1. The Basic Argument

1. 1.1. Rawls tells us that his motivation for writing Political Liberalism was the realization that, as presented in A Theory of Justice, justice as fairness was a ‘comprehensive, or partially comprehensive’ doctrine (PL: xvi).

   The serious problem is this. A modern democratic society is characterized not simply by a plurality of comprehensive religious, philosophical and moral doctrines, but by a pluralism of incompatible yet reasonable comprehensive doctrines. No one of these doctrines is affirmed by citizens generally. Nor should one expect that in the foreseeable future one of them, or some other reasonable doctrine, will ever be affirmed by all, or nearly all, citizens. Political liberalism assumes that, for political purposes, a plurality of reasonable yet incompatible comprehensive doctrines is the normal result of the exercise of human reason within the framework of free institutions of a constitutional regime.

   (PL: xvi)
   Elsewhere Rawls goes so far as to claim that there exists a plurality of ‘perfectly reasonable’ comprehensive doctrines (PL: 24n). This reasonable pluralism of comprehensive views renders them unacceptable as bases for the justification of political power. ‘[P]olitical power is always coercive power backed by the government’s use of sanctions’ (PL: 136). Now, says Rawls, according to the ‘liberal principle of legitimacy’: ‘our exercise of political power is fully proper only when it is exercised in accordance with a constitution the essentials of which all citizens as free and equal may reasonably be expected to endorse in the light of principles and ideals acceptable to their common human reason’ (PL: 137). Thus, it would seem that because there exists a reasonable plurality of comprehensive doctrines, basing the justification of political power on any single doctrine – or subset of comprehensive doctrines – would violate the liberal principle of legitimacy.

This leads Rawls to seek a political conception that ‘all affirm’ (PL: 38) and is ‘shared by everyone’ (PL: xix). Such a conception would be supported by, or at least not conflict with, the diverse reasonable comprehensive doctrines that characterize our democratic societies. This political conception is a ‘module’ (PL: 12–13, 144–5)Footnote ² that fits into our many reasonable, yet irreconcilable, comprehensive views. And because this political conception can be affirmed by all reasonable comprehensive doctrines, oppressive state power is not required to uphold it (PL: 37). Justice as fairness is offered as such a political conception: ‘If justice as fairness were not expressly designed to gain the reasoned support of citizens who affirm reasonable although conflicting comprehensive doctrines … it would not be liberal’ (PL: 143).

1. 1.2. Let us make this argument more precise. Rawls argues:

   1. (1) The Principle of Liberal Legitimacy (LL): The exercise of political power is legitimate only if it accords with a constitution the essentials of which all free and equal citizens may reasonably be expected to endorse.

   2. (2) In our democratic societies, there exists a reasonable pluralism of comprehensive religious, philosophical and moral views.

   3. (3) If (i) free and equal citizen Alpha holds a reasonable comprehensive view C_α, and (ii) if citizen Beta’s reasonable comprehensive view C_β is ‘irreconcilable’ with C_α, then (iii) Alpha cannot reasonably be expected to endorse C_β.

   4. (4) If Alpha cannot reasonably be expected to endorse C_β, he cannot reasonably be expected to endorse a constitution whose justification requires endorsing C_β.

   5. (5) Therefore, a constitution relying on C_β as in step 4 violates LL (step 1).

   6. (6) Given step 2, for every reasonable comprehensive view C_x there exists another reasonable comprehensive view held by some free and equal citizen that is irreconcilable with it.

   7. (7) Therefore, there exists no constitution satisfying LL that requires the endorsement of any specific comprehensive view.

   8. (8) However there exists a political conception P such that there exists no reasonable comprehensive view C_x, where it is the case that C_x is irreconcilable with P.

   9. (9) Given step 8, a constitution relying on P for the justification of political power does not violate LL.

   Step (3) is necessary. Fundamental to Rawls’s political liberalism is the (uncontentious) claim that it is unreasonable to expect a person to endorse an otherwise reasonable comprehensive view that is irreconcilable with his own reasonable view. Step (8) also seems necessary; unless (8) holds, a constitution depending on P is open to the same objection as a constitution depending on a comprehensive view. Interestingly, Rawls sometimes qualifies the claim in (8). He tells us, for example, that consensus on the political conception should include ‘all the reasonable opposing religious, philosophical, and moral doctrines likely to persist over generations and to gain a sizable body of adherents’ (PL: 15).Footnote ³ This suggests an alternative to step 8:

   1. (8*) However there exists a political conception P such that there exist few reasonable comprehensive views C_x, where it is the case that C_x is irreconcilable with P.

   As we will see, whether we adopt (8) or (8*) has important implications for Rawls’s political liberalism.

2. Comprehensive Views and the Political Conception

1. 2.1. Rawls’s core argument – and so his entire political liberalism – apparently depends on the contrast between comprehensive views and the political conception. Rawls repeatedly describes as ‘comprehensive’, ‘philosophical’, ‘moral’, and ‘religious’ ‘doctrines’ (PL: xxv, 4, 36, 38, 160) or ‘beliefs’ (PL: 63). Indeed, so often does Rawls characterize comprehensiveness in terms of moral, religious and philosophical doctrines or beliefs that a reader may be tempted to conclude:

   C is comprehensive if and only if it is a moral, religious or philosophical doctrine or belief.

   This would make sense of Rawls’s insistence that ‘political liberalism applies the principle of toleration to philosophy itself’ (PL: 10). Just as a traditional liberal political order tolerates a variety of religious views and does not invoke any in the justification of laws and policies, Rawls apparently seeks to tolerate all reasonable philosophical and moral doctrines while abjuring appeal to any in the justification of constitutional essentials. But though tempting, this would be wrong. Rawls is clear that:

   the distinction between the political conception and other moral conceptions is a matter of scope; that is, the range of subjects, to which a conception applies and the content a wider range requires. A moral conception is general if it applies to a wide range of subjects, and in the limit to all subjects universally. It is comprehensive when it includes conceptions of what is of value in human life, and ideals of personal character, as well as ideals of familial and associational relationships, and much else that is to inform our conduct, and in the limit to our life as a whole. A conception is fully comprehensive if it covers all recognized values and virtues within one rather precisely articulated system; whereas a conception is only partially comprehensive when it comprises a number of, but by no means all, nonpolitical values and virtues and is rather loosely articulated. Many religious and philosophical doctrines aspire to be both general and comprehensive.

