Rawls,reasonableness,and international toleration

1. Introduction

To what extent should liberal societies be tolerant of non-liberal societies in their foreign policy? What form should any such toleration take? Should they criticize or even invade liberal-rights-violating states in order to encourage the establishment of more liberal regimes there? Alternatively, are there ways to organize non-liberal societies so that they are, from a liberal point of view, worthy of toleration and even respect?

It is a familiar fact that Rawls’s The Law of Peoples gives what have seemed to many liberals to be rather disappointing answers to these urgent questions. Rawls advocates toleration for ‘decent peoples’ (non-liberal societies that meet certain minimal conditions) whose human rights records and aims may fall far short of those of liberal societies. Such toleration extends beyond mere non-intervention in decent peoples’ domestic affairs to ensuring that decent peoples can accept the principles that regulate liberal societies’ interactions with them. As it happens, the principles of the ‘Law of Peoples’ that Rawls advocates, and works out in the first instance by appeal to a choice situation involving only representatives of liberal peoples, turn out to be principles that could also be accepted in the relevant sense by decent peoples. So, Rawls concludes, the liberal Law of Peoples also expresses an appropriate degree of toleration for decent peoples. ¹

Some find these conclusions reprehensibly unambitious. Critics suspect, in particular, that it is not merely a happy coincidence that decent peoples could accept the principles of the Law of Peoples. Rawls, they believe, has diluted the principles of the Law of Peoples precisely in order to accommodate decent peoples. They see this as an unprincipled concession to unreasonableness. ²

In fact, as I will argue, there are two distinct objections underpinning these critics’ concerns. Both they and defenders of Rawls fail to distinguish these objections adequately. The first is that the choice situation involving liberal peoples to which Rawls appeals in the first instance to justify the Law of Peoples justifies, in fact, a stronger set of principles (and not those of the Law of Peoples). This stronger set of principles would justify greater criticism of or intervention in decent peoples’ domestic affairs. In that sense, then, a properly liberal law of peoples should be less tolerant of decent peoples than Rawls supposes.

The second objection is that it is inconsistent of Rawls, given his liberalism at the domestic level, to seek principles that are acceptable to decent (that is, illiberal) peoples. According to this objection, even if the correct liberal law of peoples turns out to be non-interventionist, for one reason or another, it is irrelevant to its justification whether any decent peoples accept it or not. In this second sense, then, the Rawlsian strategy of toleration is mistaken.

Prominent defenders of Rawls tend to focus on the first objection, arguing that Rawls’s choice-situation argument for the Law of Peoples does in fact justify what critics see as the excessive non-interventionism of that set of principles. But even if these defenders are right, that does not address the second objection. It may be that Rawls’s arguments do justify a non-interventionist law of peoples. But why, at least as a matter of ideal theory, should we also care whether or not decent peoples could accept it? Thus, even if the defenders of Rawls can rebut the first objection, the concern that Rawls’s view is too tolerant of illiberal societies remains in the form of the second.

I will argue that Rawls can be defended from this charge. My argument, in outline, is as follows. The explanation for Rawls’s efforts to show that decent peoples could accept the Law of Peoples is not, as many suppose, that decent peoples are the kind of people who ought to be tolerated in that way, on some prior conception of which kinds of people ought to be tolerated in that way. If that explanation were correct, then suspicions that Rawls dilutes fully liberal principles in order to ensure that they will be acceptable to decent peoples would be difficult to dispel. The explanation is, rather, that there is a fundamental methodological principle underlying Rawls’s approach to political justification (both domestically and internationally) according to which liberals owe justification, as a matter of liberal principle, to those who comply with liberal principles of justice that apply to them. Therefore, if such liberal principles can be complied with (by which I mean internalized and acted from) ³ by agents who nevertheless cannot accept fully liberal justifications for those principles, then liberalism itself requires liberals to seek justifications for them which these non-liberal compliers can accept.

Decent peoples are, on my reading, just those non-liberal societies who nevertheless comply with liberal principles for international political institutions that apply to them. So, if defenders of Rawls against the first objection are right that a fully liberal justification does not entail a greater degree of interventionism than Rawls’s Law of Peoples implies, so that decent peoples thus defined are possible, then according to the methodological principle decent peoples are owed a justification for the Law of Peoples that they can accept. This is the explanation for Rawls’s strategy of toleration. Unlike the rejected strategy above, it does not rely on an independent conception of decency to explain why decent peoples are owed toleration. Decency is, rather, a concept that is internal to liberal political justification at the international level.

My argument is structured as follows. In Section 2 below, I take a closer look at the strategy of toleration in The Law of Peoples and distinguish in more detail the two objections that I claim are implicit in prominent critiques of that strategy. In Section 3, I examine what I call ‘international reasonableness’: that characteristic which both liberal peoples and decent peoples share, in Rawls’s view, and which marks them out as worthy of toleration in the sense that they are owed justifications in terms that they can accept for the Law of Peoples. In the following sections, I argue for my understanding of international reasonableness (and so decency) as an ‘internal conception’ that is not independent of the content of the Law of Peoples. I do this by arguing for an analogous understanding of (domestic) reasonableness in Rawls’s theory of domestic justice, as advanced in Political Liberalism, in Sections 4 and 5, and then explaining, in Section 6, how the same approach underlies The Law of Peoples. I then reappraise The Law of Peoples, in Section 7, in light of this new understanding of international reasonableness. I conclude that Rawls’s approach to international justice is both more coherent and more attractive than is often supposed.

2. A closer look at Rawlsian toleration

According to Rawls, a liberal people is a territorially bounded society that has a ‘reasonably just constitutional democratic government that serves [its members’] interests’, ‘citizens united by what Mill called “common sympathies”’, and ‘a moral nature’ (Rawls, 1999a: 23). ⁴ A ‘reasonably just’ constitutional democratic government is understood here as one that accords with a reasonable liberal conception of justice such as Rawls’s own ‘justice as fairness’.

In Part I of The Law of Peoples, Rawls argues that the eight principles of the Law of Peoples are justified as principles for the regulation of the relations between different liberal peoples. The justification appeals to what I will call the ‘liberal peoples’ original position’ (LPOP). This is a hypothetical choice situation that ‘models what we would regard … as fair conditions under which the parties [in that choice situation, who are] the rational representatives of liberal peoples, are to specify the Law of Peoples, guided by appropriate reasons’ (Rawls, 1999a: 32). ‘We’ who would regard the conditions in question as fair are members of well-ordered liberal societies, which means that we accept liberal domestic political principles (Rawls, 1999a: 32–3). ⁵ As such, we would, according to Rawls, accept that the choice made by the parties in the LPOP should determine principles for the regulation of liberal peoples’ relations. What they would choose is the Law of Peoples (Rawls, 1999a: 39–44). When the Law of Peoples regulates liberal peoples’ relations, those liberal peoples are said to form the ‘Society of Peoples’ (Rawls, 1999a: 3).

This argument complete, Rawls turns his attention in Part II to the question of whether any non-liberal societies should be admitted to the Society of Peoples. Admission would entail a high degree of toleration of the new members’ ways of organizing themselves internally. Members of the Society of Peoples would be prohibited by the Law of Peoples itself from intervening by force in other non-liberal members’ domestic affairs, even if the latter implemented illiberal domestic policies, such as prohibiting homosexual acts or denying women the right to vote. ⁶ Moreover, political and economic sanctions in response to such policies would also be restricted, as would official incentives intended to persuade members to alter them, so long as those members continued to abide by the Law of Peoples (Rawls, 1999a: 59, 62, 84–5, 121–2). All members of the Society of Peoples, Rawls writes, are to be recognized as ‘equal participating members in good standing’ (1999a: 59).

The ideal of public reason operative among members of the Society of Peoples means, however, that appropriate conduct towards other members is not only a matter of this kind of non-interventionist toleration. It also requires that the Law of Peoples be justifiable to each member in terms that each member accepts – that each member be included in the constituency of justification of the Law of Peoples. So, if non-liberal societies were to be admitted to the Society of Peoples, the Law of Peoples would have to be shown to be justifiable to them, in this sense, as well as to the liberal peoples who are the Society’s original members. As Rawls says, the idea would not be ‘to prescribe principles of justice for them, but to assure ourselves that liberal principles of foreign policy are also reasonable from a decent nonliberal point of view’ (1999a: 58).

