The Democratic Peace is Not Democratic: On Behalf of Rawls' Decent Societies

Two Constitutional Republics: Liberal Democratic and Decent (Hierarchical)

Rawls does not argue that all non-democratic, non-liberal societies deserve full and good standing in the international community. He argues only that constitutional republics do. But, he claims, there are at least two kinds of constitutional republics: liberal democracies and non-liberal, non-democratic ‘decent’ societies. ⁴ Both deserve a seat at the international table, because both are artificial corporate moral persons that have a right to a significant measure of self-determination. Both are what Rawls calls ‘peoples' (LP, pp. 23–30). Peoples, both democratic and decent, deserve full and good standing in the international community.

Democracies and decent societies are artificial corporate moral persons because they possess several characteristics (LP, pp. 23–30). First, they possess sufficient cultural unity to allow them to secure identifiable and determinate ends as a group. Their members have a feeling of nationality that results from, for example, shared language, shared religion, geographical limits, a shared national history and a shared sense of their achievements and failures as a group. Second, they have the institutional means to further their ends. For instance, they are not too decentralized or poor to develop, make known and pursue their interests effectively in the world. Third, they possess a moral character. They have the capacity to regulate pursuit of their national ends according to principles of international justice, and are willing to offer and abide by fair terms of cooperation. Finally, the members of these societies constitute and govern themselves as a body politic through a public and (largely) voluntarily affirmed conception of justice, without excessive coercion, manipulation or deception (i.e. in Rawls' terms they are ‘well ordered’; LP, p. 4, p. 19, pp. 64–7). They recognize that justice is centered on the inviolability of persons and understand their society to be a system of cooperation aimed at and justified by the good of all individual members, and not by some notion of aggregate or corporate good. None of this is meant to imply controversial ontological claims (Costa, 2005, p. 50). It simply reflects the fact that we often talk about different collectives as behaving fairly or unfairly with others.

Not all societies deserve full and good standing in the international community. Outlaw states do not (LP, p. 4, p. 90). Outlaw states will go to war simply to advance their own interests, sacrificing the reasonable interests of other states in the process. They often reject international law, overstep traditional restrictions on the right to go to war, violate basic human rights and disrespect other societies. Nor do ‘burdened societies' deserve full standing (LP, p. 4, p. 90). These societies exist under unfavorable historical, social or economic conditions that prevent them from achieving and maintaining the characteristics of corporate persons. Burdened societies deserve assistance, but they cannot exercise the kind of self-determination that merits full standing in the international community. Last are ‘benevolent absolutisms' (LP, p. 4, p. 63). These societies are ruled by benevolent dictators. These societies honor most human rights, but are not well ordered, because members are denied any say in political decisions. For Rawls, this means that the law in benevolent absolutisms does not generate genuine normative duties for citizens, and thus that these societies are not legitimate structures of political authority. ⁵ This negates their good standing in the international community (though they represent, in many ways, the least of our worries).

In liberal democracies, individual persons are seen as free and equal citizens. These societies: guarantee for all citizens basic liberal rights and opportunities; assign special priority to these rights and opportunities against claims of the general good or perfectionist values; and provide social security sufficient to ensure the means to make effective use of individual freedoms (Rawls, 1999a, pp. 581–2). These features define a generic liberalism that is fleshed out in particular societies through public political discourse and real-time political processes. Many societies meet these generic requirements, including the US, Canada and Great Britain.

Decent societies are non-liberal and non-democratic, but they are well ordered, secure certain conditions of political right and justice for members and honor the principles of justice that constitute the law of peoples (LP, p. 37). As constitutional republics, they meet minimal conditions of justice and legitimacy. They recognize that constitutional republics, as corporate moral persons, are free and independent. They honor basic human rights, recognize a duty of non-intervention and a right to war only in self-defense, observe restrictions in the conduct of war and recognize a duty to assist individuals whose societies are not corporate moral persons.

Do decent societies exist? Identifying some is not vital for my purposes here, but the democratic peace thesis is compelling because of the empirical evidence that supports it. Thus we must eventually identify some decent societies, if we are to vindicate the peace argument. So what might count as an instantiation of a generically decent society? Most modern democracies were once decent. For instance, eighteenth-century Britain had some respect for human rights, a constitutional government and gave the people some measure of political representation, but it was clearly not fully democratic. Other democracies have had similar historical trajectories. One form of decent society, then, is the historical, developmental stage prior to full democracy. However, we should not limit decent societies to those moving toward democracy. Ken Jowitt (1996, p. 5) argues that such claims are as ideological as they are theoretical. Societies might be decent, even if they are not moving toward democracy. In any case, Oman might be a decent society (Reidy, 2004, p. 315). ⁶ China, with its non-liberal democratic centralism, might become a decent society, if certain reforms were made (Angle, 2005). Some have even suggested that Rawls' notion of the decent society is too restrictive, and ought to be expanded to include, for example, Egypt, Malaysia, Saudi Arabia, Singapore and Thailand (Brown, 2000, p. 132). Thus, many societies may satisfy the generic requirements of decent societies.

