1. INTRODUCTION
Over the past few years, political philosophers have been expressing increasing dissatisfaction with the dominant, Rawls-inspired, methodological paradigm in theorizing about justice. Much of their frustration with this paradigm stems from its perceived inability to deliver principles capable of guiding action in real-world circumstances. Rawls's focus on the ideally just society, so the critics argue, is practically idle and potentially counter-productive. If political philosophy is meant to help us orient our actions in the real world, Rawlsian ‘ideal theorizing’ is just the wrong way to think about the subject (see e.g. Baier Reference Baier1985; McCarthy Reference McCarthy, Benhabib and Fraser2004; Mills Reference Mills2005).Footnote 1
Amartya Sen's most recent work, culminating in the monograph The Idea of Justice, offers one of the most forceful and authoritative articulations of this general dissatisfaction with Rawls-inspired political philosophy (Sen Reference Sen2006, Reference Sen2009). Despite his admiration for Rawls's work,Footnote 2 Sen argues that political philosophy should move beyond the Rawlsian methodological outlook – which Sen calls ‘transcendental institutionalism’ – towards a different, more practically oriented, approach to justice: ‘realization-focused comparison’. Is Sen's call for a paradigm shift in thinking about justice warranted? In this paper, I argue that it is not. Most of Sen's criticisms are in fact either based on a misrepresentation of the Rawlsian approach, or correct but of little consequence. What political philosophy needs is not a paradigm shift, but a more nuanced understanding of the paradigm Sen and others criticize.
The paper is structured as follows. In section 2, I offer a brief overview of Sen's arguments against transcendental institutionalism, specifically focusing on three key complaints: (a) comparative judgements are what we want from a theory of justice, but transcendental institutionalism is neither necessary, nor sufficient for comparisons; (b) transcendental institutionalism is inherently parochial, and (c) transcendental institutionalism is inflexible. In section 3, I criticize Sen's characterization of the Rawlsian paradigm as a form of transcendental institutionalism. With a clearer picture of the Rawlsian paradigm in mind, I respond, in sections 4–6, to Sen's three criticisms, and show that they are either trivial or misguided.Footnote 3 I conclude that Sen's complaints are based on unfortunate mischaracterizations of Rawls's method, which are pervasive in the existing literature. Sen's arguments do not show that the Rawlsian paradigm is fundamentally defective, but rather, that it needs to be better understood and further developed.
I should note that the argument I offer in this paper is negative. I do not discuss Sen's own path-breaking contributions to political philosophy, whose value is not in dispute, but simply argue that his recent critique of Rawls's approach to the subject misses its target. Although this article is mostly negative, by showing that Sen's critique rests on a misdescription of the Rawlsian enterprise, my hope is to lay the foundations for a better-informed methodological debate in political philosophy.
2. TRANSCENDENTAL INSTITUTIONALISM AND ITS FLAWS
Central to contemporary theorizing about justice, Sen observes, is the question: ‘What is a just society?’ Those who, like Rawls, put this question at the heart of political philosophy subscribe to what Sen calls transcendental institutionalism. On the one hand, their approach is ‘transcendental’ because it aims to identify an ideal of a perfectly just society. On the other, it is ‘institutionalist’ because it attempts to establish what perfect institutional arrangements would be like, without paying much attention to the conduct of individuals. In short, transcendental institutionalists seek to identify a set of perfectly just social institutions. For them, societies in the real world are unjust to the extent that they fail to exhibit such institutional perfection (Sen Reference Sen2009: Chapter 1).
Although this approach finds its origins in the works of Hobbes, Locke, Rousseau and Kant, Sen (Reference Sen2009: 7–8) sees Rawls's theory of justice as ‘[t]he most powerful and momentous exposition’ of transcendental institutionalism. Recall that Rawls's theory contains two key principles:
1. Each person has an equal claim to a fully adequate scheme of equal basic rights and liberties, which scheme is compatible with the same scheme for all; and in this scheme the equal political liberties, and only those liberties, are to be guaranteed their fair value.
2. Social and economic inequalities are to satisfy two conditions: first, they are to be attached to positions and offices open to all under conditions of fair equality of opportunity; and second, they are to be to the greatest benefit of the least advantaged members of society (Rawls Reference Rawls1993: 5–6).Footnote 4
In Sen's view, this theory exhibits the following distinctive marks of transcendental institutionalism: (i) it delivers a unique and definitive set of principles, (ii) these principles select a particular set of institutions, and do not apply to individual behaviour, and (iii) a society whose institutions satisfy these principles is perfectly just.
