Abstract
This article distinguishes between three methodologies for thinking about justice: principle-based, model-based and ‘realist’, concentrating mainly on the differences between the first two. Both model-based and realist approaches pride themselves on taking institutions seriously and argue that institutions make a fundamental difference to justice. This claim is at best not proven, and it may be possible to account for the difference that institutions make to what justice requires while retaining a non-institutional account of what justice is.
Introduction
It is understandable that political theorists and political philosophers aspire to ‘relevance’, whatever that is. After all, what is the point of abstract theorising about politics, of all things, if that theorising is not to some purpose? What is the point of a theory of justice, if it has no purchase on the ‘real world’ and has no chance of being put into practice? Many political philosophers were drawn to the subject by an interest in and a commitment to making the world a better place, and it can be hard to see that pure theory, disconnected as it is from the concerns of politicians and the exigencies of policy, contributes to the very goal that motivated us in the first place. There are therefore all kinds of reasons why political philosophers should be drawn towards ‘relevance’, to ‘realism’ and to ‘non-ideal theory’. However, this desire for relevance should not and cannot affect the truth of matters involving the relationship between practice and value, between facts and principles, between reality and utopia. Those are truths (if there are truths to be had) come what may, and the wishes, aspirations and anxieties of theorists and philosophers cannot alter them. It is our job to discover what they are, and our task should not be influenced by wishing they were otherwise.
This article aims to map some of the conceptual space in the debate between so-called realists and moralists and around ideal and non-ideal theory. It argues that one important distinction is not between realists and moralists but within the moralist camp. It is a distinction between, on one hand, theories that purport to take political institutions seriously and which claim that their existence makes a fundamental normative difference and, on the other, theories that see the basic principles of justice as being, in some sense, prior to institutions (even if institutions can, by their existence, change what justice requires). Those ‘moralist’ theorists who concur with the realists that institutions make a fundamental normative difference typically employ models of the just society – so-called ideal theory – in ways that are constitutive of the principles of justice they think ought to obtain. By contrast, other ‘moralist’ theorists attribute no such constitutive role to models and seek to investigate what pre-institutional principles of justice and other rights and powers of individuals require, given the way the world contingently is. The use of idealising models, slightly paradoxically, makes moralist model-theorists in some sense less ‘realistic’ than their principle-based rivals, given that the latter have no need to consider what might be required under idealised counterfactual conditions in order to address what justice requires in the actual world. The article also suggests, provocatively, that idealised models risk serving as ideological alibis for policies that are actually unjustified.
Three different methodological approaches
In the literature on justice generally and by extension on global justice, we can distinguish three broad types of views concerning how we should go about doing our thinking on the issue. 1 First, there are theories based on the application of general moral principles to particular subjects. These can be of simple form, such as, say, act utilitarianism, or they can invoke a plurality of normative principles. Second, there are theories where an ideal model of society plays some kind of constitutive role in establishing what the relevant duties and obligations are. Both of these theories are ‘moralist’ theories in a sense employed by a third group of theorists who deem themselves ‘realists’ and who claim that the political realm generates normative standards that are entirely sui generis. I shall refer to the first group as ‘principle-based’, the second as ‘model-based’ and the third as ‘realist’. Both principle-based and model-based approaches are ‘moralist’ views, from the perspective of the realists, and both model-based and realist approaches take themselves to be ‘more realist’ views compared to the principle-based one since they hold that facts about the way the world is are at least partly constitutive of basic normative principles for that world. 2
By ‘principle-based’ theories, I intend to denote those moral and political theories that conceive of normative political theory as consisting in the application of some principle or principles to a particular subject, the social and political order, where those principles in some sense hold independently of that particular social and political order. By ‘application’ I intend quite a broad set of possible relations, which may include the idea that, say, some political institution or other is necessary to further the realisation of some value, that it is necessary to respect some value, that it is permissible according to some value and so on. By ‘hold independently’ I mean only to assert their non-dependence on those particular contingent facts or institutions; I do not mean to assert the truth of moral realism, leaving those deeper metaethical issues unaddressed and unresolved. I therefore rely on the manifest appearance of the moral in folk-conceptions of it, without prejudice to the issue of whether those folk-conceptions are in error. The church of ‘principle-based’ views is a broad one: it includes both consequentialisms and TM Scanlon’s (1998) contractualism, the deontological libertarianism of Nozick (1974) and the perfectionist liberalism of Raz (1986); it includes most versions of luck egalitarianism and the radical pluralism of GA Cohen (2008) and many other views besides.
