisions, but also as an actual institution which is the only means to measure how certain goods are valued by individuals, in the absence of actual markets there may be no way for an interpersonal comparison of resource-endowments to take place, thus no way for finding out what requirements follow from Dworkin's egalitarian theory of justice in actual practice. Principles of justice would then be so underdetermined as to be inapplicable, since without existing institutions we may be unable even to define what a just distribution is. Absent such a definition, principles of justice might not be capable of assessing alternative distributions, or guiding and evaluating individual action and institutional design, in which case they cannot be applied.Footnote 15 Thus, existing market institutions constrain the applicability of principles of justice in this theory.Footnote 16
V. POLITICAL INSTITUTIONS
We have seen how a specific kind of institutions – markets – can be necessary for making principles of justice determinate by specifying the otherwise indeterminate notion of a fair share of resources. This might be thought an idiosyncrasy of Dworkin's theory, which requires a special place for a market pricing mechanism to determine an interpersonally comparable value for distributive shares when goods are heterogeneous and individuals have varying preferences over them. The applicability of these considerations, however, is not unique to Dworkinian theory. The case for a constitutive role of institutions can be generalized to all theories of distributive justice that include a significant element of procedural justice. To see how institutions play this role, in the remaining part of the article I focus on political institutions.
For the purposes of the argument, I will assume political institutions are characterized by the following features. They possess the capacity to issue authoritative rules and decisions for persons within their territory. Because of the coercive powers available to them, they are also capable of enforcing these rules and decisions. Authoritative settlement and enforcement takes place through a public system of law which binds all law-subjects and defines a unique set of rights and obligations for all subjects.Footnote 17
Political institutions play a crucial role in the realization of principles of justice. They have an important role to play in filling out the content of justice by translating abstract principles of justice into specific rights and obligations for individuals by way of law-making and policy-making. In what follows, I outline two considerations supporting this claim. On the one hand, political institutions make the otherwise underdetermined principles of justice sufficiently determinate by specifying fair distributive shares. On the other hand, they regulate strategic interaction between large numbers of people whose actions might have an impact on the lives of numerous others. They are necessary for determining what individuals ought to do about justice even if fair distributive shares could be determined in the abstract.
V.1. Political institutions specifying fair distributive shares
We have good reasons to think that under any plausible conception of egalitarian justice, fair shares that individuals are entitled to hold are underdetermined by the principles of justice. Indeterminacy can persist for various reasons. We already saw one case that concerned the allocation of a given stock of goods among a specific number of individuals.
A second type of indeterminacy about justice in distribution, affecting a much wider range of theories, arises when we introduce a dynamic perspective into our discussion of distributive justice, taking into account the need for regulating the social and economic system in which individuals are to interact. Individuals hold the resources they are entitled to by justice as their property, which is circumscribed by public rules. These rules include, for example a property law regime, as well as laws in criminal law against force and fraud, public health law, labor regulations, etc., which together govern the ownership, production, exchange and use of resources. Such rules are necessary for maintaining social cooperation to produce and maintain the resources individuals need. They affect distributive shares by influencing what goods get produced through social cooperation and how they are distributed. However, these rules governing property are vague and hence represent another source of otherwise inescapable indeterminacy that institutions can overcome.Footnote 18 To take an example, the rule that any transfer of property through sale that was made under duress is void would be incorporated into any plausible conception of what counts as just transfer of property. However, the term ‘under duress’ is inescapably vague and open to conflicting interpretations. In abstraction from existing institutions, the rule cannot have sufficient specificity that would allow us to apply it to a range of cases. Consequently, we would not be in the position to give a full specification of what our fair shares of resources are.
