Which of the following two theories of global inequality are closest to being opposite?

Current views on global justice often hold that affluent states are under at least two duties: a duty to reduce socioeconomic inequalities at home and a duty to reduce extreme poverty abroad. Potential duty conflicts deriving from resource scarcity can be solved in broadly two principled ways. The ‘lexical’ principle requires all disputed resources to be allocated to the weightiest duty. The ‘proportionality’ principle requires resources to be distributed between the two duties according to their relative weight (the weightiest duty receives the largest resource share, but the less weighty duty receives a share too). I argue that the proportionality principle is morally preferable. I show that it is sensitive to a number of factors that are intuitively relevant when solving duty conflicts: the number of affected individuals, the size of the benefits each individual could get, and the time it could take to eventually comply with the less weighty duty. Some argue that the lexical principle should nevertheless be preferred because domestic egalitarian duties are duties of justice, and they are therefore lexically prior to mere humanitarian duties to reduce global poverty. I reject this view by showing that duties of justice are not necessarily lexically prior to humanitarian duties, and that (even if they were) duties to reduce global poverty can be regarded as duties of justice too.

Most current theories of global justice agree that affluent states have a duty to alleviate extreme global poverty. At the same time, many of those theories also hold that affluent states have a duty to reduce inequality within their own territories.1 If both duties actually exist, they seem to be currently unfulfilled to some extent. Severe poverty affects millions around the globe, and domestic socioeconomic inequalities persist even in the wealthiest states.2 Because fulfilling both duties can be costly, and resources are scarce, there is potentially a conflict between them. In some circumstances, fulfilling one might hinder fulfilling the other.

In this article, I discuss several principles for solving this duty conflict. The first section locates this discussion in the broad debate on global justice. The second section shows that a conflict between domestic and global duties is possible. The third section introduces two types of principles for solving duty conflicts: proportionality principles and winner-take-all principles. The fourth section assesses which of these principles are appropriately sensitive to a number of factors that are intuitively relevant when solving duty conflicts: the number of individuals which are involved in the conflict, the size of the benefits each individual could get, and the time it could take to eventually comply with both duties. I argue that proportionality principles are more sensitive (than winner-take-all principles) to those factors. The fifth section shows that winner-take-all principles cannot be defended by appealing to the distinction between duties of justice and humanitarian duties.

This article focuses on a particular disagreement between different egalitarian views in the global justice debate. A view is egalitarian if it claims that distributive shares of certain goods among certain subjects should comply with a certain standard that is appropriately sensitive to share inequalities.3 There is strong disagreement between egalitarians regarding the relevant goods (income, capabilities, opportunities, and rights) and subjects (individuals and certain types of group), as well as regarding the proper standard (strict equality, moderate inequality, and prioritarianism) to be met. However, these disagreements are not the subject here.

Another disagreement that is not the subject here is the disagreement about the scope of egalitarian duties. David Wiens and Christian Barry helpfully distinguish between the following two views on scope. ‘Restricters’ claim that only inequalities between compatriots should meet egalitarian standards, while ‘extenders’ argue that inequalities between all human beings should (also) meet those standards.4 In other words, restricters claim that duties to reduce inequality are domestic in scope, while extenders claim that their scope is global. This debate is however mostly irrelevant to my discussion.

In order to identify the particular disagreement that I focus on here, we need first to identify a series of agreements in the egalitarian literature on global justice. All restricters and many extenders agree that each state is under a duty to reduce inequality among its residents. The reason why extenders agree with this seemingly contradicting claim can be circumstantial (‘domestic inequalities matter given the shape the world has today, but ideally they do not’) or principled (‘domestic inequalities matter regardless of the shape the world has’).5 But the important point is that many extenders care not only about global inequality but also about domestic inequality.

Restricters and extenders also agree that affluent states are under a duty to reduce global poverty. But they agree with this claim for different reasons. Restricters agree because they usually adhere to a sufficientarian principle demanding that affluent states reduce the number of human beings unable to meet their basic needs.6 Extenders, on the contrary, agree because they claim that reducing global poverty is required by an egalitarian principle that is global in scope.7

We are now in a position to identify this article's main question: How should affluent states address potential conflicts between their duty to reduce domestic inequality and their duty to reduce global poverty? Notice that this is a conflict between two duties that both restricters and (many) extenders recognize as real.

There are other related duty conflicts which I do not address here, such as conflicts involving domestic duties of basic justice. These are state duties to secure residents’ fundamental freedoms and rights, including the right to subsistence. Most extenders and all restricters agree that states have domestic duties of basic justice which are weightier than global duties to reduce global poverty. If a state has to choose between securing food to a number of residents or to the same number of nonresidents, residents should be prioritized.8

Before turning to the next section, it is important to introduce some terminology. In this article, state duties to reduce domestic inequality will be called ‘domestic duties’ for short. State duties to reduce global poverty will be called ‘global duties’ for short. And the conflict between these two duties will be called ‘domestic–global conflict’ for short.