According to Rawls, then, admission to the Society of Peoples entitles a member to toleration in two senses. It entitles the member to toleration in the sense that other members are not to intervene in its domestic affairs either by force or via official sanctions or incentives (that is, sanctions or incentives set up at the level of the Society of Peoples). It also entitles the member to toleration in the sense that the Law of Peoples is to be not only applied but also justified to that member. Henceforth, I will reserve the term ‘toleration’ for the first of these senses alone – partly for clarity, but partly also because the second sense seems rather distant from an ordinary conception of toleration and is thus only confusingly so labelled.

Should any non-liberal societies be admitted to the Society of Peoples? Rawls argues that some should indeed, ostensibly because it is a given that there are ‘other acceptable ways … of ordering society’ besides liberal ones (1999a: 59). Societies that organize themselves in these other ways are worthy of the toleration and inclusion in the justificatory constituency that admittance to the Society of Peoples represents (Rawls, 1999a: 67). This suggests that Rawls has in mind some independent criterion that marks out certain non-liberal societies as worthy in this way, although this may be no more than an appeal to intuitive judgement. ⁷ Not to admit such societies (for liberal peoples to govern their relations with such societies according to different norms, less tolerant of or less justifiable to them than those regulating the Society of Peoples) would count, for this reason, against the justice of the global institutional order.

Rawls’s strategy, then, on what I will call the ‘standard reading’, seems to comprise the following two steps. First, describe and justify on fully liberal grounds an institutional order to regulate relations between liberal peoples. Second, argue that such an order should be tolerant of and justified to societies independently identifiable as worthy of such toleration and justification.

As we have seen, the first step involves the LPOP, in which representatives of liberal peoples, equally situated behind a veil of ignorance that deprives them of knowledge which might inappropriately bias their choice, affirm the principles of the Law of Peoples. ⁸ The second step, on this reading, comprises two sub-steps, which Rawls describes in Part II of The Law of Peoples. In the first, Rawls outlines the criterion, which I will call ‘international reasonableness’, which marks out certain societies as worthy of inclusion in the Society of Peoples. For a non-liberal society to be worthy of inclusion (for it to be internationally reasonable), it must qualify as a people and it must be what he calls ‘decent’ (I will say more about these below). Decent peoples, then, are internationally reasonable and so are to be admitted to the Society of Peoples. They are therefore owed the duty of justification that this implies. In the second sub-step, Rawls tries to discharge that duty by showing that the principles of the Law of Peoples would be acceptable to such societies. He does this by arguing that decent peoples would accept regulation by principles chosen in a second international original position in which representatives of decent peoples (and not liberal peoples) are themselves equally situated and deprived of knowledge of certain facts about their societies, and that the principles that would be chosen are those of the Law of Peoples. I will call this second international original position the ‘decent peoples’ original position’ (DPOP). ⁹

It turns out, then, that the liberal global order of the Law of Peoples is, fortuitously, acceptable not only to liberal peoples, but also to decent peoples, who are, on this reading, independently identifiable as worthy of inclusion in the constituency of justification (or ‘justificatory inclusion’ for short). Rawls presents this as a kind of discovery (see, for example, 1999a: 60, 69). Critics suspect, however, that the Law of Peoples does not in fact represent a happy coincidence between the principles that liberal peoples would, in appropriate circumstances, choose for the regulation of their own interactions and the principles that non-liberal peoples who happen to be independently identifiable as worthy of justificatory inclusion would, in appropriate circumstances, choose for the regulation of their interactions. They think that representatives of liberal peoples in the LPOP would choose more demandingly liberal principles than Rawls says they would. ¹⁰ They diagnose Rawls as watering down truly liberal global principles in order to pander to illiberal societies – as presenting, in effect, a modus vivendi as if it were a just peace. ¹¹ Moreover, they argue that it is incoherent for Rawls to do this, given his intolerance of illiberalism in his domestic theory of justice. ¹²

There are two strands in this critique that it is important to distinguish. The first strand focuses on the choice of principles in the LPOP. The critics’ concern here is with the first step in Rawls’s argument as I outlined it above. They argue that the principles of the Law of Peoples are not liberal enough, in the sense that they do not require those governed by them to be as liberal as the critics suppose they should be. ¹³

The second strand focuses not on the choice of principles, but on the idea that non-liberal peoples are worthy of the justificatory inclusion that is required by admission to the Society of Peoples. This, then, is a concern with the second step in Rawls’s argument – in particular, with the second step’s first sub-step. The critics argue that decent peoples, by virtue of the fact that they deny their members important liberal rights, show themselves to be unworthy of justificatory inclusion. ¹⁴

These two concerns are closely connected. Suppose that the principles of the Law of Peoples were as liberal in content as advocates of the first objection think they should be – requiring, for example, that all members of the Society of Peoples be organized domestically according to liberal principles such as those of justice as fairness. (I will say that in that case they would be ‘fully content-liberal’.) In that case, internally illiberal peoples would be necessarily excluded from the Society of Peoples for as long as they remained illiberal. There could be no question of ‘extending’ the Law of Peoples to them. The question that Rawls discusses in the second step of his argument (whether there are any non-liberal societies that should be admitted to the Society of Peoples, with the burden of justification to them that that implies) would be idle.

Although they are closely connected in this way, the two concerns should, on the standard reading, nevertheless be kept distinct. For they are independent of each other, and give rise to two distinct objections. You could object that the Law of Peoples is insufficiently content-liberal (perhaps you think its account of human rights is too minimal) and yet maintain that some non-liberal societies are worthy of justificatory inclusion. In that case, you might be an optimist, supposing that the more content-liberal account of human rights that you favour could in fact be shown to be acceptable to at least some illiberal societies. (To see this, imagine that the more content-liberal law of peoples that you envision is nevertheless not so liberal in content that only liberal societies could conform to it. That law of peoples might be acceptable to those non-liberal societies that could conform to it.) Or, instead, you might be a pessimist, supposing that there is a deep tension between international justice, requiring liberal human rights, and international legitimacy, requiring justifiability to both liberal and some non-liberal peoples.

Alternatively, you could at once object that Rawls is wrong to think that justificatory inclusion of some non-liberal peoples is necessary and accept that the principles of the Law of Peoples are sufficiently content-liberal. In that case, you would be in the position of those, discussed by Rawls in section 7.2 of The Law of Peoples, who say that ‘there is no need for the Law of Peoples to develop … an idea of toleration … citizens in a liberal society should judge other societies by how closely their ideals and institutions express and realize a reasonable liberal political conception’ (1999a: 60). In their dealings with non-liberal peoples, you might think, liberal peoples might adopt principles that are more or less content-liberal than those of the Law of Peoples or they might stick with the latter – either way, there would be no need to justify their conduct to non-liberal peoples.

One possible objection, then, focuses on the content of the Law of Peoples; the other focuses on the justificatory constituency for the Law of Peoples. Neither Rawls’s opponents nor his defenders distinguish these objections very clearly, principally (I believe) because all of the opponents argue that the parties in an appropriately constructed LPOP would choose fully content-liberal principles (requiring liberalism at the domestic level) for the law of peoples. The opponents recognize that for Rawls justificatory inclusion is something that is due only to those admitted to the Society of Peoples. But as I said above, if the law of peoples were fully content-liberal, then illiberal societies would necessarily be excluded from the Society of Peoples, so the second objection appears moot.

In fact, however, if the standard reading of Rawls’s argument that I just described is correct, then there is room to press the second objection independently of the first. This can be seen from the fact that the pessimistic view that I described a moment ago is not eliminated by the supposition that the parties in the LPOP would choose fully content-liberal principles. Indeed, the second objection at some level clearly animates some of Rawls’s critics. The charge of incoherence that is levelled at him by Caney, Pogge, and Tan, in particular, focuses not only on the contrast between his liberal domestic principles and his less liberal international principles, but also on the contrast between the respectful attitude that Rawls takes towards decent peoples at the global level and the dismissive attitude that he takes towards illiberal individuals at the domestic level (see, for example, Caney, 2002: 106–7; Pogge, 2001: 247, 2006: 208; Tan, 2000: 29–30).