Both the ‘realist’ theorists and those ‘moralists’ who employ models think that the practice of institutions in the actual world has some kind of constitutive significance for the kinds of normative principles that hold. The model-theorists, following Rawls (1999a, 1999b), standardly proceed by testing their ideas against a ‘realistic utopia’ or ‘well-ordered society’. In effect, this is a specific kind of thought experiment, asking whether a set of principles would be functional and stable in an imaginary society whose citizens were committed to those principles and hence motivated to comply with them. These imaginary societies are like ours in recognisable respects, containing similar institutions (markets, legal systems, systems of political representation) but have been adjusted for the purposes of normative theorising. These shifts from reality are in two respects: first, unlike most actual societies, the hypothetical societies in these models are actually trying to realise normative principles, and second, they have been cleansed of citizens with an insufficient commitment to the principles in questions and are now peopled with committed citizens who yet still have ‘realistic’ constraints on the balance between their own self-interest and the realisation of justice. These hypothetical citizens are not expected to make extraordinary sacrifices. 3
A number of interesting questions arise about the relationship between these models of ‘realistic utopia’ and the actual world. The most obvious of these concerns the applicability of principles that hold in the model society to actual societies, an issue to which I return below. But the normative fit works in both directions: not only do the prescriptions obtained in the ideal model have significance for the actual world, the values that are explored in that model also have their source in the values and practices of particular institutions. Although this was less obvious in, say, Rawls’ early work such as A Theory of Justice, his later work on public reason disclosed the fact that the source of the values worked up and cleansed for the idea model was to be found in values implicit in the practices and history of actual liberal democracies, such as the United States. 4 The realistic utopia, then, also functions as a kind of mirror, in which we see our societies as they would be, if only they were true to their own supposed values, while being divested of other values that are empirically present in them but which the philosopher deems irrelevant or undesirable from the point of view of justice. This move, based as it is on a moral interpretation of existing practices, is open to emulation for a range of different subjects. So we can ask not only about the values supposedly embedded in, say, liberal democracies but also about the values embedded in the international order, the system of global economic institutions and so forth. Some model-based theories accentuate this interpretive aspect more than others, but it is clearly the basis for the family of approaches that goes under the label ‘practice dependence’ and gives those approaches a slightly relativistic flavour (e.g. James, 2013; Ronzoni, 2009; Sangiovanni, 2007, 2008).
It is primarily within these model-based approaches that the contrast between ideal and non-ideal theory becomes an issue. Specifically, the question arises of how the principles that are generated for the realistic utopia (ideal theory) have relevance and application in the real world, where the constitutive assumptions of that realistic utopia do not obtain (non-ideal theory). One answer to that question is to say that the realistic utopia defines a standard that can orient us in the actual world and that the work of people who are committed to justice must be to work out ways to effect a transition to the kind of society defined by the realistic utopia. One kind of non-ideal theory, then, is concerned with this problem of transition. 5 Some of Rawls’ writing on non-ideal theory seems to have this character, as when he is writing of the terms of settlement of a just war and does this with the goal of the coexistence of reasonable peoples in mind, but at other times, say when endorsing ius in bello prohibitions, it is far from clear that the ideal model is playing much of a role. The difficulty, though, with looking on the realistic utopia as an ideal and seeing the goal of non-ideal theory as being about the strategy and tactics of transition to it is that it isn’t at all clear how that task relates to making the actual world more or less unjust. When thinking about tax policy, for example, should we be trying to reform it so it is more like the tax policy that would obtain in a realistic utopia? Or does justice require something else, perhaps something more pragmatic that looks to incremental improvements without such a goal as a standard?