There are a number of ways to deny that domestic and global duties may conflict. One simple way is to claim that at least one of these duties does not exist. For example, some libertarian authors deny that there is a duty to reduce global poverty.9 Others deny that domestic duties of socioeconomic equality are at all defensible.10 Another way is to claim that both duties exist, but one of them is merely an instrument for discharging the other. For example, in the ‘moral division of labor’ theory of domestic duties, these duties are mere conventions that serve as a means for complying with our general duties to humanity at large.11 If this theory is correct, whenever compliance with a domestic duty would hinder compliance with a global duty, the domestic duty ceases to exist, because the very reason why it exists is to promote global duties.12

Addressing these two challenges would require dealing with fundamental issues in moral philosophy—such as the issue whether there are basic rights to subsistence and whether consequentialist theories are more justifiable than deontological theories—which I cannot address here. Therefore, the rest of this article should be read as asking the following conditional question: if both domestic and global duties actually exist (and none of them is a mere instrument for complying with the other), how should conflicts between them be solved? As mentioned in the last section, this is a relevant question for most egalitarians.

Another way to deny that conflicts are possible is by showing that any theory containing domestic and global duties at the same would be incoherent. Consider, for example, Samuel Scheffler and David Miller's views. Similarly to most authors, they claim that there are impartial moral principles that ground general duties. But Scheffler and Miller argue that those general duties coexist with independently grounded special duties which arise purely from special relationships.13 These views have an element of inconsistency, given that to some extent they adhere to the impartialist idea that all human beings matter equally, while at the same time they seem to claim that some individuals (those who stand in special relationships with someone) matter more than others.14 There are other theories, however, that are capable of avoiding this inconsistency while still claiming that humans are simultaneously under general and special duties. There are, that is, fully impartialist views that are capable of grounding domestic and global duties at the same time. Michael Blake, for example, showed how the universal principle of autonomy can ground different duties in different contexts, some of which have a general scope (such as the duty to reduce global poverty) and some of which are only local in scope (such as the duty to reduce domestic inequalities).15

There is yet another way to deny that conflict is possible which focuses on the difference between relative and absolute distributive duties. Notice that duties to reduce extreme global poverty are concerned with the absolute level of goods controlled by individuals, while domestic egalitarian duties are concerned with relative shares. But then a state could first comply as much as possible with its global duties, and then use whatever remaining goods to comply with its domestic duties. Since domestic duties are only concerned with relative deprivation, it does not matter how much is left to be distributed, but how it is distributed. To illustrate, suppose that at t1 affluent state S has achieved equality, and at t1 every resident has 100 units of goods (to simplify things, assume that the relevant egalitarian ideal is strict equality). At t2, state S decides to comply with its global duties, so it taxes its residents to obtain goods and transfer them to nonresidents (assume for the moment that taxing and transfers are effective means to reduce global poverty global poverty). Is domestic equality compromised by the new tax? Not necessarily, because state S could tax every resident by the same amount (say, 75% of their goods), so at t3 each is left with, say, 25 units of goods. Every individual would be left with the same amount of goods, so strict equality is preserved. It is true that at t3 every resident is worse off than at t1, but equality is not compromised because what matters for strict equality is relative shares, not absolute shares.16

This scenario could be altered to make it more realistic. For example, imagine another state T which has also achieved equality at t1. Again, when trying to comply with its global duties at t2, state T taxes each resident by the same amount. However, due to contingent differences between each resident's circumstances, the tax does not have the same effect on everyone, which results in inequality at t3. Now, it could be argued that domestic duties and global duties do not conflict even in this scenario. It is still possible for state T to implement yet another tax in order to correct inequalities and achieve strict equality at t4. This new tax could be imposed on the better off only, and then the goods could be transferred to those who were left less well off. Even if the amount left to each individual at t4 is very low, strict equality is nevertheless restored.

This objection may be answered in two ways. First, the objection may be sound when directed against some egalitarian conceptions such as strict egalitarianism, but certainly not against all such conceptions. For instance, the objection is unsound when applied to prioritarianism, according to which the position of the worse off individual (defined in terms of the relevant currency) should be maximized. In the scenario involving state S, the position of the worse off individual is worse at t3 than at t1, which means that imposing a tax at t2 in order to reduce global poverty would be forbidden by a prioritarian theory of domestic justice.17 Therefore, if we interpret prioritarianism as part of the broad egalitarian family, reducing global poverty can be in tension with preserving domestic equality.

Second, the objection may be answered by reference to non-ideal theory. Let us grant that strict egalitarianism (or another non-prioritarian egalitarian conception) is the relevant egalitarian ideal and that this ideal is therefore equally satisfied at t1 and t3 (in the scenario involving state S). But now consider the scenario involving state T. Achieving equality at t4 may take some time. There may be a transition period of inequality between t3 and t4 which should be computed as a cost in terms of equality. There are two reasons why restoring equality may take time. First, redistributive measures, including many kinds of tax, often do not generate their intended effects immediately. For a progressive income tax to generate redistributable revenue, for example, some time needs to pass by until the wealthy actually declare and pay for their income.18 Second, achieving institutional changes such as tax reforms is often politically costly and slow. It could be the case that restoring equality (or at least reducing inequality) at t4 by changing the tax scheme (or by any other means) is extremely infeasible given current social and political circumstances at t3. Given that the transition period should be computed as a cost in terms of equality, we should conclude that even under strict egalitarian theories domestic–global conflicts are possible.