Despite its popularity, Sen (Reference Sen2006, Reference Sen2009) argues, the transcendental institutionalist paradigm fails to give us ‘what we want from a theory of justice’. That is, it fails to deliver theoretically sound conceptual tools that can help us advance justice in the real world.Footnote 5 In particular, Sen puts forward three main complaints against transcendental institutionalism, which I outline below.Footnote 6
2.1 Comparative judgements of justice are what we want, but transcendental institutionalism is neither necessary nor sufficient for comparisons
Transcendental institutionalism, Sen argues, allows us to make judgements about what we might call ‘absolute’ justice. From a transcendental institutionalist perspective, a society is either perfectly just, or it is unjust. However, he continues, these are not the sorts of judgements we should be interested in, especially when it comes to the issue of advancing justice in the real world. Much more important are comparative judgements of justice and injustice, and for those, an answer to the question ‘what would a fully just society look like?’ is neither necessary, nor sufficient (Sen Reference Sen2009: 98–106).Footnote 7
Firstly, to know that the ‘iniquities of hunger, illiteracy, torture, arbitrary incarceration, or medical exclusion’ are sources of injustice, notes Sen, one need not have a detailed account of what qualifies as a perfectly just society (Sen Reference Sen2006: 218; Reference Sen2009: 96). We can establish whether a society is more or less just by reference to these criteria, without appealing to the higher-order ideal of a fully just social system. ‘[T]he injustice of continuing famines in a world of prosperity, or of persistently grotesque subjugation of women’, can be easily detected without a complete and exhaustive picture of what full justice requires (Sen Reference Sen2009: 103).
Secondly, knowing what a perfectly just society looks like does not automatically allow us to make comparative judgements of justice (Sen Reference Sen2006: 219–221; Reference Sen2009: 98–101). To make such judgements, we also need a metric to evaluate which social arrangements are furthest away from the ideal and what improvements would bring them closer to it. If we are interested in comparisons, theories like Rawls's are thus radically incomplete. They offer us an ideal, but give us no tools to establish how far any specific social arrangement is from it.
In sum, since transcendental institutionalism is neither necessary nor sufficient for comparisons, its intellectual and practical contributions are unsatisfactory, and a different approach to theorizing about justice is needed.Footnote 8
2.2 Transcendental institutionalism is parochial and status-quo biased
Sen (Reference Sen2009: 24–27, and Chapter 6) complains that transcendental institutionalism unduly limits the scope of justice. This is because its demanding ideal of perfect justice can only be realized where state-like institutions exist. Only institutions such as those of the modern state can engage in the comprehensive redistributive policies advocated by most contemporary theories of justice. Since ought implies can, on this view, outside the state, principles of justice become irrelevant. This is why, Sen says, theorists such as Rawls (Reference Rawls1999b) and Thomas Nagel (Reference Nagel2005), to whom he ascribes this institutionalist paradigm, deny that principles of distributive justice apply to the global arena, and limit the scope of reasoning about justice to domestic political communities. In short, the search for perfect justice renders transcendental institutionalism oblivious to some of the gravest injustices plaguing our world: international ones.
2.3 Transcendental institutionalism is inflexible
Aiming at the identification of the perfectly just society, Sen (Reference Sen2009: 106–107) further argues, transcendental institutionalists tend to ignore the ‘inescapable plurality of competing principles’ that any plausible approach to justice should acknowledge. Rawls's original-position reasoning, for instance, is said to lead to the selection of a unique set of principles. It is unclear, however, whether all rational or reasonable persons would really assent to the theory of justice Rawls proposes. There may be a plurality of permissible principles, and the ambition to pick out one set, and one only, is misguided and counter-productive (Sen Reference Sen2009: 46).
Moreover, subjecting principles of justice to ‘some radical surgery that reduces them all into one tidy box of complete and well-fitted demands’ can be hubristic and myopic (Sen Reference Sen2009: 46). When designing a theory of justice, Sen (Reference Sen2009: 107) suggests, we should always be open to revising our conclusions. For instance
[w]e often think, if only implicitly, of the plausibility of principles in a number of specific cases . . . But once the principles are formulated in unconstrained terms, covering inter alia a great many cases other than those that motivated our interest in those principles, we can run into difficulties that were not foreseen earlier, when we signed up, as it were, on a dotted line. We then have to decide what has to give and why.