Rawls holds that decent societies can exhibit different institutional forms, but he gives just one example, the decent consultation hierarchy (or decent hierarchical society). ⁷ These societies are committed to a ‘common good idea of justice’ (LP, p. 71). They may be thought of as societies attempting to instantiate some comprehensive conception of the human good, i.e. a philosophical, moral or religious understanding of the good life, but within the bounds of a minimal (from the liberal perspective) commitment to justice and legitimacy. This latter commitment is captured by three characteristics. First, they secure human rights for their members, including (but not limited to) rights to subsistence and security, to liberty (freedom from slavery and freedom of conscience sufficient for freedom of thought and religion), to property and to formal equality. Second, the society's system of laws fits its common good idea of justice in a way that generates real obligations and duties. The members of decent societies are seen as rational and responsible persons (though not as free, equal citizens) who are cooperating members of their society. They must be able to recognize the relationship between their laws and their society's common good, in a way that allows them to see their political duties and responsibilities as contributing to the common good. Finally, the officials who administer the legal system must sincerely and not unreasonably believe that the laws are guided by their common good conception of justice.

Decent hierarchical societies are associationist (LP, p. 64, p. 68). ⁸ Both decent societies and democracies regard the basic structure of society as a genuine scheme of social cooperation between persons, but each organizes persons into a body politic differently. In decent societies, group membership mediates between persons and political authority, while in democracies there is no such mediation. Members of decent societies are seen as individual persons, each of whom is capable of understanding their common good idea of justice. However, from the perspective of public life, they are viewed fundamentally as members of different associations or groups, not as free and equal citizens. In decent consultation hierarchies, persons are associated with groups, the groups are organized into a body politic and each group is represented in the legal system. ⁹ Thus, group membership mediates between individual persons and political authority. Individual persons have a right to express political dissent at points in the political process, for example in selecting the group's representatives to the government, but there is no commitment to the ideal of ‘one person, one vote’. Representatives express the interests of each group before government officials. The government must listen to representatives, take dissent seriously, offer conscientious replies to complaints and allow representatives to renew dissent, when such protest is not unreasonable or inconsistent with their common good idea of justice. ¹⁰

As an example of a decent hierarchical people, Rawls describes an idealized Islamic society called ‘Kazanistan’ (LP, pp. 75–8). Kazanistan does not separate church and state. The favored religion is Islam, so only Muslims are allowed to hold the highest government positions. However, other religions are not only tolerated, but encouraged. Kazanistan has an enlightened view of non-Islamic religions, believing (inter alia) that religious differences are divinely willed, that God alone should punish false belief and that different religious communities should respect one another. Kazanistan is non-aggressive, having rejected military interpretations of jihad in favor of a moral and spiritual one. The Muslim rulers believe that individuals want to be loyal to their country, and will stay loyal if they are not treated unfairly. The government allows non-Muslim members to join the military and hold relatively high positions, in order to strengthen their loyalty.

Kazanistan has a decent consultation hierarchy, which changes over time in response to changes in its common good idea of justice. Representative bodies from all groups are consulted about political decisions, and each member of society is represented by at least one of these groups. The representative bodies include some members of the groups they represent. The rulers must weigh the views of each group before making final decisions, and offer not unreasonable explanations for decisions when these are called for by representatives. Decisions should be made according to the special priorities of Kazanistan, but within the limits imposed by the idea of social cooperation and the common good conception of justice; i.e. individuals cannot be sacrificed for the common or aggregate good. These special priorities would include, for example, establishing a decent Muslim society that respects the religious minorities that are part of it.

The Democratic Peace Thesis

The democratic peace thesis involves two related claims. The descriptive claim holds that war ¹¹ is much less likely to occur between (1) two democracies ¹² than it is to occur between either (2) a democracy and a non-democratic society, or (3) two non-democratic societies. This claimed correlation between democracy and peace is well supported by empirical evidence and is widely accepted by international relations scholars (e.g. Ray, 1998). ¹³ Rawls holds that ‘the absence of war between major established democracies is as close as anything we know to a simple empirical regularity in relations among societies' (LP, pp. 52–3; see also Levy, 1989). The second claim of the democratic peace thesis is causal. It holds that some feature(s) of democracy cause(s) the peace. Different theorists offer different accounts of the causal claim. Rawls' preferred causal account is a satisfaction account. Other common causal accounts are constitutional constraint accounts, re-election constraint accounts, normative accounts and interdependence accounts.

Most causal accounts of the democratic peace make no essential reference to democratic elections. This is true even of re-election constraint accounts, which do not depend on elections as such, but rather on the existence of institutional means of removing from office those political leaders who lose wars. This fact about causal accounts opens the door to the peace argument, because non-democratic societies might also have norms and institutions sufficient to support or produce a ‘democratic’ peace. Whether Rawls' decent peoples do is the question at the heart of this article.