The claim that conflicts are possible in non-ideal scenarios is strengthened if one considers which policy instruments would be actually effective in the fight against global poverty. Economists often argue that resource transfers from affluent countries to developing countries (as in the two scenarios above) are unable to reduce global poverty and may even be counterproductive.19 Perhaps a more effective and feasible means is to remove restrictions on trade and immigration from poor to affluent countries. However, this type of measure may have a greater negative impact on the less well off in affluent countries than on the better off, thereby creating or increasing inequality. This seems to be the case in the following example.20 In 1965, the United States reformed its immigration law, which became more accepting of worse off nonresidents. Apparently, this had the effect of lowering wages in the United States, especially in the (less well educated) bottom 9% of the workforce but not so much in the rest of the population.21 This means that immigration seems to have increased inequality, because the negative impact of immigration on wages was especially felt by the worse off (non-immigrant) Americans, not by the relatively better off. On the other hand, the open immigration policy probably reduced global poverty. This is because it improved the situation of new immigrants—and of their families through remittances—, some of which had suffered severe poverty in their countries of origin.22

Consider one final way to deny that domestic–global conflicts are real. It is sometimes claimed that affluent states have a ‘large surplus of opulence’ that enables them to comply with domestic and global duties simultaneously.23 This claim is often backed by pointing to expensive domestic projects engaged by affluent states such as the Millennium Dome in Britain. What is striking about these projects is that they do not seem to be aimed at addressing local inequalities, but at providing amusement and other luxuries. This motivates the impression that affluent states have solved or are capable of easily solving their local inequalities, and that they would still have a surplus of money to spend on other ends such as reducing poverty abroad.24 The argument here is not that domestic–global conflicts are impossible, but that conflicts do not currently occur. This argument does not fare very well. The fact that affluent states engage in pharaonic projects (not directly aimed at inequality reduction) does not imply that it is feasible for those states to spend resources elsewhere. Political constraints may make it hard to redirect those resources to global poverty reduction. And even if such feasibility constraints were not in place, recall that international resource transfers seem to be a counterproductive way to reduce global poverty. And effective, feasible ways to reduce global poverty (such as trade and immigration liberalization) are prone to conflict with domestic egalitarian duties in the world as it is today.

In this section, I present several types of principle for solving duty conflicts. The idea of ‘solving’ a conflict does not refer here to what would be the most prudent or pragmatically expedient thing to do in that conflict (although this is a valid question as well). In this article, a principle properly solves a conflict when it correctly states what is the morally required thing to do to address that conflict.

Duty conflicts will be addressed using the following theoretical framework. First, two or more duties are in conflict when one of them (or a part of it) cannot be fulfilled if the other duty (or a part of it) is fulfilled. The reason why two duties may conflict is that their fulfillment typically uses up ‘goods’. (Again, the idea of a good is taken here in a wide sense which includes income, capabilities, opportunities, rights, etc.) Relevant goods can be scarce, making it infeasible to fulfill the two duties at the same time.25

Second, I employ the useful metaphor that duties have ‘weight’ (or force, stringency, and importance). How do we know how much a duty weights? ‘Intuitionist’ conceptions assign weight to conflicting duties by means of mere intuition, while ‘theoretical’ conceptions determine weight by means of an explicit standard or rule.26 In what follows I will adopt a particular theoretical conception, according to which weight can be a function of two things. On the one hand, weight is determined by the reason or ground why the duty is a duty in the first place.27 This is the fundamental or basic weight assignment. On the other hand, basic weight can be adjusted for a variety of case-specific factors such as the number of individuals who would benefit from the duty being fulfilled, the size of that benefit, the degree of responsibility (primary or derived), etc.

Third, principles for solving duty conflicts have two parts. (1) A rule for assigning weights. As mentioned, basic weight is a function of the ground why the duty is a duty, but weight can also be sensitive to case-specific factors. A principle is more ‘strict’ if it allows for less case-specific factors to influence duty weight. More ‘moderate’ principles make room for case-related factors that can in some circumstances tip the scales in favor of an initially less weighty duty (i.e. a duty that would lose the contest if only its basic weight were considered). (2) The actual rule or principle—I take them as synonyms here—for solving the conflict. There are two possible principles: winner-take-all and proportionality.28

The distinction between winner-take-all principles and proportionality principles can be introduced using the following example. Suppose, I have two duties today: a duty to take care of my baby daughter (feed her, play with her, and put her to sleep) and a duty to water my neighbor's huge garden while she is on holiday. There is, however, a duty conflict: watering the whole garden would require all day, and properly taking care of my daughter would require all my time as well. Assuming that the duty toward my daughter is weightier there are two possible principles to guide me in how to distribute my scarce goods (my time). I can adopt a winner-take-all rule, in which the weightiest duty receives all the goods under dispute—so my neighbor's garden does not get watered at all. Or I can adopt a proportionality rule, in which each duty gets an amount of goods in proportion to its relative weight—so, I spend most of my time taking care of my daughter, and I spend at least some time (although less time) watering my neighbor's garden.

The strictest form of winner-take-all principle is the lexical principle.29 A principle is lexical when all case-related factors (such as number of targeted people and benefit size) are completely irrelevant for determining weight. What matters is only each duty's ground, and then the contest is solved by winner-take-all.