Defenders of Rawls, meanwhile, typically argue that the parties in the LPOP have good reasons to endorse the less content-liberal principles of Rawls’s Law of Peoples and that this outcome is not as implausible as the critics suppose (see, for example, Freeman, 2007a: 275–80; Reidy, 2004, especially section V). But even if that is true, the second objection stands.

It is this objection, the objection to Rawls’s justificatory inclusion of decent peoples, that I want to focus on. Even though both critics and defenders of Rawls seem to accept the standard reading of Rawls’s argument (in particular, the idea that the extension to decent peoples involves identifying those peoples as independently worthy of both toleration and justificatory inclusion), this objection has not received much attention in its own right. For example, David Reidy, a prominent defender of Rawls, more or less ignores the question, simply accepting Rawls’s insistence that ‘the extent to which liberal democratic peoples may demand internal reform [of a non-liberal people] is a matter of international justice, the principles of which are to be arrived at … from a moral point of view common to all peoples’ (Reidy, 2004: 306–7). But the only reason he gives for endorsing this ‘overlapping consensus’ approach in the first place is that for liberal democracies to regard the class of those societies to whom justice is owed ‘as exhausted by liberal democratic bodies politic alone would be nothing short of manifest hypocrisy’ (Reidy, 2004: 298) – for liberal democracies were themselves once non-liberal, and nevertheless suppose that they were at that time the same people, owed justice in their relations with other peoples (Reidy, 2004: 296–7).

To be owed justice, however, is not to be owed justifications in terms that one accepts. Compare Rawls’s domestic theory of justice. Unreasonable individuals are owed justice: their liberal rights are to be respected just as those of their reasonable co-citizens are (see Quong, 2004). But they are not owed justifications in terms that they accept. Rawls’s political liberalism, as I said above, makes no attempt to include adherents of unreasonable views in the overlapping consensus. If we follow both critics and defenders of Rawls in supposing that in Part II of The Law of Peoples Rawls identifies a criterion for justificatory inclusion that is independent of the argument that precedes it in Part I, then it is no answer to those who dispute that criterion to argue that it would be hypocritical of liberal peoples not to seek to justify their conduct to them. That simply begs the question. ¹⁵

The second objection, then, has not been adequately dealt with. I will argue, however, that it can be.

3. International reasonableness as a ground for toleration

For those who accept the second objection, the obvious culprit in Rawls’s theory is the criterion of international reasonableness – in the case of non-liberal peoples, their decency. Since it is in virtue of a non-liberal people’s decency that it is regarded as worthy of justificatory inclusion, and yet decency does not imply domestic liberalism, it makes sense to suppose that this criterion is the source of the supposed incoherence. Moreover, Rawls’s reasons for taking decency to be a ground for toleration seem rather vague. ¹⁶ He writes:

The reader has to judge whether a decent people, as given by [the criteria outlined by Rawls (1999a: 64–7)], is to be tolerated and accepted as a member in good standing of the Society of Peoples. It is my conjecture that most reasonable citizens of a liberal society will find peoples who meet these two criteria acceptable … Not all reasonable persons will, certainly, yet most will. (Rawls, 1999a: 67)

Peoples are in this sense, then, group agents. This is the first necessary condition of international reasonableness. The second is that they are moral agents. Rawls (1999a: 25–7) writes:

As reasonable citizens in domestic [liberal] society offer to cooperate on fair terms with other citizens, so (reasonable) liberal (or decent) peoples offer fair terms of cooperation to other peoples. A people will honor these terms when assured that other peoples will do so as well … [Peoples] are not moved solely by their prudent or rational … interests, the so-called reasons of state.

These, then, are the conditions of international reasonableness, in virtue of which a society merits justificatory inclusion in the foreign policy of a liberal people. They are conditions that are met by both liberal and decent peoples (Rawls, 1999a: 63). Rawls spends a great deal of time enumerating these conditions. Yet the reasons that he gives for supposing any society meeting them to be worthy of justificatory inclusion are, as I have said, rather hazy. This gives proponents of the second objection an easy target. As I will argue in the next few sections, however, Rawlsian liberalism contains the materials for a much more satisfying basis for justificatory inclusion, which neither Rawls nor his defenders bring out. This involves defending Rawls’s conception of international reasonableness by appeal to a reading of what Rawls (1999a: 30–35, 67) explicitly claims is its domestic analogue: the reasonableness of individuals.

4. Reasonableness in political liberalism

Let us start, then, by considering the idea of reasonableness within Rawlsian political liberalism. Reasonableness enters the picture there as follows. A question for liberals of any stripe is whether liberalism is consistent with its own realization under favourable conditions (full compliance, a large population with good representation at all ages, a good stock of resources, and so on) (see Rawls, 1993: xix). If it is not, that is surely a serious flaw, to say the least. Moreover, you might worry that it is not, for the following reason. Under even the most favourable conditions, liberal institutions would engender diversity in reasonable conceptions of the good affirmed by citizens: this is ‘the normal result of the exercise of human reason within the framework of the free institutions of a constitutional democratic regime’ (Rawls, 1993: xviii). ²¹ If such pluralism meant that large numbers of individuals would not be able to see the liberal principles regulating their society as justified, then the liberal institutions might tend to instability because support for them would be undermined (see Rawls, 1993: xlii–xliii).

So, liberals must show that the reasonable pluralism that liberal institutions would engender would not tend to undermine the stability of those institutions. They can, according to Rawls, do this. For liberalism can be presented in a way that does not bring its justification into conflict with reasonable comprehensive doctrines, which permits adherents of those doctrines to affirm liberal political principles without prejudice to their other moral beliefs. How? Well, it can be presented as justified by appeal to certain fundamental ideas in such a way that there can be an ‘overlapping consensus’ on the ideas and the justification. That overlapping consensus includes adherents of all reasonable comprehensive conceptions of the good (see Rawls, 1993: xlvii and Lecture IV).

Taking this question and the response seriously means taking stability seriously. More than that, it means taking seriously the idea that a person can be reasonable without being a comprehensive liberal. For suppose that you thought that no one could be reasonable without being a comprehensive liberal. In that case, you might worry about stability, but not what Rawls (1993: xxxix, 143) calls stability for the right reasons. For the right reasons would be guaranteed, as far as you were concerned, by a comprehensively liberal justification such as the one that Rawls offers in A Theory of Justice. ²² So long as a comprehensively justified liberalism was stable under the conditions of its own realization, there would be no problem.

Now, the argument of A Theory of Justice seems to show, at least to the likely satisfaction of an early Rawlsian comprehensive liberal, that a comprehensively justified liberalism will be more or less stable under the conditions of its own realization. Those raised under liberal institutions, members of a well-ordered society, will come to internalize liberal principles as part of their sense of justice (see Rawls, 1999b: ch. 8), and that sense of justice will be congruent with their own good (where the latter is understood —unproblematically, for the comprehensive liberal—in terms of Rawls’s ‘Aristotelian Principle’ and certain Kantian ideas about autonomy, among other things) (see Rawls, 1999b: ch. 9).

The only difficulty arises if some people begin to worry that their allegiance to liberal institutions might be the product of state indoctrination (Rawls, 1999b: 451). In response to this difficulty, Rawls (1999b: 451) says that we can lay before them the full, comprehensively liberal justification for the institutions under which they were brought up. But, of course, if they do not accept that justification because they are not themselves comprehensive liberals, for example, then we will fail to convince them that their sense of justice is not the product of state indoctrination. All the same, however, we (who are, as we are supposing, comprehensive liberals) will know that it is not; so, as long as these people are not so numerous or so upset by what they see as indoctrination as to undermine the stability of liberal institutions, we need not view our failure to convince them as a problem. (And surely they will not be that numerous – the kind of reflection in question is pretty abstract, and seems unlikely to occupy most people very much, if at all.)

Rawls does, however, think that it is reasonable not to be a comprehensive liberal. He thinks that a person can reasonably reject comprehensive liberal views. More than that, he thinks that it is one of the basic ideas of liberalism itself that any given comprehensive doctrine, including liberal ones, can reasonably be rejected. In consequence, the concern with stability for the right reasons becomes problematic, since we cannot simply assume that a comprehensively liberal justification gives us the right reasons any more. (It could reasonably be rejected.) This is a major step away from the comprehensive liberal approach I have described in the foregoing paragraphs. Why should we accept it?