In the following sections, I discuss the main causal accounts of the democratic peace, in order to draw out the conditions (of the peace) that each highlights. But the resulting set of peace conditions is not a list of necessary and sufficient conditions. In fact, many researchers argue that their account identifies the most important condition of the peace. Rawls thinks, for example, that the satisfaction account is sufficient (LP, pp. 44–54). In any case, few think it necessary to invoke all of the conditions of the different causal accounts to explain the peace. It might be secured through one or a couple of the relevant conditions. Nor is there reason to see the various peace conditions as mutually exclusive. It is likely that several conditions (though not necessarily all of them, and not necessarily the same sub-set in each case) are at work in any historical instance of democratic peace. I argue in the end that decent societies meet enough of the conditions identified in the democratic peace literature to justify the claim that their existence would not threaten the democratic peace.

But before I discuss the causal accounts, I want briefly to explain the basis and nature of the democratic peace thesis, in order to avoid some common misunderstandings. Some claim, for instance, that democracies are such a recent social phenomenon, historically speaking, that there cannot be much empirical support for the democratic peace thesis. And even if there were much empirical support, almost everyone can think of a counter-example, i.e. an example of a democracy engaging in war. How can the democratic peace thesis even get off the ground? These are reasonable questions, but they do not do justice to the work put into the democratic peace literature by international relations scholars and historians. So I will briefly describe the empirical research that is thought to support the democratic peace thesis, and discuss the significance of counter-examples. I do not claim, of course, that the democratic peace thesis cannot be challenged. I claim only that those who find it compelling draw on a vast, deep, historically and statistically sound literature and, thus, that any serious challenge to the democratic peace thesis must recognize this.

Empirical research on the democratic peace thesis is generally conducted by tracking pairs of states (‘dyads’) through time, often on a yearly basis, and comparing rates of warfare in jointly democratic dyads (pairs of democratic states) to rates in other types of dyads (pairs that include at least one non-democratic state). Some studies (e.g. Bremer, 1992; Maoz and Abdolali, 1989) track these dyads back to the early 1800s. Others (e.g. Maoz and Russett, 1993) hold that the post-Second World War era is most relevant to democratic peace research, since there were so few democracies around before 1939. It is not uncommon for democratic peace researchers to analyze statistically over 200,000 ‘observations' (a dyad tracked through one year is often counted as one observation). Zeev Maoz and Bruce Russett (1993) restricted their analysis to the post-Second World War period, and to ‘politically relevant’ dyads (those dyads that included geographically contiguous states or at least one major power), ¹⁴ and they still analyzed over 29,000 observations. They found no wars in jointly democratic dyads for the period 1946–86 and 32 wars in other kinds of dyads. They conclude that non-jointly democratic dyads are 0.1 per cent more likely to engage in war than are jointly democratic dyads. This difference is statistically significant and is comparable to the difference between the proportion of cigarette smokers who develop cancer and the proportion of non-smokers who develop it (Ray, 1998, p. 35). So it is unfair simply to dismiss the democratic peace thesis as unsupportable, because there are not enough democracies to study.

Democracies do engage in wars with other societies. The recent US war with Iraq is one example. Other examples are noted in the democratic peace literature: Hitler came to power by democratic means, the US engaged in imperialistic wars against Native Americans and the US was almost destroyed by its Civil War. ¹⁵ Does this not refute the democratic peace thesis? No. Democratic peace theorists claim that democracies are less likely to fight other democracies than they are to fight other kinds of regimes, and that other kinds of regimes are more likely to fight all others than are democracies. ¹⁶ Wars between democracies and non-democracies do not refute the democratic peace thesis; nor would a war between democratic states. Democratic peace researchers claim only to find a significant disposition toward peace among democracies. The question is not whether democracies fight wars – no-one claims otherwise – but whether democratic dyads are less warlike than other kinds of dyads. The literature suggests they are.

Satisfaction Accounts

Satisfaction accounts hold that democracies and the individuals that comprise them are satisfied, so they have no reason to go to war with each other (Aron, 1966; Rawls, LP, pp. 44–54). They are satisfied in three ways. First, the basic needs of individual members are met. Second, members of democracies, understood now as bodies politic, are satisfied with and have proper pride in the decency of their shared history and social and political institutions. Third, the interests of democracies, again understood as wholes, are compatible with the interests of other democracies.

According to Rawls' satisfaction account, reasonably just liberal constitutional democracies secure for their citizens: (1) fair equality of opportunity; (2) a decent distribution of income and wealth, such that all are assured the all-purpose means necessary for effective use of basic freedoms; (3) social security in the form of meaningful work provided by society as a last resort; (4) basic health care; and (5) public financing of elections (LP, pp. 48–51). By securing these goods, democracies protect basic liberties and prevent inequalities from becoming excessive. This satisfies individual members, which produces internal stability or domestic peace. Further, individual members need not seek to satisfy their interests at the expense of any other society; nor would any external observer expect to see internal unrest.