In the aforementioned daughter case, a relatively strict winner-take-all principle is employed, because size of benefit is not relevant. Indeed, a more moderate winner-take-all principle would produce a different outcome. Given the fact of decreasing marginal utility of most goods, it is likely that the last minutes spent with my daughter would add little to the goal of taking care of her, while those minutes spent in the garden would make a big difference. Therefore, there is a point after which an extra minute spent on my daughter will weigh less than a minute spent on the garden. A moderate winner-take-all principle which is sensitive to benefit size would therefore require me to spend at least a few minutes watering the garden.30 The lexical principle, recall, required me to spend even those last few minutes with my daughter.

Reasons need to be given in favor of using one principle or the other, given that each principle has substantially different practical consequences. For example, consider the domestic–global conflict. If the lexical principle were used, one of those two duties would be left completely unattended until the other is fully complied with. If, on the contrary, a proportional principle were used, both duties would be complied with at the same time (although probably to different degrees). The next two sections assess arguments in favor of different principles and show that a proportionality principle should be preferred.

It is important to warn from the start against a tempting fallacy.31 When facing a duty conflict, it is natural to reason this way: ‘Duty A and B are in conflict; but duty A is weightier than duty B; therefore, duty A is lexically prior to duty B’. In the domestic–global conflict, an argument with that form would look like this: ‘duties to reduce global poverty are weightier than domestic duties; therefore, global duties are lexically prior to domestic duties’. But the conclusion does not necessarily follow. As explained above, the fact that a duty outweighs another duty could elicit two types of response: a winner-take-all (lexical or non-lexical) solution or a proportional solution. Given that both are equally possible ways to solve duty conflicts, an additional argumentative step is needed to show that the lexical solution is the morally required solution.

In this section, I argue that the right principle for solving the domestic–global conflict is a proportionality principle. I assess alternative principles and show that they are insufficiently sensitive to a number of concerns that are intuitively relevant.

To simplify matters, principles will be tested in a hypothetical duty conflict scenario involving only two individuals. In this scenario, a state S has to choose between reducing resident A's inequality or reducing nonresident B's extreme poverty. Both choices are obligatory. There are no other alternatives. And both choices use up disputed goods that can be infinitely fractioned. (When necessary this basic scenario will be modified to involve more than two individuals.)

Consider first the strictest form of winner-take-all principle, the lexical principle. Several authors claim that global duties are lexically prior to domestic duties.32 This means that in cases of conflict, domestic duties should be left completely unfulfilled. Only if (or once) there is no conflict, and complying with domestic duties causes absolutely no impediment to complying with global ones, can domestic duties be complied with.

The lexical principle has many shortcomings. First, it is insensitive to benefit size. Imagine that there are only two feasible options: to reduce B's poverty by an extremely minimal amount or to reduce A's inequality by an enormous amount. The lexical principle requires that we choose the first option. But this is unreasonable, given that A would receive comparatively huge benefits if we chose the second option.

Second, the lexical principle is insensitive to time. Imagine that fully complying with global duties to B takes a long time (all or almost all the time that is left in A's life). The lexical principle requires state S to completely postpone its duties to A until its duties to B are fully complied with. This is problematic, because sometimes it takes long to comply with complex duties such as the duty to reduce global poverty. If duties to A are taken seriously they should be at least partially discharged at some point in the near future, or at the very least before A's death. But if the lexical principle is adopted, there is no guarantee that duties will be complied with on time, not even partially.

Third, the lexical principle is insensitive to numbers. Imagine a modified scenario in which there are only two feasible alternatives: reducing B's poverty by a moderate amount or reducing inequality for millions of S's residents by a moderate amount. The lexical principle unreasonably requires that we choose the first alternative.

These three unintuitive insensitivities may be summarized by the idea that the lexical principle is insufficiently sensitive to the less weighty duty. By assigning all disputed goods to the winning duty regardless of numbers, benefit size and time, the lexical principle does not to take the losing duty seriously enough.33

Moderate forms of winner-take-all principles are able to overcome some of the difficulties with the lexical principle. Consider, for example, David Miller's proposal.34 He believes that duties toward compatriots are ceteris paribus weightier than duties toward nonresidents.35 And he seems to adhere to a winner-take-all principle.36 However, his principle is not lexical because for him a duty's ground is not all that matters for determining its weight.37 On the contrary, Miller claims that relative weight can be influenced by several factors.38 First, duty weight depends on the number of individuals that are targeted by the duty.39 Miller's solution is therefore sensitive to numbers. Second, Miller claims that weight should be sensitive to benefit size.40 The bigger the benefit, the stronger the duty. So Miller's proposal is capable of overcoming two important objections to lexical principles: insensitivity to benefit size and to numbers.41

Miller's moderate winner-take-all principle, however, is still insensitive to time. This is a problem that no winner-take-all principle can overcome: if we adopt that principle, it is at least theoretically possible that the losing duty will never be discharged during its target's life. Suppose, for example, that inequalities in Sweden could be eliminated if foreign aid were suspended for 20 years. And suppose that reducing inequality for one additional individual in Sweden is not excessively costlier than (1) reducing poverty to a very large extent for a single nonresident or (2) reducing poverty to a moderate extent for a large amount of nonresidents. In these circumstances, Miller's principle would require suspending foreign aid for 20 years. Now, for some people living in poor countries this suspension would mean that their claims to subsistence are going to be left completely unattended for the rest of their lives, because many will die before the 20 years have passed.

