One aim of this lecture is to sketch – in a short space, I can do no more than that – how the law of peoples¹ may be developed out of liberal ideas of justice similar to but more general than the idea I called justice as fairness and presented in my book A Theory of Justice (1971). By the law of peoples I mean a political conception of right and justice² that applies to the principles and norms of international law and practice. In section 58 of the above work I indicated how from justice as fairness the law of peoples might be developed for the limited purpose of addressing several questions of just war. In this lecture my sketch of that law covers more ground and includes an account of the role of human rights. Even though the idea of justice I use to do this is more general than justice as fairness, it is still connected with the idea of the social contract: the procedure of construction, and the various steps gone through, are much the same in both cases.

A further aim is to set out the bearing of political liberalism once a liberal political conception of justice is extended to the law of peoples. In particular, we ask: What form does the toleration of nonliberal societies take in this case? Surely tyrannical and dictatorial regimes cannot be accepted as members in good standing of a reasonable society of peoples. But equally not all regimes can reasonably be required to be liberal, otherwise the law of peoples itself would not express liberalism’s own principle of toleration for other reasonable ways of ordering society nor further its attempt to find a shared basis of agreement among reasonable peoples. Just as a citizen in a liberal society must respect other persons’ comprehensive religious, philosophical, and moral doctrines provided they are pursued in accordance with a reasonable political conception of justice, so a liberal society must respect other societies organized by comprehensive doctrines, provided their political and social institutions meet certain conditions that lead the society to adhere to a reasonable law of peoples.

More specifically, we ask: Where are the reasonable limits of toleration to be drawn? It turns out that a well-ordered nonliberal society will accept the same law of peoples that well-ordered liberal societies accept. Here I understand a well-ordered society as being peaceful and not expansionist; its legal system satisfies certain requisite conditions of legitimacy in the eyes of its own people; and, as a consequence of this, it honors basic human rights (part IV). One kind of nonliberal society satisfying these conditions is illustrated by what I call, for lack of a better term, a well-ordered hierarchical society. This example makes the point, central for this argument, that although any society must honor basic human rights, it need not be liberal. It also indicates the role of human rights as part of a reasonable law of peoples.

I How a Social Contract Doctrine is Universal in Its Reach

I begin by explaining the way in which a social contract doctrine with its procedure of construction is universal in its reach.

Every society must have a conception of how it is related to other societies and of how it is to conduct itself toward them. It lives with them in the same world and except for the very special case of isolation of a society from all the rest – long in the past now – it must formulate certain ideals and principles for guiding its policies toward other peoples. Like justice as fairness, the more general liberal conception I have in mind – as specified in Part III – begins with the case of a hypothetically closed and self-sufficient liberal democratic society and covers only political values and not all of life. The question now arises as to how that conception can be extended in a convincing way to cover a society’s relations with other societies to yield a reasonable law of peoples. In the absence of this extension to the law of peoples, a liberal conception of political justice would appear to be historicist and to apply only to societies whose political institutions and culture are liberal. In making the case for justice as fairness, and for similar more general liberal conceptions, it is essential to show that this is not so.

The problem of the law of peoples is only one of several problems of extension for these ideas of justice. There is the additional problem of extending these ideas to future generations, under which falls the problem of just savings. Also, since the ideas of justice regard persons as normal and fully cooperating members of society over a complete life, and having the requisite capacities to do this, there arises the problem of what is owed to those who fail to meet this condition, either temporarily or permanently, which gives rise to several problems of justice in health care. Finally, there is the problem of what is owed to animals and the rest of nature.

We would eventually like an answer to all these questions, but I doubt that we can find one within the scope of these ideas of justice understood as political conceptions. At best they may yield reasonable answers to the first three problems of extension: to other societies, to future generations, and to certain cases of health care. With regard to the problems which these liberal ideas of justice fail to address, there are several things we might say. One is that the idea of political justice does not cover everything and we should not expect it to. Or the problem may indeed be one of political justice but none of these ideas is correct for the question at hand, however well they may do for other questions. How deep a fault this shows must wait until the question itself can be examined, but we should not expect these ideas, or I think any account of political justice, to handle all these matters.

Let’s return to our problem of extending liberal ideas of justice similar to but more general than justice as fairness to the law of peoples. There is a clear contrast between these and other familiar views in the way they are universal in reach. Take, for example, Leibniz’s or Locke’s doctrines: These are universal both in their source of authority and in their formulation. By that I mean that their source is God’s authority or the divine reason, as the case may be; and they are universal in that their principles are stated so as to apply to all reasonable beings everywhere. Leibniz’s doctrine is an ethics of creation. It contains the idea of morals as the imitatio Dei and applies straightway to us as God’s creatures endowed with reason. In Locke’s doctrine, God having legitimate authority over all creation, the natural law – that part of God’s law that can be known by our natural powers of reason – everywhere has authority and binds us and all peoples.