The use of a hypothetical model of society does not, in itself, mark a distinction between model-based and principle-based theories. The important thing to note is that the two approaches employ such models in different ways. On one hand, we can use the idea of a hypothetical society as a purely epistemic device, where we think about how an ideal family, or society, or international system could work can be a way of testing our principles and working out what our fundamental values actually are. 6 This is basically nothing more than the old philosophical method of examining superficially plausible principles in the light of counterexamples and being willing to revise them if they appear to yield unacceptable counterexamples. For example, the fact that act utilitarianism appears to sanction the exemplary execution of the innocent counts against it, thus forcing us to revise or abandon the principle or, if we are feeling very brave, to ‘bite the bullet’. Many people engaged in ‘ideal theory’ are therefore best seen as principle-based theorists engaged in testing and refining their principles using hypothetical examples. But for them those principles don’t depend constitutively or ontologically on those models or thought experiments. The theorists whom I designate as model-theorists, on the other hand, assign not just an epistemic but also a constitutive role to models: for them, part of what makes a valid principle a valid principle is the fact that it would be implemented and respected in the well-ordered counterpart to the institutions under investigation.
According to their protagonists, both realist and model-based theorists take institutions seriously in ways that principle-based theories do not. What this means and the extent to which it is true are explored below, but first I need to introduce some further context concerning the global justice debate.
The ‘global justice’ debate: An outline of the territory
To get some traction on these methodological issues in relation to the ‘global justice’ debate, I need to say something about that debate, which has proceeded in the normal way, that is, by a proliferation of positions and stances, each one subtly different from the next. However, at least within the camp of people who think of themselves as roughly egalitarian and liberal, we can distinguish three broad families of views, differentiated along two different dimensions. 7 First, there is distinction between so-called relational and non-relational views, that is between views that hold that strong duties of distributive justice apply to people who stand in some specific kind of relation to one another, and those who hold that such duties apply to all human beings (or even all persons, if there are non-human persons) in virtue of their common humanity or personhood. Among so-called relational views, there is wide disagreement about the kinds of relationships or associations that trigger distributive-justice duties and much of the micro-differentiation of theories flows from differences on this question. Second, there is the dimension of scope. Since non-relational views hold among persons generally, and there are persons all over the planet, it follows that non-relational theories have global scope. 8 For relational views, by contrast, the key question is whether the supposed distributive-justice triggering condition holds globally or merely locally. Theorists who hold that productive co-operation via a division of labour is the correct triggering condition can point to the fact of globalisation and the existence of a global division of labour; those whose claim picks out some more local condition, such as subjection to a common political authority, will favour a theory with narrower scope, arguing perhaps that strong egalitarian duties hold only among co-citizens or residents of particular states.
Notice that hybrid views of various kinds are possible and even common. So, for example, some relational views support egalitarian duties of graduated strength, depending on the nature of the ties in question, so that more equality is required among co-citizens than with foreign co-operators, but where there is some egalitarian normative pressure to reduce inequalities with those foreign co-operators, although not by as much. 9 Similarly, those committed to strong egalitarian duties within states often concede that there are humanitarian duties of assistance to destitute outsiders, quibbling perhaps over whether those humanitarian duties count as duties of justice in some proper sense of that word. They thereby include within their theory of global justice a recognition of some non-relational grounds of justice and use that to justify a sufficientarian duty.