The answer, I think, involves implicit appeal to the following plausible liberal methodological principle (LMP).

LMP: If your comprehensive liberal moral view commits you to principles of justice which are such that those who comply with them under ideal conditions need not at the same time subscribe to the comprehensive liberal moral view in question, then you are committed to justifying the resultant liberal political order to such people in terms that they can accept. ²³

Notice that this explanation does not appeal to any independent ideas about the reasonableness of subscribing to non-liberal comprehensive doctrines. Liberal respect for non-liberal doctrines is not prescribed on the grounds that they are reasonable according to some conception of reasonableness that liberals privilege even above their liberalism. That way of thinking would give too much critical force to the joke about the liberal being someone who cannot take his own side in an argument. More seriously, it would make what Jonathan Quong (2005) calls the ‘asymmetry objection’ a serious problem for liberals even before they begin. ²⁶ The asymmetry objection asks why the factors that make it reasonable to affirm non-liberal comprehensive doctrines do not make it equally reasonable to affirm non-liberal conceptions of justice. Without a satisfactory answer to this question, liberals cannot justify their liberalism at a political level, let alone at a comprehensive moral one. If a genuinely independent (that is, non-gerrymandered) conception of reasonableness is the deepest foundation of liberalism, a satisfactory answer looks unlikely.

I will say a bit more about the asymmetry objection later. For now, note that it is not immediately suggested by the explanation of liberal respect for non-liberal comprehensive doctrines that appeals to the LMP. For that explanation does not employ a conception of reasonableness that is independent of and assumed prior to the argument for liberal principles themselves, and which acts, as a result, as a constraint on that argument. (I will call a conception of reasonableness that is independent in this way an external conception of reasonableness.) Rather, the reasonableness of a comprehensive moral view is constituted, according to the LMP-based explanation of liberal respect for reasonable such views, by the fact that it is compatible with full compliance with liberal political principles. ²⁷ That makes it what we can call an internal conception. ²⁸

If we accept an external conception of reasonableness, then, that justice as fairness turns out to be rejectable by people whom we can independently identify on that conception as reasonable will be sufficient to show that it is not stable for the right reasons, even if it is stable simpliciter. On an internal conception of reasonableness, by contrast, the test of stability for the right reasons does not involve appeal to the beliefs of any constituency apart from the members of a well-ordered liberal society – one, that is, in which everyone accepts, complies with, and knows everyone else accepts and complies with liberal principles (see Rawls, 2001: 8–9).

It is worth pointing out that a person’s compliance with, and so in the relevant sense her acceptance of, a given set of liberal principles of justice need not imply that she takes those principles to be the ideal principles of justice. The internal conception of reasonableness, in other words, does not preclude reasonable disagreement about which are the most appropriate principles of justice. For one can see the fact that some principle is generally complied with and applies to one as a non-instrumental moral reason to act in accordance with it (so that one’s so acting would constitute compliance) without thereby supposing that this situation is better than a situation in which some other principle is generally complied with and applies to one (see Gaus, 2011: 398–400). Supposing that these principles are not the best possible principles, then, does not preclude complying with them as opposed to merely conforming to them. Nevertheless, as I will argue in the next section, a person’s acceptance of the liberal principles of justice in question does restrict the range of principles that she could see as better justified than those.

5. Two objections to the internal conception of reasonableness

On my reading, then, Rawls thinks that it is one of the basic ideas of liberalism itself that any given comprehensive doctrine can reasonably be rejected, because it might be rejected by members of a well-ordered liberal society. The LMP says that such people are nevertheless owed justification in terms that they can accept. As a result, a necessary condition of a sound liberal theory of justice is that it can be justified to all reasonable individuals, regardless of the diversity in their comprehensive moral views.

The conception of reasonableness that is being invoked here is, to repeat, an internal conception, for what makes a person reasonable is simply that she is a member of a well-ordered liberal society and what makes a comprehensive doctrine reasonable is simply that it might be held by such a person. ²⁹ Now, it is tempting to be suspicious of my employment of this conception in the stability-for-the-right-reasons test, for the following two reasons. First, it seems that people could accept liberal political principles and institutions for all sorts of absurd reasons. If a person is reasonable if she accepts such principles and institutions, as the internal conception of reasonableness has it, and if all reasonable people in the well-ordered society are owed justifications in terms that they can accept, as the LMP has it, then Rawlsian political liberals may find themselves required to give justifications that are acceptable to people who accept liberal principles and institutions, but whose acceptance of those liberal principles is based (only) on absurd reasons. But surely political liberals’ justifications of liberal principles do not have to be compatible with such absurd reasons. I will call this the ‘absurd justifications’ objection.

Second, if a person is reasonable if, and only if, she accepts liberal political principles and institutions, as the internal conception of reasonableness has it, then surely there is no value in showing that liberalism is stable for the right reasons under conditions of reasonable pluralism. For, since everyone accepts liberal principles and institutions, it will be by definition stable for the right reasons. Following Quong (2011: 147) again, I will call this the ‘spare wheel’ objection, since if it is sound, the appeal to stability for the right reasons and the acceptability of liberal principles to reasonable people becomes trivial – a mere ‘spare wheel’. ³⁰

I do not have space to discuss these objections in as much detail as they merit, but it is worth explaining briefly how the view I am recommending would deal with them. Let us begin with the absurd-justifications objection. It will help to have an example of an absurd reason for acceptance of liberal principles and institutions. Here is such a reason: ‘A little magic bird that sometimes speaks to me tells me to act in accordance with liberal principles and institutions, and I do whatever the magic bird tells me to do.’ Call someone who accepts this reason an ‘ornitholator’, and let her be representative of all those whose acceptance of liberal principles is based on similarly absurd reasons. According to the absurd-justifications objection, it is not plausible to suppose that liberals’ justifications for liberal principles must be acceptable to an ornitholator. But if an ornitholator would count as reasonable on the internal conception, then Rawlsian liberals who accept the internal conception are committed to justifying liberal principles to her in terms that she can accept. Therefore, the internal conception, and with it my reading of Rawlsian political liberalism, should be rejected.

The following options suggest themselves as possible replies to this objection.

Accept that we are committed to justifying a liberal political order to ornitholators in terms that they can accept, but insist that this is not absurd.

Option 1 is unattractive even if we concede that it is not absurd to be committed to justifying the political order to ornitholators. For if we are committed to justifying a liberal political order to the ornitholator in terms that she can accept, then the kinds of justification that we will have to offer to her cannot be justifications that anyone else could accept. So, the reasons that people have for seeing the political order as justified will not be shared reasons. ³¹

Why should this matter? Well, the citizens of a stable society must be able, when it faces new circumstances and countenances new proposals, to discuss the merits and demerits of these, and then to rank them, by reference to its conception of justice. But if the only point of convergence in citizens’ judgements about justice is the principles of justice themselves (and not, in particular, any of the values that these express), then there may be no way to frame discussions and suggestions about ranking proposals one way rather than another except in terms that others (reasonably) reject. There may thus be no consensual way forward. Nor can this problem be resolved simply by putting everything to the vote, since democratic forms themselves are subject to the same problems. Convergence on liberal principles of justice alone, then, is not enough. Political liberalism’s ‘public reason’ must be, as Rawls (1999a: 144–7) puts it, ‘complete’.

Option 2.1 would be a helpful answer if it were clear what warranted the assumption that, as a matter of empirical fact, everyone in the well-ordered society would accept the Rawlsian ‘fundamental ideas’ of society as a fair system of mutually advantageous cooperation, citizens as free and equal, and so on, from which a liberal political conception of justice might be constructed (see Rawls, 1993: 3–46, especially sections 2–5). For in that case, we would have no need to ask ourselves what to say to the ornitholator: she simply would not be present. But it is not clear what would warrant the assumption.

Let us turn, then, to option 2.2. This option, of course, entails 2.1, and so would also give us the explanation that we wanted for the empirical expectation. What could justify 2.2? The thought would have to be that it is conceptually impossible to internalize the principles of justice without at the same time accepting the fundamental ideas. So, once again, we would have no need to ask what might be said to the ornitholator: she would not be present because she could not be a member of the well-ordered society.

