1
CHANGING CONSTRUCTIONS
Onora OâNeill
CONTRACTS AND CONSTRUCTIONS
The publication of John Rawlsâs Political Liberalism surprised many of his admirers. It defended the same substantive principles of justice that Rawls had put forward in A Theory of Justice in 1971 but introduced major changes in the arguments by which they were supported. There is of course nothing wrong about offering distinct considerations in support of a single set of claims, yet it startled many readers to find Rawls proposing such large revisions in justification of substantive claims about justice that he had put forward for many years. Here I shall reflect on some of the innovations and the continuities in Political Liberalism that seem to me of most interest and address some of the problems that I think it leaves unresolved.
From the time of writing A Theory of Justice, Rawls saw his work as continuing the social contract tradition while aiming to avoid its recurrent problems. In the preface to A Theory of Justice he writes that he aims to carry âto a higher level of abstraction the familiar theory of the social contract as found, say, in Locke, Rousseau and Kantâ (TJ viii) and in the opening paragraph that he aims to offer âa theory of justice that generalizes and carries to a higher level of abstraction the traditional conception of the social contractâ (TJ 3).1 But whereas in A Theory of Justice Rawls had mainly described his work as contractarian, in Political Liberalism the metaphors of the social contract were mostly replaced by the idea that principles of justice can be constructed by reasonable procedures. The approach taken in Political Liberalism defends a form of political constructivism that seeks to identify principles of justice that would be reached by citizens who reasonably deliberate together, where reasonable citizens are seen as proposing and abiding by fair terms of social cooperation among equals (PL 93ff.).
This specific focus on citizens and their deliberation had not been prominent in A Theory of Justice but can be seen as developing rather than replacing the earlier approach. A Theory of Justice proposes a theory of justice for a society that is thought of as âa more or less self-sufficient association of personsâ (TJ 4) and âas a closed system isolated from other societiesâ (TJ 8). It was clear from its first pages that John Rawls had bracketed questions about international justice (TJ 8), despite assuming that there are many societies. Rereading his initial account of the context of justice with hindsight, it seems that many of the issues raised by his assumption that justice is internal to bounded societies, with their explicit inclusions and exclusions, were present and might have led to earlier questions about these aspects of his arguments.
Yet widespread discussion of the implications of these inclusions and exclusions began only after the publication of Political Liberalism in 1993 and of The Law of Peoples in 1998. I shall argue that the assumption that we are dealing (at least in the first instance) with a bounded society of fellow citizens became more central to Rawlsâs approaches to justifying principles of justice in the years between A Theory of Justice and Political Liberalism, despite the reality that globalization of many sorts was already underway, and that it was becoming less and less plausible to assume that issues of âdomesticâ justice could be handled in isolation. A focus on citizens and citizenship incorporates specific views of who counts in matters of justice and also of who does not count or counts for less. Claims about the scope of justice are fundamental to arguments about its demands, and Rawlsâs views about who counts in various contexts are of central importance to understanding his changing position.
WHO COUNTS IN THE ORIGINAL POSITION?
Rawls claimed in A Theory of Justice that his method is âmore abstractâ than that used in traditional social contract theories.2 In certain obvious ways this is true. For example, he does not assume that there was some historical moment at which the members of each society actually contracted or that (by some unknown means) their successors are now bound by the terms of that original contract. However, denying that the social contract was a historical event would hardly have been an innovation: the point had been familiar since the Enlightenment. But if the terms of the social contract are not historical but hypothetical, how and why do they bind actual persons who have never agreed to them? (See TJ 21). The more abstract approach that Rawls took in A Theory of Justice was intended to continue the social contract tradition by constructing an argument that provided reasons why members of any society of citizens should accept specific principles of justice,3 without drawing on (or inventing?) fictitious accounts of past occasions when they or some of their predecessors agreed to do so, but also without assuming that actual persons have accepted or must accept the terms of a hypothetical agreement or contract. The approach was to be philosophical and constructive rather than historical or quasi legal.