But, problematically, these trade-offs seem to be inadmissible within Rawls-inspired, transcendental theorizing, with its insistence on ‘exacting and highly demanding rules’ (Sen Reference Sen2009: 107). Once we have identified what perfect justice requires, we can no longer revise that ideal. We remain trapped, so to speak, in the realm of perfection.
In short, transcendental institutionalism seems ‘practically irrelevant’ on three key dimensions. First, it is neither necessary nor sufficient for establishing what counts as more or less just in real-world circumstances. Second, it prevents us from applying principles of justice to many contexts, most importantly the global one, for which we feel such principles are most relevant. Third, it is inflexible and thus ill-suited to offer a framework for thinking about justice in a world where agents might not all agree on what justice requires.
What could a Rawlsian respond to these charges? An easy response would be to point out that practical relevance is not what we want from a theory of justice. A theory of justice, on this view, is first and foremost a theoretical exercise, and its value is largely independent of its practical import. This response would be readily available to philosophers like G.A. Cohen (Reference Cohen2003: 243), who believe that justice is not primarily about what we ought to do, but about what we ought to think. However, this response is not available to the Rawlsian philosophers Sen is criticizing, since they believe that a sound theory of justice should be action-guiding.Footnote 9 To defend themselves, Rawlsians cannot dismiss the accusations of lack of guidance capacity as irrelevant, but must show that they are ill-founded. This is my aim in the remainder of this paper.
3. TRANSCENDENTAL INSTITUTIONALISM AND THE RAWLSIAN PARADIGM
Let me begin with a few clarifications regarding Sen's description of the Rawlsian paradigm as a form of transcendental institutionalism. First, there is an apparent ambiguity in Sen's use of the adjective ‘transcendental’. At first blush, by calling Rawls's theory ‘transcendental’, Sen may be taken to suggest that the theory sets out principles for a perfectly just society, transcending the limits of human existence. This, of course, would make Rawls's theory eminently non-practical.
Although Rawls does indeed attempt to identify principles for a fully just society, those principles are far from being ‘transcendental’ in this sense. On the contrary, Rawls is keen to make sure that the principles he defends are consistent with Humean moderate scarcity and limited altruism, that the ‘ideal society’ he envisages is not beyond reach.Footnote 10 Of course, one might argue that, substantively, Rawls has failed to accomplish this task, and that his favoured social arrangements are beyond the limits of human practical possibility. But this would not make Rawls's approach in any way transcendental, it would only make his theory less plausible than it would otherwise be by Rawls's own non-transcendental standards. In short, it is fair to say that Rawls's theory is ‘ideal’, but inaccurate to say that it is transcendental in this sense.
As it turns out, on a charitable reading, this is not the sense of ‘transcendental’ Sen has in mind. For Sen (Reference Sen2009: 6), a theory is transcendental when it is ‘absolute’, namely non-comparative. While comparative principles of justice take roughly the form ‘Society X is more just than society Y if (or iff) . . .’, transcendental ones are formulated in the categorical form ‘Society X is perfectly just iff . . .’. From this more formal perspective, Sen is correct in describing Rawls's theory as transcendental (although he could have chosen a better label, such as ‘categorical’). Its two principles tell us what it takes for a society to be perfectly just, rather than for it to be more just than another. Whether this feature of Rawls's theory is genuinely as problematic as Sen thinks is something I shall discuss later in the article. I now turn to the characterization of Rawls's paradigm as institutionalist.
Rawls, unlike Sen, is indeed an institutionalist, but not in the sense Sen seems to attribute to him.Footnote 11 Sen complains that institutionalists are concerned with perfectly just institutions, but not with overall ‘social realizations’. In his words, ‘transcendental institutionalism concentrates primarily on getting the institutions right, and it is not directly focused on the actual societies that would ultimately emerge’ (Sen Reference Sen2009: 6). This description of the Rawlsian paradigm is somewhat misleading. Rawls is an institutionalist insofar as his principles of justice are meant to apply to the basic structure of society (i.e. to its most important legal, political and economic institutions), rather than to individual behaviour. But Rawls's principles do not single out one set of perfect institutions. Rather, on a Rawlsian view, the institutional arrangements which make a society just vary depending on the nature of the society in question, the character of its citizens and so forth.