Rawls claims also that a global society of reasonably just democracies would be stable. They have no state religions, even when their citizens are religious, so they are not moved to conquer or dominate other societies for religious reasons. Democracies also possess proper self-respect, based on the freedom of their citizens, the decency of their institutions and their cultural achievements in the civic and public spheres. They ‘are not inflamed by what Rousseau diagnosed as arrogant or wounded pride or by lack of due self-respect’ (LP, p. 47). Nor would any other democracy perceive a need to intervene in the internal affairs of another stable democracy. On satisfaction accounts, there is nothing for democracies to go to war over.

The claim that the members of a society are satisfied does not imply that dissenters are non-existent. In fact, every satisfied society will have dissenters. In democracies there are individuals who reject even the most fundamental of institutional arrangements. For instance, some reject winner-take-all democracy, regarding it as incompatible with genuinely shared political authority. The existence of dissenters, however, does not produce instability. In democracies, individuals are allowed to express dissent, and reform is always possible. Actual reform may be unlikely, but the mere possibility keeps dissenters from despair and violence. Further, it is not clear how dissenters might secure reforms through wars with other states. Still, dissenters might seek aid from other states to mount coups, if they became hopeless and desperate.

Decent societies, and the individuals who comprise them, are satisfied too. The basic human rights of members are secured. Members understand their society to be consistent with their conception of justice. They thus feel a proper pride in their accomplishments as a people, for example for securing human rights for members, having decent institutions and having a vital civic and public life. This makes decent societies internally stable and makes it unnecessary for other societies to intervene in their domestic affairs. Decent societies respect the law of peoples and see all peoples as equal and free. They are non-aggressive and non-expansionist, and respect the duties of non-intervention and aid. Thus the interests of decent societies are compatible with those of democratic and other decent societies.

Critics of Rawls' LP have argued that his account of international justice is too limited. ¹⁷ This criticism involves an important set of issues, which I cannot address here. But I do want to address one issue that relates to satisfaction. Rawls' list of basic human rights is short, and some of the rights he lists are not distributed equally in decent societies. One might wonder, then, how members of decent societies could be satisfied. This objection has two parts, so my response does too. First, Rawls' view of human rights, while clearly minimalist, is more robust than many critics have recognized, and includes a right to some measure of political representation for all members (LP, pp. 35–44). Second, decent societies are associationist (LP, p. 64, p. 68). Given these two facts, the human rights situation in decent societies will not produce a destabilizing dissatisfaction among members.

Rawls' view of human rights is limited, especially when compared to the Universal Declaration of Human Rights of 1948 (hereafter UDHR). Further, he views many of the rights contained in UDHR as liberal aspirations, not human rights proper (LP, p. 80). However, Rawls' view of rights is more robust than many realize (Reidy, 2006). Rawls identifies eight principles that constitute his law of peoples (LP, pp. 35–44). Principle six states that ‘peoples are to honor human rights’. Many critics hold that Rawls' view of human rights is exhaustively described in the list he offers in section 8.2.2. a of LP, where he affirms rights to subsistence and physical security, to freedom from slavery, to freedom of religion and thought, to property and to formal equality (LP, p. 65). But Rawls says that these rights are ‘among the human rights’, indicating that this list is not exhaustive. In section 10, Rawls expands this list to include the rights specified in or entailed by articles 3–18 of UDHR (LP, p. 80). There he affirms other key elements of due process and the rule of law, a right against cruel, inhumane or degrading punishment, a right against torture, the right to refuse non-consensual marriage, a right to seek asylum, a right to freedom of movement and a right to a national identity. Rawls does reject parts of UDHR as human rights. For instance, he does not include a general right to non-discrimination, or a right to non-discrimination in employment. And given his commitment to non-democratic decent peoples, he implicitly rejects a human right to universal and equal suffrage. Nevertheless, the rights he affirms put limits on allowable discrimination. For instance, while decent societies are not committed to the ideal of ‘one person, one vote’, they do not exclude members from political representation based on group membership. Political leaders must consult representatives of each group before political decisions are made. This is the central feature distinguishing decent peoples from benevolent absolutisms. Thus decent peoples are committed to giving every member some measure of political representation, small though it may be.

So the human rights situation in decent societies is not as bad as critics often suggest. Members of decent societies may have reasons to complain about their situation, especially when viewed from the perspective of liberal democracies, but they are protected from gross, systematic human rights abuses. And since members of decent societies are guaranteed political representation, they have some (admittedly small) opportunity to lobby for reform. Thus the human rights situation, while not perfect, is also not likely to cause destabilizing forms of dissatisfaction. Of course, some individuals will be dissatisfied. But many members of democratic societies are dissatisfied with features of their societies too. This dissatisfaction is not likely to lead to violent or other destabilizing unrest.