We would face similar problems if instead of Miller's principle we adopted the opposite moderate winner-take-all principle: a principle claiming that global duties are weightier than domestic duties. Indeed, imagine that the Sustainable Development Goal of ending poverty can be met by 2030 if at least some affluent states postpone their domestic egalitarian duties until then.42 In these circumstances, the principle would require those states to postpone their domestic duties until 2030. But for some people in affluent states (especially those over 65 years old), this would mean that the state will leave their claims to more equality completely unattended for the rest of their lives (given that overall life expectancy is around 78 years in the EU and 79 years in North America).

Proportionality principles can overcome all shortcomings of winner-take-all principles. Much in the same way as moderate winner-take-all principles, proportionality principles can incorporate rules requiring duty weight to vary according to how many people would be benefited and according to how much bigger these benefits are when compared with benefits provided by competing duties. Now, proportionality principles have a crucial advantage: they are inherently sensitive to time. In every duty conflict, each side gets at least one portion of the disputed goods. Therefore, resident A does not have to wait a long time to see his or her claims at least partly fulfilled. Proportionality principles thus take losing duties relatively more seriously: they receive due recognition by always being assigned a proportional part of the disputed goods.43

A proportionality principle for the domestic–global conflict would look like this: ‘each affluent state must comply with its domestic and global duties to the maximum possible extent; in case of duty conflict, states must satisfy each duty in proportion to its relative weight; weight should be adjusted for the number of individuals who benefit from compliance, and for the size of this benefit’. At this point the question may arise, what is the relative weight of domestic and global duties (before adjusting for numbers and benefit size)? This article remains neutral regarding this question and expects its conclusions to apply to all competing views about relative weight.

It is of course possible to bite the bullet and argue that there are strong theoretical grounds for adopting the lexical principle, whatever its unintuitive consequences. In this section, I present and reject an argument that is based on the usual distinction in practical philosophy between duties of justice and humanitarian duties, and on the related distinctions between perfect and imperfect duties, enforceable and non-enforceable duties, and duties that are correlative to rights and duties that are not.44 Although the following discussion refers to the lexical principle, the conclusions apply indistinctly to all winner-take-all principles.

Consider the following argument: (a) duties to reduce domestic inequality are duties of justice, while (b) duties to reduce global poverty are humanitarian duties; (c) duties of justice are lexically prior to humanitarian duties; therefore, (d) domestic duties are lexically prior to global duties.45

Claim (c) could mean at least two different things. We could make first a ‘definitional’ interpretation, according to which that claim would mean the following: (c.1) ‘the proper use of the term “duty of justice” refers to a duty that is lexically prior to humanitarian duties; therefore, for something to count as a duty of justice, it must be true that it is lexically prior to humanitarian duties (among other conditions)’. Under this interpretation, it is clear that the reasoning from (a) to (c) is insufficient to reach conclusion (d). It needs to be shown first that domestic egalitarian duties really are duties of justice, that is, that domestic duties truly are lexically prior to humanitarian duties (and that global duties really are humanitarian, that is, lexically inferior to duties of justice). But this is what the argument was trying to show (conclusion (4)). Therefore, the definitional interpretation of claim (c) renders the argument superfluous.

But claim (c) could also have the following ‘implicational’ interpretation: (c.2) ‘the proper use of the term “duty of justice” refers to a duty that has has attributes X, Y, Z; attributes X, Y, and/or Z imply lexical priority over humanitarian duties’. Under this interpretation, claim (c) could clearly support an argument for lexically prioritizing duties.

There are, however, some decisive obstacles against this move. First, it is not clear what are those ‘attributes’ in duties of justice that imply lexical priority over humanitarian duties. Duties of justice are usually defined as being enforceable and correlative to rights.46 Now, enforceability does not seem to imply lexicality. At most, an enforceable duty can be regarded as weightier than a duty that is not enforceable, but we have learned above that this cannot be enough to prefer a lexical solution over a proportional solution. Perhaps, lexicality is implied by correlativity to rights. Duties which are correlative to rights seem to be lexically prior to, or ‘trump’, duties which are not.47 But what does it really mean that ‘duties which are correlative to rights are lexically prior’? Again, there could be a definitional and an implicational interpretation. Under the definitional interpretation, we have the same problem as with (c.1) above: it still remains to be shown that domestic duties really are correlative to rights (in a lexicality-implying sense). Under the implicational interpretation, it is not clear which attributes possessed by right-correlative duties could imply lexical priority. And, of course, it cannot be argued that ‘right-correlative duties are lexically prior because they are duties of justice’, because this would lead to circular reasoning (given that we started by asking why duties of justice are lexically prior to humanitarian duties).

There is a second obstacle to the implicational interpretation. Even if it could be shown that attributes such as correlativity to rights and enforceability do imply lexicality, it could be argued that global duties possess those attributes as well (so premise (b) is false). In other words, it could be argued that the domestic–global conflict is not a conflict between a duty of justice and a humanitarian duty, but between two duties of justice. This is a complex point to make, and I will here provide just a few supporting remarks.

