It's not that democracies are pacifist by nature. Democratic countries, acting alone or in concert, do go to war with nondemocratic countries from time to time, for example the United States and others against Saddam Hussein's Iraq and NATO against former Yugoslavia over the attempted ethnic cleansing of Kosovo.

Moreover, the record of peace among democracies is not without its asterisks. As the neorealist scholar Kenneth N. Waltz has noted, Germany on the eve of the First World War was, by the standards of the day, "democratic." German "militarism" in the late 19th century was not an authoritarian imposition on the German people but something they and their elected representatives supported--as indeed going to war in 1914 was popular in Britain as well. Reclassifying a country as "nondemocratic" because it has chosen to make war on a democracy is a temptation to which the "democratic peace" thesis may give rise.

Young democracies also pose problems; for example, Ecuador (less than three years democratic) and Peru (less than one year) went to war in 1981. The Kargil war between India and Pakistan in 1999 exemplifies another category of exception: India's democratic bona fides are beyond challenge, but Pakistan, though it had had a democratically elected president and a civilian government since 1988, was a distinctly illiberal democracy. Indeed,

Pakistan's current ruler, Pervez Musharraf, took power in a military coup within a few months of the war's end. And there are other objections: Critics of the democratic peace thesis often advance the argument that the number of instances on which to base generalizations is small.

The Bush administration, of course, has invested heavily in the idea of the beneficial effects of democracy. As the president put it in his second inaugural address, "It is the policy of the United States to seek and support the growth of democratic movements and institutions in every nation and culture, with the ultimate goal of ending tyranny in our world." Bush pointed to two reasons for this policy, one metaphysical ("no one is fit to be a master, and no one deserves to be a slave") and one practical: "as long as whole regions of the world simmer in resentment and tyranny--prone to ideologies that feed hatred and excuse murder--violence will gather, and multiply in destructive power, and cross the most defended borders, and raise a mortal threat." Now, one can agree with the metaphysical claim, and even with the long-term prognosis--as well as the inescapable fact that there is no way to get to the long term except via the short term--without relinquishing reasonable qualms about who might be empowered by free elections in the short run.

But for all the imperfections of a thesis claiming that democracies do not go to war with each other, we are left with one big and inescapable fact: A large number of mature, liberal democracies have no intention whatsoever of going to war with each other, not now, not ever. Such disputes as arise among them, they will settle by peaceful means. And one of the reasons they will be able to settle disputes peacefully is that they have learned they have no "vital interests" that conflict. That's no small thing historically.

Moreover, the number of such countries is growing. The 20th century saw several "waves" of expansion of democracy, the most recent in Central and Eastern Europe, extending in the new millennium as far east as the Orange Revolution in Ukraine and the Rose Revolution in Georgia. Freedom House listed 122 "electoral democracies" as of 2005.

Citizens of democracies and their elected leaders tend to regard democratic political arrangements--government with the consent of the governed--as uniquely legitimate. As they see it, democracy is not one choice among many legitimate types of regime but the best choice for those capable of it. Democracies therefore have values in common--an affinity that constrains conflict and constitutes the true underpinning of the "democratic peace."

Democracies also have a pretty good record of speaking up for free and fair elections, human rights, and other liberal values in nondemocratic states and states in transition. Liberal democracies regard the spread of liberal democracy as in their collective interest. The United States and Europe stood in solidarity in support of the Orange Revolution in Ukraine, for example, and the united front contributed to the pressure on the streets to force the government to nullify the results of the election it had stolen. It's striking that the leading outside opponent of the Orange Revolution was none other than Vladimir Putin, Russia's increasingly autocratic leader, who values Russian influence in its near-abroad so highly that he is happy to make common cause with such dictators as Alexander Lukashenko in Belarus.

