But as dreary reports of sectarian violence continue to arrive daily and Americans lose the will to "stay the course," such an Iraq appears more and more elusive.

A constitutional-democratic Iraq requires first having an Iraqi state -- an entity with, in Max Weber's words, a "monopoly on the legitimate use of force." The United States destroyed the Iraqi state in 2003 and ever since has been trying to reassemble the rubble and implement constitutional democracy at the same time. America set itself a difficult task. Constitutional democracy is about reason and consensus; states are usually built by force.

In a country such as Iraq -- really a set of societies that are mutually mistrustful, penetrated by foreign entities, competing for valuable resources and bristling with arms -- it may be that relentless, pitiless coercion is needed. The United States, a liberal democracy in an age of instant telecommunications, cannot be relentless or pitiless enough. By insisting on keeping troop levels low, the Bush administration has shown that it is not willing to pay very much to rebuild the Iraqi state.

As Washington pursues a receding goal, support for disengagement grows. Quitting Iraq would likely worsen the ongoing Iraqi civil war. This intensified war could attract direct intervention from Iran, and perhaps Syria, no longer deterred by U.S. troops and keenly desiring a Shiite victory. That, in turn, would place pressure on other Iraqi neighbors -- Jordan, Saudi Arabia, Turkey and Israel -- to counter-intervene.

Averting or ending a regional war would require a return of U.S. troops. If an exhausted America refused -- think the fall of Saigon in 1975 -- the end would likely be the triumph of the Iraqi Shiites, who make up approximately 60 percent of Iraq's population and have well-armed and well-motivated militias.

How bad would a Shiite-dominated Iraq be for U.S. interests? Such an Iraq would likely be authoritarian, inasmuch as Sunni Arabs and Kurds would not accept it without coercion. As a majority-Arab state with its own interests, this Iraq would not be a puppet of Persian Iran. Indeed, during the U.S.-led occupation many prominent Shiite politicians have walked a delicate line between Washington and Tehran. Yet a Shiite Iraq would likely lean eastward, which in turn would solidify the nascent and much-discussed "Shiite Crescent" extending from the Gulf of Oman to the Mediterranean.

A Shiite Crescent is alarming on its face, particularly in the threat it would pose to Israel. But not only Israel would feel compelled to take counter-measures. Iranian ascendancy would alarm Sunni actors -- both states such as Saudi Arabia and terrorist networks -- who aspire to unite the umma under their leadership rather than that of the Shiites. These actors would have some interest in drawing together to counter-balance Iranian leadership.

A purely realist reading would suggest that the region will enter some sort of equilibrium. Indeed, America and other Western powers might induce Syria -- a country ruled by Shiite-related Alawites but whose population is 70 percent Sunni -- to abandon Iran, which would complicate Iranian plans for regional hegemony. Still, this outcome relies heavily on spontaneous power balancing based upon divisions within the Muslim world and also on heroic restraint by Israel.

The risks of quitting Iraq, then, are very great. What other course is open to the United States? Some prominent foreign policy thinkers have favored the partition of Iraq into Shiite-Arab, Sunni-Arab and Sunni-Kurdish rump states. The remarkable rise of Iran over the past few years compels us to take the three-state solution seriously.

The partitioning of Iraq would be dangerous -- more like that of India or Yugoslavia than that of the Soviet Union or Czechoslovakia. Much Iraqi territory, including Baghdad itself, is home to members of more than one ethnic group. Many of Iraq's oil deposits and facilities straddle the current ethnic boundaries, and areas that are exclusively Sunni-Arab have little oil.

The risk is great that some oil fields, or the Sunni-Arab rump state itself, would fall into jihadi hands. Hence, outsiders must assist with the partition, and Iraq's neighbors would insist on being involved. America's ally Turkey is adamantly opposed to an independent Kurdish state, and working with Iran would be a delicate matter indeed.

Should all of these difficulties be managed, however, a three-state solution might better maintain a balance of power in the region between Iran and Sunni actors. A model would be the fragmented Holy Roman Empire (Germany) after the 1648 Peace of Westphalia. Some German states gravitated toward France, others toward its competitors.

The system did break down on those occasions when Louis XIV and later Napoleon attempted to subjugate the German states. But in general, the fragmentation of Germany that lasted until 1871 reassured the great powers, and indeed they took pains to preserve it. In the case of a partitioned Iraq, the rump Shiite state would tilt toward Iran, the rump Sunni-Arab one toward Jordan and Saudi Arabia and the rump Kurdish one (perhaps) toward the United States.

If the regional powers agreed to the independence of the three states, they would develop an incentive to uphold the status quo. Furthermore, insofar as the rump states were ethnically homogeneous, they would not themselves dissolve into civil wars, and hence not entice neighbors to intervene on behalf of "their" people.

It is the nature of the Tehran regime that forces us to think about Iran when thinking about Iraq. Were the long-awaited liberal reform or revolution in Iran to come, that country's goals would probably overlap more with those of the United States. Unlike in Sunni states such as Egypt and Saudi Arabia, much of the Iranian public -- particularly the educated urban young -- is oriented toward the West. Their country, unlike Egypt and Saudi Arabia, has actually tried Islamism, and has found it wanting. There are reports that even conservative Iranian forces are worried about the costs that Ahmadinejad's radical policies and statements have inflicted on the country. A constitutional-democratic Iran would still want regional influence, but that influence would be more consistent with U.S. interests.

So regime change in Iran is to be hoped for and encouraged, but not counted on, and certainly not attempted by force. In the meantime, we must proceed in Iraq so as to contain Iran. Three-and-a-half years after the fall of the Baathist regime, America's least bad option in Iraq may well be to begin organizing, with Iraqis and other external actors, the partition of the country -- a country cobbled together by the British in 1921 and held together ever since by coercion and repression.

Partition would not be a happy ending to Operation Iraqi Freedom. If it is too risky, then Washington should send thousands more troops now. Riskiest of all may be continuing the status quo: We are losing, Iran is winning, and we are enemies. One of those three facts must change.

(John M. Owen IV is associate professor of politics at the University of Virginia and author of Liberal Peace, Liberal War: American Politics and International Security (1997). This article was reprinted with permission from The National Interest.)

(United Press International's "Outside View" commentaries are written by outside contributors who specialize in a variety of important issues. The views expressed do not necessarily reflect those of United Press International. In the interest of creating an open forum, original submissions are invited.)

Related Links
Iraq: The first techonology war of the 21st century