Consider first correlativity to rights. It is often argued that in order to meaningfully speak of a right, one must be able to identify a correlative perfect duty, that is, one must be able to identify a specific duty bearer, with specific responsibilities to specific right bearers. Now, it seems that duties to secure basic needs to every human being are not perfect: there are no institutions assigning specific responsibilities to provide those goods and services required to satisfy the basic needs of every human being.48 Henry Shue has convincingly answered this challenge by providing a more complex picture of the universal right to secure basic needs. According to him, this right is correlative to ‘mediating duties’, which are imperfect duties to create (or reform) institutions that assign perfect duties.49

Duties to secure basic needs to every human being can also be regarded as enforceable. This depends, however, on what is meant by saying that a duty is enforceable. If it means that there are actual institutional mechanisms in place for forcing compliance or penalizing non-compliance, then duties to secure basic needs are not enforceable. International institutions are currently incapable of forcing compliance or issuing credible threats to back the relevant duties (and these duties are not even clearly assigned yet). Now, if enforceability simply means a moral permission or requirement to enforce a duty, then duties to secure basic needs can be regarded as enforceable. There are reasons of fairness for this. If mediating duties required creating institutions for distributing perfect duties without at the same time requiring the creation of enforcing mechanisms, free rider problems would ensue. Those unwilling to comply with their duties would bear fewer costs than those who are willing, which is unfair. Moreover, those willing to comply may even be required in non-ideal situations to shoulder the duties of those who are unwilling, which is even more unfair.50 Mechanisms coercing compliance or penalizing non-compliance help solve these fairness concerns.

In sum, the argument built around the distinction between duties of justice and humanity is incapable of showing that the lexical principle should be preferred to the proportionality principle. Given that the lexical principle (and all moderate winner-take-all principles) also has unintuitive consequences as well, there are strong reasons for solving the domestic–global conflict in a proportional way.

In this article, I have analyzed several principles for solving conflicts between domestic duties to reduce inequality and global duties to reduce extreme poverty. The lexical priority principle is insensitive to a range of important considerations: the number of people who would benefit from compliance with a duty, the size of those benefits, and the time it takes to comply with a duty. Insensitivity to time is not exclusive of the lexical principle, but affects all winner-take-all principles. I have shown that proportionality principles are immune to all these problems, so they are better suited as a solution to duty conflicts. I have also assessed an argument using the distinction between duties of justice and humanitarian duties, which aimed to show that domestic duties are lexically prior to global duties. This argument was found lacking because nothing in duties of justice or their defining attributes implies that they are lexically prior to humanitarian duties. The conclusion is that domestic duties and global duties should be complied with simultaneously, dividing available goods between both duties in a way that matches their relative weight.

Research leading to this article was financed by a doctoral scholarship from the Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET). I am grateful to Marcelo Alegre, Nahuel Maisley, Ernest Mbonda, Julio Montero, Martín Oliveira, Cristián Rettig, Eduardo Rivera López, and assistants to the International Seminar on Human Rights and Justice organized by Universidad Torcuato Di Tella and the Political Philosophy Group (GFP) for helpful comments.

Notes

1 For references to authors who agree that both duties exist simultaneously, see notes 5–8 below.

2 On the extent of severe global poverty, see Francisco Ferreira et al., ‘A Global Count of the Extreme Poor in 2012: Data Issues, Methodology and Initial Results’ (Washington, DC: Policy Research working papers, World Bank Group, 2015), 1–66.

The United States is an example of an affluent state with important socioeconomic inequalities in terms of educational opportunities, health care, and job opportunities. On inequalities in educational opportunities in the United States, see Joseph Altonji and Richard Mansfield, ‘The Role of Family, School, and Community Characteristics in Inequality in Education and Labor-Market Outcomes’, in Whither Opportunity? Rising Inequality, Schools, and Children's Life Chance, eds. Greg Duncan and Richard Murnane (New York: Russell Sage Foundation, 2011), 339–58; in health care, see Centers for Disease Control and Prevention, ‘CDC Health Disparities and Inequalities Report: United States, 2011’, Morbidity and Mortality Weekly Report 60 Suppl. (2011): 1–113; and in job opportunities, see Heidi Shierholz, John Schmitt, and Lawrence Mishel, ‘Assessing the Job Polarization Explanation of Growing Wage Inequality’ (presentation at the Labor Economics Seminar at the University of California, Berkeley, CA, March 7, 2013), http://eml.berkeley.edu/~webfac/moretti/e251_s13/mishel.pdf (accessed July 20, 2015).

3 This definition of egalitarianism and of the possible stances in the global justice debate is based on David Wiens and Christian Barry, ‘What Second-Best Scenarios Reveal About Ideals of Global Justice’, in Oxford Handbook to Global Justice, ed. Thom Brooks (Oxford: Oxford University Press, Forthcoming), 2–7, https://www.academia.edu/15589866/What_Second-best_Scenarios_Reveal_about_Ideals_of_Global_Justice (accessed July 20, 2015).

4 Wiens and Barry, ‘What Second-Best Scenarios Reveal About Ideals of Global Justice’, 4.

5 Extenders accepting domestic egalitarian duties on contextual grounds include Gillian Brock, Global Justice: A Cosmopolitan Account (Oxford: Oxford University Press, 2009), 282; and possibly Kok-Chor Tan, Justice without Borders: Cosmopolitanism, Nationalism, and Patriotism (New York: Cambridge University Press, 2004), 160–1, while extenders accepting those duties on principled grounds include Brian Barry, ‘Statism and Nationalism: A Cosmopolitan Critique’, Nomos 41 (1999): 12–66; and Darrell Moellendorf, Cosmopolitan Justice (Boulder, CO: Westview Press, 2002), 39–43.