Most familiar philosophical views – such as rational intuitionism, (classical) utilitarianism, and perfectionism – are also formulated in a general way to apply to us directly in all cases. Although they are not theologically grounded, let’s say their source of authority is (human) reason, or an independent realm of moral values, or some other proposed basis of universal validity. In all these views the universality of the doctrine is the direct consequence of its source of authority and of how it is formulated.

By contrast, a constructivist view such as justice as fairness, and more general liberal ideas, do not begin from universal first principles having authority in all cases³. In justice as fairness the principles of justice for the basic structure of society are not suitable as fully general principles: They do not apply to all subjects, not to churches and universities, or to the basic structures of all societies, or to the law of peoples. Rather, they are constructed by way of a reasonable procedure in which rational parties adopt principles of justice for each kind of subject as it arises. Typically, a constructivist doctrine proceeds by taking up a series of subjects, starting, say, with principles of political justice for the basic structure of a closed and self-contained democratic society. That done, it then works forward to principles for the claims of future generations, outward to principles for the law of peoples, and inward to principles for special social questions. Each time the constructivist procedure is modified to fit the subject in question. In due course all the main principles are on hand, including those needed for the various political duties and obligations of individuals and associations⁴. Thus, a constructivist liberal doctrine is universal in its reach once it is extended to give principles for all politically relevant subjects, including a law of peoples for the most comprehensive subject, the political society of peoples. Its authority rests on the principles and conceptions of practical reason, but always on these as suitably adjusted to apply to different subjects as they arise in sequence; and always assuming as well that these principles are endorsed on due reflection by the reasonable agents to whom the corresponding principles apply.

At first sight, a constructivist doctrine of this kind appears hopelessly unsystematic. For how are the principles that apply to different cases tied together? And why do we proceed through the series of cases in one order rather than another? Constructivism assumes, however, that there are other forms of unity than that defined by completely general first principles forming a consistent scheme. Unity may also be given by an appropriate sequence of cases and by supposing that the parties in an original position (as I have called it) are to proceed through the sequence with the understanding that the principles for the subject of each later agreement are to be subordinate to those of subjects of all earlier agreements, or else coordinated with and adjusted to them by certain priority rules. I shall try out a particular sequence and point out its merits as we proceed. There is in advance no guarantee that it is the most appropriate sequence and much trial and error may be needed.

In developing a conception of justice for the basic structure or for the law of peoples, or indeed for any subject, constructivism does not view the variation in numbers of people alone as accounting for the appropriateness of different principles in different cases. That families are smaller than constitutional democracies does not explain why different principles apply to them. Rather, it is the distinct structure of the social framework, and the purpose and role of its various parts and how they fit together, that explain why there are different principles for different kinds of subjects. Thus, it is characteristic of a constructivist idea of justice to regard the distinctive nature and purpose of the elements of society, and of the society of peoples, as requiring persons, within a domain where other principles leave them free, to act from principles designed to fit their peculiar roles. As we shall see as we work out the law of peoples, these principles are identified in each case by rational agents fairly, or reasonably, situated given the case at hand. They are not derived from completely general principles such as the principle of utility or the principle of perfectionism.

II Three Preliminary Questions

Before showing how the extension to the law of peoples can be carried out, I go over three preliminary matters. First, let’s distinguish between two parts of justice as fairness, or of any other similar liberal and constructivist conception of justice. One part is worked up to apply to the domestic institutions of democratic societies, their regime and basic structure, and to the duties and obligations of citizens. The other part is worked up to apply to the society of political societies and thus to the political relations between peoples⁵. After the principles of justice have been adopted for domestic justice, the idea of the original position is used again at the next higher level⁶. As before, the parties are representatives, but now they are representatives of peoples whose basic institutions satisfy the principles of justice selected at the first level. We start with the family of societies, each well-ordered by some liberal view meeting certain conditions (justice as fairness is an example), and then work out principles to govern their relations with one another. Here I mention only the first stage of working out the law of peoples. As we shall see in part IV, we must also develop principles which govern the relations between liberal and what I shall call hierarchical societies. It turns out that liberal and hierarchical societies can agree on the same law of peoples and thus this law does not depend on aspects peculiar to the Western tradition.