The indeterminacy of lower-order rules governing the ownership and use of things infects the principles of justice with an element of indeterminacy we cannot individually overcome.Footnote 19 In order to do this, we need political institutions capable of prescribing an authoritative interpretation of such terms, which is to an extent an arbitrary fact about the rules or directives of institutions. Again, our fair shares of resources are substantially underdetermined, since a just distribution cannot be determined independently of just institutions.Footnote 20
Yet another reason for substantive indeterminacy may be found in theories of justice that specify a range of permissible outcomes in distributive matters rather than a unique point. Rawls's theory is indeterminate in this way too, in at least two of its elements: the rate of just saving required by the theory, and the weight of self-respect in the index of primary goods.Footnote 21 The outcomes occupying the permissible range entail mutually exclusive arrangements of rules and policies. Society must coordinate on exactly one of these arrangements, but such coordination cannot take place through decentralized voluntary agreements due to the large-scale nature of the setting. Authoritative institutions are necessary to carry out successful coordination. If it is a range of outcomes and corresponding rules that is marked out by principles of justice as permissible, then political institutions are to make an authoritative determination of which point within the range society must pursue. Here too, fair shares of resources are partially determined by the rules or directives of political institutions, so there would be no way of knowing them and meeting the requirements of justice prior to and independently of these institutions.
V.2. Political institutions governing strategic interaction between individuals
We have seen why political institutions are indispensable for determining the content of justice in cases where fair individual shares of resources could not, even in principle, be determined independently of and prior to their operation. I now turn to a second reason for holding political institutions important in determining the content of justice. This consideration, however, applies even to cases when fair individual shares of resources could be determined without institutions. It focuses on the role of the institutions of the basic structure in regulating strategic interaction among individuals, and emphasizes the collective nature of various moral requirements, and especially that of requirements of justice.
Moral agency is strategic. The outcome and moral evaluation of our actions are often conditional on the actions of others. For this reason, when making a decision about how to act, we have to take into account the likely actions of others as a background for our choice, which we do not control. However, what actions others will take is in turn a function of our own action. As a consequence of this strategic feature, there is often no way to determine which course of action one ought to take since there is no information that would be available about the likely actions of others, taken in isolation from one's own future action. In such cases, when individual decisions about moral action are inescapably strategic, coordination is needed among individuals to single out one specific set of actions, indicating that other agents will act in certain ways. Institutional rules are an effective means to perform the requisite coordination, thus they enable individuals to achieve a morally required or permissible outcome when this is possible only if everyone in a group or a sufficiently large number of people follow the same course of action.Footnote 22
In addition to necessitating concerted action, moral requirements are often subject to pervasive disagreement. This creates an additional assurance problem since there will be no guarantee that others adhere to a decision about a joint course of action, were such a decision to take place. In cases when individuals have to act in concert but nevertheless disagree about the morally best joint action, resolution and enforcement are required. Political institutions, by yielding and enforcing authoritative decisions, are capable of supplying assurance to individuals subject to them, and hence of resolving conflicts. Correspondingly, parties are morally required to comply with institutional rules or directives, even if they judge some alternative course of action morally superior.Footnote 23
Principles of distributive justice encounter the same problems, and thus call for institutional settlement. Suppose again we accept a broadly egalitarian conception of justice, which aims to limit the range of socio-economic inequalities among individuals. However, egalitarian distributive principles can be realized in more than one way, even when we can determine fair distributive shares ex ante. The importance of justice being done nonetheless requires that individuals act in a concerted manner and have grounds for forming reasonable expectations about the actions of others involving large stakes. Furthermore, they need assurance that their reasonable expectations are going to be met. To illustrate, I now briefly introduce two problems concerning distributive justice that call for institutional settlement.
One reason why justice requires institutional settlement is that social and economic systems are made up of a number of interrelated institutions and policies such as property regulations, welfare provisions, educational and health care systems, among others. These constituents admit of several possible combinations capable of generating the same distributive outcome. Thus, there might be several combinations of institutions and policies, corresponding to sets of actions, that are equally acceptable under our conception of justice. In such cases one particular combination must be singled out on what are essentially arbitrary grounds.Footnote 24 Political institutions specify the underdetermined demands of egalitarian justice by setting a unique set of distributive rules.