A society characterized by racial prejudices, for instance, may very well require affirmative action policies to realize Rawlsian fair equality of opportunity. The very same policies, however, would violate fair equality of opportunity in a society in which no prejudices existed. In short, on a plausible Rawlsian approach, there is no such thing as an ‘ideal set of institutions’, but rather a set of lexically ordered principles (equal basic liberties, fair equality of opportunity, and the difference principle) which can be realized by a number of different institutional arrangements. As Thomas Pogge (Reference Pogge2000: 165 emphasis original; Reference Pogge2004) points out, ‘Rawls's criterion of justice assesses a basic structure by the distribution it would tend to produce in the actual social system it organizes.’ The institutional features of a just basic structure change depending on the particular social system they regulate precisely because what matters is ‘the actual society that would ultimately emerge’.
With this clearer picture of the Rawlsian paradigm in mind, we can move on to assess Sen's three challenges.
4. RAWLS-STYLE PRINCIPLES ARE NOT WHAT WE WANT FROM A THEORY OF JUSTICE
Sen's first challenge is that, since an account of perfect justice is neither necessary nor sufficient for making comparative judgements of justice across different societies, the Rawlsian paradigm is seriously defective. Even though the paradigm allows us to make judgements about absolute justice – i.e. about whether a social system is perfectly just – the ability to make these judgements is not ‘what we want from a theory of justice’. Comparative judgements are of far greater importance, especially when it comes to advancing justice in the real world.
For this challenge to be successful, two claims have to be true: (i) there is little theoretical and (especially) practical value in knowing what perfect justice is, and (ii) the Rawlsian paradigm is neither necessary, nor sufficient for comparisons. In what follows, I take up and discuss each claim in turn.
4.1 Rawlsian perfect justice is of little relevance
While I agree with Sen that an account of perfect justice is not all we want from a theory of justice, I also think it is part of what we want from any comprehensive such theory. In particular, an account of the principles which would govern a fully just society is necessary to establish when a society is ‘completely’ just. Sen significantly downplays the importance of this particular aspect of Rawlsian theorizing. For instance, he suggests with some scepticism that
it may well turn out that in a comparative perspective, the introduction of social policies that abolish slavery, or eliminate widespread hunger, or remove rampant illiteracy, can be shown to yield an advancement of justice. But the implementation of such policies could still leave the societies involved far away from the transcendental requirements of a fully just society (since transcendence would have other demands regarding equal liberties, distributional equity, and so on) (Sen Reference Sen2006: 217).
But why is it problematic that, from a Rawlsian perspective, a slavery-free society would still not qualify as fully just? Surely, Sen would agree with Rawls that justice requires a lot more than the abolition of slavery, and therefore that there can be societies that are significantly more just than the slavery-free ones mentioned in the present example.
More generally, the claim that a society is (or can be) perfectly just says something important about both the nature of the value of justice, and how to orient our actions in the real world. With respect to the former, there is a crucial structural difference between a conception of justice, such as utilitarianism, in which justice can always be increased (in the form of greater sum-total utility), and one where justice cannot be increased indefinitely as a matter of principle.Footnote 12 If part of the point of a theory of justice is (plausibly) to describe the nature of the value of justice, objecting to the theory that it is not ‘intrinsically’ comparative in the way utilitarianism, for instance, is, is somewhat beside the point. For it may be that the correct account of justice does have a cut-off point beyond which the idea of an increase in justice simply makes no sense. Unless Sen is prepared to deny this substantive claim, he cannot downplay the value of theorizing about perfect justice quite so easily.
Moreover, from a practical perspective, if it is true that justice has a cut-off point, then we have an interest in knowing what that point is. This will enable us to decide whether we should strive to change existing circumstances – insofar as they depart from the ideal – or preserve the status quo. We should better know if our society is just, and no more justice-improvements are necessary or possible.Footnote 13
So far, I have defended the theoretical as well as practical significance of perfect justice. This already diminishes the import of Sen's accusation that Rawls-style principles are neither necessary nor sufficient for comparative judgements of justice. If what Rawls does is necessary for something else, and this is also of some importance, the call for a paradigm shift in justice theory appears overstated.
4.2 Rawlsian justice is both unnecessary and insufficient for comparisons
Let me begin with the claim that Rawlsian ‘perfect’ justice is unnecessary for justice comparisons. In one respect, this claim is correct. For instance, a society in which people are arbitrarily arrested is obviously more unjust than one in which, all other things equal, they are not. Similarly, a society in which women are subjugated is clearly more unjust than one in which, all other things equal, they are not. No account of perfect justice is needed to make these kinds of judgements.