Rawls claimed both in A Theory of Justice and later to show that members of any society had reason to accept certain principles of justice. He notes at the end of A Theory of Justice that âproof is not justificationâ since âproofs simply display logical relations between propositionsâ but justifications work only âonce the starting points are mutually recognizedâ (TJ 581). The thought that principles of justice must be justified to those who will be bound by them is present throughout Rawlsâs work on justice, but it takes different forms in A Theory of Justice and in his subsequent work.
In A Theory of Justice the first step that Rawls takes is to distance reasonable agreement on principles of justice from ordinary, real-world agreements and disagreements by setting out a distinctively characterized Original Position. Agreements in ordinary life reflect differentials in power and knowledge, so they donât offer a good basis for identifying principles of justice. Rawls seeks to correct for this problem by arguing that specific principles of justice would be chosen by rational agents4 in a hypothetical Original Position in which âno one knows his place in society, his class position, or social status, nor does anyone know his fortune in the distribution of natural assets and abilities, his intelligence, strength and the like ⌠their conceptions of the good or their special psychological propensitiesâ (TJ 12).
Despite Rawlsâs description of this approach as âcontractarian,â the term cannot be taken literally. Any claim that parties in the original position âagreeâ or âcontract,â or even that they endorse a hypothetical âagreementâ or âcontract,â is highly metaphorical because there is nothing to differentiate one person from another in the Original Position. Rawls points out that âwe can view the choice in the original position from the standpoint of one person selected at randomâ (TJ 139).5 No actual persons take part in or count in the Original Position.
So this hypothetical âagreementâ cannot be understood as a solution to disagreements of the sort that in actual situations might be resolved by bargaining or contracting, since by hypothesis nothing differentiates the parties. There is nothing that they need to agree on, or that they can agree on, since the terms of the hypothetical âcontractâ are reached simply by obliterating sources of disagreement. The Original Position provides a way of generating or identifying principles of justice that articulate an ideal of fairness, but it is not used to justify them to those who (supposedly) are to be bound by them.
WHO COUNTS IN REFLECTIVE EQUILIBRIUM?
In A Theory of Justice Rawls deals with the limitations of the Original Position by embedding it in a conception of Reflective Equilibrium that specifically addresses questions of justification to actual people. The Original Position is a device for generating tentative principles of justice, but their justification to âus,â who are to live by or under them, is not that they would be chosen in a situation that we arguably should, but in fact may not, regard as fair, but that they are in reflective equilibrium with âourâ best considered judgments about justice. âWeâ confirm the claim that the principles hypothetically agreed in the Original Position are principles of justice because they are (supposedly) âthe same judgments about the basic structure of society that we now make intuitively and in which we have the greatest confidenceâ (TJ 19, cf. 579). Reflective Equilibrium rather than the Original Position is Rawlsâs fundamental strategy of justification in A Theory of Justice:
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 duly pruned and adjusted. This state of affairs I refer to as reflective equilibrium.
(TJ 20)
So justification in the end is not to be hypothetical, and we should (as Rawls was later to put it) view the Original Position merely as a âdevice of representationâ (PL 24ff.) or as an âartifice of reasonâ (PL 75). This awareness of the limitations of the Original Position was something Rawls held to throughout his discussion of justification in A Theory of Justice, where he notes that âMoral philosophy is Socratic: we may want to change our present considered judgments once their regulative principles are brought to lightâ (TJ 49, see also 578). The deep strategy of A Theory of Justice is coherentist, and the device of the Original Position is used to fix the content rather than the justification of principles of justice.
However, subordinating the Original Position to Reflective Equilibrium for purposes of justification comes at a price. This is not because coherentist justifications are invariably problematic. Many of Rawlsâs colleagues and contemporaries used coherentist strategies of justification for a variety of purposes. Chomsky looked for coherence between his proposed deep principles of grammar and intuitive judgments of grammaticalness made by native speakers (cf. TJ 47); Quine argued in many works that justification cannot be linear or foundationalist but must aim to show coherence between a plurality of claims. However, introducing coherentist justificatory strategies unavoidably raises questions of scope: whose considered judgments are to be brought into reflective equilibrium with principles hypothetically generated by the device of the Original Position? Who exactly counts in Reflective Equilibrium? In A Theory of Justice Rawls left his answer to this question open: he appeals to âourâ considered judgments but leaves it unclear whom he has in mind. In later works, including âJustice as Fairness: Political not Metaphysicalâ and Political Liberalism, he says more about who counts for justification.