It may be objected that to proceed in this way is to accept the state as traditionally conceived, with all its familiar powers of sovereignty. These powers include first, the right to go to war in pursuit of state policies – Clausewitz’s pursuit of politics by other means – with the aims of politics given by a state’s rational prudential interests⁷. They include second, the state’s right to do as it likes with people within its own borders. The objection is misapplied for this reason. In the first use of the original position domestic society is seen as closed, since we abstract from relations with other societies. There is no need for armed forces and the question of the government’s right to be prepared militarily does not arise, and would be denied if it did. The principles of domestic justice allow a police force to keep domestic order but that is another matter, and although those domestic principles are consistent with a qualified right of war in a society of peoples, they do not of themselves support that right. That is up to the law of peoples itself, still to be constructed. And, as we shall see, this law will also restrict a state’s internal sovereignty, its right to do as it likes to people within its borders.

Thus, it is important to see that in this working out of the law of peoples, a government as the political organization of its people is not, as it were, the author of its own power. The war powers of governments, whatever they should be, are only those acceptable within a reasonable law of peoples. Presuming the existence of a government whereby a people is domestically organized with institutions of background justice does not prejudge these questions. We must reformulate the powers of sovereignty in light of a reasonable law of peoples and get rid of the right to war and the right to internal autonomy, which have been part of the (positive) international law for the two and a half centuries following the Thirty Years War, as part of the classical states system⁸.

Moreover, these ideas accord with a dramatic shift in how international law is now understood. Since World War II international law has become far more demanding than in the past. It tends to restrict a state’s right to wage war to cases of self-defense (this allows collective security), and it also tends to limit a state’s right of internal sovereignty⁹. The role of human rights connects most obviously with the latter change as part of the effort to provide a suitable definition of, and limits on, a government’s internal sovereignty, though it is not unconnected with the first. At this point I leave aside the many difficulties of interpreting these rights and limits, and take their general meaning and tendency as clear enough. What is essential is that our elaboration of the law of peoples should fit – as it turns out to do – these two basic changes, and give them a suitable rationale.

The second preliminary matter concerns the question: In working out the law of peoples, why do we start (as I said above) with those societies well-ordered by liberal views somewhat more general than justice as fairness? Wouldn’t it be better to start with the world as a whole, with a global original position, so to speak, and discuss the question whether, and in what form, there should be states, or peoples, at all? Some writers (I mention them later) have thought that a social contract constructivist view should proceed in this manner, that it gives an appropriate universality from the start.

I think there is no clear initial answer to this question. We should try various alternatives and weigh their pluses and minuses. Since in working out justice as fairness I begin with domestic society, I shall continue from there as if what has been done so far is more or less sound. Thus I build on the steps taken until now, as this seems to provide a suitable starting point for the extension to the law of peoples. A further reason for proceeding thus is that peoples as corporate bodies organized by their governments now exist in some form all over the world. Historically speaking, all principles and standards proposed for the law of peoples must, to be feasible, prove acceptable to the considered and reflective public opinion of peoples and their governments.

Suppose, then, that we are (even though we are not) members of a well-ordered society. Our convictions about justice are roughly the same as those of citizens (if there are any) in the family of societies well-ordered by liberal conceptions of justice and whose social and historical conditions are similar to ours. They have the same kinds of reasons for affirming their mode of government as we do for affirming ours. This common understanding of liberal societies provides an apt starting point for the extension to the law of peoples.

Finally, I note the distinction between the law of peoples and the law of nations, or international law. The latter is an existing, or positive, legal order, however incomplete it may be in some ways, lacking, for example, an effective scheme of sanctions such as normally characterizes domestic law. The law of peoples, by contrast, is a family of political concepts with principles of right, justice, and the common good, that specify the content of a liberal conception of justice worked up to extend to and to apply to international law. It provides the concepts and principles by which that law is to be judged.

This distinction between the law of peoples and the law of nations should be straightforward. It is no more obscure than the distinction between the principles of justice that apply to the basic structure of domestic society and the existing political, social, and legal institutions that actually realize that structure.

III The Extension to Liberal

The three preliminary matters settled, I turn to the extension of liberal ideas of justice to the law of peoples. I understand these ideas of justice to contain three main elements: (i) a list of certain basic rights and liberties and opportunities (familiar from constitutional democratic regimes); (ii) a high priority for these fundamental freedoms, especially with respect to claims of the general good and of perfectionist values; and (iii) measures assuring for all citizens adequate all-purpose means to make effective use of their freedoms. Justice as fairness is typical of these conceptions except that its egalitarian features are stronger. To some degree the more general liberal ideas lack the three egalitarian features of the fair value of the political liberties, of fair equality of opportunity, and of the difference principle. These features are not needed for the construction of a reasonable law of peoples and by not assuming them our account has greater generality.