6 See, for example, Michael Blake, ‘Distributive Justice, State Coercion, and Autonomy’, Philosophy and Public Affairs 30 (2001): 271; David Miller, National Responsibility and Global Justice (Oxford: Oxford University Press, 2007), 164, 166; and Thomas Nagel, ‘The Problem of Global Justice’, Philosophy and Public Affairs 33 (2005): 118.

7 See, for example, Darrell Moellendorf, ‘Absolute Poverty and Global Inequality’, in Absolute Poverty and Global Justice: Empirical Data, Moral Theories, and Initiatives, eds. Elke Mack et al. (Burlington, VT: Ashgate, 2009), 122. Extenders may also adhere to a global sufficientarian principle, in which case states’ duties toward the global poor would be overdetermined.

8 Extenders arrive at this conclusion through a variety of reasons, instrumental (some claim that addressing severe poverty at home is more efficient) and principled (failing to prioritize residents’ survival over nonresidents’ survival shows insufficient respect for residents). For the instrumental path, see Robert E. Goodin, ‘What Is So Special about Our Fellow Countrymen?’, Ethics 98 (1988): 663–86; for the principled path, see Richard Miller, ‘Cosmopolitan Respect and Patriotic Concern’, in The Political Philosophy of Cosmopolitanism, eds. Gillian Brock et al. (Cambridge: Cambridge University Press, 2005), 127–47; and Thomas Pogge, World Poverty and Human Rights: Cosmopolitan Responsibilities and Reforms (Cambridge: Polity, 2002), 132–3.

Another important agreement between restricters and extenders regards the prohibition to violate fundamental freedoms of nonresidents in order to promote domestic basic justice or domestic egalitarian justice. Miller, National Responsibility and Global Justice, 44.

9 See, for example, Jan Narveson, ‘Welfare and Wealth, Poverty and Justice in Today's World’, The Journal of Ethics 8 (2004): 305.

10 See, for example, Charles Jones, ‘Patriotism, Morality, and Global Justice’, in Global Justice, eds. Ian Shapiro and Lea Brilmeyer (New York: New York University Press, 1999), 134.

11 Goodin, ‘What Is So Special about Our Fellow Countrymen?’, 679.

12 Tan, Justice without Borders, 167–8.

13 David Miller, Citizenship and National Identity (Cambridge: Polity Press, 2000); Samuel Scheffler, Boundaries and Allegiances: Problems of Justice and Responsibility in Liberal Thought (Oxford: Oxford University Press, 2002).

14 Arash Abizadeh and Pablo Gilabert, ‘Is There a Genuine Tension Between Cosmopolitan Egalitarianism and Special Responsibilities?’, Philosophical Studies 138 (2006): 359.

15 Michael Blake, ‘Distributive Justice, State Coercion, and Autonomy’, Philosophy and Public Affairs 30 (2001): 259.

16 It could even be the case that at t3 each resident is left with 0 goods. Yet, strict equality would still be preserved. Of course, if residents are left with 0 goods, then basic justice is compromised. But this is a different question (see the first section).

17 The claim that the position of the worse off individual is worse at the end than it was at the beginning is true only if it refers to individuals within the same state. If its reference were wider, including all human beings, the claim could be false.

18 Instead of introducing a tax reform at t3, another way in which equality may be restored is by letting time pass by. It might be the case there is no need for the state to issue new regulations. The sum of multiple interactions may gradually restore equality by itself. Nevertheless, it would still be true that the time in between should be computed as a loss in terms of equality.

19 This is argued by economists such as William Easterly in The White Man's Burden: Why the West's Efforts to Aid the Rest Have Done So Much Ill and So Little Good (Oxford: Oxford University Press, 2006), cited in Wiens and Barry, ‘What Second-Best Scenarios Reveal’, 8.

20 This example and its description is extracted from Stephen Macedo, ‘When and Why Should Liberal Democracies Restrict Immigration?’, in Citizenship, Borders, and Human Needs, ed. Rogers M. Smith (Philadelphia, PA: University of Pennsylvania Press, 2011), 303–4.

21 George Borjas, Economic Research on the Determinants of Immigration. Lessons for the European Union (Washington, DC: World Bank, 1999).

22 Angus Deaton, The Great Escape: Health, Wealth, and the Origins of Inequality (Princeton, NJ: Princeton University Press, 2013).

23 Robert van der Veen, ‘Reasonable Partiality for Compatriots and the Global Responsibility Gap’, Critical Review of International Social and Political Philosophy 11 (2008): 422. For a similar view, see Kok-Chor Tan, ‘Nationalism and Global Justice: A Survey of Some Challenges’, in Sovereign Justice: Global Justice in a World of Nations, eds. Diogo P. Aurelio, Gabriele De Angelis, and Regina Queiroz (New York: Walter de Gruyter, 2011), 14–15.

24 van der Veen, ‘Reasonable Partiality for Compatriots and the Global Responsibility Gap’, 422.