   (PL: 13)
   Comprehensive and general doctrines cover a wide range of topics, values and ideals applicable to various areas of life while, in contrast, the scope of the political is narrow.

However, Rawls tells us later on that ‘[m]ost people’s religious, philosophical, and moral doctrines are not seen by them as fully general and comprehensive, and these aspects admit of variations and of degree’ (PL: 160). It is not certain whether Rawls believes that most people are correct in this self-conception, but it seems clear that they must be. Few people have all-embracing philosophies of life that provide a single, coherent perspective on questions of value, human character and social life.Footnote ⁴ So most people do not actually possess, and so cannot rely on, (fully) comprehensive moral, religious and philosophical doctrines.

1. 2.2. Once we recognize that Rawls’s argument is against comprehensive views qua systems of thought that are wide in scope and rich in content, ranging over many areas of life, we see that the argument does not exclude moral, religious or philosophical beliefs – as opposed to comprehensive views or general theories – from serving as the basis for an exercise of political power that passes the test of Liberal Legitimacy. That a belief is moral, religious or philosophical does not itself show that it is comprehensive or general. Indeed, Rawls himself indicates that the political conception has moral, epistemological and metaphysical elements (PL: 10, 11, 13, 62). Moral, religious and philosophical beliefs need not be, and very often are not, comprehensive or general. They may cover a limited topic and stand alone. Consider, for example, the beliefs that: (A) God exists; (B) the external world is real, and a proposition is true if it describes the world accurately; (C) children should respect their parents; (D) stealing is wrong; (E) people who work hard deserve more than those who do not.

For each of these beliefs, it could be argued that: (1) it is narrow in scope, and is embraced by a wide variety of people with different comprehensive views and (2) it is the object of a wide, though certainly not complete, consensus in the United States. (A), for example, meets the test of (8*). Belief (B) is interesting; it is philosophical, and is debated by philosophers, but there is probably much more consensus on it than on any political claim that Rawls makes. Beliefs (C)–(E) are all very widely shared moral beliefs, though it seems that Rawls is one of the few who denies (E).Footnote ⁵

It might be objected that, although (A)–(E) are simply beliefs and not comprehensive doctrines, they are inevitably embedded in some comprehensive doctrine; thus appeal to them necessarily brings some comprehensive doctrine into play. In reply, though, we can first query whether this is so. Many people have abstract and isolated intuitions that there is a deity, that true statements in some way correspond to a real external world or that certain actions simply are morally right or wrong. But second, Rawls’s doctrine of an overlapping consensus (see §3.1) relies on comprehensive doctrines supporting a shared political conception. So that a belief such as (A) is endorsed by a wide variety of religious and philosophical views hardly constitutes an objection to advancing it in the justification of constitutional essentials: it shows that it is an object of an overlapping consensus.

This may seem to be making too much of a small point: all Rawls has to do is be clear that the test of Liberal Legitimacy requires the rigorous step (8) rather than the looser (8*). For while (A)–(E) may be endorsed by the huge majority of reasonable citizens, each is rejected by some reasonable citizen. However, we still see from all this that it is not comprehensive doctrines that Rawls must object to, but simply appeal (in the justification of constitutional essentials) to any reasonable belief b when it is the case that some citizen entertains a reasonable belief that is irreconcilable with b. The relevant distinction is not between the comprehensive and the political, but the reasonably disputed and the not reasonably disputed. Thus, it would appear that we should reformulate step (3) as:

But this is still not quite correct. As we have seen, Rawls himself apparently admits that most people do not possess fully comprehensive doctrines. Suppose Alpha is one of these citizens whose views do not hang together into a highly coherent scheme; but suppose that, while he does not obtain the integration of a fully comprehensive view, he still has various reasonable beliefs, and b is inconsistent with them. His system of beliefs is ‘partially comprehensive’ insofar as it does form some sort of system, but it has no single or few leading idea(s), and a number of issues are not covered. It still would seem unreasonable to expect him to endorse b. Thus, we have:

Some might object that downgrading the source of Alpha’s objection to b from that premised on a comprehensive view of life to a mere reasonable belief undermines the conviction that it is wrong to expect Alpha to endorse b. But it is hard to see why this should be so. Alpha is not a schizoid personality without an integrated belief structure; he has obtained sufficient integration such that, having considered his other beliefs and values, he has come to a reasonable belief that b_α. To disregard this still seems to ignore his status as a free and equal person. Recall, moreover, that Rawls does not think that most people possess fully comprehensive doctrines (§2.2): to insist that only fully comprehensive doctrines are the grounds for genuine complaints based on LL implies that most citizens are precluded from appealing to LL. Surely this would be an unwelcome result for a political liberal.

1. 2.3. Having freed ourselves of the confusing and distracting reference to comprehensive doctrines, Rawls’s claim must be that only political beliefs pass the implied test of (3*), and so can satisfy LL. But now it is hard to see how this can be explained simply as a difference in scope; although the political may well have a more limited scope than a religious, philosophical or moral fully comprehensive doctrine, it is implausible to suppose that, inherently, any religious or moral belief is of greater scope than any political belief. Why should we believe that?

We have seen that sometimes Rawls indicates (1) that the political conception is a ‘module’ that fits into many comprehensive views (PL: 12–13, 144–5); the implied contrast might be to (2) a religious or moral belief that is inherently non-modular, necessarily presupposing an entire comprehensive doctrine. However, we already rejected the second claim (§2.2). Regarding (1) I have argued elsewhere that conceptions of the political can themselves be deeply influenced by one’s moral, religious and philosophical beliefs and so are themselves open to dispute.Footnote ⁶ Employing the familiar distinction between a concept and a conception,Footnote ⁷ it seems that the concept of the political perspective is characterized by competing conceptions. A Marxist has very different ideas about the nature of politics from that of a political liberal, or a libertarian or a Fundamentalist. It is important to stress that this is not the banal point that the Marxist, political liberal, libertarian, and Fundamentalist all have different political proposals - it is the more interesting point that they entertain different conceptions of the political. As Habermas notes, ‘the boundaries between public and private’ [by which he means the political and non-political] are ‘historically shifting’ and ‘in flux’Footnote ⁸ – indeed, they have been one of the main sources of dispute between different ‘comprehensive’ theories of theories of self and society – i.e. what we often call ‘political theories’. The concept of politics, it has been argued, is ‘essentially contested’, being composed of a number of dimensions that can be ordered differently, producing different conceptions, each of which is a reasonable interpretation of the concept.Footnote ⁹