Without question, the post-Cold War "enlargement" agenda of NATO and the European Union fostered democratic development in Central and Eastern Europe and the former Soviet Union. Both NATO and the E.U. emphasize democratic development as a precondition for beginning the process that leads to membership. The prospect of access to Western institutions, including the security they offer and the economic opportunities they create, provided an incentive to young (and in some cases unsteady) democracies for good policy choices, political development, and cooperative rather than competitive strategies with their neighbors. There is no better illustration of this than the decision in 2000 of countries aspiring to join NATO to work together for membership for all who qualified rather than for each to try to establish its bona fides by invidious comparisons highlighting the shortcomings of the others. That cooperative effort led to a far bigger round of NATO enlargement in 2004, seven countries in all, than anyone had anticipated only a few years earlier.

Likewise, the E.U. increasingly points with considerable justification to the regional success of its "soft power," to use Joseph Nye's term for the ability of a state to get others to want what it wants. In truth, it's a lot more work to get into the E.U. than to qualify for NATO membership; the amount of national law that has to be brought into harmony with the E.U.'s acquis communautaires is vast. But the benefits are also quite substantial in terms of market access, standardization, and investment flows. Poorer countries are also eligible for assistance from Brussels, not only as new members but also as aspiring members. The hard political choices required to fight corruption, increase transparency, and develop a democratic political process thus make one eligible for rewards both long- and short-term.

It's easy now to be complacent about Central and Eastern Europe, because of the extraordinary tales of political and economic success that have unfolded there. But suppose Western institutions hadn't reached out to Central and Eastern Europe in the 1990s. Suppose, instead, they had made it clear that the E.U. was an institution for rich Western European countries alone and that NATO was closed to new membership, if indeed it wasn't going out of business. Does anyone really think that 15 years after the fall of the Berlin Wall, the outcome would have been the same in terms of the region's peaceableness, democratic development, and economic growth? In an international culture of self-help rather than cooperation and integration, what would have kept Slobodan Milosevic's bloody-minded nationalism from becoming not an aberration but the norm?

The attractive power of NATO and the E.U. have led to proposals for enlargement beyond the original geographical origins of each. Ivo Daalder and James Goldgeier have recently made the case for a "Global NATO" in Foreign Affairs, and Jan Ole Kiso and Adrian Taylor recently offered a European perspective along the same lines in Europe's World. The E.U. is grappling with the question of "European" identity in assessing future enlargement. Both NATO and the E.U. have expressed support for robust policies of engagement with nonmembers in their "neighborhood." Nevertheless, there is significant resistance among current members of the E.U. and NATO to the transformation of either by enlargement outside the limits of Europe or the Euro-Atlantic, however construed. The defeat by referendum of the E.U.'s proposed Constitutional Treaty in France and the Netherlands in 2005 has given rise to much talk of "enlargement fatigue."

No one expects any change in that sentiment anytime soon, leaving Turkey's candidacy in a state of limbo or worse. The joke making the rounds on the transatlantic circuit in fall 2006 was that the optimists think Turkey will join the E.U. during the Albanian presidency, whereas the pessimists think Albania will join the E.U. during the Turkish presidency. Although NATO's door remains ajar, the cases for entry are clearly getting harder. At some point soon, we begin to rub up against the practical geographical limits of the "attractive power" of NATO and the E.U. for promoting democracy and good governance.

So where does that leave the rest of the world? While the Euro-Atlantic community is relatively well served in terms of its capacity for democratic states to cooperate when they wish to do so through NATO and the E.U., other geographical regions lack comparable institutional infrastructure. Major regional organizations typically do not enforce criteria for membership any more than the United Nations enforces the requirements in its Charter for members to adhere to human rights and other norms. If you are there, you are in. Needless to say, this doesn't create incentives for reform.

Multilateral organizations and institutions that are not geographically oriented often lead would-be members or beneficiaries to make better policy choices, or at least officially favored choices. Here, the World Trade Organization and the World Bank come to mind. But, of course, democracy promotion by these means is at best indirect, usually through measures aimed at increased economic and government transparency, anticorruption efforts, and the like.