25 The notion of ‘feasibility’ is complex and admits of degrees: something can be highly feasible or highly infeasible (see Pablo Gilabert, ‘The Feasibility of Basic Socioeconomic Human Rights: A Conceptual Exploration’, Philosophical Quarterly 59 (2009): 659–81).

26 This distinction is based on John Rawls, A Theory of Justice (Cambridge, MA: Belknap Press of Harvard University Press, 1971), 34.

27 Rawls, A Theory of Justice, 42.

28 I take the distinction between winner-take-all and proportionality from Mark R. Reiff, ‘Proportionality, Winner-take-all, and Distributive Justice’, Politics, Philosophy & Economics 8 (2009): 6–7. There are additional rules for solving conflicts, such as the equality rule (‘all duties should be assigned an equal share of goods’), which are not directly relevant for this article.

29 The idea of a ‘lexical’ ordering of normative considerations was made famous by Rawls, A Theory of Justice, 42.

30 This example shows that moderate winner-take-all principles tend to produce similar results to those of proportionality principles (although not quite the same results, as will be shown in the next section).

31 This fallacy seems to be committed by, for example, Mathias Risse, On Global Justice (Princeton, NJ: Princeton University Press, 2015), 330–1.

32 Carl Knight, ‘In Defence of Cosmopolitanism’, Theoria 58 (2011): 19–34; Tan, Justice without Borders, 160–1; Robert van der Veen, ‘Limiting the Scope of the Weighting Model: A Reply to David Miller’, Critical Review of International Social and Political Philosophy 13 (2010): 553.

33 In a similar vein Mark Reiff claims that winner-take-all principles treat losing claims with unequal respect and consideration (Reiff, ‘Proportionality, Winner-take-all, and Distributive Justice’). If Reiff is right there are strong moral reasons for preferring proportionality principles (or at least strong pro tanto reasons).

34 Miller's proposal can be found in his article ‘Reasonable Partiality Towards Compatriots’, Ethical Theory and Moral Practice 8 (2005): 63–81, in his book National Responsibility and Global Justice, and in a series of articles in the volumes 11 and 13 of the Critical Review of International Social and Political Philosophy, where he discussed with Robert van der Veen, Gillian Brock, Leif Wenar, Veit Bader, among others: David Miller, ‘A Response’, Critical Review of International Social and Political 11 (2008): 553–67; and David Miller, ‘In Defence of Weighting: A Reply to Robert van der Veen’, Critical Review of International Social and Political Philosophy 13 (2010): 561–6.

35 Miller, National Responsibility and Global Justice, 45; and Miller, ‘A Response’, 562.

36 Miller uses a series of terms such as ‘trump’, ‘outweigh’, and ‘take precedence’ that suggest the idea that the weightiest duty receives all disputed goods (see, for example, Miller, National Responsibility and Global Justice, 563, 564; and Miller, ‘In Defence of Weighting: A Reply to Robert van der Veen’, 562). Moreover, Miller provides several examples to illustrate how his view is supposed to work, and in those examples the weightiest duty collects all disputed goods (Miller, National Responsibility and Global Justice, 44–6).

37 Ibid., 49.

38 Ibid., 50.

39 Ibid., 45–6.

40 Miller, ‘A Response’, 564. Miller speaks of amounts of harm to be prevented, not of benefits to be provided, but the idea sufficiently is similar.

41 Miller's view is more complex than may seem from how I present it. According to him, there are several other factors that influence duty weight. For instance, Miller claims that duties are weaker when they arise from a failure of responsibility by a third party (such as the government in a poor country), and they are stronger when they arise from no such failure (Miller, National Responsibility and Global Justice, 50).

42 For an introduction to the Sustainable Development Goals, see Sustainable Development Goals, http://www.un.org/sustainabledevelopment/sustainable-development-goals/ (accessed February 17, 2016).

43 A similar idea can be found in Reiff, ‘Proportionality, Winner-take-all, and Distributive Justice’, 14–5.

44 In this article humanitarian duties are taken as synonymous with duties of beneficence, assistance, and charity.

45 A similar argument can be found in Miller, ‘Reasonable Partiality towards Compatriots’, 76.

46 Laura Valentini, ‘Justice and Assistance: Three Approaches and a Fourth One’ (Odense: Working Papers Series, Centre for the Study of Social Justice, 2009), 5.

47 Ronald Dworkin, ‘Rights as Trumps’, in Theories of Rights, ed. Jeremy Waldron (Oxford: Oxford University Press), 153–67.

48 Onora O'Neill, Bounds of Justice (Cambridge: Cambridge University Press, 2000), 103–5.

49 Henry Shue, ‘Mediating Duties’, Ethics 98 (1988): 699–704. Perhaps, the argument for lexically prioritizing domestic duties could be saved by appealing to a different sense of the distinction between perfect and imperfect duties. Perfect duties may be defined as not allowing the agent any degree of discretion regarding their fulfillment, while imperfect duties allow for such discretion. But this move hardly does the trick. Even if it is granted that non-discretionary duties are weightier than discretionary ones, this is not enough to show that the former are lexically prior to the latter. Why cannot proportionality be the proper principle for solving weight differences between discretionary and non-discretionary duties?

50 Pablo Gilabert, From Global Poverty to Global Equality: A Philosophical Exploration (Oxford: Oxford University Press, 2012), 41–2.