Perhaps I have misunderstood Rawls. Like Habermas, I have read Rawls as positing a logically basic (in Habermas’s words, an ‘a priori’) distinction between the political and non-political spheres.Footnote ¹⁰ And at times Rawls certainly suggests this. For example, in the above quotation (§2.1) Rawls explains the difference in scope between comprehensive doctrines and the political conception by pointing out that ‘[a] conception is fully comprehensive if it covers all recognized values and virtues within one rather precisely articulated system; whereas a conception is only partially comprehensive when it comprises a number of, but by no means all, nonpolitical values and virtues and is rather loosely articulated’. The italicized phrase suggests that the distinction between the political and the non-political is logically prior to the distinction between the comprehensive and the political: Rawls explicates a comprehensive doctrine as one that appeals to nonpolitical values. Elsewhere Rawls seems to posit a basic distinction between the ‘political’ and the ‘social’, the latter being the realm of ‘comprehensive doctrines of all kinds – religious, philosophical, moral’ (PL: 14). It would seem that on this view we need to know what are the boundaries of the political before we can know whether a doctrine is comprehensive. Habermas would thus seem correct that Rawls’s political liberalism relies on a logically basic contrast between the concepts of the political and non-political (or social), one that on reflection seems dubious and contentious. That the political is focused on the justice of the basic structure (PL: Lecture VIII) is a reasonable – perhaps the correct – view, but it is by no means an uncontentious conception of the political endorsed by all reasonable citizens.

An alternative, however, is to understand the political conception as constructed out of that which we share. On this reading the nonpolitical is, by definition, those matters on which our use of reason leads us to different, reasonable conclusions. It is, by its very nature, the realm of reasonable pluralism. In contrast, we can define the political as those matters on which human reason converges, and so necessarily generates constitutional principles that satisfy the Principle of Liberal Legitimacy.Footnote ¹¹

This approach to identifying the political avoids the problems I have thus far been canvassing. It does not rely on the ideas of a ‘comprehensive doctrine’, or a priori or uncontroversial (within the limits of reasonability) notions of what is inherently political, moral, philosophical or religious. To be sure, it implies that what prima facie appears to be a moral belief can end up part of the political, but Rawls expressly allows this: the political conception, he tells us, ‘is a moral conception worked out for a specific kind of subject, namely, for political, social, and economic institutions’ (PL: 11).Footnote ¹² Let us explore, then, this constructed understanding of the domain of the political.

3. The Political: Where Human Reason Agrees?

1. 3.1. Fred D’Agostino has identified two ways in which human reason can agree: consensus and convergence.Footnote ¹³ A consensus argument seeks to show that everyone has reason R to accept b. Such an argument seeks to show that we share a reason for endorsing b. In contrast, a convergence argument seeks to show that we have different reasons for endorsing b, though we all have some reason for endorsing it. Rawls employs both types of arguments in his case for agreement on the political.

Rawls argues that the political conception can be justified as free-standing (PL: 10): it is based on a conception of persons as reasonable and rational, free and equal – a conception that is said to be implicit in our democratic society, and so shared by all. Justice as fairness thus expresses a ‘shared reason’ (PL: 9). Rawls thus argues that justice as fairness is a justified political conception as it articulates the requirements of the concepts of the person and society that all reasonable citizens in our democratic societies share. However, Rawls does not believe that this exhausts justification. In later stages of justification – what he refers to as ‘full’ and ‘public’ justification’ – citizens draw on their full range of beliefs and values and find further reasons for endorsing the political conception (RH: 142–3). Thus ‘overlapping consensus’ constitutes a convergent public justification, drawing on our various ‘comprehensive doctrines’.

1. 3.2. The key idea, then, is that the political conception exemplifies a consensus and convergence of the powers of our reasoning. The content of this political conception is ‘broadly liberal in character’:

   By this I mean three things: first, it specifies certain basic rights, liberties, and opportunities (of the kind familiar from constitutional democratic regimes); second, it assigns a special priority to these rights, liberties and opportunities, especially with respect to claims of the general good and of perfectionist values; and third, it affirms measures assuring all citizens adequate all-purpose means to make effective use of their basic liberties and opportunities.

   (PL: 223)
   Justice as fairness, we are told, is simply one such liberal conception; because ‘each of these elements can be seen in many different ways, so there are many liberalisms’ (PL: 223). This is significant: Rawls believes that there are diverse interpretations of the basic concept of a liberal political order. Indeed, he insists that ‘it is inevitable and often desirable that citizens have different views as to the most appropriate political conception; for the public culture is bound to contain different fundamental ideas that can be developed in different ways’ (PL: 227).Footnote ¹⁴ Rawls also accepts that citizens arguing in good faith and employing public reason will not accept ‘the very same principles of justice’ (PL: 214).

Now in many ways this is puzzling. If citizens entertain ‘different views as to the most appropriate political conception’ a society cannot be well-ordered. In a well-ordered society ‘everyone accepts, and knows everyone accepts, the very same principles of justice’ (PL: 35). It thus seems that reason will not itself lead us to a well-ordered society; the reasoning of free and equal citizens may lead them to all accept the liberal concept of justice, but will not lead them to all embrace justice as fairness, the ‘very same principles of justice’ or the same views of constitutional essentials. And no evidence indicates that Rawls believes disputes about the favored political conception and the principles of justice are a sign that some citizens are either irrational or unreasonable. Indeed, it seems an instance of the ‘burdens of judgment’, which was originally introduced to show why we disagree about moral, philosophical and religious matters. In his account of the burdens of judgment, Rawls stresses the complexity of value disputes, and the different way of ordering and weighing values (PL: 54–8). It seems that it is precisely this complexity in ordering and weighing ‘political values,’ and the complexity of developing democratic ideals, that leads to competing reasonable political conceptions. Reasonable pluralism does, after all, apply to political conceptions. At only the most abstract level – the level of the very concept of a liberal order – does Rawls indicate that the exercise of the powers of human reason produces agreement. At more specific levels – and by ‘specific’ here I mean something as abstract as justice as fairness – our use of reason leads to reasonable disagreement.Footnote ¹⁵

1. 3.3. Now it may be thought that Rawls’s claim is that: (1) justice as fairness is a reasonable liberal political conception of justice and (2) citizens living under it will tend to develop allegiance to it, and thus (3) a society ruled by justice as fairness will move toward being well-ordered as citizens come to see that it coheres with their moral, religious and philosophical views. Thus, we might interpret Rawls as saying that, while reason alone does not now produce consensus on his favored liberal political conception of justice, the long-run tendency of a society living under justice as fairness is to converge on it. As Rawls stresses, there is a path to an overlapping consensus on justice as fairness that, through a series of steps, leads to a well-ordered society (PL: 158–68).

