Second, as many have warned, a civil war would bring the Turkish military across the border and into the northern provinces. Turkey's primary concern would be to prevent Iraqi Kurds from declaring independence and encouraging further unrest in Turkey's own Kurdish regions. But neither Turkish Prime Minister Recep Erdogan nor the country's military leadership favor the deployment of troops into the strategically crucial Iraqi city of Kirkuk. They know the move would badly damage relations with the United States--to say nothing of Turkey's bid to join the European Union.

We could expect Turkey to move a substantial number of troops a few kilometers inside Iraq to establish a buffer zone against Kurdish militants and Islamist insurgents. This limited move, which U.S. and European leaders would probably quietly accept, would roil Turkish markets and block the cross-border trade that has boosted both the Turkish and Kurdish economies. But Turkey has no interest in involving its troops directly in an Iraqi civil war.

Third, though such a conflict would heighten the risk of terrorist attacks in the region, it is unlikely to provoke any sharp sudden rise in their number or intensity. Iraqi insurgents have upgraded both their weapons and the technical skill with which they wield them. But states like Saudi Arabia, Syria and Iran have the means to limit the risk of large-scale terrorist attacks within their borders, even if they struggle to quell unrest in areas dominated by minorities.

So a civil war would be costly, it would be messy, it would be complicated. But much like Bosnia's civil wars during the 1990s, an Iraqi civil war could also be contained. The challenge for U.S. policymakers is to assess whether the costs of preventing a civil war in Iraq outweigh those of containment--and the international political fallout that would attend a failure to keep Iraq whole.

The presence of substantial numbers of U.S. troops is essential for the near term because civil war in Iraq is not inevitable. Civil war can be averted if the new government clears a series of legislative and political hurdles. But if the new government cannot take the hard decisions needed to keep the country together, what should the United States do?

Stability and Security

The Bush Administration hopes the destruction of the Askariya mosque and the continuing violence it has incited have given Iraqis a good look at the costs of civil war--and that all sides, sobered, will renew their commitment to form a viable central government.

But the recent violence also demonstrates that some Iraqis believe there is something to be gained from bloodshed. There are Sunnis who consider destruction and chaos preferable to political domination by Shi'a. There are Shi'a who believe that Sunni violence can only be answered by men with guns. Both groups remain a distinct minority within their communities. But in the new Iraq, an armed minority can inflict lasting damage on the aspirations and interests of all. Once in motion, violence takes on a logic and momentum of its own, and the risk is great that no one can stop it.

With a continued military presence and the diplomatic leverage it provides, the United States can continue to try to put out fires long enough to give the new government its best near-term chance for success. But coalition troops alone cannot guarantee long-term stability.

All Iraqis know that the foreign soldiers will eventually leave. As that day approaches, U.S. influence in Iraq can only diminish. As many commentators have recently reminded us, Abraham Lincoln once said of the U.S. civil war, "All dreaded it, all sought to avert it. And the war came." The White House should now define what civil war means for Iraq, the region and the United States--and plan how best to prepare for the day it might come.

Ian Bremmer is president of Eurasia Group, a political risk consultancy, and a contributing editor to The National Interest. He is author of The J Curve: A New Way to Understand why Nations Rise and Fall (2006).