Suppose next that, even though the specific circumstances of the society determine a unique optimally egalitarian institutional scheme, there is disagreement about what this is, given various possible schemes. There are two schemes under consideration, both of which are likely to yield a reasonably just distribution in the society. Scheme A would rely more heavily on a progressive income tax and would keep taxes on consumption low, whereas scheme B would operate with higher consumption taxes and would tax incomes less heavily. Some people think it is scheme A that best serves justice, others think it is scheme B. Governments authoritatively settle debates about distributive justice, and they apply and enforce these decisions on their subjects. By doing so they determine a unique set of rules and provide assurance that they will be adhered to. Using their effective coercive powers, they will enforce laws even on those subjects who disagree with these. Given that both schemes are reasonably just and one of them, A, is enacted as law, in the face of a need for coordination and conflict resolution justice requires one to comply with the rules of A, regardless of whether one thinks it is the best possible egalitarian scheme. This holds even for those who are correct in their judgment that an alternative scheme would be preferable if others followed suit.Footnote 25 Thus, in non-ideal theory political institutions determine what individuals ought to do with regard to justice because of disagreement and the need for conflict resolution in a society.Footnote 26
In both cases, without institutional settlement it would be impossible to tell what justice requires us to do.Footnote 27 Institutions are necessary for determining the content of justice even in cases when fair individual shares of resources could, in principle, be determined without institutions.
The political institutions of the basic structure are capable of filling out the content of principles of justice due to the powers they possess. They coordinate the conduct of individuals and supply assurance by providing authoritative settlement and enforcement through a public system of law which binds all their law-subjects. The legal system defines a unique set of rights and obligations for all subjects, backed up by the capacity of political institutions to enforce authoritative rules and decisions within their territory.
VI. THE SIGNIFICANCE AND PERVASIVENESS OF INDETERMINACY ABOUT JUSTICE
I have argued that basic political and economic institutions play a special role in justice because of a number of indeterminacies in the content of principles. In the absence of existing institutions, the content of distributive principles is substantially underdetermined. The information contained in the principles, together with information about principles of social theory, general economic and social facts, and information about individual preferences and resources available for distribution, are sometimes insufficient to determine what fair shares individuals are entitled to and what actions they ought to take in accordance with the principles of justice.Footnote 28 Underdetermined principles, however, must be specified before they can guide action and institutional design.
Institutions of the basic structure are capable of specifying the underdetermined principles of justice. They determine fair resource shares by singling out a unique set of distributive rules and by enabling an interpersonal comparison of resource holdings. They also define a unique set of rights and obligations for all law-subjects that are sufficiently determinate to guide their conduct.
Therefore, I take the political and economic institutions of the basic structure as playing a constitutive role in determining the content of principles of justice. This contrasts with the alternative view I sketched earlier, which regards institutions as the most effective means for discharging principles of justice applying prior to and independently of the working of these institutions. The view defended in this article does not deny that institutions often provide more effective means for bringing about a just distribution and discharging justice-based requirements, but it denies that it is all there is to them. It is not the case that institutions merely make independently existing information about fair distributive shares accessible for individuals who, because of individual limitations, cannot themselves gather and process it. Institutions play a constitutive role in determining the content of principles of justice by marking out a unique set of just distributive rules, coordinating individual conduct and providing assurance in cases of disagreement about justice. They make the otherwise indeterminate requirements of justice sufficiently determinate by subjecting individual judgment to rules or directives.