However, before a society converges on justice as fairness, there will be some period in which free and equal citizens exercising their reason will disagree whether justice as fairness is the favored political conception. Indeed, it is very difficult to believe that this period will not extend indefinitely. That the reasoning of free and equal people will someday lead everyone to accept justice as fairness seems, at best, a controversial prediction. During this period – however long it lasts – the exercise of political power on the basis of a constitution justified by appeal to justice as fairness violates the criterion of Liberal Legitimacy. Recall (from §§1.2, 2.2):

Now suppose citizen Alpha believes that the most reasonable liberal political conception of justice enshrines private property, allows for a social provision of a minimum income (with no further provision of equality), and seeks to award people differentially on the grounds of economic desert. We can assume that this reasonable articulation of the liberal concept of justice – which is in fact a popular oneFootnote ¹⁶ – departs in important ways from justice as fairness. Assume further that the majority accepts justice as fairness as the favored political conception; on that basis, they adopt a constitution that allows socialism, and the legislature proceeds to institute a market socialist regime. Citizen Alpha, however, has a reasonable belief that private property ought to be protected by a just constitution; it is thus unreasonable to expect him to endorse a constitution that allows socialism. Consequently, if we demand allegiance to the Principle of Liberal Legitimacy, a society cannot start on the path to being well-ordered under justice as fairness.

1. 3.4. It might be replied in Rawls’s defense that, while there may be reasonable disagreement as to whether justice as fairness is the favored political conception, rational and reasonable citizens can reach consensus on constitutional essentials. As Rawls notes, Kurt Baier suggests that Americans already have broad consensus on these matters (PL: 149).Footnote ¹⁷ The Principle of Liberal Legitimacy does not require consensus on a conception of justice, but only on constitutional essentials. Moreover, in Rawls’s steps to a well-ordered society based on justice as fairness, a constitutional consensus is prior to an overlapping consensus based on justice as fairness. So, as long as rational and reasonable free and equal citizens endorse the same constitutional essentials, the Principle of Liberal Legitimacy is satisfied.

Rawls, however, explicitly tells us that reasonable and rational free and equal citizens disagree about constitutional essentials. ‘A vote can be held on a fundamental [constitutional] question as on any other; and if the question is debated by appeal to political values and citizens vote their sincere opinion, the ideal is sustained’ (PL: 241). Public reason, Rawls tells us, rarely leads to close agreement, even on matters of constitutional essentials and basic justice (PL: 241). And, again, this is essentially because of what we might call the ‘burdens of political judgment’ (see §3.2). The political values relevant to constitutional essentials are multiple and complex, and so free and equal citizens exercising their powers of practical rationality and reasonability come to good-faith different answers about their proper weighing, leading to diverging views of justified constitutional essentials. Even in the political – in this case, constitutional – sphere reasonable pluralism manifests itself. If citizen Alpha has a reasonable belief that clause x is essential to a just constitution, then employing political power under a constitution that contains y, where y is irreconcilable with x, violates the Principle of Liberal Legitimacy. And, as a matter of fact, such debates occur in constitutional deliberations in the United States. Some American liberals insist that constitutional clauses upholding freedom of contract and preventing the taking of private property are constitutional essentials (that have been ignored); others follow Rawls in insisting that the protection of extensive private property rights is not a constitutional essential.Footnote ¹⁸

To be sure, here too Rawls believes that a series of steps can lead to a constitutional consensus (PL: 158–64). But, as we saw above, it must be true that before there is such a consensus the exercise of political power is illegitimate. As long as the Principle of Liberal Legitimacy is honored, the process cannot get under way. If we view LL as a side-constraint on the exercise of political power, it not only blocks a constitution premised on ‘comprehensive’ doctrines, it blocks justice as fairness – and indeed any specific liberal conception – as well.

1. 3.5. Rawls is aware of these problems; his response is to implicitly weaken the Principle of Liberal Legitimacy to allow for such reasonable political pluralism. Two alterations are important.

   1. (1) Rawls seems to exploit an ambiguity between strong and weak senses of what it is ‘reasonable’ to endorse. Rawls often tells us that the ‘political conception is a reasonable expression of the political values of public reason and justice between citizens seen as free and equal’ (PL: 247; see also xx, 243, 246, 253) or gives ‘reasonable’ answers to questions about how to weigh political values and constitutional essentials (PL: 225). We are also told that citizens may ‘reasonably accept’ the terms of cooperation specified by the political conception (PL: 16). Now this is much less than Rawls required of ‘comprehensive doctrines.’ Recall our initial step (3):

      If (i) free and equal citizen Alpha holds a reasonable comprehensive view C_α, and (ii) if citizen Beta’s reasonable comprehensive view C_β is ‘irreconcilable’ with C_α, then (iii) Alpha cannot reasonably be expected to endorse C_β.

      The political parallel would be:

      If (i) free and equal citizen Alpha holds a reasonable political view of constitutional essentials, P, and (ii) if citizen Beta’s reasonable political view P_β is ‘irreconcilable’ with P_α, then (iii) Alpha cannot reasonably be expected to endorse P_β.

      If we accept this, the claim that P_β is ‘reasonable’ or gives ‘reasonable answers’ is in no way sufficient to show that it passes the test of Liberal Legitimacy; simply put, on this stronger criterion, it is not reasonable to expect a citizen to endorse a doctrine just because it is a reasonable doctrine. Indeed the whole problem of reasonable pluralism is that there are numerous reasonable views that are irreconcilable with other reasonable views; political liberalism’s search for consensus in the domain of the political was intended as a response to this very problem: to ‘resolve the impasse in our recent political history … that there is no agreement on the way basic social institutions should be arranged if they are to conform to the freedom and equality of citizens as persons’ (PL: 300). Yet Rawls often seems content to rely on the claim that his favored political conception is simply a reasonable political view. But that would imply the following Principle of Weak Liberal Legitimacy:

      WLL: The exercise of political power is legitimate if it accords with a constitution the essentials of which all free and equal citizens can see as reasonable.

      Weak Liberal Legitimacy suggests that it is reasonable to expect a citizen to endorse a political view just because it is a reasonable political view. This, though, would allow various reasonable comprehensive views – such as Mill’s liberalism – as bases for the legitimate exercise of political power.