There are two main stages to the extension to the law of peoples and each stage has two steps. The first stage of the extension I call the ideal, or strict compliance, theory, and unless otherwise stated, we work entirely in this theory. This means that the relevant concepts and principles are strictly complied with by all parties to the agreements made and that the requisite favorable conditions for liberal or hierarchical institutions are on hand. Our first aim is to see what a reasonable law of peoples, fully honored, would require and establish in this case.

To make the account manageable, we suppose there are only two kinds of well-ordered domestic societies, liberal societies and hierarchical societies. I discuss at the first step the case of well-ordered liberal democratic societies. This leads to the idea of a well-ordered political society of societies of democratic peoples. After this I turn to societies that are well-ordered and just, often religious in nature and not characterized by the separation of church and state. Their political institutions specify a just consultation hierarchy, as I shall say, while their basic social institutions satisfy a conception of justice expressing an appropriate conception of the common good. Fundamental for our rendering of the law of peoples is that both liberal and hierarchical societies accept it. Together they are members in good standing of a well-ordered society of the just peoples of the world.

The second stage in working out the law of peoples is that of nonideal theory, and it also includes two steps. The first step is that of noncompliance theory. Here we have the predicament of just societies, both democratic and hierarchical, as they confront states that refuse to comply with a reasonable law of peoples. The second step of this second stage is that of unfavorable conditions. It poses the different problem how the poorer and less technologically advanced societies the world can attain historical and social conditions that allow them to establish just and workable institutions, either liberal or hierarchical. In actual affairs, nonideal theory is of first practical importance and deals with problems we face every day. Yet, for reasons of space, I shall say very little about it (parts VI and VII).

Before beginning the extension we need to be sure that the original position with the veil of ignorance is a device of representation for the case of liberal societies. In the first use of the original position, its function as a device of representation means that it models what we regard – you and I, and here and now¹⁰ – as fair conditions for the parties, as representatives of free and equal citizens, to specify the terms of cooperation regulating the basic structure of their society. Since that position includes the veil of ignorance, it also models what we regard as acceptable restrictions on reasons for adopting a political conception of justice. Therefore, the conception the parties would adopt identifies the conception of justice that we regard – you and I, here and now – as fair and supported by the best reasons.

Three conditions are essential: First, the original position represents the parties (or citizens) fairly, or reasonably; second, it represents them as rational; and third, it represents them as deciding between available principles for appropriate reasons. We check that these three conditions are satisfied by observing that citizens are indeed represented fairly, or reasonably, in virtue of the symmetry and equality of their representatives’ situation in the original position. Next, citizens are represented as rational in virtue of the aim of their representatives to do the best they can for their essential interests as persons. Finally, they are represented as deciding for appropriate reasons: The veil of ignorance prevents their representatives from invoking reasons deemed unsuitable, given the aim of representing citizens as free and equal persons.

At the next level, when the original position is used to extend a liberal conception to the law of peoples, it is a device of representation because it models what we would regard – you and I, here and now¹¹ – as fair conditions under which the parties, this time as representatives of societies well-ordered by liberal conceptions of justice, are to specify the law of peoples and the fair terms of their cooperation.

The original position is a device of representation because, as before, free and equal peoples are represented as both reasonably situated and rational, and as deciding in accordance with appropriate reasons. The parties as representatives of democratic peoples are symmetrically situated, and so the peoples they represent are represented reasonably. Moreover, the parties deliberate among available principles for the law of peoples by reference to the fundamental interests of democratic societies in accordance with, or as presupposed by, the liberal principles of domestic justice. And finally, the parties are subject to a veil of ignorance: They do not know, for example, the size of the territory, or the population, or the relative strength of the people whose fundamental interests they represent. Although they know that reasonably favorable conditions obtain that make democracy possible, they do not know the extent of their natural resources, or level of their economic development, or any such related information. These conditions model what we, as members of societies well-ordered by liberal conceptions of justice, would accept as fair – here and now – in specifying the basic terms of cooperation between peoples who, as peoples, regard themselves as free and equal. We use the original position at the second level as a device of representation as we did at the first.

I assume that working out the law of peoples for liberal democratic societies only will result in the adoption of certain familiar principles of justice, and will also allow for various forms of cooperative association among democratic peoples and not for a world state. Here I follow Kant’s lead in Perpetual Peace (1795) in thinking that a world government – by which I mean a unified political regime with the legal powers normally exercised by central governments – would be either a global despotism or else a fragile empire torn by frequent civil strife as various regions and peoples try to gain political autonomy¹². On the other hand, it may turn out, as I sketch below, that there will be many different kinds of organizations subject to the judgment of the law of democratic peoples, charged with regulating cooperation between them, and having certain recognized duties. Some of these organizations (like the United Nations) may have the authority to condemn domestic institutions that violate human rights, and in certain severe cases to punish them by imposing economic sanctions, or even by military intervention. The scope of these powers is all peoples’ and covers their domestic affairs.