Early incarnations of the global justice debate often couched the disagreement between people who contended that principles of distributive justice ought to have global scope and those who thought that egalitarian duties held mainly among compatriots (for example) as a disagreement between cosmopolitans and nationalists. But this way of expressing the disagreement turned out to be controversial because it appeared to suggest that ‘nationalists’ were committed to a thesis about the comparative intrinsic worth of human beings, to the view that some people matter more than others. In fact, all theorists who are broadly in the egalitarian liberal ‘family’ of views are committed to an egalitarian thesis about the moral status of all human beings. Given this commitment, and their parallel commitment to some egalitarian view or other about distribution, it follows that those who favour giving principles of distributive-justice-limited scope face a puzzle about how they can both affirm the equal worth of persons and argue that it is morally permitted (indeed required) for some people or institutions to show greater concern for the interests of some of those persons than for others. Hence, there is the requirement for theorists who want to limit the scope of distributive principles to invoke some associative or relational reason that justifies this discrimination in favour of ‘insiders’.
Relational views come in a number of different flavours. Prominent among them are theories that stress the importance of states or peoples as self-determining collectives that must take responsibility for their choices (for good or ill); theories that claim that the key triggering factor is when people impose a framework of coercive law with distributive implications on one another; and theories which claim that when people co-operate closely in the production of a limited range of important collective goods under some co-ordinating authority, this gives rise to egalitarian obligations that don’t otherwise exist. Although such theories vary importantly in their details, they all share a commitment to the claim that institutions make a difference and that the importance of institutions for distributive justice is not limited to being instruments for the realisation of some antecedently given moral principle, but that by their existence and practice, they generate a new set of duties that were not around before. Moral argument around such institutions has both an inside and an outside. We can ask about the justification of the institution as a whole, whether it realises certain important human goods, refrains from imposing significant harms and so on. We can also ask about the internal norms that constitute it and bind its constituent parts, be they institutions or persons. Sometimes those internal norms will be explicit terms of association of the collective in question, sometimes they will be a matter of its custom and practice and sometimes our more critical focus will be on revising norms and practices better to fit in with what we take the meaning and moral purpose of the relationship, association or institution to be.
How does the classification of global justice views relate to the distinction among methodological approaches that I outlined in the previous section? While there may be some cases which are not easy to classify, as a rough generalisation, non-relational globalist views about justice are instances of principle-based views, whereas relational views which restrict the scope of distributive justice to some group, entity or set of institutions tend to be model-based views. All of these views are, in the eyes of ‘realists’, such as Galston or Rossi, moralistic. The key divide here, at least in the eyes of those advancing it, is that institutions make a difference to what is normatively required. It is going to be crucial, then, to see whether those accused of denying the normative significance of institutions do (or need) deny it or whether they also can account for that putative significance within their own approach.
The ‘institutions-make-a-difference’ thesis
The claim that ‘realists’ think distinguishes them from moralists and which model-based theorists (including Kantian justice and practice-dependence theorists) in turn believe distinguishes them from a variety of other moral views is, then, the institutions-make-a-difference thesis, the thesis that the existence of and participation in a system of rule-governed practices of some kind makes a really big difference to the normative principles that apply.
William Galston (2010) has a typical statement of this kind when he writes,
Political moralists, … tend to view institutions simply as means for the realization of antecedently established principles and aims. To be sure, there is something to this means/ends thesis: a flat-footed reading of the Declaration of Independence (‘to secure these rights’) or of the Constitution (‘in order to’) tends to support it. But, say the realists, there is much more to institutions than their instrumental function. (2010: 393)
In similar vein, Samuel Freeman (2007a), who is in Galston’s terms a ‘political moralist’, writes,
[C]osmopolitans do not recognize that there are distinct and independent principles of social and political justice … Social principles of justice, if they exist at all, are derivative from allegedly more basic principles of cosmopolitan justice. (2007a: 423)
Where this claim is made by model-based theorists and is applied to reasons of justice, it has one of its modern sources in Rawls’ A Theory of Justice and comes in strong and weak forms. The strong form claims that institutions of some kind are an existence or applicability condition for claims of justice; the weak form claims that institutions create reasons of justice which were not there before, but which may coexist or supplement some other reasons of justice.