How pervasive is the indeterminacy of principles of justice? The cases discussed earlier show that the requirement of institutional settlement is likely to be a necessary feature of any plausible egalitarian theory of justice. Both disagreement about the demands of justice and the indeterminacy of their content are a permanent characteristic of politics. As for disagreement, John Rawls plausibly argues that ‘a diversity of conflicting and irreconcilable comprehensive doctrines’ is ‘not a mere historical condition that may soon pass away; it is a permanent feature of the public culture of democracy’.Footnote 29 Reasonable pluralism, Rawls claims, results from ‘the work of free practical reason within the framework of free institutions’.Footnote 30 Moral concepts, including the concept of justice, involve various ‘burdens of judgment’ which make disagreement a permanent feature of life even under free institutions.Footnote 31 Also, the complexities of contemporary social life render the demands of egalitarian justice indeterminate. Principles of justice depend for their application on large-scale social coordination, including rules circumscribing property and political processes specifying a unique set of institutions and policies. So, it is in the framework of the basic structure that we make claims of justice since its institutions make the demands of justice determinate for us and resolve disagreement about the right principles and their application.
This function of the basic structure also illuminates the significance of the distinction Rawls draws between an institution as an abstract object, i.e. as ‘a possible form of conduct expressed by a system of rules’, and an institution as an actual practice, i.e. the way these rules are realized ‘in the thought and conduct of certain persons at a certain time and place’, when he claims his principles apply to the latter.Footnote 32 The foregoing arguments provide one rationale for holding that, even though there may be several possible configurations of the basic structure considered as an abstract object, the principles of justice can sometimes apply only in the context of institutions realized in the conduct of individuals composing them, because of the indeterminacy of these principles and the strategic character of the interaction between individuals upholding the institutional scheme.
VII. THE INDETERMINACY AND APPLICATION OF PRINCIPLES OF JUSTICE
Political and economic institutions are indispensable for specifying the content of principles of distributive justice. Therefore, in the absence of existing institutions principles of distributive justice might not be applicable, for the reason that these principles are not determinate enough for assessing alternative distributive shares, and even when they could do so, they are not determinate enough for guiding individual conduct. They do not allow judgments about the justice or injustice of some distributions, actions or institutions.
We should consider two potential objections to this argument, both of which would rely on arguments advanced by G. A. Cohen in other contexts. First, Cohen argues that fundamental principles of justice do not depend on facts. Thus, facts of indeterminacy and disagreement, as well as the need for coordination and assurance, are irrelevant for determining the content of fundamental principles of justice. These considerations might be relevant for justifying rules of regulation to govern interaction among agents, but they have no bearing on what fundamental principles of justice are. Cohen would thus be content with underdetermined principles of justice, since he thinks it is not necessarily part of their job to guide action or institutional design.Footnote 33 In response to this objection, I have two replies to make. First, I hope to have shown that the content of egalitarian principles of justice is substantially underdetermined, and we might not even be able to tell what equality itself – as distinct from other values – consists in without the operation of institutions. Second, however, whatever we think of Cohen's claim about the content of justice, it does not defeat my claim that principles of justice may not apply in the absence of institutions, if the application of principles requires that they are capable of evaluating distributions, guiding action or institutional design. It may be true that principles of some sort exist in the absence of institutions, but they may not be applied to any site under certain circumstances. The claim this article makes is limited to the application of principles of justice, thus unless one is willing to grant that principles apply even when they cannot possibly assess distributions, guide action or institutional design, the main thesis is not undermined by Cohen's point.
One could further object, however, that the conclusion about the limited applicability of principles of justice is too strong. The indeterminacy of principles, the objection would run, does not preclude making at least some judgments about the justice or injustice of distributions, actions, or institutions. Surely, we do not need working markets or political institutions to see that a world which contains billionaires as well as people below the poverty line is by any egalitarian standard unjust. Since we can make judgments about such flagrant injustices even in the absence of institutions, the objector would conclude, principles of justice do apply in the absence of institutions even when they are indeterminate.Footnote 34
In response to this objection, I should emphasize that the thesis of this article is limited in the following sense. The foregoing argument established only that some principles of justice, under some circumstances, fail to be applicable in the absence of institutions. I was not defending the implausibly strong claim that no principles of justice ever apply without institutional specification. This article shows that, under some circumstances, principles of justice are insufficiently specific to guide individual action or institutional design, or to evaluate distributions; therefore they cannot be applied to adjudicate distributive questions under these circumstances. Take the case when principles of justice specify a range of permissible distributive outcomes rather than a unique point: we know even in the absence of institutions that values lying outside this range are unjust. However, when we have to decide about the justice or injustice of a distribution within the permissible range, principles of justice do not yield a sufficiently determinate answer in the absence of working institutions. I contend that in such circumstances, for such values, these principles of justice are not applicable. That is, they cannot guide us in evaluating whether any such distribution is unjust, and they cannot guide action and institutional design with regard to these.