But why not? Surely it would be possible to conform to the principles of justice without accepting the fundamental ideas. ³² The thought has to be that it is not possible to internalize them (this being necessary for compliance as opposed to mere conformity) without accepting the fundamental ideas. One possible explanation for this might be that it is not possible to internalize a principle without a sense and acceptance of the values that underlie it. But then if the ornitholator sees (non-instrumental) value in conformity to the pronouncements of the magic bird (as, perhaps, a theist might see value in conformity to the pronouncements of God), there is no reason to suppose that her conformity to the principles of justice could not also be compliance.

It would be better if we could take option 2.3 and simply claim that the little bird justification of the ornitholator’s acceptance of the principles of justice is not a sufficiently good justification. But we would want to do that in a way that was consistent with accepting that at least some theistic justifications of a theist’s acceptance of the principles could be good justifications (for otherwise reasonableness would become much more restrictive than we were aiming to make it). Now, this would not mean that we would have to grant that any theistic justification is a good justification for acceptance of the principles of justice. There could be something of justificatory value that only some theists’ justifications for acceptance of the principles of justice have and that the ornitholator’s justification lacks. But we cannot just stipulate that what that is is the intermediary fundamental ideas. For if we do that, then the notion of a sufficiently good justification smuggles in, effectively, an external conception of reasonableness which entails acceptance of the fundamental ideas. An external conception with a substantive component of that sort is both implausible and suspicious (for reasons that I explain below).

So, what might feature in a theist’s acceptance of the principles of justice as a result, ultimately, of her faith or belief in God that would not feature in the ornitholator’s acceptance of the principles of justice? I propose that any reasonable view must include (if only implicitly) a coherent story which connects its most fundamental commitments (for example, to God’s authority) to the principles of justice. If this is a necessary condition of the reasonableness of a comprehensive doctrine, then merely replacing the little bird with God, for example, will not make the doctrine any less unreasonable, since there would still be no story which connects God with the principles of justice. By contrast, a story about how much God (or, indeed, the little bird) loves humans might well allow the doctrine to meet the condition by connecting them in a coherent way – in a way that enables a person to understand, for example, why God’s (or the little bird’s) commands are to be obeyed. ³³

It is important to distinguish between a coherence condition of the sort I have in mind (what we might think of as a requirement that views are reasonably held) and a necessary condition of reasonableness which specifies, by content, some view that it is unreasonable not to hold. Clearly, the latter offers an unattractive option for my purposes. It runs against the grain of the internal conception of reasonableness that I am defending here. The point of the LMP approach is that what counts as reasonable is primarily defined by compliance with and so internalization of liberal principles of justice. To specify in advance some conception of reasonableness that rules out the need to justify the principles even to some of those who comply with them under ideal conditions, because those people do not accept some specified principle or set of ideas (such as the fundamental ideas), is to endorse an external conception of reasonableness that looks, as I said above, both implausible and suspicious. It looks implausible because so long as we do not conceive of reasonableness as an internal conception (as defined in terms of the liberal account of justice rather than independently of it) then it seems unlikely that to be reasonable will turn out to require adherence to just one particular evaluative view. It looks suspicious because if this external conception of reasonableness is used to ground the denial that liberals are required to justify a liberal political order to some (as per option 2.3), then it will be hard to avoid the appearance of gerrymandering. For it would seem too great a coincidence that the particular normative ideas, acceptance of which is entailed by reasonableness, on this external and therefore supposedly independent conception, should be precisely such as to guarantee acceptance of the liberal political order. ³⁴

The view I propose, then, is that LMP-accepting liberals are not committed to justifying the political order in terms that they can accept to people whose views, though they do support compliance with the principles of justice, are not appropriately coherent. Someone meets the coherence condition when there is a story to tell (a story which she at least implicitly accepts or would accept under appropriate conditions) that pulls together both the principles of justice and her fundamental commitments in a way that makes sense of their relations to each other. Note that this condition is stronger than the condition that a person’s compliance with the principles of justice be comprehensible in light of her fundamental commitments. The ornitholator meets this weaker condition. But she does not, without a story of the sort that I am describing, meet the stronger condition. Her compliance is comprehensible in light of her commitment to obedience to the little bird, but the view as a whole does not include sufficient resources for satisfying answers (even from the point of view of someone who accepts it) to non-sceptical questions that any adequate view should be able to give (Why these principles? What is the point of obedience? What makes you think that the little bird is to be obeyed rather than ignored?).

This requirement that there must be a story that makes sense of the relations between a person’s fundamental commitments and the principles of justice is not a requirement that reasonable comprehensive doctrines satisfy strict norms of rationality. A story can make sense of the relations between fundamental commitments and principles of justice even though elements in the story are not clearly consistent. The Christian doctrine of the Trinity, as Martha Nussbaum points out, ‘asks the believing Christian to believe a contradiction’ (2011: 26). So, those Christians whose endorsement of the principles of justice is founded ultimately on Christian commitments which include affirmation of the doctrine of the Trinity would not have a story to tell about the relations between their fundamental commitments which satisfied strict norms of rationality. But if we can, nevertheless, make sense of such Christians’ endorsement of political principles in light of their fundamental commitments (as Nussbaum’s own discussion of the doctrine of the Trinity (2011: 26) suggests that we can), then they satisfy the coherence condition that I am proposing. ³⁵

The coherence condition is a necessary condition of reasonableness, since it excludes those who do not meet it from the scope of liberal justification. But it is not itself either an internal or an external conception of reasonableness. This is because conceptions of reasonableness for the purposes of political liberalism give us sufficient conditions for inclusion in the scope of liberal justification. The relevant type of coherence is, rather, akin to citizens’ rationality or the capacity to cooperate. It is also like these in that it sets a minimal and uncontroversial bar to qualification for inclusion in the constituency of justification (membership of the well-ordered society) that does not threaten to undermine liberals’ commitment to liberalism itself, as an external conception of Rawlsian reasonableness does (see Section 4 above).

We have then, grounds to deny that political liberals who accept the LMP are committed to justifying liberal political principles to people who comply with those principles for absurd reasons. Members of the justificatory constituency specified by the LMP (reasonable citizens) meet a coherence condition which ensures that their fundamental commitments and their acceptance of the principles of justice are related by intermediary ideas which provide the resources for the kinds of satisfying answers to non-sceptical questions that any adequate comprehensive doctrine should be able to give. Thus, we can affirm option 2.3. In light of this, we have reason to affirm option 2.2 and, therefore, option 2.1 as well.

When I rejected option 1 I did so on the grounds that it gives us an inadequate picture of political liberalism’s public reason. Do options 2.1–2.3 fare any better in this respect? One might worry that they do not. The coherence condition does not appear to require any particular set of intermediary ideas and values as components in the story that links individuals’ fundamental moral commitments and their adherence to principles of justice. So, we might expect a wide range of inconsistent ideas and reasons to be the grounds of individuals’ acceptance of such principles. Thus, there would be no guarantee that there will be a set of shared reasons sufficient for public reason to be complete.

This conclusion would be too hasty. It is true that the coherence condition on its own does not tell us anything about the values that operate as intermediaries between a person’s fundamental commitments and the principles of justice she accepts. However, it does license conclusions about the ideas and values that operate as intermediaries between a person’s fundamental commitments and the specifically liberal principles of justice that regulate a well-ordered society. For those principles are evidently expressive of or derived from the fundamental ideas of fairness, individual freedom and equality, and of society as a system of cooperation, as Rawls says they are. (To affirm this is not to deny that there may be other principles which also express those ideas or to contend that Rawls’s derivation is uncontroversial.) Moreover, it is hard to see how the principles could be expressive of or derived from ideas and values that differed, in any substantial way, from these. If I am right about this, then the coherence condition, in the context of a well-ordered society, gives us reason to suppose that each person accepts the fundamental ideas. Individuals who accept and comply with the principles of justice will also accept, and see the principles as expressive of or based on, the fundamental ideas. Consequently, the overlapping consensus will include those ideas together with the principles of justice; so, the public political culture will include shared resources adequate for public reason to be complete. ³⁶

Lest all this should sound overoptimistic, it is worth emphasizing that the aim here is to show that the stability for the right reasons of the well-ordered society is likely, not that it is guaranteed. Its being likely does not imply, of course, that there could never be a member of the well-ordered society who both met the coherence condition and yet did not accept the fundamental ideas. (As I said above, it is hard to see how the principles of justice could be expressive of or derived from ideas and values that differed substantially from those appealed to by Rawls, but I do not see either how this could be ruled out altogether.) The existence of such a person would threaten the completeness of public reason and so the stability for the right reasons of the well-ordered society. But acceptance of the remote possibility of such a person is much less worrying for a political liberal than acceptance of the idea that ornitholators might be members of the well-ordered society and included, therefore, in the constituency of justification. It is perhaps unfortunate that political liberalism, on my reading of it, cannot be shown to be consistent with its own realization under all conceivable circumstances, even if this can be shown to be likely. But if it were an implication of the LMP that ornitholators were owed justification of the political order, that would be far worse.