Further, in a world society of decent societies and democracies, massive violations of basic human rights would not occur. Decent societies and democracies are committed to ridding the world of, for example, aggressive wars, genocides and widespread domestic oppression (LP, p. 113). This goal falls well short of securing for all individuals the full set of rights and liberal aspirations detailed in UDHR. But the point is that there would be no need for societies to intervene (politically, militarily or economically) in the domestic affairs of other societies, if all societies were either decent or democratic. Democratic societies might feel obligated to engage in a politics of persuasion, hoping to convince decent peoples of the values of liberalism and democracy, but there would be no compelling justification for more invasive forms of intervention. Decent peoples might also feel obligated to engage in this politics of persuasion, in order to extol the virtues of the values central to their own common good conceptions of justice. But for similar reasons, decent societies will not have good justification for intervening in the domestic affairs of democratic societies either.

But what about the fact that some rights are distributed unequally in decent societies? For instance, though all members of decent societies have some measure of freedom of conscience, this freedom is not equal (LP, p. 66, n. 2, pp. 72–5). In decent societies, one religion can legally control much of the government, while other tolerated religions are refused access to certain political positions. How can the members of tolerated minority religions be satisfied, when they are not equally free?

This unequal distribution of rights will not lead to destabilizing forms of dissatisfaction, because decent peoples are associationist (LP, p. 64, pp. 72–5). The members of decent societies regard themselves politically as members of different associations or groups, and not as free and equal citizens. Rawls suggests that decent peoples might hold a view similar to Hegel's in Philosophy of Right (1821): persons belong first to groups, i.e. to estates, corporations and associations. The members of decent societies might think, for instance, that when each person has one vote, his or her interests shrink and become focused on merely private concerns, to the detriment of the community, but that when groups are represented in a consultation hierarchy, the voting members of the groups take more seriously the interests of society as a whole. The members of decent societies might also think that the idea of one person, one vote, is mistakenly based on the idea that persons, as atomistic units, have the right to equal participation in political discussions. Rawls' point is that associationism is an important way that members of decent societies differ from members of liberal societies. Persons who regard themselves politically as members of groups, and not as free and equal citizens, would not expect equal political freedoms. Rather, they would expect political freedom consistent with their group's place in the consultation hierarchy and their common good conception of justice. There will, of course, be dissenters in decent societies. However, just as dissenters in democracies are compatible with the satisfaction of democracies as wholes, so too are dissenters in decent societies compatible with the satisfaction of decent societies as wholes. ¹⁸

Constitutional Constraint Accounts

Constitutional constraint accounts base the democratic peace on constitutional and legal restrictions on executive action typically found in democracies (e.g. Owen, 1994). These restrictions, coupled with free public debate, slow the march to war, giving the societies involved time to find peaceful solutions to disagreements. The restrictions and debates also make it difficult for democracies to launch surprise attacks on others. This reduces the fear felt by both sides when two democracies find themselves in conflict.

The term ‘constitution’ has two meanings (Freeman, 1992). It refers to written documents that proclaim certain fundamental social rules. The US has a constitution in this sense, while England does not. And there is a prior, deeper meaning, which Samuel Freeman calls the ‘institutional’ meaning (Freeman, 1992, p. 6). An ‘institutional’ constitution is a set of fundamental, shared, publicly recognized and commonly accepted rules for making and applying laws. The individuals who comprise some political society literally constitute themselves as a body politic through these shared political norms. These norms define the offices and positions of political authority in a political system, their qualifications, rights, powers, duties and so on, and the procedures for making, applying and enforcing laws. Written constitutions announce important social norms, and can play an important role in societies that have them. But they presuppose institutional constitutions. Both the US and England have institutional constitutions. Institutional constitutions do several things (Preuss, 1998). Most importantly, they restrict the arbitrary rule of leaders. Constitutional societies are governed by the rule of law – no-one is above it. Both the rulers and the ruled in constitutional societies have internalized shared fundamental norms and thus accept as binding the fundamental rules that define their political system. Although the rules are not always followed, most people take them to be standards of behavior that should be upheld and that provide grounds for criticism and sanction (including self-criticism and sanction) of violations.

Decent societies have institutional constitutions. Members have internalized, to a significant degree, a shared set of fundamental social norms, including their shared common good idea of justice. This is part of what it means to be well ordered. Members of decent societies have internalized a respect for the inviolability of persons, and for basic human rights. They also understand their laws to generate genuine obligations and duties. This is so because they have internalized their social norms and see their laws as consistent with them, that is, as being grounded by their shared political values, i.e. by their constitution. Rulers will feel a responsibility to uphold their shared social norms, and will feel obligated to justify departures from them in ways not unreasonably inconsistent with them. Non-ruling members will similarly feel obligated to uphold the rules, and will want to question departures and irregularities, and will expect reasons not inconsistent with those rules whenever changes are proposed.

Decent societies are constitutionally committed to fundamental norms that explicitly restrict aggressive behavior, e.g. rules defining jus ad bellum and jus in bello, and others that implicitly restrict such conduct, e.g. norms regarding basic human rights and the equality of peoples. These commitments bind members to certain political institutions, practices and procedures, all of which reduce the chance of war. Further, these commitments support a willingness to seek non-lethal remedies to disputes.