Bilateral or regional initiatives can make a difference too. The Bush administration's Millennium Challenge Corporation provides development assistance ranging to the hundreds of millions of dollars to governments that meet reform criteria, including a component devoted to "Ruling Justly" that pushes in the direction of democratic development. The National Endowment for Democracy is able to point to a number of successes over the course of its 24-year history of channeling assistance to democratic reformers and revolutionaries. In the case of Belarus, Northern European neighbors from Sweden to the Baltics have been especially active in helping a democratic opposition to the Lukashenko regime get organized.

But we are still waiting for anything like the "attractive power" of the E.U. and NATO on a global scale. The best effort so far was a Clinton administration brainchild, the Community of Democracies. The CD is a loose affinity organization that first met in 2000 in Warsaw, where participating nations, typically represented by their foreign ministers, adopted the "Warsaw Declaration" pledging their commitment to democracy and the promotion of democracy. The problem is that a number of nondemocratic countries, such as Jordan, Algeria, Tunisia, and Bangladesh, participated and signed the declaration. The CD has met on several occasions since but has yet to develop much in the way of institutional capacity. Worse, how you might transform it from what it is to what one would like it to be is an exquisite diplomatic challenge, entailing, as it necessarily would, kicking people out.

Some have proposed greater cooperation and coordination among democracies at the United Nations. A wide array of NGOs favor establishing a "Caucus of Democracies" within the U.N. in order to encourage promotion of democracy and human rights. The 2005 United States Institute of Peace Task Force on U.N. Reform (the Gingrich-Mitchell task force) strongly endorsed efforts to strengthen the Caucus of Democracies at the U.N. But the odds against using the United Nations to promote democracy are formidable, as the ongoing depredations of the Human Rights Council, the "reformed" successor to the widely discredited Human Rights Commission in Geneva, make painfully clear. As long as the informal mechanism providing for rotation within regional blocs remains entrenched--thus giving each dictatorship its day in the sun--the U.N. will be largely ineffectual in promoting human rights.

Efforts to strengthen cooperation among democracies are chiefly motivated by the view of proponents that democracies acting in concert have a special capacity to legitimize international action because the governments have a legitimate claim to be speaking for the people of their countries. Ivo Daalder and James Lindsay have been at the forefront of this argument, urging in the American Interest the creation of a "Concert of Democracies." An earlier incarnation of the "Concert of Democracies" idea became a marquee recommendation of the Princeton Project on National Security. Some proponents, such as Daalder, regard the legitimacy obtained by agreement among democracies as superior to the legitimacy represented by agreement of groups of nations that include nondemocracies, such as the U.N. Security Council or General Assembly. Others, such as Princeton's Anne-Marie Slaughter and G. John Ikenberry, co-directors of the Princeton Project, prefer to work through the U.N. to seek legitimacy but see democracies acting in concert as an alternative in the event of the inability of the U.N. to take effective action (e.g., NATO in Kosovo, Darfur).

The question of how to give democratic states the operational capacity to act collectively has remained a difficult one. Granting the fact that mature, liberal democracies live in peace with each other, agree on the unique legitimacy democratic governance provides, have a track record of assisting countries making a transition to democracy, and might wish to collaborate on at least some issues in a global forum that excludes the worst human rights abusers, tyrants, and authoritarians from the deliberations, maybe it's time to make a clean break. Maybe it's time for the United States to join other democracies in adopting a new Treaty of the Democratic Peace.

The parties to such a treaty would reaffirm, consistent with the United Nations Charter and their other international obligations, their commitment to democratic governance; note their long practice of living peacefully among themselves; affirm their intention to continue to do so permanently and to settle all matters between them peacefully; and commit to the extension of the democratic peace by pledging assistance to other states in the development and improvement of their practice of democratic self-government.

The treaty would create a council--it could indeed be called the Concert of Democracies or, as Kiso and Taylor propose, the Organization of Democratic States, or something else--charged with implementation of the treaty provisions. The council would be its decision-making body. The treaty would also create a secretariat to advise the council on matters of relevance to the treaty organization and to implement decisions of the treaty council. The treaty would provide for the accession of additional states upon invitation of the council and ratification by national governments; it would also have to grant contracting parties the right to withdraw within a short interval of renouncing the treaty, and it should include a mechanism to ensure that any member that abandons democracy can be excluded from future participation.