If all this is sound, I believe the principles of justice between free and democratic peoples will include certain familiar principles long recognized as belonging to the law of peoples, among them the following:

1. Peoples (as organized by their governments) as free and independent and their freedom and independence is to be respected by other peoples.
2. Peoples are equal and parties to their own agreements.
3. Peoples have the right of self-defense but no right to war.
4. Peoples are to observe a duty of nonintervention.
5. Peoples are to observe treaties and undertakings.
6. Peoples are to observe certain specified restrictions on the conduct of war (assumed to be in self-defense).
7. Peoples are to honor human rights.

This statement of principles is of course incomplete; other principles would need to be added. Further, they require much explanation and interpretation, and some of them are superfluous in a society of well-ordered democratic peoples, for instance, the sixth regarding the conduct of war and the seventh regarding human rights. The main point is that given the idea of a society of free and independent democratic peoples, who are ready to recognize certain basic principles of political justice governing their conduct, principles of this kind constitute the charter of their association¹³. Obviously, a principle such as the fourth – that of nonintervention – will have to be qualified in the general case. Although suitable for a society of well-ordered democratic peoples who respect human rights, it fails in the case of disordered societies in which wars and serious violations of human rights are endemic. Also, the right to independence, and equally the right to self-determination, hold only within certain limits, to be specified by the law of peoples for the general case. Thus, no people has the right to self-determination, or a right to secession, at the expense of the subjugation of another people¹⁴; nor can a people protest their condemnation by the world society when their domestic institutions violate the human rights of certain minorities living among them. Their right to independence is no shield from that condemnation, or even from coercive intervention by other peoples in grave cases.

There will also be principles for forming and regulating federations (associations) of peoples, and standards of fairness for trade and other cooperative arrangements. There should be certain provisions for mutual assistance between peoples in times of famine and drought, and were it feasible, as it should be, provisions for ensuring that in all reasonably developed liberal societies people’s basic needs are met¹⁵. These provisions will specify duties of assistance in certain situations, and they will vary in stringency depending on the severity of the case.

An important role of a people’s government, however arbitrary a society’s boundaries may appear from a historical point of view¹⁶, is to be the representative and effective agent of a people as they take responsibility for their territory and the size of their population, as well as for maintaining its environmental integrity and its capacity to sustain them. The idea here appeals to the point of the institution of property: Unless a definite agent is given responsibility for maintaining an asset and bears the loss for not doing so, that asset tends to deteriorate. In this case the asset is the people’s territory and its capacity to sustain them in perpetuity; the agent is the people themselves as politically organized. They must recognize that they cannot make up for irresponsibility in caring for their land and conserving their natural resources by conquest in war or by migrating into other people’s territory without their consent¹⁷.

These remarks belong to ideal theory and indicate some of the responsibilities of peoples in a just society of well-ordered liberal societies. Since the boundaries of peoples are often historically the outcome of violence and aggression, and some peoples are wrongly subjected to others, the law of peoples in its nonideal part should, as far as possible, contain principles and standards – or at least some guidelines – for coping with these matters.

To complete this sketch of the law of peoples for well-ordered liberal societies only, let’s consider under what conditions we can reasonably accept this part of the law of peoples and regard it as justified.

There are two conditions beyond the three requirements earlier noted in discussing the original position as a device of representation. These requirements were: that the parties (as representatives of free and equal peoples) be represented as reasonably situated, as rational, and as deciding in accordance with appropriate reasons. One of the two further conditions is that the political society of well-ordered democratic peoples should itself be stable in the right way¹⁸. Given the existence of a political society of such peoples, its members will tend increasingly over time to accept its principles and judgments as they come to understand the ideas of justice expressed in the law among them and appreciate its benefits for all liberal peoples.

To say that the society of democratic peoples is stable in the right way is to say that it is stable with respect to justice, that is, that the institutions and practices among peoples always more or less satisfy the relevant principles of justice, although social conditions are presumably always changing. It is further to say that the law of peoples is honored not simply because of a fortunate balance of power – it being in no people’s interest to upset it – but because, despite the possibly shifting fortunes of different peoples, all are moved to adhere to their common law accepting it as just and beneficial for all. This means that the justice of the society of democratic peoples is stable with respect to the distribution of fortune among them. Here fortune refers not to a society’s military success or the lack of it, but to other kinds of success: its achievement of political and social freedom, the fullness and expressiveness of its culture, the economic well-being of all its citizens.