   2. (2) More formally, Rawls tells us:

      Let us say that we honor public reason and its principle of legitimacy when three conditions are satisfied: (a) we give very great and normally overriding weight to the idea it prescribes; (b) we believe public reason is suitably complete, that is, for at least the great majority of fundamental questions, possibly for all, some combination and balance of political values alone reasonably shows the best answer; and finally, (c) we believe that the particular view we propose, and the law or policy based thereon, expresses a reasonable combination and balance of those values.

      (PL: 241)Footnote ¹⁹

Clause (b) is more demanding than the test implied in WLL. It requires us to assume that there is a uniquely reasonable best answer to the political question. Surprisingly, however, clause (c) does not require a person to believe that the answer she is proposing is the best answer supposed in (b), only that it is a reasonable answer (thus reverting to the view I considered in 1 above). It is not clear that this is simply an oversight. Rawls tells us that the answer provided by public reason ‘must at least be reasonable, if not the most reasonable’ (PL: 246). Again, this suggests that any reasonable answer is sufficient. However, even supposing that one interprets (c) to require a good-faith belief that one’s reasonable answer is the uniquely most reasonable answer, if we apply this criterion to ‘comprehensive doctrines’, it would allow, say, a Millian to advocate a perfectionist constitution. For the Millian could claim: (1) that he believes that there is a uniquely best answer to this question (the Millian one), though of course there are other reasonable views too, and (2) he is advocating the doctrine which, in good faith, he believes is the uniquely most reasonable one. If we reduce the demands of Liberal Legitimacy to requiring simply a good faith belief that one’s reasonable view is the best or most reasonable one, constitutions relying on ‘comprehensive’ doctrines are legitimate.

Rawls, however, insists that the answer provided by public reason ‘must at least be reasonable, if not the most reasonable, as judged by public reason alone’ (PL: 246).Footnote ²⁰ Thus Rawls can insist that in the case of a Millian constitution, its reasonability is not judged by public reason alone, but in reference to the Millian comprehensive view. This reply brings us full circle, for it supposes a basic (what Habermas called an ‘a priori’) contrast between the political and the non-political (§2.3). If we could distinguish in a reasonably uncontentious manner the properly political from the non-political or social, then we would be in a position to distinguish what is inconclusive on political grounds from what is non-politically inconclusive. However, having given up that attempt, we have been seeking to construct the notion of the political out of the reasons we can share. If both Millian liberalism and justice as fairness are reasonable views, that their adherents believe to be correct but which cannot be shown to be uniquely reasonable to others, we do not have the conceptual resources to say that one appeals to the properly political while the other does not.

1. 3.6. We now can see the dilemma of political liberalism. If Rawls could identify a uniquely reasonable political point of view – one which manifestly excluded Millian and other ‘comprehensive’ liberalisms as reasonable political doctrines – he could identify a realm of reasonable though conflicting political opinions that was restricted to a small family of political conceptions, one of which would be justice as fairness. Thus, the idea of a basic contrast between ‘comprehensive’ and ‘political’ doctrines. We have seen, though, that this basic contrast cannot be maintained. We have reasonable differences about what is properly political, while many beliefs associated with comprehensive doctrines are widely shared. The alternative, then, is to abandon any logically basic contrast between the political and the social, and to instead construct the notion of the political out of the reasons we share. But since the use of human reason leads us to reasonable disagreement about conceptions of justice and constitutional essentials, the political qua shared is limited to the abstract concept of a liberal political order.

1. 3.7. I have been stressing here the way in which reasonable pluralism characterizes the thinking of free and equal citizens on matters of basic justice and constitutional essentials in Rawls’s political liberalism. As such, I have argued, political liberalism fails to satisfy its own principle of legitimacy, and so is self-defeating. It should be stressed that LL is itself a weak – and in my view much too weak – principle of legitimacy. As long as an exercise of political power is consistent with a constitution the essentials of which all reasonable citizens can be expected to endorse, this power is justified. As Rawls sees it, this allows ‘citizens and legislators’ to ‘properly vote their more comprehensive views when constitutional essentials and basic justice are not at stake’ (PL: 235, but cf. 252).Footnote ²¹ Suppose, then, a matter is before us that does not involve a constitutional essential, say, whether we should have a government-provided education system or one which is government funded but which is provided by private schools (e.g. through vouchers). Now suppose that the main argument given by those supporting government provision is that a single, government-run system will be better able to ensure that citizens are raised to endorse certain controversial views: overall, the system will be more favorable to, say, multiculturalism and environmentalism. It will not go so far as to repress competing views, for it will be careful to remain within the bounds of the basic liberal constitution; but within those bounds the majority explicitly advocates the use of state power to uphold its own comprehensive views.

It is hard not to see this as an illiberal and oppressive policy. Some citizens are to be subjected to coercive state enactments that are designed to further (reasonable) doctrines that are irreconcilable with their own reasonable views. It is hard to see why any citizen should reasonably be expected to accept a coercively-imposed law when this law has been justified by appeal to comprehensive doctrines that the citizen reasonably opposes. To be sure, a Rawlsian state will not be grossly oppressive, as it must respect the publicly justified essentials; it does, though, allow many small coercive impositions that are explicitly justified on what seem manifestly non-public grounds. In its day-to-day operations, political liberalism sanctions the majority’s use of state power to advance its ‘comprehensive doctrines’.

We can see why Rawls is driven to this conception of politics. If the political is the realm of the respect for the freedom and equality of our fellow citizens because it manifests our agreement, the political obviously cannot be instantiated in day-to-day politics, which is, first and foremost, about the ways in which we differ. Hence Rawls must accept a radical dualistic conception of politics, sharply distinguishing the constitutional, which (at least at times) Rawls depicts as a matter of shared reasoning, and the normal business of politics, which is about the ways in which we differ and which constitutes a hostile arena for the use of public reason.

4. Five Compelling Ideas

I have thus far been critical of Rawls’s political liberalism. It is important to stress, though, that its problems are important, as they stem from five compelling ideas.

1. (1) Respect for the freedom and equality of our fellow citizens requires that the state’s exercise of coercive authority must be justified to each and every citizen. This, of course, is the core idea of the Principle of Liberal Legitimacy (LL). Difficulties arise when we add the second compelling idea:

2. (2) The free exercise of human reason leads us to disagree on a wide variety of issues concerning value, goods, ideals of the good life and so on. One citizen’s reasonable views are often reasonably rejected by others. Consequently:

3. (3) Many of the beliefs we hold most dear are not available to us in our efforts to meet LL, as they are the subject of reasonable disagreement.