VIII. THE SCOPE OF JUSTICE VERSUS THE SCOPE OF PRINCIPLES OF JUSTICE
An implication of the constitutive view of the basic structure defended here is that the scope of application of these principles is conditioned by existing institutions. Does this leave us with a relational view of justice? It might seem so for the following reason. We just saw that, in order to overcome the essential indeterminacy of the content of justice and to coordinate the actions of a large number of individuals, we need political institutions capable of issuing and enforcing authoritative rules and decisions. This capacity is predicated on the coercive force political institutions exercise over their subjects, i.e. all those within their territory. It might seem that this commits us to a relational conception of justice along the lines of the arguments of Thomas Nagel and Michael Blake, who hold that requirements of distributive justice arise only within the confines of coercively enforced schemes of political institutions.Footnote 35 The existence of a certain type of relation, that of being subject to a common set of political and economic institutions, may after all be necessary in order for principles of justice to be applicable to specific distributive problems.Footnote 36
It would be a mistake to equate these two views of political institutions, however. Coercive political institutions enter the two theories at different levels. Relational theories such as Blake's and Nagel's regard standing in a practice-mediated relation as a necessary condition for any kind of justice-requirement to arise. There would be no occasion for justice in their absence. The view I am defending, however, is compatible with holding that some requirements of justice can exist independently of and prior to coercive political institutions. To take one example, the requirement to promote the establishment of just relations can exist even in the absence of institutions.Footnote 37
To relate this discussion to a term I introduced in the beginning of the article, the basic structure is not viewed here as foundational in grounding justice. Justice is not grounded in existing institutional relations, even though the application of principles of justice is conditioned by these. The position I defend in the article occupies an intermediate ground between those who claim that institutions play a foundational role in the emergence of the requirement to promote justice, and those who argue that institutions are mere devices for more effectively carrying out what justice requires us to do anyway.
One qualification is in order at this point. The non-relational theory I started out with assumes that some kind of requirements with regard to justice exist independently of and prior to the existence of institutional relations. What this background position implies for cases where common political and economic institutions are non-existent is that everyone has a duty to work toward establishing them so that a just distribution – whatever it turns out to be once institutions start to operate – can be brought about. However, the function of political and economic institutions in determining the content of justice I have defended in this article is neutral with regard to the ground of justice. The arguments I presented are applicable to both relational and non-relational theories; the two distinctions cut across each other. If, for reasons independent of the considerations I have outlined, we can show that justice is grounded in practice-mediated relations, political and economic institutions still play a constitutive role in defining principles of justice.Footnote 38 However, my arguments do not imply a relational view of institutions.
It might be suggested, however, that the way this theory treats the scope of justice would nevertheless be extensionally equivalent to relational theories. Political institutions limit the scope of justice because of the role they play in its implementation. They delimit the application of principles of justice and therefore they condition the scope of distributive requirements. In the absence of global political institutions relevantly similar to nation-states principles of justice may not apply to global distributive questions. Even though there is an occasion for justice globally even in such cases, the currently existing international mechanisms for enforcing rules, policies and decisions might not be capable of coordinating across the globe and providing assurance that international norms will be effectively enforced. This would make it unlikely that egalitarian distributive principles can be implemented in the international domain.
