The second possibility-that national boundaries might be considered revisable-may therefore make more sense. We are already living through a period of fragmentation-in 1945, for example, there were 74 independent nations; today, there are 195 (counting Taiwan and Palestine). National boundaries were extensively rewritten in the wake of the collapse of Soviet communism, as new states (Ukraine, Kazakhstan) and old (Estonia, Armenia) emerged from the wreckage of the Soviet Union and as former communist states in central Europe (Czechoslovakia, Yugoslavia) divided along linguistic, ethnic or sectarian fault lines. Why not divide Rwanda, for example, into two countries, or Iraq into three?

Consider the Iraqi case more closely. Perhaps as a result of the Gulf War, which was fought to preserve the territorial integrity of Kuwait, the United States and its coalition do not seem to have seriously considered partitioning Iraq so as to bring the political institutions of that territory into closer alignment with the distribution of its main ethnic (Arab and Kurdish) and sectarian (Shi‘a and Sunni) groups. Indeed, the status of the United States and the United Kingdom as "occupying powers" under Security Council Resolution 1483 (2003) and its successors signals a determination that Iraqi sovereignty and borders are to remain intact. Yet even in the recent past, the United States has not treated national boundaries as immutable. In 1999 and after, the United States and its allies have in effect been party to severing Kosovo from its nominal sovereign, Serbia; indeed, with the creation of a "no-fly" zone in northern, Kurdish-dominated Iraq in 1992 and its maintenance for over a decade after, the United States encouraged an already-powerful demand for Kurdistan's secession and independence. Given that the boundaries of Iraq have never corresponded to any underlying realities of language, ethnicity or faith, and given also that sectarian and ethnic conflicts threaten to tear the flimsy fabric of Iraqi nationhood apart, why should Iraq's boundaries be considered unalterable?

Of course, partitioning Iraq into three new countries could invite outside intervention from Iran, Turkey or elsewhere. If the United States gave no guarantees for the security of the three successor states, they would probably either band together or be conquered. Even after most of our forces withdrew, the United States would have to supply the three new nations with the security against foreign threats that was originally provided by the larger Iraq. But protecting the successor states from their external enemies would require a much smaller force than what is needed for counterinsurgency, and one more suited to the kind of wars our military is best at fighting.

Downsizing states may be desirable (or inevitable) anyway. Some economists have argued that the size of nations is determined by the trade-off between benefits of larger scale-such as defense, law and order, and free trade markets-and the costs created by heterogeneity of preferences within a nation. While larger nations can provide public goods at lower per capita costs, at some point they will become so large that heterogeneous groups within the nation would rather split apart than live under policies which they oppose. This has, indeed, been one of the main reasons historically why large, internally diversified nations and empires (like the Austro-Hungarian Empire or the Soviet Union) have collapsed: The costs to Estonians of remaining inside the Soviet empire, or to Czechs of remaining inside the Austrian one, came to outweigh any benefits they derived from the efficiencies of greater national scale.

For more than half a century, the United States has supported an international system that has supplied most of the globe with external security and free trade-the very public goods that larger nation-states had existed to provide. It follows that, as the threat of interstate war recedes and free trade agreements expand, nation-states should tend to become at once smaller and more homogeneous: The benefits of size will be provided by the international system, not by the (large) nation-state. By playing the "hegemonic" role of principal supplier of international public goods, the United States has created the exact conditions to encourage national fragmentation-a development it now paradoxically opposes by attempting to keep nations intact. One could view the blood and treasure spent in Iraq as the costs of opposing broader trends in history toward smaller, more homogeneous nations and governments, trends in fact accelerated if not put in motion by American foreign policy itself.

Partitioning, to be sure, is not necessarily a panacea for political violence. The two great twentieth-century partitions undertaken by a disintegrating British Empire-those of Ireland and of India-have not produced peace in either land. Even where partitioning ostensibly "worked", as in Germany between 1945 and 1989 or in Korea from the 1953 armistice until the present, it was maintained only because it was a solution imposed by the great powers. Further, partition and other forms of boundary revision could provoke mass deportation or ethnic cleansing, as the partition of India in 1948 did, with calamitous results, and as the shift westward of Poland's boundaries in 1945 also did-on that occasion, however, with benign long-term effects for the peace of central Europe. The partitioning of Iraq is, therefore, a risky, if also defensible, policy option.

But Iraq is only one example of a deeper reorientation that should occur in American foreign policy, away from the democratic nation-state model and toward more flexible forms of governance. The United States should reject the notion that it must bring failed states to democracy, national independence and full sovereignty. Perhaps in partnership with other nations, it should explore other forms of governance that before all else provided internal security, an impartial justice system and functioning markets. Striving first for the grail of constitutional democracy and national sovereignty may be self-defeating. In any case, the more ambitious project raises the costs of intervening to restore order in failed states to what can swiftly become unacceptable levels. Holding down the costs of American intervention is a vital consideration: It will allow us to remove more territory from the grip of dangerous tyrants and terrorists.

The perfect should not become the enemy of the good. The United States should seek to achieve more modest goals of restoring sufficient order in a failed state to prevent it from becoming a breeding ground for terrorism. While the United Nations has generally proven a poor resource for achieving these ends, the United States might benefit from working with nations with large militaries and populations that could specialize in providing order within failed nations. For example, the United States might try to work out arrangements with countries like India, Turkey or Egypt under which they would be paid for providing sufficient military personnel to repress insurgencies and maintain order in a failed state, thus enabling the United States to concentrate on other tasks.

Second, the United States and its allies should consider more seriously the prospect of fragmenting failed states into more governable parts. If the United States and its allies continue to supply international peace and security, and current efforts to promote free trade succeed, existing nations may often be larger than necessary. The United States should reject efforts to keep failed states together, and instead allow them to break up along ethnic or religious lines.

Robert Delahunty is an associate professor of law at the University of St. Thomas School of Law, Minneapolis, MN. John Yoo is a law professor at the University of California, Berkeley, a visiting scholar at the American Enterprise Institute and the author of War by Other Means(Grove/Atlantic, 2006).