There is also public debate in decent societies. It is not the sort we find in democracies, but it can slow the march to war. Decent societies require leaders to consult representatives before making political decisions. Representatives must consult members of the groups they represent before meeting political leaders. This is the consultation hierarchy. Leaders must take dissent seriously and make conscientious efforts to respond to complaints. And they must allow representatives to renew dissent, if they are not satisfied by the government's response (LP, p. 72). All of this discussion must be consistent with their common good idea of justice. While this is not the kind of public debate we find in democracies, it would slow the march to war. The consultation hierarchy and other constitutional constraints make it unlikely that decent societies will go to war, except in self-defense.

Re-election Constraint Accounts

Re-election constraint accounts hold that democratic leaders want to be re-elected and are thus sensitive to the political consequences of losing wars (e.g. Bueno de Mesquita et al., 1999; Ray, 1995). In democracies, there is a good chance that politicians will lose their offices after a lost war, especially if they are thought to have initiated the war (Ray, 1998, p. 41). This tends to make democratic leaders cautious in selecting opponents, willing to fight extra hard and willing to spend great amounts of resources in the course of war. Democratic leaders may find it politically expedient to keep fighting to the bloody end. As a result, democratic leaders will see other democracies as formidable opponents, and will seek peaceful resolutions in disputes with them.

In democracies, the power of rulers is always subject to the will of the citizenry. Rulers who find war necessary will have to develop a compelling case for it, in order to persuade a substantial majority of citizens. But individual democratic citizens are likely to be divided, to have conflicting interests and to make different judgements, about the need for war. Given this, there is little chance that a democratic body politic will reach a substantial consensus over going to war. Hence, rulers who find war necessary will often have to pursue it without broad public support. But few rulers wish to pursue war against the will of the people. Those who are willing to do so can be removed from office by citizens in elections, so democratic rulers are less likely to initiate wars than leaders of other kinds of societies.

There are three components to this account of the democratic peace. Two can be brought out by considering this question: how does the will of the people in democracies constrain the power of democratic leaders? First, leaders fear the loss of political office, and what their political positions bring them – power, prestige or authority. The desire to keep their positions causes leaders to seek broad popular support, which they will not often find in the case of war. Second, leaders who do not fear the loss of office and who make unpopular decisions can be removed from power in democracies relatively quickly. Regular elections are a feature of democracies. They give citizens the means to exert their collective will on their leadership, by putting candidates they favor in office, and by removing from office those politicians who have lost the confidence of the people. A third key to this account of the democratic peace is the broad nature of the support that must be garnered by leaders in democracies. In democracies, citizens must be convinced that war is necessary. But the more voters there are, and the more disparate their interests, the harder it is to convince a majority that war is necessary.

Thus, this account of the democratic peace raises three questions about decent societies. Are there institutional means in decent societies for removing unpopular leaders from office, such that: (1) leaders will fear to act against the will of the citizenry; and (2) leaders who do not fear to act against the will of the citizenry can be removed for failing to respond to that will? And (3) is that portion of the body politic that leaders must respond to sufficiently diverse in decent societies to serve as a brake on the will of leaders to wage war?

Decent societies have institutional constitutions. This implies well-entrenched norms restricting the use of political power and regulating and defining the transfer of political power. Any society committed to basic human rights and the equality of peoples must have constitutional restrictions on the ability of leaders to violate these commitments. And a society's commitment to such norms implies a commitment to some potential means for reining in rogue leaders, and for replacing them and transferring their power to someone else. This goes some way toward satisfying the conditions of the re-election constraint account. Such institutions give rulers in decent societies reason to fear the loss of office, and hence reason to respond to the will of the people, and give citizens generally the potential means to replace leaders who fail to respond to their collective will.

Decent consultation hierarchies ensure that representatives of all groups are heard before political decisions are made, and allow dissent to be expressed. But how much dissent can we really expect? A lot. Decent societies can be seen as societies organized in accord with and pursuant to some comprehensive philosophical, moral or religious doctrine. (Here the similarity between Rawls' decent societies and traditional accounts of republican states is most evident.) So consider Christianity. Christians have long been divided over the true nature and purpose of the Christian faith, despite the fact that Christians share a history and a common set of ideas. It is not surprising, then, that we find different ideas in the Christian community about the need for and proper conduct of war, and that there is almost never a consensus in the Christian community about the need for any particular war. Decent societies are like this diverse Christian community, with some other religions and philosophies included too. Decent societies will not be as diverse as democratic ones, but there is still good reason to believe that substantial consensus regarding war will be hard to reach in decent societies too.

So decent societies have the disposition and means to remove unpopular hawks from office. Further, we can expect a great deal of disagreement in decent societies about the need for war. But there is another reason to worry. The diversity of opinions on war likely to be found in decent societies may not put as much pressure on leaders to conform as it does in democracies, because decent leaders are insulated from members in two ways.