WHO COUNTS IN POLITICAL LIBERALISM?
The view of justification that John Rawls introduced in âJustice as Fairness: Political not Metaphysicalâ and developed in Political Liberalism is specifically political and appeals to a distinctive conception of public reason.6 As Rawls argued from the mid-1980s onward, when it comes to justification, citizens count, and noncitizens do not. He summarizes his revised position in the claim that
Public reason is characteristic of a democratic people: it is the reason of its citizens, of those sharing the equal status of citizenship ⌠[it is] is public in three ways: as the reason of citizens as such, it is the reason of the public; its subject is the good of the public and matters of fundamental justice; and its nature and content is public, being given by the ideals and principles expressed by societyâs conception of political justice, and conducted open to view.
(PL 213, cf. 10)7
This conception of public reasoning takes specific, distinctive, and interconnected views of the context, the subject matter, and the procedures of public reasoning. I shall discuss each in turn.
The Context and Audiences of Public Reasoning
The context and audience of public reason, as understood in Political Liberalism, is once again a society conceived of as one among many. However, the societies in which such reasoning is seen as possible are now explicitly confined to bounded, liberal, and democratic political societies, each conceived of as a âcomplete and closed social system ⌠[where] entry into it is only by birth and exit from it is only by deathâ (PL 40â41).8 These strikingly strongâand markedly unrealisticârestrictions on the context of public reasoning raise several problems, in particular because they ignore reasoning undertaken in nondemocratic societies and reasoning that crosses the borders between different societies. Both of these themes gained some importance in Rawlsâs later work, but both were marginalized in Political Liberalism, which is entirely clear about who counts for purposes of political justification: fellow citizens count, and others do not.
The Nature and Content of Public Reasoning
The context of public reason that Rawls assumes in Political Liberalism has decisive implications for his claims about its nature and content. The thought that public reasoning takes place among fellow citizens in bounded societies is fruitful because those citizens can be taken to share certain fundamental commitments. Their shared civic commitments to liberalism, to democracy, and to the continued existence of their bounded society provide a basis for debating or justifying its specific political arrangements.
Yet the very fact that citizens are assumed to share these commitments may also hamper the justification of fundamental political arrangements. If commitments to liberalism, to democracy, and to the continued existence of a bounded society are taken as given, the answers to certain fundamental questions about justice, such the justice of secession or annexation, or the justice of liberalism or democracy, are prejudged. Even coherentist justifications must do more than reiterate commitments. The thought that public reasoning can take place only among fellow citizens may seem inadequate because it is excludes reasoning that crosses boundaries or engages with those who are not fellow citizens. It seems as if Rawls assumed more than is advisable if fundamental political arrangements are to be justified but also less than is needed for justifying political arrangements in a globalizing world.
However, Rawlsâs discussions of the nature and content of public reasoning in Political Liberalism do not focus on the implications of assuming political boundaries but on the importance of a quite different boundary between political and nonpolitical questions, in particular religious questions. At the end of âThe Idea of Public Reason Revisitedâ he wrote: âThroughout, I have been concerned with a torturing question in the contemporary world, namely: Can democracy and comprehensive doctrines, religious or non-religious be compatible? And if so how?â (CP 611).
Rawls was more worried by the threat to liberal democratic societies that can arise from within, when religious and cultural views threaten to eclipse the public domain, than he was by questions about the injustice created by external boundaries and the exclusions they maintain. He concludes that public reasoning among fellow citizens cannot resolve all questions on which they may disagree but only those on which their fundamental political commitments bear. Citizens in pluralistic, liberal democratic societies may hold a great variety of religious, ethical, and social views that remain impervious to processes of public reasoning. Consequently, persisting pluralism of outlooks and ideologies within each society is a natural outcome of the free public use of reason.9 In Political Liberalism Rawls therefore explicitly sets aside aspirations to justify any comprehensive moral position in favor of a strictly political view of what can be justified by public reason....