Strong and weak forms of the institutional thesis
Right at the beginning of A Theory of Justice, John Rawls (1999a) tells us that ‘justice is the first virtue of social institutions’ (1999a: 1). The natural way to read his statement is simply as a declaration that justice is of paramount importance for social institutions. Indeed, it is a rather extreme declaration of that importance, given that Rawls does not appear to allow that the virtue of justice be compromised, for institutions, in order to allow for the value of other things like well-being or efficiency. What Rawls does not say in that famous passage is that justice only applies within institutions, that is, it is in some constitutive sense an institutional virtue. Indeed, he leaves open the possibility that justice, as a virtue, applies to other contexts and subjects, including extra-institutional ones. It is not, then, ‘justice’ that is constitutively limited to institutional contexts, but, at most ‘Rawlsian justice’, a subset of justice, worked through for a particular subject, the basic structure of society.
One reason that it is important to draw attention to this is that some subsequent writers in the Rawlsian tradition have argued as if justice does indeed only come to apply as a virtue within institutions of a specific type. For example, Thomas Nagel (2005), in his well-known article on global justice, argues that justice, properly speaking, only comes into being subsequent to the establishment of a political authority. For Nagel, it is only when a state, guaranteeing social order, holding back the bellum omnia contra omnes and securing co-operation in the production of key public goods, is established, that it makes sense to talk about justice at all. In similar vein Samuel Freeman (2007b), criticising attempts by cosmopolitans and others to specify principles of justice independent of institutions, argues that Rawls
assumes that justice and social cooperation too are not possible without governments and complicated legal systems, and that what social justice involves, in large part, are principles for structuring and defining the powers of political institutions … For Rawls, our considered convictions of justice arise within the practices and institutions we live with, and are attuned to the structure and demands of those institutions. (2007b: 271–272)
The reason behind this claim derives historically from Kant (or from a particular reading of Kant, at any rate) and involves a rather restricted understanding of what justice is, taking it to be essentially about rights over property. Kant argues, according to this interpretation, that without the state there can be no property rights, properly speaking, and it follows from this that there can be no justice or rather that talk of justice has neither sense nor application. Kant’s thought here builds on Hobbes. Without a sovereign in place, there can be no property rights because there is nobody authoritative to judge in cases where the extent and nature of those rights are indeterminate, and, second, without the sword of Leviathan in place, nobody can have an assurance that their possessions are secure against the predations of others, and in the absence of such an assurance they have no reason to refrain from taking and using the possessions of others. To this, Kant adds his own distinctive element, the omnilateral will. To claim property rights in a state of nature is unilaterally to impose duties on others, and this nobody has the authority to do. Hence, there is the need to establish an entity with such an authority, an entity that can regulate everybody’s claims from an impartial perspective, that of an omnilateral will (Kant, 1996). 10
This very strong claim, that having institutions in place and institutions of a particular type (namely, a state and a legal system), is a necessary condition for justice and social co-operation seems very unsatisfactory. It rules out by definitional or conceptual fiat concerns that many people have about all kinds of issues that standardly get talked about in terms of justice. If we want to ask whether it is unfair that one person is wealthier or happier than another, through no fault of their own or whether this or that co-operative arrangement involves a fair distribution of benefits and burdens, it will be very frustrating to be told that our concerns involve some kind of category mistake. More plausible than the idea that legal and political institutions are an absolutely necessary precondition for talk of justice to have sense is the idea that institutions make a difference to our judgements and to the applicable principles. This idea seems entirely plausible, but I shall suggest below that this weaker claim can be accommodated by principle-based theorists and need not, despite appearances, mark a distinctive type of view.
Responding to the institutional thesis
There are two ways of responding to the institutional claim an opponent might employ. The first is to bite the bullet and to deny that institutions make a difference for justice. This response demarcates the territory just as the more realist theorist wants, but then denies the truth of the proposition. The second way of responding is to concede that institutions make a difference, but to defuse the claim by arguing that this doesn’t threaten the opponent’s view in the way that the more realist theorist think that it does and that it is possible to absorb the more realist view into a wider, more complete, account of justice. In other words, in affirming that institutions matter, those who make this claim do not, contrary to the view they take of themselves, assert anything that their opponents are committed to denying.