The historical record suggests that, at least so far as the principle against war is concerned, this condition of stability would be satisfied in a society of just democratic peoples. Although democratic societies have been as often involved in war as nondemocratic states¹⁹ and have often vigorously defended their institutions, since 1800, as Michael Doyle points out, firmly established liberal societies have not gone to war with one another²⁰. And in wars in which a number of major powers were engaged, such as the two World Wars, democratic states have fought as allies on the same side. Indeed, the absence of war between democracies is as close as anything we know to an empirical law in relations between societies²¹. This being so, I shall suppose that a society of democratic peoples, all of whose basic institutions are well ordered by liberal conceptions of justice (though not necessarily by the same conception) will be stable in the right way as above specified. The sketch of the law of such peoples therefore seems to meet the condition of political realism given by that of stability for the right reasons.

Observe that I state what I call Doyle’s law as holding between well-established and well-ordered liberal democracies that are significant if not major powers. The reasons for this law’s holding (supposing it does) are quite compatible with actual democracies, marked as they are by considerable injustice and oligarchic tendencies, intervening, often covertly, in smaller countries whose democracies are less well established and secure. Witness the United States’ overturning the democracies of Allende in Chile, Arbenz in Guatemala, Mossadegh in Iran, and, some would add, the Sandinistas in Nicaragua. Whatever the merits of these regimes, covert operations against them can be carried out by a government bureaucracy at the urging of oligarchic interests without the knowledge or criticism of the public, and presenting it with a fait accompli. All this is made easier by the handy appeal to national security given the situation of superpower rivalry in the Cold War, which allowed those democracies, however implausibly, to be cast as a danger. While democratic peoples are not expansionist, they do defend their security interest, and this an oligarchic government can easily manipulate in a time of superpower rivalry to support covert interventions once they are found out²².

The last condition for us to accept this sketch of the law of democratic peoples as sound is that we can, as citizens of liberal societies, endorse the principles and judgments of this law on due reflection. We must be able to say that the doctrine of the law of peoples for such societies, more than any other doctrine, ties together our considered political convictions and moral judgments at all levels of generality, from the most general to the more particular, into one coherent view.

IV Extension to Hierarchical Societies

Recall from part III that the extension of liberal ideas of justice to the law of peoples proceeds in two stages, each stage having two steps. The first stage is that of ideal theory and we have just completed the first step of that: the extension of the law of peoples to well-ordered liberal societies only. The second step of ideal theory is more difficult: It requires us to specify a second kind of society – a hierarchical society, as I shall say – and then to state when such a society is well ordered. Our aim is to extend the law of peoples to these well-ordered hierarchical societies and to show that they accept the same law of peoples as liberal societies do. Thus, this shared law of well-ordered peoples, both liberal and hierarchical, specifies the content of ideal theory. It specifies the kind of society of well-ordered peoples all people should want and it sets the regulative end of their foreign policy. Important for us, it has the obvious corollary that nonliberal societies also honor human rights.

To show all this we proceed thus. First, we state three requirements for any well-ordered hierarchical regime. It will be clear that satisfying these requirements does not entail that a regime be liberal. Next, we confirm that, in an original position with a veil of ignorance, the representatives of well-ordered hierarchical regimes are reasonably situated as well as rational, and are moved by appropriate reasons. In this case also, the original position is a device of representation for the adoption of law among hierarchical peoples. Finally, we show that in the original position the representatives of well-ordered hierarchical societies would adopt the same law of peoples that the representatives of liberal societies do. That law thus serves as a common law of a just political society of well-ordered peoples.

The first of the three requirements for a hierarchical society to be well ordered is that it must be peaceful and gain its legitimate aims through diplomacy and trade, and other ways of peace. It follows that its religious doctrine, assumed to be comprehensive and influential in government policy, is not expansionist in the sense that it fully respects the civic order and integrity of other societies. If it seeks wider influence, it does so in ways compatible with the independence of, and the liberties within, other societies. This feature of its religion supports the institutional basis of its peaceful conduct and distinguishes it from leading European states during the religious wars of the sixteenth and seventeenth centuries.

A second fundamental requirement uses an idea of Philip Soper. It has several parts. It requires first, that a hierarchical society’s system of law be such as to impose moral duties and obligations on all persons within its territory²³. It requires further that its system of law be guided by a common good conception of justice, meaning by this a conception that takes impartially into account what it sees not unreasonably as the fundamental interests of all members of society. It is not the case that the interests of some are arbitrarily privileged, while the interests of others go for naught. Finally, there must be sincere and not unreasonable belief on the part of judges and other officials who administer the legal order that the law is indeed guided by a common good conception of justice. This belief must be demonstrated by a willingness to defend publicly the state’s injunctions as justified by law. Courts are an efficient way of doing this²⁴. These aspects of a legal order are necessary to establish a regime’s legitimacy in the eyes of its own people. To sum up the second requirement we say: The system of law is sincerely and not unreasonably believed to be guided by a common good conception of justice. It takes into account people’s essential interests and imposes moral duties and obligations on all members of society.