Without doubt, any proposal for a major new international institution has an idealistic and aspirational component to it. This has been true at least since Tennyson mused about "a Parliament of man, a Federation of the world" ushering in an era of common sense and universal law in "Locksley Hall," not to mention the fond hopes Woodrow Wilson pinned on his League of Nations as well as the similar hopes animating the drafters of the United Nations Charter. And of course the annals of 20th-century diplomatic history feature such notorious misfires as the Kellogg-Briand Pact of 1928, signed by Germany, Italy, and Japan as well as the United States, France, and the U.K., renouncing no less than "war as an instrument of national policy"--which remind us of the regular failure of such initiatives to live up to the aspirations of their proponents.

This treaty proposal certainly has its aspirational element as well. Yet it begins not with a dream but with the fact of democratic peace. It is not merely aspirational. It differs from the Kellogg-Briand Pact in a number of decisive ways. It is at best unclear that the Kellogg-Briand signatories were sincere in their undertakings, and in any case, little more than a decade had passed since a number of the signatories were at war with each other. In addition, the governments of a number of the parties to Kellogg Briand were not democratic or were at best illiberal. India was "represented" by British imperial authorities. Italy still had its monarchy, Japan its emperor. By contrast, the parties to the Treaty of the Democratic Peace would be states that have long lived peacefully with each other, in some cases for two generations or more, and expect to continue to do so. They are mature, liberal democracies whose internal democratic processes have been tested by internal and external political shocks without disruption. They have a track record of working cooperatively on matters of mutual interest.

Given the current configuration of power politics internationally, the United States is a lonely "hyperpower" (in the coinage of former French foreign minister Hubert Védrine). Having explored the limits of unilateralism in the first Bush term, the second Bush administration has placed a much higher value on working with others, and future administrations are likely to make even more pronounced efforts along those lines. Decisions made in Washington ramify around the world in ways that are often discomfiting or worse to those affected by them but who have no say in them. This is never going to be a source of American popularity, including among democracies. Others want influence over the United States, and who can blame them? For them, the treaty body offers the forum in which, as Daalder and Lindsay note, the United States is most likely to be influenced: among like-minded democratic states seeking a basis for cooperative action.

Nevertheless, the historical record argues for a cautious approach and limited aspirations. The treaty should explicitly assign only the most general role to the treaty council and the secretariat. The institutional point is not to assume the treaty council and the secretariat will be playing major roles on the world stage, but to create these entities in order to respond to the needs of members as they see fit. It is quite possible that the treaty council and the secretariat will be rather sleepy places for some time. But they will be available for members to take action whenever the members themselves find utility in acting as a body of democracies. The NATO alliance was hardly founded with the expectation that member-states would one day convene at its Brussels headquarters to decide to undertake a combat mission in Afghanistan against Islamist extremists. That institution continues to prove its utility for members even though the threat the alliance was created to defend against is no more.

Insofar as a Treaty of the Democratic Peace affirmed a commitment to the spread of democratic principles and liberalization, there would seem to be an organic role for the council and secretariat to play in supporting states making transitions to democratic governance. It would be hard to imagine the council turning down a fledgling democracy's request for technical assistance and "best practices" guidance. A parallel example would be the process by which NATO became involved in Darfur. In early 2005, the North Atlantic Council found itself blocked from considering support for humanitarian operations there because of the view of some allies (notably France) that NATO had no business in Africa. Yet when a formal request for NATO's assistance came in from the African Union, which under U.N. mandate was providing a peacekeeping force in Darfur, the position of those allies opposed in the abstract to the idea of a NATO role in Africa became untenable.

Beyond offering such assistance when asked, the treaty council also might want to involve itself in work to promote democracy and liberalization and to support those working peacefully toward those ends. Here, of course, we enter a more controversial sphere of activity, as it is by no means clear that all the member states (or whatever majority of them would constitute the basis for a council decision) would want to risk antagonizing nondemocratic states by promoting activities that autocrats deem subversive. Nevertheless, in certain instances they might. The case of Ukraine 2004 comes to mind.