4. (4) If we are to meet LL, we must thus restrict the beliefs or considerations to which we appeal, restricting ourselves somehow to those reasons which we all share.

5. (5) However, as we have seen, even in politics our reason leads us to disagree. We cannot go very far by understanding the political as characterized by the absence of reasonable disagreement.

5. Reasonable Disagreement Reexamined

1. 5.1. To accommodate all five of these ideas, while still showing how a liberal regime can meet LL, we need to better grasp the idea of reasonable disagreement, and how it leads to problems in satisfying the Principle of Liberal Legitimacy. Rawls always assumes a close link between (1) being a reasonable person and (2) entertaining reasonable beliefs or doctrines (PL: 59). Rawls supposes that reasonable people have reasonable beliefs. If we grant this, and also accept that the Principle of Liberal Legitimacy requires that a sound justification of coercive state power must not be irreconcilable with any citizen’s reasonable beliefs (see (3*), §3.3), it is not hard to see why LL appears to require a consensus of reasonable people on constitutional essentials. Any reasonable person’s veto apparently demonstrates that she has a reasonable belief that is irreconcilable with the justification, hence the justification does not meet LL.

I have tried to show elsewhere that reasonable people often have unreasonable beliefs.Footnote ²² A number of empirical studies indicate that intelligent, normal reasoners who manifestly qualify as reasonable people – they are, in general, disposed to reasonable beliefs – can entertain strikingly unreasonable (indeed, irrational) beliefs in specific cases. Reasonable people, for example, very often refuse to believe what is manifestly highly credible, instead persevering in irrational beliefs. P. C. Wason and P. N. Johnson-Laird report an experiment in which subjects were given three numbers, and were told that the experimenter had in a mind a rule for the generation of a series of numbers, and that these three numbers fit the rule. The task of the subjects was to ‘test hypotheses’ about what the rule might be by asking the experimenter questions of the form ‘Do numbers n, m, o fit the series?’ where the subjects themselves supplied various candidates. The aim was to test hypotheses about what the rule might be; when the subjects felt sure they knew what the rule was, they were to announce it to the experimenters. Wason and Johnson-Laird found many subjects poor at such hypothesis testing, but of immediate interest are some of the responses of subjects once they discovered that their confident announcements about the rule regulating the series were wrong.

In some cases, subjects simply remained immune to the fact they were wrong, even to the point of simply insisting that they must be right: they refused to accept the manifestly credible statement that they had guessed wrong. As Wason and Johnson-Laird point out, the evidence showing they were wrong was clearly before them and indeed they obviously had good reason to accept it. And this is by no means an unusual reaction: the research of Deanna Kuhn and her colleagues found that at least half her subjects share an epistemic attitude, central to which is a personal certainty that their own beliefs are correct, even when they are not based on extensive knowledge.Footnote ²⁴ Not only is it the case that reflection often fails to induce reasonable people to abandon irrational beliefs, but reflection can induce them to abandon perfectly rational ones. For example, Stephen Stich and Richard Nisbett report that it is not difficult to teach people the gambler’s fallacy and induce them to jettison sound views about the probability of independent events.Footnote ²⁵

1. 5.2. Once we appreciate that the notions of reasonable people and reasonable beliefs diverge in this way, we can see that satisfying the Principle of Liberal Legitimacy should not lead us to seek the consensus of reasonable people, but to seek arguments that are not irreconcilable with reasonable beliefs. We seek a justification for the exercise of political power that is not open to reasonable objections, although it may be rejected by many reasonable people. Constitutional politics, then, is not the realm of consensus, but of conclusive justifications – those not open to reasonable doubt. The domain of constitutional politics is not to be grasped in terms of the quasi-sociological notion of a possible consensus, but in terms of an epistemology that reveals to us the conditions that render it unreasonable to reject a justification. A liberalism appreciating this is justificatory rather than political – it rests on a theory of justified, reasonable, belief and argument. This is crucial: the application of LL presupposes a criterion of reasonable belief. Moreover, this criterion cannot itself be defined as the conception of reasonable belief that all reasonable people would accept, or which can itself be the object of reasonable consensus; all these ideas are logically derivative of the epistemic criterion of reasonability.Footnote ²⁶

1. 5.3. For a justificatory liberalism, then, satisfying LL requires a justification of coercive authority that is not open to reasonable doubt. This leads us to Rawls’s third compelling idea: few of the beliefs around which we construct our lives achieve this level of justification. For the most part, our guiding beliefs are credible but not conclusively justified. Given the relevant evidence, adopting them is typically reasonable. However, our judgments on these matters are almost always inconclusive; many of our most cherished and important beliefs involve complex matters involving a number of values and considerations, and the evidence on which we make decisions is almost always incomplete; and even if it is not, our cognitive ability to process all the evidence and weigh the various considerations is inadequate to the task. Moreover, people’s belief systems differ: considerations that are salient and important in one person’s system may have little relevance in another’s: thus, even if I am sure that I have made the correct choice for me, it is dubious whether it is surely the correct choice for others. Consequently, because most of our beliefs are reasonably held, but can be reasonably rejected by others, arguments that require others to accept these beliefs are inadequate from the perspective of LL.

1. 5.4. Note that this analysis does not suppose any basic distinction between comprehensive doctrines and the political, or between, on the one hand, the moral, the religious, the philosophical and, on the other, the political. The basic contrast is between justifications that are open to reasonable objections (because they are irreconcilable with a reasonable, justified, belief) and those that are not. Any of one’s beliefs that are not open to reasonable doubt by others – they do not entertain justified beliefs incompatible with them – can be legitimate parts of a justification meeting the demands of LL. Although there is no a priori reason why such beliefs cannot be religious, it is certainly a modern liberal conviction that the history of religious political debate has shown that all candidates proposed – all the religious beliefs that have been advanced as beyond reasonable doubt – have been shown to be lacking. This has led liberals to conceive of religion as inherently personal and based on faith rather than public reason; in this way it has been eliminated as a basis for public justification. Pace Rawls, moral and philosophical beliefs are different from the religious.Footnote ²⁷ We cannot say in advance that a justification which necessarily relies on a philosophic doctrine is, just for that reason, inconsistent with LL. Insofar as philosophers seek to present public ‘knock-down’ arguments, they are endeavoring to live up to the ideal of public justification and LL. To be sure, philosophers and moral theorists may typically fail, but they are engaged in public reasoning, not articulating articles of faith or personal preferences, and as such are not to be equated with those seeking to justify the use of political power on their view of God and his commands.