I reject, then, the absurd-justifications objection. I will be briefer in addressing the spare-wheel objection. That objection, recall, runs as follows. If a person is reasonable if, and only if, she accepts liberal political principles and institutions, as the internal conception of reasonableness has it, then surely there is no value in showing that liberalism is stable for the right reasons under conditions of reasonable pluralism. For, since everyone accepts liberal principles and institutions, it will be by definition stable for the right reasons.

The spare-wheel objection is mistaken. There is value in showing that liberalism can be stable for the right reasons under conditions of reasonable pluralism, and that is because the internal conception of reasonableness does not entail acceptance of any liberal justification. It entails only acceptance of and compliance with liberal political principles, that is, liberal principles for the institutions of the basic structure. Acceptance of such liberal principles does not entail acceptance of comprehensive liberal justifications for these liberal principles. (That is precisely the point of the fact of reasonable pluralism.) So, the possibility of stability that is not for the right reasons (the possibility that individuals who accept and comply with liberal principles for institutions may, nevertheless, not be in a position to see them as justified) remains open. To foreclose it, it must be shown that justifications for these liberal principles can be given that are acceptable to people who endorse different comprehensive doctrines despite their acceptance of liberal principles for the basic structure. This means showing that there can be an overlapping consensus that guarantees that each member of it sees the liberal principles as justified. This task is a substantial one which involves showing, first, that the liberal principles can be justified by appeal to the fundamental ideas (Rawls’s argument for this (1993: 89), which goes via the original position, is an instance of what he calls ‘political constructivism’) and, second, that acceptance of the fundamental ideas can legitimately be assumed. I tried to make some headway with this second task in the foregoing response to the absurd-justifications objection. As should be clear enough from that alone, showing that liberal principles will be stable for the right reasons is not a simple matter of definition.

6. An internal conception of international reasonableness

Armed with the LMP-based understanding of the role of reasonableness in political liberalism, we can now articulate a parallel and more satisfying defence of Rawls’s justificatory inclusion of internationally reasonable societies. This way, we can vindicate not only that inclusion, but also Rawls’s insistence (1999a: 33) that his approach in The Law of Peoples is analogous to his approach in Political Liberalism. The idea is to interpret international reasonableness not, as is standard, as a rather sketchily defended external conception, but as an internal conception.

Recall that at the domestic level, Rawls thinks that a person can reasonably reject comprehensive liberal views, and hence that stability for the right reasons can be secured only by justification in terms that people who do reasonably reject comprehensive liberalism can accept. The supposition that a comprehensive liberal cannot simply assume the right reasons and concern herself only with the stability simpliciter of liberal institutions called for explanation. I argued above that the explanation lies in Rawls’s adherence to the LMP.

The same principle, mutatis mutandis, can be appealed to in order to explain Rawls’s justificatory inclusion of decent peoples. In the international context, it holds that if the principles of the law of peoples that is worked out by appeal only to liberal ideas (which I will refer to as a liberally justified law of peoples) are such that those societies that comply with them, under ideal conditions, need not at the same time subscribe to the liberal ideas that were appealed to in their construction in the first place, then adherents of the liberal ideas are committed not only to governing their relations with such societies by these principles, but also to justifying them to such societies in terms that they can accept. Just as liberal domestic principles would engender a pluralism of comprehensive views even when everyone complied with those principles, so liberal international institutions, according to this argument, might engender (or at least be consistent with) a pluralism of domestic conceptions of justice even when every society complied with the principles regulating those international institutions. In addition, if Rawls is right to suppose that a liberally justified law of peoples is not a fully content-liberal law of peoples, then that is precisely the situation that we, as adherents of those ideas, face. In light of the methodological principles, therefore, Rawls is committed to justifying the Law of Peoples to decent peoples, who are simply those non-liberal peoples who, nevertheless, comply with the principles of the Law of Peoples.

This is why it is necessary for Rawls to show that decent peoples would accept ‘the original position as fair among peoples, and would endorse the Law of Peoples adopted by their representatives [in the DPOP] as specifying fair terms of political cooperation with other peoples’ (1999a: 69). The possibility was always open, of course, that the Law of Peoples would turn out not to engender or be consistent with the relevant sort of pluralism – in other words, that there would be no societies which could at once be non-liberal and yet comply with the principles of the Law of Peoples. (In that case, ‘internationally reasonable’ would have turned out to be coextensive with ‘internally liberal’.) Rawls’s discussion is intended to show, however, both that there could be such societies and that the Law of Peoples could be justified to them.

It is worth stressing that the societies whose acceptance of the Law of Peoples might or might not be compatible with domestic non-liberalism are to be conceived of as peoples, with everything that that term implies (see Section 3 above). Otherwise, you might be suspicious of the suggestion that a society can comply with the Law of Peoples. The first necessary condition of a society’s counting as a people is that it is a scheme of social cooperation in Rawls’s sense (that is, as distinct from mere socially coordinated activity), so that the members of that society may be seen as acting collectively, and capable of accepting and acting from any principles that they might collectively adopt. It is because of this that if a society is a people, its compliance with the Law of Peoples may be viewed as more than merely the compliance of the de facto authority in that society with the Law of Peoples, for example.

It might be objected to my account that the fact of reasonable pluralism gets explained, in Rawls’s domestic theory, as the inevitable consequence of the operation of human reason under free (that is, liberal) institutions, in light of what Rawls calls the ‘burdens of judgment’ (1993: 54–8). By contrast, we have been offered no analogous explanation of what might be called the ‘fact of internationally reasonable pluralism’, that is, the supposed fact that the institutions of the Law of Peoples, under favourable conditions, would engender or be consistent with pluralism in societies’ domestic conceptions of justice (compare Caney, 2002: 105–6).

In response, let me make two points. First, it is important to see that on my reading of Rawls’s domestic theory, the question of why the realization of comprehensive liberalism, even under ideal conditions, is consistent with the existence of reasonable people who are not themselves comprehensive liberals is secondary to the question of what to do about it. The LMP requires that liberal institutions be justified to such people regardless of whether we have a satisfying explanation of how they come to endorse different comprehensive doctrines, such as the burdens-of-judgement explanation that Rawls indeed goes on to give. On an external conception of reasonableness, by contrast, it makes sense to ask for explanations of the reasonableness of non-liberal views before one bothers to think about justifying liberal institutions to people who hold such views. For there is no point justifying those institutions to them if it cannot be shown that it is reasonable to be non-liberal. So, the priority of the two questions is reversed. ³⁷

As I read Rawls, he is not operating with an external conception of reasonableness, either domestically or internationally. Consequently, the lack of an analogous explanation of internationally reasonable pluralism does not pose a particular problem for Rawlsian justificatory inclusion at the international level. It needs only to be plausible to suppose that there might be such pluralism. I take Rawls’s description of the decent people ‘Kazanistan’ to be an effort to show that there might (1999a: 75–8).

It is not surprising, moreover, that little empirical explanation for the fact of internationally reasonable pluralism is available. For we have no experience of such pluralism. In order to explain the fact of (domestic) reasonable pluralism, Rawls can perhaps avail himself of the evidence that surrounds him, since we live in pluralistic times under more or less liberal institutions. But we do not live under the institutions of the Law of Peoples. Empirical explanations of the pluralism that such institutions might engender are correspondingly more difficult to identify. ³⁸ One might suspect that something like the burdens-of-judgement explanation applies to peoples as well as to individuals. If that does not seem plausible, then until a liberal law of peoples is implemented, we must content ourselves with the intuitive plausibility of assuming diversity in domestic conceptions of justice under such a global order. But on my reading of Rawls, this is not particularly problematic.