First, the consultation hierarchy has a ‘funneling’ effect on diversity of opinion. In decent societies, citizens express themselves to their representatives in the consultation hierarchy. These representatives will not be able to relay all of the views of those they represent, and will likely disagree with some of them. So representatives will present only certain views to the government. Thus, the number of different opinions held by the citizenry will be funneled down to a significantly smaller number of opinions actually presented by representatives to the leadership. This may make consensus easier to gain in decent societies than in democracies.

Second, it is not clear how much say members of decent societies have in the selection of their leaders. In democracies, it is generally the case that the candidate who gets the most votes wins. But decent leaders are not responsible directly to the citizenry, because decisions about the fitness of rulers are made by some favored sub-set of the population. Imagine a decent society, the majority of whose citizens belong to non-ruling groups. The ruling group cannot ignore the representatives of non-ruling groups when selecting rulers, but it is not clear that the members of non-ruling groups have much influence over this process. The ruling group may overrule non-ruling groups, as long as this is not unreasonably inconsistent with their shared social norms. In the domestic case for decent societies, reasonable pluralism actually seems to make rulers less responsible to the people generally. This is not so in democracies, where reasonable pluralism is accounted for by the ideal of ‘one person, one vote’.

So decent societies meet some of the conditions of the democratic peace identified in re-election constraint accounts, but maybe not all of them. If the conditions identified in this account were the sole cause of the democratic peace, I do not know if peace could be secured between decent societies and democracies. But since each example of democratic peace likely results from several factors working together, this is not a fatal flaw.

Normative Accounts

Normative accounts are based on the idea that democracies share cultural or normative features that lead them to perceive each other as peaceful and non-threatening (Chan, 1997; Maoz and Russett, 1993). Maoz and Russett (1993, p. 625) argue that democracies have an atmosphere of ‘live and let live’, and are consequently stable at the ‘personal, communal, and national level’, because political conflicts are settled through compromise and not physical struggle. Democracies externalize these domestic political norms and expectations when dealing with other democracies. Democracies thus see other democracies as committed to liberal ideals, and as reasonable, predictable and trustworthy. This promotes peace.

Since political conflicts in decent societies are also settled through compromise rather than by fighting, a similarly stable atmosphere should develop in them. (This is, in part, what satisfaction accounts claim.) However, it is not clear that democracies will ever come to see themselves mirrored in decent societies, in the way that normative accounts highlight. The two kinds of society are very different. Decent peoples are associationist. They see group membership as properly mediating between individuals and political authority. Democracies reject this mediated view for an ideal of one person, one vote. Decent common good conceptions of justice are rooted in comprehensive doctrines. Liberal conceptions of justice are based on ideals of politically free and equal citizens. And while the human rights situation in decent societies is not as bad as critics often suggest, the rights that characterize decent societies fall well short of the rights that characterize democracies. So it is not clear that democracies will externalize their domestic political norms when dealing with decent societies.

Rawls claims that, through time, as peoples honor the law of peoples, mutual trust and confidence will develop (LP, pp. 44–5). He is probably right. There is an overlapping consensus among decent societies and democracies regarding the law of peoples. This gives us reason to be optimistic. But this is not the same as saying that decent societies and democracies will some day see their values mirrored in one another. Perhaps this mutual understanding will never develop beyond the mere expectation that each will honor the law of peoples. We should wonder, then, if the source of the peace identified in normative accounts is not missing between liberal and decent peoples.

Interdependence Accounts

Interdependence accounts are based on the idea that international trade promotes peace in several ways (e.g. Oneal and Russett, 1997). Democracies trade with one another more frequently than with other types of states because democracies are (perceived to be) better than other kinds of states at guaranteeing trade terms and the continual flow of resources and money. International trade increases contacts between states, contributes to mutual understanding and creates ties that are better served by peaceful conflict resolution than fighting. Trade also tends to benefit both sides, and since these mutual benefits must often be sacrificed in times of war, trade discourages conflicts. Finally, the mutual benefits of trade reduce the potential benefits of conquest. Since war threatens trade, trading partners do not want to fight over disagreements.

Rawls' law of peoples does not require decent societies and democracies to engage in trade or other cooperative activities that might produce the mutual understandings and benefits that drive interdependence accounts. The law of peoples requires that societies maintain their status as corporate moral agents, or else they lose the ability to make justice claims (as groups or peoples) on the international stage, but Rawls views each people as autarkic with respect to the means necessary for the development and maintenance of its moral status as a people (LP, p. 108). Few societies in the world lack the resources to become well ordered. Each is in principle capable of securing, on its own, the two fundamental interests of peoples – independence and self-respect (LP, p. 34). So any cooperative activity between peoples is, strictly speaking, voluntary. Each society must maintain its corporate moral status, but each can do this on its own. It is true, though, that each society depends on others to maintain policies of non-interference, and to refrain from engaging in practices that might threaten the background conditions necessary for the development of corporate moral agency worldwide. And so while societies are autarkic with respect to their own moral standing, they are perhaps not completely independent, even in this limited sense. Further, taking our world, here and now, as an example, there is every reason to believe that societies will engage in trade and other cooperative activities. Still, Rawls' law of peoples does not require them to do so as a matter of international justice.