This second requirement can be spelled out further by adding that the political institutions of a well-ordered hierarchical society constitute a reasonable consultation hierarchy. They include a family of representative bodies, or other assemblies, whose task is to look after the important interests of all elements of society. Although in hierarchical societies persons are not regarded as free and equal citizens, as they are in liberal societies, they are seen as responsible members of society who can recognize their moral duties and obligations and play their part in social life.

With a consultation hierarchy there is an opportunity for different voices to be heard, not, to be sure, in a way allowed by democratic institutions, but appropriately in view of the religious and philosophical values of the society in question. Thus, individuals do not have the right of free speech as in a liberal society; but as members of associations and corporate bodies they have the right at some point in the process of consultation to express political dissent and the government has an obligation to take their dissent seriously and to give a conscientious reply. That different voices can be heard is necessary because the sincere belief of judges and other officials has two components: honest belief and respect for the possibility of dissent²⁵. Judges and officials must be willing to address objections. They cannot refuse to listen to them on the grounds that they think those expressing them are incompetent and cannot understand. Then we would not have a consultation hierarchy but a purely paternalistic regime.

In view of this account of the institutional basis of a hierarchical society, we can say that its conception of the common good of justice secures for all persons at least certain minimum rights to means of subsistence and security (the right to life)²⁶, to liberty (freedom from slavery, serfdom, and forced occupations) and (personal) property, as well as to formal equality as expressed by the rules of natural justice²⁷ (for example, that similar cases be treated similarly). This shows that a well-ordered hierarchical society also meets a third requirement: it respects basic human rights.

The argument for this conclusion is that the second requirement rules out violations of these rights. For to satisfy it, a society’s legal order must impose moral duties and obligations on all persons in its territory and it must embody a reasonable consultation hierarchy which will protect human rights. A sincere and reasonable belief on the part of judges and other officials that the system of law is guided by a common good conception of justice has the same result. Such a belief is simply unreasonable, if not irrational, when those rights are infringed.

There is a question about religious toleration that calls for explicit mention. Whereas in hierarchical societies a state religion may be on some questions the ultimate authority within society and control government policy on certain important matters, that authority is not (as I have said) extended politically to other societies. Further, their (comprehensive) religious or philosophical doctrines are not unreasonable: They admit a measure of liberty of conscience and freedom of thought, even if these freedoms are not in general equal for all members of society as they are in liberal regimes²⁸. A hierarchical society may have an established religion with certain privileges. Still, it is essential to its being well-ordered that no religions are persecuted, or denied civic and social conditions that permit their practice in peace and without fear²⁹. Also essential, and this because of the inequality of religious freedom, if for no other reason, is that a hierarchical society must allow for the right of emigration³⁰. The rights noted here are counted as human rights. In part V we return to the role and status of these rights.

An institutional basis that realizes the three requirements can take many forms. This deserves emphasis, as I have indicated only the religious case. We are not trying to describe all possible forms of social order consistent with membership in good standing of a reasonable society of peoples. Rather, we have specified three necessary conditions for membership of a reasonable society of peoples and then shown by example that these conditions do not require a society to be liberal.

This completes the account of the requirements imposed on the basic institutions of a well-ordered hierarchical society. My aim has been to outline a conception of justice that, although distant from liberal conceptions, still has features that give to societies regulated accordingly the moral status required to be members in good standing in a reasonable society of well-ordered peoples. It is important to see, as I have noted, that an agreement on a law of peoples ensuring human rights is not an agreement only liberal societies can make. We must now confirm this.

Hierarchical societies are well-ordered in terms of their own conceptions of justice³¹. This being so, their representatives in an appropriate original position would adopt the same principles as those sketched above that would be adopted by the representatives of liberal societies. Each hierarchical society’s interests are understood by its representatives in accordance with or as presupposed by its conception of justice. This enables us to say in this case also that the original position is a device of representation.

Two considerations confirm this. The first is that, in view of the common good conception of justice held in a hierarchical society, the parties care about the good of the society they represent, and so about its security as assured by the laws against war and aggression. They also care about the benefits of trade and assistance between peoples in time of need. All these help protect human rights. In view of this, we can say that the representatives of hierarchical societies are rational. The second consideration is that they do not try to extend their religious and philosophical doctrines to other peoples by war or aggression, and they respect the civic order and integrity of other societies. Hence, they accept – as you and I would accept³² – the original position as fair between peoples and would endorse the law of peoples adopted by their representatives as specifying fair terms of political cooperation between them and other societies. Thus, the representatives are reasonably situated and this suffices for the use of the original position as a device of representation in extending the law of peoples to hierarchical societies³³.