An essential element of the democracy promotion possible under the treaty would be its openness to new members. A state that demonstrates a commitment to democratic governance and declares its commitment to the democratic peace should be eligible for participation in a process that leads eventually to an invitation to join. This process should not be too hasty, insofar as the Treaty of the Democratic Peace would have its origins in the actual, demonstrated commitment of democratic states to live in peace with each other. But states should also receive benefits and encouragement for a genuine demonstration of an intention to accede. The treaty council could designate states in train for accession as eligible for observer status. With such status might come funds to assist with democratic transitions and democracy-building. Here is the potential for the global extension of the "attractive power" incentives of the E.U. and NATO.

Since the accession process would be intended to help and encourage states in transition to deepen their commitment to democracy, the treaty council, acting through the secretariat, might want to advise aspirant states about measures they should consider to improve their democratic governance. The council could also set conditions for eligibility and a process to evaluate potential members' candidacies on a country-by-country basis, as NATO and the E.U. have done.

In the end, since accession would be by invitation of the treaty council and ratification by national governments, the parties themselves would have final say on whether a country met the test of being truly democratic and truly committed to the democratic peace. Admission would be by democratic "peer review." The evaluation each national government would perform, taken collectively, would be a better test of how democratic an aspirant was than the application of an abstract set of criteria. This mechanism would likely prevent states about which genuine questions remain with regard to their commitment to democracy and the democratic peace from acceding to the treaty and diluting its essential character.

It would also be useful for the treaty to include a "sunset" provision on membership, according to which each country's accession would expire after a certain term of years, say seven or ten. Upon expiration of membership, and given a stated intention to rejoin, a country would take observer status pending a renewed invitation from the treaty council and re-ratification by national governments. Such a novel provision would provide an automatic policing arrangement for the treaty council in the event of a state's abandonment of either democracy or peaceableness. This would, in my view, be an improvement over any possible expulsion provision. The U.N. Charter has provisions (Articles 6 and 18) allowing the General Assembly, on recommendation of the Security Council, by two-thirds vote to expel a member for "persistently violat[ing] the Principles contained in the present Charter." Needless to say, the mechanism has never been deployed against even the most egregious aggressor or human rights violator. An expulsion provision in the Treaty of the Democratic Peace would likely be similarly toothless: Expulsion could never be automatic according to "objective" or technical criteria; it would require the positive action of the treaty council, which the U.N. experience suggests would be seldom obtained. A sunset provision on membership, on the other hand, would require an affirmative act of re-invitation according to those same procedures, followed by re-ratification by national governments. The ability of the council to let a state that has automatically dropped off remain out until genuinely qualified for readmission would be a strong enforcement mechanism and a powerful incentive for good behavior among members.

A question on which the success of the project as a whole would ultimately hinge is which states would be founding parties to the Treaty of the Democratic Peace. To be blunt, an effective means of excluding those states that would seek to participate without being truly committed to democracy or the democratic peace (even if they professed both) is essential. One plausible criterion is time: The initial contracting parties would be those that had been democratic and had lived at peace with one another for some lengthy period. How long would be critical. A 25-year test would exclude South Korea, for example. A test of longer than 12 years would exclude South Africa, but a 12-year test would raise the question of inclusion of the Russian Federation. Without attempting to be exhaustive, it seems to me that critical states for early buy-in would be India, Japan, South Africa, Brazil, South Korea, Australia, Mexico, and if possible France (to which the project should be presented as the vindication of Aristide Briand's vision)--as well as, obviously, the U.K., Canada, Germany, and other Euro-Atlantic allies.

But perhaps rather than trying to get exactly the right group around the table, no more and no less, a better approach would be to start with a very small group and have others come in through the accession mechanism. Suppose the initial group consisted of one nation from each continent. Such an approach would also provide a bypass around the potentially dicey problem of trying to evolve the overly inclusive Community of Democracies into a treaty body.

