What Next?

In Iraq, it is not clear which proxy would be the most effective militarily. Many communities are divided, fighting against one another more than against their supposed enemies. Iraq's Shiites may go the way of the Palestinians or the various Lebanese factions, who generally killed more of their own than of their declared enemies.

Manage the Kurds

Should chaos engulf Iraq, the Kurds will understandably want out, but this risks inspiring secessionists elsewhere in Iraq and throughout the region. In return for the Kurds agreeing to postpone formal secession, Washington should offer them extensive economic aid, assistance with refugees and security assurances (perhaps backed by U.S. troops) -- as well as promising support for their eventual independence when Iraq is more stable.

Buffer the Borders

One of Washington's hardest tasks would be to prevent the flow of dangerous people across Iraq's borders in either direction -- refugees, militias, foreign invaders and terrorists.

One option might be to create a system of buffer zones and refugee collection points inside Iraq staffed by U.S. and other coalition personnel. These collection points would be located on major roads, preferably near airstrips along Iraq's border -- thus on the principal routes that refugees would take to flee, providing a good logistical infrastructure to house, feed and otherwise care for tens or hundreds of thousands of refugees. Iraqi refugees would be gathered at these points and held there. In addition, coalition military forces would defend the refugee camps against attack, pacify and disarm them, and patrol large swaths of Iraqi territory nearby.

These zones would serve as "catch basins" for Iraqis fleeing the fighting, offering a secure place to stay within the nation's borders and thus preventing them from destabilizing neighboring countries. At the same time, they would serve as buffers between Iraq and its neighbors, preventing other forms of spillover -- such as militia movements, refugee flows out of Iraq and invasions into Iraq.

The catch-basin concept, while potentially useful, faces at least one big problem: Iran. Unlike Iraq's borders with Jordan, Kuwait, Saudi Arabia and Syria, the Iranian border is too long and has too many crossing points for it to be policed effectively by smaller numbers of coalition troops. Iran will never allow the United States the access across its territory, let alone logistical support, that would be necessary to make catch-basins along the Iran-Iraq border realistic. Thus, this scheme could make it look as though the United States was turning Iraq over to the Iranians, with the catch-basins effectively preventing intervention by Iraq's Sunni neighbors while doing nothing to deter Iran. For this reason, the United States's clear red lines to Iran about not intervening (at least overtly) would have to be enforced assiduously.

Perhaps most important, the catch-basin proposal requires Americans to endure significant long-term costs -- both in blood and treasure -- in Iraq. The United States would still need to deploy tens of thousands of troops to the nation (albeit on its periphery), as well as supplies to feed and care for hundreds of thousands of refugees. The United States would still occupy parts of Iraq, and the U.S. presence would remain a recruiting poster for the jihadist movement. Finally, all of these costs would have to be endured for as long as the war rages; recall that refugees from the wars in Afghanistan lived away from their homes for more than 20 years.

* * *

No country in recent history has successfully managed the spillovers from a full-blown civil war; in fact, most attempts have failed miserably. Syria spent at least eight years trying to end the Lebanese civil war before the 1989 Taif accords and the 1991 Persian Gulf War gave it the opportunity to finally do so. Israel's 1982 invasion was also a bid to end the Lebanese civil war after its previous efforts to contain it had failed, and when this also failed, Jerusalem tried to go back to managing spillover. By 2000, it was clear that this was again ineffective and so Israel pulled out of Lebanon altogether.

Withdrawing from Lebanon was smart for Israel for many reasons, but it did not end its Lebanon problem -- as the latest conflict showed all too clearly. In the Balkans, the United States and its NATO allies realized that it was impossible to manage the Bosnian or Kosovar civil wars and so in both cases they employed coercion -- including the deployment of massive ground forces -- to bring them to an end.

That point is critical: Ending an all-out civil war typically requires overwhelming military power to nail down a political settlement. It took 30,000 British troops to bring the Irish civil war to an end, 45,000 Syrian troops to conclude the Lebanese civil war, 50,000 NATO troops to stop the Bosnian civil war, and 60,000 to do the job in Kosovo. Considering Iraq's much larger population, it probably would require 450,000 troops to quash an all-out civil war there. Such an effort would require a commitment of enormous military and economic resources, far in excess of what the United States has already put forth.

How Iraq got to this point is now an issue for historians (and perhaps for voters in 2008); what matters today is how to move forward and prepare for the tremendous risks an Iraqi civil war poses for this critical region. The outbreak of a large-scale civil conflict would not relieve us of our responsibilities in Iraq; in fact, it could multiply them. Unfortunately, in the Middle East, one should never assume that the situation can't get worse. It always can -- and usually does.

dlb32@georgetown.edu

kpollack@brookings.edu

Daniel L. Byman is director of Georgetown University's Center for Peace and Security Studies. Kenneth M. Pollack is research director at the Saban Center for Middle East Policy at the Brookings Institution.