Despite being correct, this observation is also inconsequential. There are many judgements of justice – both absolute and comparative – we make confidently and intuitively. These include the judgements about arbitrary arrest, destitution, illiteracy and severe human suffering Sen invokes in support of his claim that transcendental principles are not what we want from a theory of justice. Problematically, however, Sen seems to ignore that no theory is needed to formulate such judgements in the first place. If anything, those judgements constrain any plausible normative theory: a theory of justice that permits child torture, abuse of women and arbitrary arrest is obviously absurd. By arguing that, to make straightforward comparative judgements, we do not need a theory of perfect justice, Sen proves too much. As it turns out, to make those judgements, we need no theory.
What, then, is a theory of justice supposed to do? Why do we need a theory of justice to begin with? Rawls's practically oriented answer is that a theory of justice is necessary to establish how to distribute scarce goods (including liberties, opportunities, wealth and the social bases of self-respect) we all need to lead lives worth living. If there were no conflicts over goods, there would be no need for a theory of justice in the first place.Footnote 14 Similarly, there would be little point in theorizing about justice if we were already certain about how goods ought to be distributed and we all agreed on the relevant distributive criterion, and the reasons in support of it. But this is not the situation in which we find ourselves, both individually and collectively.
Individually, we often have strongly held convictions about what counts as just and unjust, but no overarching criterion helping us to check their mutual consistency and to understand their relations to one another. Collectively, we often disagree about what counts as just or unjust, and when we agree in judging some actions or states of affairs as just or unjust, we typically rank them differently. Is a society that arbitrarily arrests some of its citizens more or less unjust than one in which part of the citizenry lives in conditions of poverty and destitution? Is a society with the death penalty more or less just than one in which the death penalty has been abolished but where serious crimes are committed much more frequently?
These are the sorts of questions on which people's intuitions are either shaky or divergent (or both). It is because we find ourselves in situations of uncertainty, disagreement, and confusion that we engage in abstract theorizing about justice along the lines suggested by Rawls. In Rawls's (Reference Rawls1999a: 508) own words, when we theorize about justice we produce ‘argument addressed to those who disagree with us, or to ourselves when we are of two minds’. Normative justification ‘presumes a clash of views between persons or within one person’. Unless we want to content ourselves with our unsystematic and diverging intuitive judgements, Rawlsian-style higher-order moral reasoning becomes unavoidable (Rawls Reference Rawls1999a: 42ff.).Footnote 15
From this perspective, designing an overall conception of what justice requires is necessary for practical purposes – a view Sen also endorses. We need to develop a general account of justice to become clear about our priorities, about what is more or less important as a matter of justice, and about how our seemingly conflicting intuitions fit together.Footnote 16 The process of achieving greater clarity and coherence in our judgements of justice gradually delivers a picture of what justice requires. Without such a picture, coherent guidance in matters of justice would seem to be impossible.
At this point, Sen may agree that general principles of justice, worked out in abstraction from many contingent aspects of human existence, are necessary to make systematic and action-guiding judgements of justice and injustice. However, he might further argue, the real problem with Rawls's theory is that its abstract principles are formulated in absolute, rather than comparative, terms. Instead of claiming that ‘Society X is more just than society Y if (or iff) . . .’, Rawls's principles state ‘Society X is perfectly just iff . . .’ and this renders the principles unsuitable for making comparative judgements of justice.
There are two things Rawlsians can respond to Sen's charge. First, as I have already noted, it is by no means obvious or evident that the value of a theory of justice lies entirely in its conduciveness to comparative judgements. If there is such a thing as absolute justice, then a good theory of justice should, among other things, tell us what this is. That said, a good theory of justice should also enable us to compare different social systems from the viewpoint of justice. Is Rawls's theory so useless in this respect? I believe not, and this leads me to the second response.
Although Rawls's theory is formulated in absolute terms, important tools for comparison can be extrapolated from it. For example, its appeal to the lexicographic priority of the basic liberties vis-à-vis fair equality of opportunity and the difference principle already enables us to establish that a society in which citizens’ fundamental liberties are violated is more unjust than one in which such liberties are respected, no matter how unequal its distribution of income and wealth is. Moreover, Rawls's difference principle is comparative in nature: it allows us to assess alternative basic structures by reference to their impact on the position of the worst off.Footnote 17 Indeed, Sen (Reference Sen2009: 97) is aware that ‘Rawls's formulation of the difference principle . . . gives us ground enough to rank other alternatives in terms of the respective advantages of the worst-off’. To that extent, Rawls's ‘absolute’ theory of justice is in no way inimical to comparisons, or unable to deliver any.