To take the first response first then, it isn’t clear whether anyone ought to bite the bullet, but if anyone does it is some sort of consequentialist, an act utilitarian, for example. On such a view, we are morally required to do whatever we have to do to bring about the best consequences, and institutions are valued or not to the extent to which they help or hinder the maximisation of the good. What we have reason to do is therefore fundamentally unaffected by the existence of institutions, although of course we have to take account of them, as we take account of all contingent features of the world, in our calculation of what we have reason to do at any particular moment. I shall not discuss this view further, both because I regard it as implausible and because, although both ‘realists’ and ‘practice-dependence’ theorists may be happy to present it as a model of the kind of thing they are against (as the quotes from Galston and Freeman above suggest), it does not characterise accurately the position of the people who they are most obviously opposed to. It is, in other words, a caricature.
A weaker, non-maximising, version of a similar view might be Robert Goodin’s (1988) ‘assigned general duties’ model of institutions. On this view, we all have certain general duties to humanity as a whole, both positive (duties of assistance, rescue) and negative (duties not to harm), and considerations of efficiency dictate that we should assign these duties to specific associations that will take responsibility for them on our behalf. So, for example, we assign the duty of providing social order and personal security to particular states, who look after the interests of citizens and others on the territory they have jurisdiction over. Similarly, the general duty to provide for the interests and education of children is assigned to families on the basis that the psychological attachment of parents to their children is the best means of getting the job done. In the case where the institutions fail to do what they are supposed to do, a residual duty then falls on everyone else. People who are persecuted or otherwise failed by the states to which they were assigned can seek protection from other states under the Refugee Convention, and those states have a duty to take them in; children who are neglected or abused by their parents get assigned to foster parents or children’s homes by the state. States and families on this view are essentially instruments for the discharge of general duties and the relationships they engender and make possible do not have free-standing moral significance.
Principles of justice and what they require
However, a second way of responding is to absorb the institutional claim into a wider normative theory. This approach would accept that institutions can make a difference to what justice requires (as opposed to what it is) but finds a place for that claim within a more comprehensive view about what ought to be done. (I have in mind here a distinction that Simon Caney (2011) makes between the principles of justice and what justice requires.) The principles of justice here may include a variety of things, perhaps cosmopolitan views about fundamental equality of moral status together with some more substantive principles (equality, priority, sufficiency) about how some types of things (resources, capability, welfare) should be distributed. Other elements of the picture would include views about the moral rights, powers and duties of individuals and about their capacities to act in ways that transform the moral landscape, including via the use of their moral powers to create institutions. In this optic, there is something misleading about model-theoretic views, including practice-dependent views, because they confuse what they are doing (working out what justice requires for a particular subject) for an investigation into the principles of justice as such. Rather, the basic normative truths hold extra-institutionally, but differences in what justice requires can arise within institutions because of different circumstances and because individuals have contingently used their moral powers in ways that set up particular patterns of duty, obligation, permission and so forth.
If we accord principles of justice and the rights and powers that people justly hold a lexical priority over other values, then this application of extra-institutional moral elements will settle the question of what justice requires for a particular subject. Alternatively, there might be a question about how these various elements combine with other values in order to produce all-things-considered judgements about what should be done or what we should value. So, we might take the view that we have to compromise between justice and some other value such as aggregate well-being.
All of this is rather messy, but inevitably so, and there is no point in hankering for an algorithm for making moral judgements if no such thing is to be had. What might be helpful, though, is some kind of illustration how this picture might work for global justice. Accordingly, in the next section, I sketch out, merely for illustrative purposes, how this might be done.
Moral powers and associative duties
The task then is to sketch how an account of global justice based on pre-institutional moral principles, yet giving an account of the same moral phenomena that model-based theories do, might go.