7. The Law of Peoples reappraised

In Section 2 above, I noted that the objection to Rawls’s justificatory inclusion of decent peoples seems to be moot if those critics who argue that a liberally justified law of peoples would require domestic liberalism are correct. For if that is true, there is no possibility that decent peoples could be members of the Society of Peoples, and so no question of a requirement of public reason that the terms of membership should be justifiable to them. As I argued, however, if the correct conception of international reasonableness is an external conception, as critics and defenders alike appear to have assumed, then in fact the objection to justificatory inclusion can be pressed regardless of the success of the objection to toleration. For an external conception of international reasonableness would imply that a morally adequate law of peoples should be justifiable to decent peoples regardless of what fully liberal justifications of it entail. As I pointed out, there is room for both optimism and pessimism about the compatibility of the liberal justifications and the justificatory inclusion required by an external conception of international reasonableness that includes decent peoples.

By way of defence against the objection to justificatory inclusion, I have argued that Rawls should be read as employing an internal conception of international reasonableness. But on such a conception, the success of the objections to toleration really would make the question of justificatory inclusion inapplicable. To see this, suppose that a liberally justified law of peoples would be fully content-liberal, that is, it would disallow any form of domestic organization apart from liberalism. According to the international analogue of the LMP (see Section 6 above), if the principles of that law are such that those societies that comply with them, under ideal conditions, need not at the same time subscribe to the liberal ideas that were appealed to in their construction in the first place, then adherents of the liberal ideas are committed not only to governing their relations with such societies by these principles, but also to justifying them to such societies in terms that they can accept. For such societies are internationally reasonable. But if a liberally justified law of peoples would disallow any form of domestic organization apart from liberalism, then there is no possibility of internationally reasonable, yet non-liberal peoples. So, liberals are not committed to showing the law of peoples could be accepted by them. And so, there is no question of justificatory inclusion for non-liberal peoples.

It transpires, then, that my defence of Rawls against the objection to justificatory inclusion of decent peoples is self-undermining if the objection to toleration turns out to be successful. A full defence of The Law of Peoples along the lines I am recommending requires, therefore, a defence of the less than fully content-liberal Law of Peoples that Rawls endorses. I do not have the space to offer a full defence, but I will sketch the main lines of what is, to my mind, the most plausible one.

The defence must show, in particular, that the LPOP is the appropriate constructivist procedure for a just law of peoples, and that the representatives of liberal peoples in the LPOP would not choose principles requiring domestic liberalism. The key points here, for our purposes, concern both the fact that the parties in the LPOP are representatives of liberal peoples rather than of individuals and the limits to a liberal peoples’ fundamental interest in securing liberal justice domestically. ³⁹ For it is in virtue of their representing peoples rather than individuals that the parties are not motivated simply to safeguard the interests of individuals as given, according to Rawls’s writing on domestic justice, by the ‘two moral powers’ – which motivation, Rawls argues there, leads to the choice of fully liberal principles. ⁴⁰ In addition, it is in virtue of limits to the fundamental interest of liberal peoples in securing liberal justice domestically that it is not a requirement of the Law of Peoples that peoples be organized domestically according to a liberal conception of justice.

The decision to make the parties representatives of peoples rather than of individuals reflects the priority that Rawls accords to what I will call the ‘basic problem of social justice’. That problem is a problem that faces societies, namely, forms of cooperative association to which there is no alternative for individuals: ‘some workable [cooperative] arrangement is a necessary condition of the good of all’ (Rawls, 1971: 103; compare 1993: 275–8). ⁴¹ The problem is to find principles appropriate to regulate this essential form of cooperative association. For Rawls, this is the most fundamental problem of justice, and the problem of international justice is to be seen through the prism of his solution to that more fundamental problem, which is given in A Theory of Justice and Political Liberalism. This is why he conceives of the Law of Peoples as principles for the foreign policy of a liberal democratic regime (Rawls, 1999a: 9–10, 55, 82–3).

The necessity of social cooperation as a condition of anyone’s good at all is what makes the basic problem of social justice fundamental. But why, one might ask, is not the Law of Peoples conceived as part of the same problem, as it should be according to cosmopolitan critics who argue for a ‘global original position’ in which each individual in the world is represented by a party in the original position? ⁴² The familiar Rawlsian answer to this question is twofold. In the first place, there is no need for the appropriately regulated cooperative association that solves the basic problem to be an association that includes all individuals in the world. The problem can be solved (justice can be done) within a much smaller association. So, a world of numerous such associations is no worse, in principle, than a world of only one. ⁴³ In the second place, there may be reasons to think that it would be better. Following Kant, Rawls (1999a: 36) suggests that ‘a world government … would either be a global despotism or else would rule over a fragile empire torn by frequent civil strife as various regions and peoples tried to gain their political freedom and autonomy’. If that is true, then our initial statement of the basic problem of social justice should not assume that it is a problem to be solved by principles for a basic structure for a single, global society, as the critics’ favoured global original position would. ⁴⁴ Only if we are already there (that is, if we already have a basic structure of the sort that Rawls has in mind as the proper realm of applicability for principles of social justice) should we conceive of global and domestic justice as the same problem. But we are not already there: the global institutions that exist today are not those of the basic structure of a single society in Rawls’s sense. ⁴⁵

The foregoing reasoning raises the following important objection. Even if a world of many societies can solve the basic problem of justice, that does not show that the basic problem of justice can be solved by principles that apply only within a single society, as those that Rawls describes in A Theory of Justice and Political Liberalism are supposed to. It may be that a world of many societies solves the problem only thanks to a combination of societal and global principles, as cosmopolitans might argue. ⁴⁶ The global principles, as an essential element in the solution to the basic problem, would in that case have priority over anything that the LPOP procedure might produce. But Rawls’s approach appears illegitimately to rule out this possibility in advance.

The response to this objection that I propose on Rawls’s behalf is as follows. ⁴⁷ It may be true that even a society-focused solution to the basic problem of justice must include, at a fundamental level, principles that are global in scope alongside the principles for societies. But these global principles need not be egalitarian. They may, for example, require only that interactions between individuals respect Lockean norms of individual self-ownership. The egalitarian element in Rawls’s own solution arises from the fact that members of a single society cooperate in the distinctive way that constitutes participation in the practice of life together as a single society and makes them a group agent. ⁴⁸ This fact makes special distributive norms to govern their relations with one another applicable that are not applicable to the relations between members of different societies.

Now, as the objection points out, the LPOP is constrained, on this view, not to produce principles that are inconsistent with the global principles already applicable as part of the solution to the basic problem. But if the global principles are less demanding than cosmopolitans tend to suppose, then this constraint need not jeopardize Rawls’s conclusions about the results of the LPOP. ⁴⁹

Suppose, then, that the appropriately regulated cooperative associations that solve the basic problem of social justice are several and that any global principles that are also required represent a constraint on the outcome of the LPOP. There remains a question of how, within those constraints, the associations should relate to each other as the group agents that they are. (Is anything more required of them than compliance with the global principles?) Hence the decision to make the parties in the LPOP representatives of liberal peoples rather than of individuals. That decision does not alone explain the less than fully liberal content of the Law of Peoples, however. For one might think that given Rawls’s specification of the fundamental interests of liberal peoples in terms of their securing liberal justice domestically, the parties would agree to institutions designed to ensure precisely that – through intervention by other peoples if necessary. ⁵⁰

But this would be to underestimate the importance of peoples’ interests in autonomy and the autonomous achievement of justice. Rawls (1999a: 34; see also 118) says that peoples ‘strive to protect their political independence’, an interest which is connected to their further interest in ‘proper self-respect of themselves as a people, resting on their common awareness of their trials during their history and of their culture with its accomplishments’. As Reidy (2004: 304) notes, domestic liberalism itself is viewed with pride by a liberal people as a historical achievement – rather as an individual who lives a good life according to a particular comprehensive doctrine that she comes to endorse might see this as her own achievement. In each case, it matters that the agent in question should have arrived at the relevant point without having been forced to arrive there by others. The value of the achievement of liberalism for a people and the value of the achievement of the good life for an individual are both subject to something like Dworkin’s ‘endorsement constraint’, which makes a person’s endorsement of a conception of the good according to which her life is going better a necessary condition of her life’s really going better. ⁵¹ A people’s self-respect is seriously damaged if its liberal democratic political institutions cannot be viewed by its members as their own collective achievement.