Of course, this does not mean that a Rawls-style theory can give us all answers to comparative questions of justice. This leads us to Sen's claim that the ‘transcendental’ theorizing Rawls and the Rawlsians engage in is insufficient for making comparative judgements of justice and injustice. Taken literally, this claim is false. As I have already suggested, it seems undeniable that principles like the ones Rawls proposes do take us a fairly long way towards the possibility of making some comparative judgements of justice and injustice. For instance, the principles clearly establish (if one accepts them) that violations of basic liberties are more serious than violations of fair equality of opportunity which, in turn, are more serious than violations of the difference principle.
A correct claim, in the vicinity of Sen's, is instead that Rawls's theory only allows us to make some comparative judgements of justice. For instance, the theory does not tell us how violations of different basic liberties are to be traded-off against one another (Sen Reference Sen2006: 220; Reference Sen2009: 99). And while there may be clear-cut cases (e.g. a small restriction in freedom of movement is less unjust than torture), there are also bound to be controversial ones, in which ranking different societies from the viewpoint of justice proves extremely difficult. However, this merely amounts to acknowledging that a Rawlsian-style theory is incomplete, and thus does not automatically deliver all solutions to all problems of justice. More work needs to be done – this is for sure – but, once again, no paradigm shift is called for.
To conclude, the complaint that Rawls-style ‘transcendental’ principles of justice are neither necessary nor sufficient for comparative judgements of justice is far from revealing the need for a paradigm shift in theorizing about justice. First, since the comparative is only part of what we want from a theory of justice, even if a particular account of justice is unnecessary for purposes of comparison, it can still be of theoretical and practical value. Second, as it turns out, Rawls's ‘absolute’ principles of justice do give us some important materials to compare alternative social arrangements. This suggests that Rawls's theory – which Sen considers to be the most important example of transcendental institutionalism – delivers much, although not all, of what Sen himself wants from a theory of justice.
5. AN ARBITRARY RESTRICTION OF THE SCOPE OF JUSTICE?
The second critique Sen mounts against the Rawlsian approach to justice concerns its restricted scope. By arguing for such demanding principles, Sen (Reference Sen2006: 226–228; Reference Sen2009: 24–27) claims, Rawls and other philosophers who follow his paradigm (for instance Thomas Nagel) inevitably limit the scope of justice to those contexts in which there already exist institutions capable of realizing it: bounded societies. This has two unpalatable implications. First, it makes talk of justice inapplicable to the global arena, where in fact most of us feel that appeals to justice are urgent and appropriate. Second, it unduly restricts the scope of the reasoning by which principles of justice are arrived at. Recall that Rawls's thought experiment, the original position, only contains the representatives of members of a particular society. But why should our reasoning about justice (even if justice is thought to be confined to the domestic arena) be limited in this way? Wouldn't foreign perspectives enrich our deliberations? Let me consider both aspects of the scope restriction in turn.
Sen is right that the fact that currently there exist no institutions capable of realizing Rawls's egalitarian distributive principles on a global scale does not suffice as a reason for denying their global moral validity. If it were possible to construct the institutions needed to realize them at reasonable costs to those involved, then such institutions ought to be constructed. What Sen seems to miss in his account of why Rawls and Nagel deny the applicability of (egalitarian) justice globally is that they are not primarily concerned with issues of feasibility, but rather, with issues of moral appropriateness.Footnote 18
On what strikes me as the most plausible interpretation of their views, stringent principles of (egalitarian) justice are morally appropriate only when certain kinds of social relations are in place. Rawlsians disagree about what the relevant relations are: some believe they are coercive relations, others think they concern reciprocity in the production and distribution of certain social goods and so forth.Footnote 19 Settling these disagreements is not necessary for present purposes. The important point is that, on a plausible reading of Rawls and Nagel, what justifies their domain-restriction for principles of (egalitarian) justice is not a commitment to transcendental institutionalism. Central to their views is instead the normative claim that (egalitarian) justice only applies in the presence of certain types of social relations, coupled with the empirical claim that such relations do not exist at the global level. In sum, to criticize Rawls's and Nagel's stances on global political morality, Sen needs to engage with their normative and empirical underpinnings, but these underpinnings have very little to do with transcendental institutionalism.