A starting point for such a sketch would be to assign shares of the world to individuals and also to grant them a plausible set of rights, duties and moral powers permitting them to associate together and transform the moral landscape in various ways. Taking their initial resources and using their moral powers, they might associate together, both to further their private interests and perhaps provide a framework for the more efficient discharge of their general duties. Given their moral powers, they will plausibly acquire special duties of various kinds, including promissory and rectificatory duties, as they associate together, sometimes more or less accidentally. Over time, associates will come to engage in more or less settled co-operative practices that will vary in contingent ways between different societies. Some will arrange their common affairs in one manner; others in another.
To say this might be to appear to cede a great deal of ground to the ‘realist’ or the ‘practice-dependence’ theorist. But this appearance would be misleading. Although associations bring new relationships and obligations into being, they do so on the basis of moral rights and powers that their members already had. They may alter their relationship to one another in various ways, but that association cannot, for example, change their normative situation with respect to outsiders, by endowing them with rights towards those outsiders they formerly lacked. 11 People might get richer by associating together because of the benefits they can get by co-operating, but it is hard to see how this can fairly be at the expense of non-members. If their co-operation does worsen the situation of others, for example, if their activity causes pollution that makes them worse off, then it looks as if associates have a special obligation of rectificatory justice towards those outsiders, an obligation that will involve either restoring the status quo ante, or compensating them for their losses.
If association does not alter the original egalitarian relationship with outsiders concerning resource shares, what about the relationship among associates? Here, model-theorists (at least those whose theory of global justice has a relational basis and restricted scope) have often argued that duties towards other associates, including distributive duties, are more demanding and egalitarian than those towards outsiders. On one way of looking at the matter, this might seem surprising. It might be that we should see less egalitarianism as an option, simply because, given their association, people might use their moral powers in order to sanction inegalitarian norms among themselves that they would have no right to impose on non-associates. However, we must not assume that the rights and powers that get people associated with one another are purely a matter of voluntary choice, and once a co-operative scheme is up and running, it is bound to be the case that new members (by birth) are subjected to a framework of authority or coercion that they did not choose. This may license a comparatively egalitarian approach.
With respect to outsiders, the important question is going to be how the initial allocation of resources translates into a story about distributive shares in the future, once co-operative associations are up and running. What we say here will depend on whether we allow the members of those schemes to hold onto the wealth accumulated by previous generations on the basis that the collective is a kind of intergenerational person, with a capacity for taking responsibility for its actions, bequeathing its assets and so on, 12 or whether the accumulations of previous generations – for which present ones are not responsible – should revert to a global commons. But whichever way the argument goes on that, the jury is out on whether institutions make the kind of difference that model-theorists and realists suppose. The principle-theorist can tell a plausible story about the normative relationships that obtain, both among people in the world in general and among fellow members of associations, a story that can in principle be constructed from a combination of general principles and justice and moral rights, powers, duties and liabilities of persons to associate in various ways that modify their moral relationships. There is no reason to suppose that the modification, in and of itself, creates something new that was not there in potentia beforehand.
Implications for the ‘real world’
Principle-based theories as realistic ones
One slightly paradoxical consequence of what I have argued is that the relationship between some sorts of ‘realism’ and some sorts of ‘moralism’ ought to be conceived differently to how it is often understood. If the key distinction is supposed to be between theorists who are concerned with getting at the truth about fundamental values and those who are keen to engage more closely with political and policy realities, then GA Cohen with his ‘fact-independent principles’, Cohen (2008) and Amartya Sen (2006) look to be at opposite poles of the debate, with Rawls and his followers sitting somewhere in the middle and the model-theoretic idea of a ‘realistic utopia’ appearing to be some sort of compromise between the purest of ideals and grubby practicalities. But actually there need be no great conflict between Cohen and Sen: an entirely practical orientation can be informed by the purest of values: if we conceive of justice as requiring equality, say, we can ask ourselves which policies in the actual world promote that value, as well as taking into account other competing values.