6. Conclusive and Inconclusive Political Justification: The Moderate Dualism of Public Reason

1. 6.1. Are there any justifications of political authority that do not essentially depend on claims that are irreconcilable with reasonable beliefs held by some? Our problem, it will be recalled, has been set by the Principle of Liberal Legitimacy – LL is a supposition of the liberal analysis. Now although Rawls applies LL only to political institutions, there is no reason why its scope should be so restricted: just as a governmental use of power that cannot be justified to some citizens manifests disrespect for their freedom and equality, a coercive imposition by one person against another manifests the same disrespect. LL is a constraint on individuals as well as states. Liberals are thus committed to what we might call the Principle of Individualized Liberal Legitimacy:

   ILL: Alpha’s coercion against Beta is legitimate only if there exists a justification for it that Beta may reasonably be expected to endorse.

   Given, then, that all liberal individuals are committed to ILL, they have but two options: (1) to abjure coercive imposition on others or (2) arrive at public justified principles that sanction such imposition. The first is not a real option. To opt for (1) would be to unilaterally renounce what Hobbes called ‘the right of nature’ – to defend ourselves.Footnote ²⁸ Rational liberal citizens are thus committed to option (2): arriving at justified principles that sanction interference. It is because we all have a moral interest in arriving at such a justification that, as Rawls puts it, we all have reason to seek a mutual accommodation (PL: 253) and meeting others half-way is a virtue of civility. Given this, liberals have insisted, it would be unreasonable to reject coercion required to support a regime of extensive equal liberty that protects the person of each from invasion, and provides a structure for property rights.

1. 6.2. We seem to have arrived at exactly Rawls’s position. He too, it will be remembered, argued that reasonable citizens will concur on the basic concept of a liberal regime (§3.2). It was ‘only’ on more specific political questions that reasonable pluralism asserts itself. And it is certainly correct that citizens will disagree about the interpretation and application of basic liberal principles. If each relies on his own (private) judgment about the best interpretation of liberal principles – What are the bounds of our rights and freedoms? What coercive interferences do these principles justify? – reasonable pluralism produces deep disagreement. Indeed, for Kant, relying on one’s individual judgment in this way characterizes the state of nature:

   Although experience teaches us that men live in violence and are prone to fight one another before the advent of external compulsive legislation, it is not experience that makes public lawful coercion necessary. The necessity of public lawful coercion does not rest on a fact, but on an a priori Idea of reason, for, even if men to be ever so good natured and righteous before a public lawful state of society is established, individual men, nations and states can never be certain they are secure against violence from one another because each will have the right to do what seems just and good to him, entirely independently of the opinion of others.Footnote ²⁹

   Kant goes on to insist that justice is absent in the state of nature because each relies on his own judgment, and thus ‘when there is a controversy concerning rights (jus controversum), no competent judge can be found to render a decision having the force of law’.Footnote ³⁰ Indeed, Hobbes, Locke, and Kant all maintain that the chief inconveniences of the state of nature arise from individuals relying on their individual, controversial, judgments about natural rights and natural law.Footnote ³¹ The chief inconveniences are two, one moral and one practical.

The moral flaw of the state of nature ruled by individual judgment is that we act without justification. If Alpha believes that b_α is the best interpretation of liberal principles while Beta believes that b_β is, where b_α is irreconcilable with a reasonable belief b_β, Alpha’s acting on b_α violates our Principle of Individualized Liberal Legitimacy: he coercively imposes on Beta even though she has a reasonable belief that leads her to reject his justification. Leaving aside its moral shortcomings, a state of nature (i.e. a regime in which people all relied on their rationally contentious judgments about the demands of liberal justice) would be characterized by uncertainty and conflict, undermining the basis for cooperation. Inconsistent interpretations of each other’s rights and responsibilities would lead to conflict and thwart the development of settled expectations. This, of course, is a familiar theme in liberal, and especially contractualist, political philosophy: Hobbes’s, Locke’s and Kant’s accounts of the state of nature all aim to establish variations of it. Although on some matters we can agree to differ, disputes engendered by competing judgments about our rights and duties will block common action.Footnote ³²

1. 6.3. Relying on our reasonable but by no means conclusive judgments thus would lead to injustice and conflict. For Kant, if one ‘does not wish to renounce all concepts of justice’, one must ‘quit the state of nature, in which everyone follows his own judgments’ and subject oneself to ‘public lawful external coercion’.Footnote ³³ Hobbes, Locke, and Kant concur that an umpire or judge is required to make public determinations of our rights and duties. Says Hobbes:

   And because, though men be never so willing to observe these laws [of nature], there may nevertheless arise questions concerning a man’s actions; first, whether it were done, or not done; secondly, if done, whether against the law or not against the law; the former whereof, is called a question of fact; the latter a question of right, therefore unless the parties to the question, covenant mutually to stand to the sentence of another, they are as far from peace as ever. This other to whose sentence they submit is called an ARBITRATOR. And therefore, it is of the law of nature, that they that are at controversy, submit their right to the judgment of an arbitrator.Footnote ³⁴

   The social contract theories of Hobbes, Locke, and Kant, first and foremost, are justifications of an ‘arbitrator’ (Hobbes), ‘umpire’ (Locke) or ‘judge’ (Kant) whose task is to provide public, definitive, resolutions of conflicting, reasonable judgments about the demands of justice.

1. 6.4. According to Hobbes, Locke, and Kant, then, citizens committed to respecting each other and gaining the fruits of social cooperation require government just because of reasonable pluralism about the political. For Hobbes, Locke, and Kant our common human reason produces some conclusions that are beyond reasonable doubt; for Kant we all have reason to see that a regime of equal liberty is justified while for Locke it is beyond reasonable doubt that a just regime must respect and protect the rights to life, liberty and property. And, as I have said, they concur that reason instructs us to submit to an impartial judge to resolve our disputes about justice. Writes Kant:

   The postulate of public Law comes out of private Law in the state of nature. It says: If you are so situated as to be unavoidably side by side with others, you ought to abandon the state of nature and enter, with others, a juridical state of affairs, that is, a state of distributive legal justice.Footnote ³⁵

   Conclusive reasoning, I have tried to show elsewhere, can take us somewhat further, specifying the broad procedures identified with the rule of law and constitutional democracy.Footnote ³⁶ However, it will not take us much further. The realm of political is characterized by reasonable dispute because our judgments differ about the demands of justice and the interpretation of liberal principles and constitutional provisions.