If we accept that peoples are group agents, as Rawls does, then this should not seem particularly implausible, particularly to liberals. Note that liberal democracy is not a necessary condition (on Rawls’s account) of group agency, so it makes sense for a liberal people to view its liberalism as its achievement, that is, as the achievement of an agent that existed prior to its becoming a liberal people. It is at least arguable, in light of these points, that the parties in the LPOP would not agree to a law of peoples that licensed or required external intervention in people’s domestic affairs to guarantee liberal institutions, even though they know that they have a fundamental interest in those institutions. For such principles would be inconsistent with peoples’ independence and self-respect. In addition, since they know that they represent liberal peoples, and we take it as a given that domestic liberalism can be stable for the right reasons, they have no reason to suppose that their liberal democratic institutions are under any particular threat from within.

This is not to grant that any external intervention would be inconsistent with peoples’ independence and self-respect. Intervention to ensure that a people remains a people in the face of internal threats of dissolution is perfectly consistent with that independence and self-respect. This is why parties in the LPOP would endorse the minimal human rights, and the intervention that these license, that are a condition of social cooperation and so group agency. There is no question of these violating a people’s self-respect, since if intervention is called for by the Law of Peoples, then the likely alternative to such intervention is that there ceases to be a people around to have any self-respect in the first place. ⁵²

It is worth noting, finally, that even if the principles that would be chosen in the LPOP would be more demanding with respect to peoples’ internal organization than those of Rawls’s Law of Peoples, that does not show that my defence of Rawlsian justificatory inclusion is self-undermining. For, so long as the interest in independence and self-respect has sufficient weight to prevent the parties going so far as to choose principles that require fully liberal domestic institutions, the possibility of non-liberal peoples who accept and act from the law of peoples remains. It is this possibility that my defence of Rawlsian justificatory inclusion depends upon.

The foregoing argument, if it is successful, explains why a liberally justified law of peoples would not permit members of the Society of Peoples to force other members to remain domestically liberal. The same kinds of considerations can also explain why a liberally justified law of peoples would not sanction other, non-coercive forms of intervention that might represent serious threats to members’ independence and self-respect. As Joseph Raz (1986: 377–8, 420) argues, manipulation is objectionable for the same reasons that coercion is objectionable: it interferes with individuals’ autonomy. I understand Rawls’s objection to liberal peoples or the Society of Peoples officially offering non-liberal decent peoples incentives to become more liberal in their domestic organization to be an objection to manipulation that is based on the same reasons as the objection to coercive interference just described. ⁵³

A further difficulty that must be dealt with is as follows. Rawls does not explicitly say that the Society of Peoples will operate according to a norm of official non-criticism of members’ internal affairs, so long as they continue to comply with the law of peoples. But I think (for reasons that I explain below) that he would endorse such a norm, even while accepting that liberal peoples may, when not speaking as representatives of the Society of Peoples, criticize decent peoples’ non-liberalism. ⁵⁴ If the content of the law of peoples is liberally justified in the way that I have outlined, however, one might wonder why a norm of official non-criticism should obtain. The reasons which explain the reluctance of the parties in the LPOP to choose principles licensing coercive intervention in each other’s affairs, except in cases of human rights abuses, are not also reasons to suppose that they would choose a principle prohibiting official criticism of peoples’ domestic non-liberalism. So, what are the grounds for an official norm of non-criticism for the Society of Peoples?

The grounds for such a norm lie in the fact that the public reason of a well-ordered society of internationally reasonable peoples would rule out any principle sanctioning criticism. Even if such a principle were affirmed in the LPOP, and so one element in a fully liberally justified law of peoples, it would be superseded by norms of public reason for the Society of Peoples that would disallow criticism based on reasons not shared by all internationally reasonable peoples. By hypothesis, not all internationally reasonable peoples would accept reasons based on the justice of domestic liberalism, since some of them would not accept that conception of domestic justice.

It might be objected that on the internal conception of international reasonableness, all internationally reasonable peoples would by definition accept those reasons, since according to the LMP to be internationally reasonable is to comply with the principles of the liberally justified law of peoples, which, we are supposing, would include the norm of criticism. But this would be to misunderstand the LMP. To see this, consider the way in which the LMP applies to a liberal theory of domestic justice. Some comprehensive liberals (for example, liberal perfectionists) might suppose it to be permissible for the government to criticize the non-liberal comprehensive views even of individuals who comply with the principles of liberal justice in their relations with others. Even though such people were doing nothing politically wrong, as we might put it, such liberals would take it to be permissible for state officials to criticize their moral views. But the LMP-endorsing justificatory liberal denies this, precisely because such people are doing nothing politically wrong by liberal lights. For acting in a way that draws perfectionist liberal criticism, but that is not, by such liberals’ lights, unjust is acting in a way that is not politically wrong. Liberals therefore owe them justification of the principles of justice in terms that they can accept. As we saw in Section 5 above, uncontroversial necessary conditions of reasonableness give us strong reasons to suppose that there will be a shared set of intermediary political ideas and values which may be appealed to in public reasoning about policies and proposals in the well-ordered society (and which enable us to say that public reason in that society is complete). The norms of public reason will prohibit state criticism of reasonable comprehensive doctrines, since no justification can be given for such criticism in terms that appeal only to the political ideas and values that make public reason complete.

In this way, the norms of public reason supersede any norms of criticism of illiberal, but reasonable comprehensive doctrines that might otherwise appear justified to liberals. The LMP thus explains the existence at the domestic level of a norm of official non-criticism of such doctrines. We can proceed analogously at the international level in order to explain the existence of a norm of official non-criticism of the illiberal internal organization of internationally reasonable peoples. Of course, it is open to justificatory liberals not to accept the LMP, or to liberals not to be justificatory liberals at all. But the former will be vulnerable, as I have argued, to the asymmetry objection, and my argument is not directed at the latter at all.

8. Conclusion

Two strands can be detected in standard critiques of Rawls’s toleration of decent peoples. The first strand represents a concern with toleration in the sense of non-intervention. The second represents a concern with justificatory inclusion. The two strands have typically not been properly distinguished. This, I have suggested, is partly because if objections based on the first concern succeed, then the second concern may appear to be moot. However, on standard readings of Rawls, it is not.

Defenders of Rawls have typically focused on the first concern. They argue that there are good reasons to suppose that the LPOP is an appropriate decision procedure for determining the principles of the law of peoples, and that representatives of liberal peoples in the LPOP would choose the principles that Rawls says they would. I accept these reasons. Even if that is so, however, the question remains why we should see non-liberal but decent peoples’ acceptance of those principles as a necessary condition of the principles’ moral adequacy, any more than we should see illiberal individuals’ acceptance of the principles of justice as fairness as a necessary condition of those principles’ moral adequacy. In both cases, we might think, seeking acceptance is pandering to unreason.

I have argued that the answer to this question lies in an understanding of the Rawlsian conception of international reasonableness (the trait that decent and liberal peoples share) as an internal conception, rather than as an external one, together with an appreciation of the liberal methodological principle that I take to underpin Rawls’s approach. International reasonableness is simply willingness, on condition of reciprocal willingness in others, to comply with the law of peoples. If it is possible to be internationally reasonable in this sense and yet not liberal, as it is if the parties in the LPOP would choose as Rawls says they would, then according to the LMP liberals are committed to showing that the Law of Peoples could be justified to non-liberal but decent peoples. It is for this reason that Rawls’s Law of Peoples not only tolerates such peoples, but also includes them in the constituency of justification.

This reading shows Rawls to be less vulnerable than his critics suppose, and it offers us an attractive alternative to these critics’ favoured cosmopolitan approaches. Beyond that, it shows there to be a satisfying unity to Rawls's work (a unity that he himself insisted on) and makes sense of what has been a real puzzle about The Law of Peoples.