Rawlsian constructivism, by contrast, seeks to define an ideal of justice for a well-ordered society, for a realistic utopia, and so we have to ask ourselves a whole series of difficult questions, first defining our principles according to what would be workable and stable under counterfactual conditions rather distant from actuality and then addressing what we ought to do in the actual world by trying to relate it to how things would be in that counterfactual one. It is not obvious either how to do this or how to relate it to policy. If the business of thinking about a ‘well-ordered society’ is merely one way of trying to get clear about what our basic values are, there can be no objection, but if we conceive of it as being definitive of what justice requires, then things are much more problematic.
Model theory as ideology
Finally, it seems to me that model-theoretic approaches try to define what justice requires in terms of how an ideally just society (a ‘realistic utopia’) face a dilemma when they encounter real-world policy: they risk either irrelevance or apologetics for the status quo. We could even talk of ‘model theory as ideology’. 13 This is a difficult charge to make stick as a strict implication of the approach, but I believe it is not unreasonable given the pragmatics of political argument. The thought is that principles that officially only have application conditionally, in ideal circumstances, get translated too easily into apologetics for policy in non-ideal circumstances in part due to the pressure not to appear irrelevant to ‘real world’.
An example may help bring this out. Immigration is a major feature of the global order, so any theory of global justice will have to take a view on it, and particularly on the question of whether (or under what circumstances) states have the right to exclude would-be migrants. One author who has written extensively on this topic is Christopher Heath Wellman, who argues that legitimate states have the discretion to exclude, a right he derives from the right to self-determination and free association of peoples, freedom of association including, as it does, the freedom not to associate. So what does Wellman’s view imply for real-world policy? Apparently, very little. After all, Wellman has a definition of what a legitimate state is:
a state has earned legitimacy if it is willing and able (a) to protect its members against ‘substantial and recurrent threats’ to a decent human life – threats such as the arbitrary deprivation of life or liberty, and the infliction of torture – and (b) to refrain from imposing such threats on outsiders.
14
Looking at this test of legitimacy, a necessary condition for a state’s right to exclude, it is quite possible to reach the conclusion that hardly any countries meet it. Some perhaps do. Maybe the Scandinavian countries and Canada, possibly Uruguay, have a good enough human rights record. A country such as the United States presents a much more difficult case as critics will point to evidence of a dysfunctional and racist justice and carceral system at home and America’s record of torture and drone strikes overseas. What does Wellman himself think? It is hard to know. If he were to agree with the negative assessment of US legitimacy, it would follow that he could not endorse the freedom of association argument for the right to exclude in that case. However, it is also easy to imagine that readers, even not particularly casual ones, who are sympathetic to states’ rights to control their borders will conclude, on the basis of an idealised representation of freedom of association, that actual states, which do not technically meet the legitimacy test, enjoy the exclusionary right.
One could find similar slippages in the work of other people who work in a model-theoretic way. Ronald Dworkin’s theory of equality of resources appears to issue policy prescriptions under very restricted and ideal circumstances, but his work also contains very concrete prescriptions on, say, welfare and unemployment benefits, that seem to require holding people responsible for their actions in the real world in a way that the austere version of the theory ought to disclaim. 15
Conclusion
The primary purpose of this article has been to mark some distinctions among different methodological approaches to justice and to argue that the dividing lines may not be in quite the positions that partisans of one approach or another think. ‘Taking institutions seriously’ does not mark a divide between realism and moralism, but cuts within the moralist camp between what I call model-based and principle-based theories, the former of which accords institutions – at least in an idealised form – a constitutive role for principles of justice. I have tried to sound a sceptical note, however, about whether institutions should play this role, suggesting that principle-based theories, which try to derive what justice requires from non-institutional principles, may be able to account for how institutions make a difference while denying models a constitutive role for principles of justice as such. Finally, I have suggested that arguments that draw on idealised models of society may play an ideological role in legitimating principles that ought not to apply in the actual world.
Footnotes
Acknowledgements
The author thanks Joanna Burch-Brown and two anonymous referees.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.