Suppose, further, that the founding North American nation was not the United States but Canada, with the United States holding back to join through the accession process. Such an act of self-restraint might help rebut the certain criticism that the project as a whole was no more than an American bid for special status in a forum in which it had more influence than others.

As to what other utility democracies might find in a body in which they acted together, one can speculate. The Princeton Project on National Security draft "Charter for a Concert of Democracies" explicitly constitutes the concert to serve as an alternative to the United Nations for legitimizing the use of force in relation to the "responsibility to protect," for example in Darfur. Here, the inability of the Security Council to take action to stop ethnic cleansing and unfolding genocide in Kosovo in 1998-99 also weighs heavily on the minds of many. NATO's decision to take military action against the Belgrade regime without U.N. authority has been, for some, a model of how democracies acting collectively provide legitimacy--albeit one that could be improved upon by an explicit charter of authority.

I think inclusion of an explicit provision related to the authorization of the use of force would make adoption of the treaty problematic if not impossible for many states. Many democracies assign the U.N. a unique role as bestower of international legitimacy and legality, and would see a treaty provision purporting to offer a backup mechanism for such legality or legitimacy as a deliberate attempt to undermine the U.N. For the same reason, my inclination would be not to favor a NATO "Article V"-style collective defense provision for the treaty. The intention behind the Treaty of the Democratic Peace is not to create a military alliance, or to legitimize U.S. military intervention.

If, indeed, the members of the treaty council ever wished to give an imprimatur to an intervention, nothing in the treaty would prevent them from doing so, so long as it was consistent with their other international obligations, including under the U.N. Charter. Likewise, if one of the parties to the Treaty of the Democratic Peace came under armed attack by a non-party, the treaty council could frame any response it wished, including an assertion that members regarded the attack as an attack on them all. For example, at a summit of ten NATO aspirant countries in Sofia, Bulgaria, in October 2001, the group jointly declared that they regarded the 9/11 attack on the United States as an attack on themselves as well: They declared themselves "allies in fact" if not yet by treaty. Nothing required them to do so but their own sense of the ethical imperative of the moment.

Many other details would need to be worked out by treaty drafters. The procedure governing the decision-making process of the council is an obvious one. Should decisions be by majority, by majority weighted for population, by supermajority, or by consensus? There must not, of course, be veto powers, à la the Security Council. NATO employs a "silence" procedure, according to which no action is taken while any stated objection is pending. Given the central objective of enhancing cooperation among democracies, perhaps a novel "consensus-minus" procedure might be appropriate, according to which the treaty council would seek silence (i.e., no overt objection), but in which a state could break silence only in concert with another state. You could not object only for yourself. The idea would be to ensure that particular local circumstances or historical grievances did not impede the function of the treaty council. The question of Turkey's view of Armenia's accession, or vice versa, comes to mind. No single state would be able to block action. Such a provision would, of course, equally prevent the United States from blocking action.

There is also the question of where the treaty council and secretariat should be headquartered. A location outside the Euro-Atlantic area would be highly desirable. Plausible contenders include Tokyo (or elsewhere in Japan), New Delhi (or elsewhere in India), and Johannesburg (or elsewhere in South Africa). Establishing a funding mechanism for the institution would likely be no less fraught than funding other international bodies--with the exception that the democratic and peaceable character of the participating states would be a source of some measure of mutual goodwill.

Again, one must be modest in one's aspirations. If the time is not right for democracies to formalize their peaceable intentions in a globe-spanning treaty that envisions working cooperatively on matters of mutual interest and concern, we will likely find out sooner rather than later from a serious diplomatic effort to promote the idea. And if the time is right, though it would surely take more than one administration to bring the treaty to fruition, we would equally likely see indications of its viability. For those committed to liberal democracy as the legitimizing principle of government at home, now is the time to put the global commitment to the democratic peace and to cooperation among the like-minded to the test.

Contributing editor Tod Lindberg is a fellow at the Hoover Institution and editor of Policy Review. He served on the staff of the Gingrich-Mitchell task force on U.N. reform and on the steering committee of the Princeton Project on National Security.