Note that I have supposed that the parties as representatives of peoples are to be situated equally, even though the conception of justice of the hierarchical society they represent allows basic inequalities between its members. For example, some of its members are not granted equal liberty of conscience. There is, however, no inconsistency in this: A people sincerely affirming a nonliberal conception of justice may still think their society should be treated equally in a just law of peoples, even though its members accept basic inequalities among themselves. Though a society lacks basic equality, it is not unreasonable for that society to insist on equality in making claims against other societies.

About this last point, two observations. One is that although the original position at the first level, that of domestic justice, incorporates a political conception of the person rooted in the public culture of a liberal society, the original position at the second level, that of the law of peoples, does not. I emphasize this fact, since it enables a liberal conception of justice to be extended to yield a more general law of peoples without prejudging the case against nonliberal societies.

This leads to a second observation. As mentioned earlier, the law of peoples might have been worked out by starting with an all-inclusive original position with representatives of all the individual persons of the world³⁴. In this case the question of whether there are to be separate societies and of the relations between them, will be settled by the parties behind a veil of ignorance. Offhand it is not clear why proceeding this way should lead to different results than, as I have done, proceeding from separate societies outward. All things considered, one might reach the same law of peoples in either case. The difficulty with an all-inclusive, or global, original position is that its use of liberal ideas is much more troublesome, for in this case we are treating all persons, regardless of their society and culture, as individuals who are free and equal, and as reasonable and rational, and so according to liberal conceptions. This makes the basis of the law of peoples too narrow.

Hence I think it best to follow the two-level³⁵ bottom-up procedure, beginning first with the principles of justice for the basic structure of domestic society and then moving upward and outward to the law of peoples. In so doing our knowledge of how peoples and their governments have acted historically gives us guidance in how to proceed and suggests questions and possibilities we might not otherwise have thought of. But this is simply a point of method and settles no questions of substance. These depend on what can actually be worked out.

One might well be skeptical that a liberal social contract and constructivist³⁶ idea of justice can be worked out to give a conception of the law of peoples universal in its reach and also applying to nonliberal societies. Our discussion of hierarchical societies should put these doubts to rest. I have noted the conditions under which we could accept the law of liberal peoples we had sketched as sound and justified. In this connection we considered whether that law was stable with respect to justice, and whether, on due reflection, we could accept the judgments that its principles and precepts led us to make. If both these things hold, we said, the law of liberal peoples as laid out could, by the criteria we can now apply, be accepted as justified.

Parallel remarks hold for the wider law of peoples including well-ordered hierarchical societies. Here I simply add, without argument or evidence, but hoping it seems plausible, that these societies will honor a just law of peoples for much the same reasons liberal peoples will do so, and that both we and they will find the judgments to which it leads acceptable to our convictions, all things considered. I believe it is of importance here that well-ordered hierarchical societies are not expansionist and their legal order is guided by a common good conception of justice ensuring that it honors human rights. These societies also affirm a peaceful society of peoples and benefit therefrom as liberal societies do. All have a common interest in changing the way in which politics among peoples – war and threats of war – has hitherto been carried on.

We may therefore view this wider law of peoples as sound and justified. This fundamental point deserves emphasis: There is nothing relevantly different between how, say, justice as fairness is worked out for the domestic case in A Theory of Justice, and how the law of peoples is worked out from more general liberal ideas of justice. In both cases we use the same fundamental idea of a reasonable procedure of construction in which rational agents fairly situated (the parties as representatives of citizens in one case and of peoples or societies in the other) select principles of justice for the relevant subject, either their separate domestic institutions or the shared law of peoples. As always, the parties are guided by the appropriate reasons as specified by a veil of ignorance. Thus, obligations and duties are not imposed by one society on another; instead, reasonable societies agree on what the bonds will be. Once we confirm that a domestic society, or a society of peoples, when regulated by the corresponding principles of justice, is stable with respect to justice (as previously defined), and once we have checked that we can endorse those principles on due reflection, then in both domains the ideals, laws, and principles of justice are justified in the same way³⁷.

[ Continued in Issue No. 3-4/1996 • 1-2/1997 ]

Webmaster: Slobodan Markovic (twiddle@eunet.yu)

All rights reserved. Belgrade Circle Journal encourages the reproduction of material appearing on its pages, provided that the source and the author are cited, except in cases where this would constitute violation of copyright held by other organisations or individuals.