Let me now turn to the second aspect of Sen's ‘scope’ critique. This concerns the variety of perspectives that can contribute to our reasoning about justice (Sen Reference Sen2009: 138–145). Of course, if the scope of principles of justice were indeed global – say because the relevant relations existed globally – then it would seem entirely arbitrary to restrict the scope of reasoning about justice to a particular society. The design of principles of justice should take into consideration the voice of everyone to whom they apply. This is precisely the rationale behind Rawls's domestic original-position thought experiment: assuming a closed society, principles of justice should be unanimously accepted by citizen representatives behind a veil of ignorance.Footnote 20 But can the scope of reasoning about justice be detached from the scope of application of principles of justice, as Sen suggests? In other words, if we are designing principles of justice for society X, should we also take into account what members of societies A, B, C think about them?
This is an interesting suggestion. In a Millian spirit, theorists of justice should be willing to test their views against as many opponents (or as many other people) as possible (Mill Reference Mill1985 [1859]: Chapter 2). Only this can guarantee the type of impartiality Rawls (and Sen) are looking for. Sen (Reference Sen2006: 235; Reference Sen2009: 149–152) is therefore right when he says that ‘the demands of objectivity not only require avoiding a “personal slant” (as Rawls noted), but also national parochialism . . .’. However, there seems to be a fundamental problem with the open public reasoning approach Sen favours, and this has to do with setting the boundaries of reasoning about justice.
In the Rawlsian (Reference Rawls1993: 26) architecture, these boundaries are set by the constraints built into the original-position thought experiment, which is meant to articulate the ideals of citizens as free and equal and of society as a fair system of cooperation. For instance, the reasons and perspectives of those who believe that different human beings have different moral worth and that some are intrinsically superior to others, are barred from entering the deliberative process via which principles of justice are selected. For Rawls, a commitment to the fundamental moral equality of persons is a necessary prerequisite for participating in that process. Sen, on the other hand, seems to reject this idea – given that many of those whom he would want to include in the deliberation process might disagree with it. There are many societies in which, for instance, women are considered inferior to men. How should the views of those who hold this conviction be factored into the process of reasoning about justice? How are they to be dealt with?
These questions raise an important challenge for Sen's view. His ideal of public and open reasoning about justice either is so inclusive as to become almost empty (given that public reasoning will contain completely irreconcilable views) or it surreptitiously implies certain substantive moral commitments which automatically exclude perspectives that are distant enough from the liberal one.
The latter alternative is probably most likely to be correct. After all, there would be an obvious clash with Sen's procedurally inclusive approach if he were not also committed to the moral equality of persons. Why should we want to extend the principle ‘audi alteram partem’ globally, if not because of a belief in the moral equality of persons? However, once we concede a belief in the moral equality of persons, we have already substantially restricted the scope of (public) reasoning about justice to those views which are consistent with that belief, in which case Sen's position is not that far from Rawls's after all. In short, anti-parochialism is laudable and important, but no coherent theory of justice can be anti-parochial all the way down. Certain fundamental commitments – such as commitments to liberty and equality – must be non-negotiable. To the extent that Sen does not wish to abandon those commitments, his view cannot boast much greater inclusiveness than Rawls's.
6. IS RAWLSIAN JUSTICE INFLEXIBLE?
It remains to analyse the last of Sen's three complaints against Rawlsian theorizing. This points to its alleged lack of flexibility and open-endedness. First, Rawls defends one set of principles as the output of the original-position thought experiment but, as it turns out, there may be a plurality of admissible principles which might be chosen in the original position. Rawls's theory, then, is too ambitious. It aims to be complete, to establish what perfect justice requires, while in fact a ‘partial’ – albeit abstract and general – ideal of justice is all we need and can plausibly reach. Second, the principles Rawls defends do not seem to be open to revision. They indicate what perfect justice requires, and are thus as demanding as they are inflexible. This makes Rawls's theory of justice unable to offer a basis for a fruitful public discussion, and reduces both its theoretical and practical appeal.