The crucial task of government, then, is to serve as an umpire, judge or arbitrator, providing a practical resolution of our reasonable disputes about justice. Umpiring is based on the suppositions that (1) there is intractable difference of opinion that calls for different courses of action; (2) to proceed with practice, there must be a practical resolution of the dispute about what to do; (3) this practical resolution about what to do need not be accepted by all the parties as being based on the correct or most reasonable judgment; however, (4) the authority of the umpire’s practical decision requires that it does its best to arrive at the most reasonable answer. Citizens committed to ILL in a world of pluralism require precisely this sort of umpiring of their disputes. It honors their commitment to Individualized Liberal Legitimacy because they do not act coercively against another simply on the basis of their own controversial reasoning: all have conclusive reason to submit their dispute to the umpire, who provides an impartial practical resolution of the dispute.

The umpire’s legitimate decision is, then, simply a reasonable judgment. Umpires are not sages who we suppose always give the best answer. It is not at all inconsistent with accepting the authority of an umpire to insist that your opinion is more reasonable than his. Rather, umpires are unique in that they alone have the authority to use coercion to support a reasonable, though contentious, interpretation of liberal principles. Note, then, that justificatory liberalism provides a coherent account supporting Rawls’s observation that the judgment of public reason ‘must at least be reasonable, if not the most reasonable’ (PL: 246). Because of their intractable disputes about what is the most reasonable interpretation of publicly justified liberal principles, free and equal individuals would embrace an umpire who is empowered to act on its reasonable, but by no means conclusively correct, judgment about these matters.

This is important. Liberal legitimacy, Rawls and I have agreed, requires that coercion must be justified in a way that is not subject to reasonable objection; because of that the justification must be strongly reasonable – not subject to reasonable dissent or objection. But because of political pluralism, normal politics can, at its best, only claim to result in weakly reasonable conclusions; in Rawls’s words, they are ‘reasonable, if not the most reasonable’ (§3.5). As we have seen, citizens can reasonably dissent from weakly reasonable laws and policies. The idea of liberal adjudication explains why the government can satisfy ILL, even though its results are only weakly reasonable. Because (1) we require a common answer on questions of justice, (2) we have conclusive reason to embrace an umpire, (3) we thus have conclusive reason to follow the directives of the umpire even though they are only weakly reasonable. In short, there is a strongly reasonable justification to follow the weakly reasonable decisions of the umpire.

1. 6.5. Democracy, I have argued, can itself be understood as an umpiring mechanism.Footnote ³⁷ In his or her deliberations each citizen presents what he or she believes is the best public justification; the voting mechanism constitutes a publicly justified way to adjudicate our deep disagreements about what is publicly justified. It does not seek political consensus, but reasoned debate about what is best justified, and procedures that do a tolerable job in tracking justification. Adjudicative democracy recognizes that the political is required just because even rough consensus is not a plausible political ideal.

Under an adjudicative conception of democracy, however, citizens are not free to draw willy-nilly on their reasonable beliefs so long as they do not violate constitutional constraints and the rights of others. In contrast to Rawls, who restricts public reason to constitutional essentials and matters of basic justice, all legitimate political deliberation must be an exercise of public reasoning. When proposing coercive laws citizens must advance arguments that they believe are the best interpretation of basic liberal principles. The telos of politics is not to allow citizens the opportunity to impose their contentious ideals of life on each other, but to adjudicate our inconclusive reasoning about the demands of justice. Liberal citizens accept an umpire because they need to impartially resolve their disputes about justice if they are to honor their commitment to the Principle of Individualized Liberal Legitimacy; they have, as Kant stressed, conclusive reason to submit these disputes to an umpire. However, they have no reason to submit their disputes about religion, ways of having sex, the good life or good beers to adjudication: on these matters it is entirely reasonable to object that no public judgment is required, and so the use of state power – even in the service of reasonable views – is illegitimate.

We must distinguish limiting and empowering conceptions of constitutionalism.Footnote ³⁸ On the limiting conception of constitutional essentials, a constitution spells out limits to the government’s rightful authority (the model here is the United States Bill of Rights). A government is unjust if it enacts legislation that oversteps these bounds by, say, seeking to establish a religion or severely curtailing freedom of speech. On the limiting conception, however, as long as the government respects these constraints, its legislation is just: within the limits of the constitution, government (or, say, the majority) may do as it wishes. Justificatory liberalism and its allied notion of adjudicative democracy reject this conception of limited government. Instead, they advocate an empowering conception of constitutionalism: the constitution empowers government to act to interpret abstract, conclusively justified, principles of justice that we reasonably interpret differently. Unless these matters are adjudicated, we cannot live moralized lives meeting the requirement of Individualized Liberal Legitimacy. If an issue is not one on which the state is empowered to act, it is outside the bounds of political dispute.

7. Conclusion: Liberal Legitimacy, Reasonable Pluralism and the Domain of the Political

I have tried to show how, under Rawls’s political liberalism, the domain of the political is a response to the reasonable pluralism of comprehensive doctrines: Rawls wishes it to provide a common point of view that allows the Principle of Liberal Legitimacy to be satisfied. However, I have insisted that (1) the domain of the political is itself characterized by reasonable pluralism and (2) Rawls does not, and I believe cannot, show how this reasonable pluralism is a distinctive political sort of pluralism that does not run afoul of LL, and so is not a worry for political liberalism.

Rawls, I believe, is correct in that there is conclusive rational consensus on the basic concept of a liberal regime, and the broad outlines of liberal principles. But consensus only characterizes liberal politics at the most abstract level. As in many other complex matters, the use of human reason under free institutions leads us to disagree. The political cannot be insulated from this, nor can the realm of public reason be convincingly constricted so as not to be affected by it. The contract theories of Hobbes, Locke and Kant are not tempted to understand the political as essentially the realm of shared judgment. For them the state’s rationale is the reasonable pluralism of political judgment, and the need to provide practical resolution of our political differences. Because the pluralism of the political manifests our different interpretations of the demands of liberal justice, the political is constituted by a plurality of articulations of public reason, of citizens seeking to construct arguments to convince each other about the best interpretation of the demands of justice. As such, political pluralism is distinct from our wider pluralism of personal beliefs and values, which we have no compelling reason to submit to the authoritative judgment of the umpire.Footnote ³⁹