Both decent societies and democracies have a duty to assist burdened societies, to help them become well ordered through the development of decent or democratic institutional frameworks (LP, pp. 105–13). This duty, however, does not require decent societies and democracies to work together. Burdened societies often lack political and cultural traditions and the human, material and technological resources necessary for well orderedness. Decent societies and democracies may, through simple prudential reasoning, decide to work together to assist burdened societies. But the law of peoples does not require this cooperation. Each decent and democratic society is free to decide for itself how to assist burdened societies. Rawls claims that the key to assisting burdened societies lies in helping them to develop the right kind of political culture, but admits that there is no easy recipe for getting this done (LP, p. 108). Thus it seems likely that different societies will disagree about how best to aid burdened societies. It is possible that democratic societies will band together to aid burdened non-associationist societies, and that decent societies will work together to aid burdened associationist societies.

If the democratic peace is secured primarily through the process described in interdependence accounts – trade, mutual understanding and mutual benefits – then the prospects for Rawls' peaceful world order will dim significantly. His law of peoples does not require interaction between decent and democratic peoples, and so any stability that must be secured in this way is not a feature of his theory of international justice. Since substantial cooperative activities are not required, some societies may refuse to participate. On this account, such abstainers would threaten the democratic peace.

Decent Peoples and the Democratic Peace

We can reasonably extend the democratic peace thesis to include decent societies, even though they do not meet all of the conditions of the democratic peace contained in the various accounts. Decent societies meet the conditions of the constitutional constraint account. Members of constitutional societies share a commitment to fundamental social norms, which ultimately commit them to several institutions and practices that slow the march to war. Further, these shared commitments are likely to produce in many individuals a willingness to seek non-lethal means of conflict resolution. But these institutions and dispositions may be overcome by, for instance, humiliation, wounded pride or lack of due respect, so the conditions of the satisfaction account must also be met. Decent societies meet these conditions. So a world society comprised of satisfied decent and democratic constitutional republics would be stable, following these two accounts. The fact that decent societies and democracies may never meet the conditions of the normative account is troubling. But as long as all societies are satisfied, and no society feels humiliated, wounded or disrespected, this lack of shared values should not prove destabilizing. The conditions identified by the re-election and interdependence accounts would strengthen the democratic peace, but the fact that decent societies only partially meet the conditions of the re-election account, and that Rawls' law of peoples ignores the conditions of the interdependence account, does not make peace impossible or even less likely. All things considered, we can conclude that international relations between democracies and decent societies would be peaceful, according to the terms of the democratic peace thesis. If all societies were democratic or decent, we would have a peaceful world order.

Whether or not one finds Rawls' proposed international order, or his decent societies, interesting or plausible, my analysis has implications for theories of international relations and global peace. Strictly speaking, the democratic peace could be maintained in a world order comprised of decent societies alone. What is essential for the democratic peace is that the world be filled with satisfied constitutional republics. Thus, democratic societies need not seek to democratize all other societies in order to produce world peace. This is good for a couple of reasons. First, democracies have a better chance of convincing non-liberal and non-democratic societies – especially those with strong religious traditions – to become non-democratic constitutional republics rather than democratic ones, at least in the short term. If peace really is the goal, this would be no small accomplishment. It would go some way toward securing respect for human rights around the world, and would provide a more secure platform for further discussion of democratic values. Second, while democracy is valuable, there are also good reasons for democracies to value the self-determination of other peoples. If the members of some society chose to constitute themselves as a decent religious republic, strong forms of intervention would not be justified. Democracies would have reason to engage them in discussion on the value of persons and liberty and so on, but little beyond such conversation seems imperative. The situation would be more complicated where two factions, one liberal and one decent, were vying for dominance in some society. But even then, as long as the struggle remained peaceful, there would not be reason for democracies to engage in more than the politics of persuasion.

Finally, the world society Rawls envisions is not a static state of affairs, or the end of human history. It represents the realization of certain core moral conditions, which, once in place, ensure that the further development of peoples may be regarded as consistent with liberty and justice, in both domestic and international spheres. Although Rawls argues that all constitutional republics are equal as such, and thus that decent societies can reasonably demand due respect, he regards democracies as superior in terms of justice (LP, p. 62). Thus, democracies will always have good reason to try to liberalize and democratize decent ones. Due respect for self-determination means that democracies cannot coercively demand that all societies be democratic, but it does not mean that they cannot engage in a politics of persuasion. Democracies must be careful not to frustrate or otherwise injure the vitality and self-respect of any decent society. Thus, they should not offer incentives (e.g. subsidies) to decent societies to encourage reform, because these are likely to cause resentment between societies and to interfere with the self-determination of decent ones (LP, p. 85). But democracies can encourage pockets of liberal individuals in decent societies through discussion. While the lines that separate appropriate from inappropriate conduct are not clear, Rawls is right that there is an important distinction here. Nevertheless, since we can extend the democratic peace to include decent societies, a concern for global peace is not a reason to try to democratize them.