However, the theory's reliance on economic and political institutions need not undermine its applicability to circumstances where there are no such actual institutions in place. The claim that at present there are obstacles to implementing principles of justice globally is insufficient to demonstrate that their full scope of application cannot eventually extend to the global domain. In order to show that, it further needs to be proven that the obstacles to putting these principles into practice are fixed and cannot be removed.Footnote 39 Otherwise, they do not delimit the potential scope of principles, and it will be the task of non-ideal theory to suggest ways to overcome the obstacles to making them operative. Even if currently there may not be political institutions at the global level that are capable of performing the same functions as nation-states, most importantly that of making and enforcing authoritative norms, such institutions can evolve in the future. If they could, our duty of justice should guide institutional reform at the global level, in order that global institutions required for maintaining justice can be created.
IX. THE NATION-STATE SYSTEM AND GLOBAL JUSTICE
Where does this leave us? Do we have to conclude that existing political institutions – territorial states as we know them – are to remain the loci of justice, in the sense that they circumscribe the potential scope of application of principles of justice, at least for the foreseeable future?
This is not very likely. Recent changes in the global institutional setup which have taken place in response to a growing demand for supranational coordination and to human rights concerns provide new opportunities for extending the scope of application of principles of justice. Territorial political authority has in practice been supplemented by a system of global governance. The result is a complex global, multilayered scheme of institutions performing supra- and transnational, regional, and local governance with a mixture of functionally and territorially defined authority. In performing their governance functions, its constituents are at least potentially capable of specifying the content of justice, and thus of enabling the application of principles of justice on a global scale. They authoritatively set, interpret and enforce rules worldwide, contributing to the determination of distributive shares and requirements.Footnote 40
Once we accept its global scope, the duty of justice requires that we rely on some of the existing elements of this scheme, reform others so that they better fit principles of justice, and establish new ones, rather than return to the system of territorially defined nation-states.
X. SUMMARY AND CONCLUSION
Justice is an institutional virtue in several respects. The existing literature makes prominent two dimensions in which institutions can be of special significance for justice. The first concerns whether they are the only site of justice, in the sense of being properly subject to justice-based evaluation.Footnote 41 The second debate centers on whether institutions of the basic structure are necessary for giving rise to requirements of justice, and if so, how they constrain the scope of justice. Thus, the typical strategy pursued by statist theories has been to limit the scope of application of principles of justice by arguing for the foundational role of the basic structure in grounding considerations of justice and pointing out the absence of basic structure at the global level. In this article I completely set aside the issue of the site of justice. I briefly outlined the debate about the ground of justice, but only to contrast it with a different way in which the basic structure can affect the scope of application of principles of justice. The argument presented in this article is novel in that it shows how the basic structure's role in determining the content of justice can affect the scope of application of principles of justice even if we embrace a non-relational cosmopolitan theory. It did so by presenting some considerations that make economic and political institutions indispensable for the realization of principles of justice. I considered two ways in which institutions may be partly constitutive of the content of principles of justice. First, I argued that fair distributive shares are substantially underdetermined, and that economic and political institutions are necessary to make them determinate. Economic markets may be required for specifying the value of resources to be distributed. Political institutions are necessary for selecting a unique value from among a range of permissible options, and they also determine fair distributive shares by specifying property. Second, I noted that even if fair shares of resources could in principle be determined without institutions, there is still a residual indeterminacy about what this implies for individual conduct because of the strategic feature of human action and disagreement about the correct interpretation of principles of justice. I argued that we need political institutions to coordinate individual conduct and provide assurance by making and enforcing authoritative decisions. These considerations showed that even if we take a non-relational view of justice, political and economic institutions play a constitutive role in determining the content of principles. Fundamental principles of justice themselves do not give a complete specification of what shares of resources individuals are entitled to and what individuals ought to do with regard to justice. Thus existing institutions may constrain the scope of application of principles of justice, since these are not specific enough to guide individual action and institutional design in the abstract. In the absence of the institutions of the basic structure they may not apply to arising distributive questions. I concluded, however, by arguing that existing nation-states do not represent the outer bounds of justice since the global institutional setup can be reformed so as to become more sensitive to the demands of global justice.Footnote 42