Are these complaints well-founded? They are not.Footnote 21 First, in his later work, Rawls himself acknowledged that there is a plurality of reasonable liberal conceptions of justice, which may be less distributively egalitarian than his own ‘justice as fairness’.Footnote 22 More generally, it is true that Rawls might have been too optimistic about the possibility of conclusively justifying a theory of justice as rich and demanding as the one he proposed in his early work. But it seems odd to criticize his methodological paradigm on these grounds, when he himself has explicitly acknowledged this shortcoming. Moreover, it is also hard to see the excessive ambition of the early Rawls as calling for a paradigm shift in theorizing about justice. Sen's argument, as I see it, does not show that we need not think about what a fully just (or reasonably just) society is. Rather, it points to the fact that our prospects of offering a plausible and robust justification for a highly specific account of justice are not as good as many theorists, including the early (but not the later) Rawls, think they are.
That said, Rawls has always remained firm on the possibility of identifying constitutional essentials and fundamental principles of justice that any reasonably just society should satisfy. I assume that Sen would not want to disagree with this weaker claim. Indeed, while it may be sensible to reduce the ambitions of our theorizing about justice, as we have already noted, taking this modesty to the extreme would make theories of justice entirely useless. A theory of justice has to say something substantive about what justice requires in order to be of interest in the first place.
Let me thus turn to the second complaint advanced by Sen. In Sen's view, the dominant, Rawls-inspired paradigm delivers principles of justice which are, in some sense, unrevisable. Although this is a charge that can perhaps be plausibly made against some contemporary political philosophers, I do not think it can plausibly be directed to those who take Rawls's methodology seriously.
Rawls's entire architectonic of justice, including the original-position thought experiment, is embedded in the holistic method of reflective equilibrium. Within such a method, no assumptions, principles or claims are taken for granted or unproblematically assumed. Everything is open to revision, and ‘fixed points’ are only provisionally fixed.Footnote 23 When trying to reach reflective equilibrium, we have to go back and forth between general principles and considered judgements in search for overall balance. Within this process, we are constantly faced precisely with those decisions about what to revise, and why, which Sen sees as never arising within a ‘transcendental’ approach to justice. Recall his quote from a previous section:
[w]e often think, if only implicitly, of the plausibility of principles in a number of specific cases . . . But once the principles are formulated in unconstrained terms, covering inter alia a great many cases other than those that motivated our interest in those principles, we can run into difficulties that were not foreseen earlier, when we signed up, as it were, on a dotted line. We then have to decide what has to give and why. Some may find social choice theory to be too permissive and indecisive . . . but the alternative, well illustrated by mainstream theories of justice, like Rawls's or Nozick's, does not give the idea of justice its due (Sen Reference Sen2009: 107).
The particular reasoning process Sen associates with his preferred approach to justice, and sees as alien to mainstream Rawlsian theorizing, in fact turns out to be nothing other than reflective equilibrium.Footnote 24
What is more, Rawls is quite explicit in his openness to revising not only his principles of justice in light of opposing intuitive judgements, but also the very conditions built into the original position. In his words:
[i]n searching for the most favoured description of this situation [i.e., the original position] we work from both ends. . . . By going back and forth, sometimes altering the conditions of the contractual circumstances, at others withdrawing our judgments and conforming them to principle, I assume that eventually we shall find a description of the initial situation that both expresses reasonable conditions and yields principles which match our considered judgments . . . (Rawls Reference Rawls1999a: 18).
Once this is appreciated, Sen's complaint that Rawls's methodology is aimed at delivering only one set of principles appears all the more bizarre. Even if the early Rawls thought that only one such set of principles could be compatible with the original-position thought experiment, that very thought experiment was itself always open to revision in light of new intuitive judgements and considerations. For Rawls and the Rawlsians, the process of theorizing about justice is inherently dynamic and open-ended. Sen's charge of inflexibility is therefore also misguided.
7. CONCLUSION
In this article I have raised some doubts about Amartya Sen's recent critique of the Rawlsian paradigm in theorizing about justice. Taken together, my arguments suggest that, at most, this paradigm needs to be better understood, and further developed, but is far from being fundamentally flawed in the ways Sen describes. On closer scrutiny, the Rawlsian paradigm delivers much of what Sen himself wants from a theory of justice. Sen has simply chosen the wrong target. The paradigm he criticizes is not the one pioneered by Rawls. If there are transcendental institutionalists, Rawls is not one of them.
That said, nothing of what I have argued should be read as an endorsement of Rawls's substantive, as opposed to methodological, views or as suggesting that these views cannot be further improved. Moreover, Sen's emphasis on the importance of advancing justice in the real world, and his focus on comparative judgements of justice are laudable. My aim has been to show that his critique of Rawls's method is, for the most part, misdirected and, in so doing, to provide the background for a more fruitful methodological debate within political philosophy.
