IF THE surge fails, it is imperative that the United States have a plan for containing the Iraqi civil war. As painful as it may be to admit that the U.S. effort to bring peace and stability to Iraq has failed, our new priority must become preventing the Iraqi conflict from spilling over and destabilizing neighboring states. This will not be easy. In fact, the history of states trying to contain the "spillover" from civil wars suggests that it will be very hard for the United States to do so. But planning now may allow the United States to mitigate spillover's worst effects.

What Spillover Means

THE COLLAPSE of Iraq into all-out civil war would mean more than just a humanitarian tragedy that could easily claim hundreds of thousands of Iraqi lives and produce millions of refugees. Such a conflict is unlikely to contain itself. In other similar cases of all-out civil war the resulting spillover has fostered terrorism, created refugee flows that can destabilize the entire neighborhood, radicalized the populations of surrounding states and even sparked civil wars in other, neighboring states or transformed domestic strife into regional war.

Terrorists frequently find a home in states in civil war, as Al-Qaeda did in Afghanistan. However, civil wars just as often breed new terrorist groups-Hizballah, the Palestine Liberation Organization, the Salafist Group for Preaching and Combat of Algeria, and the Tamil Tigers were all born of civil wars. Many such groups start by focusing on local targets but then shift to international attacks-starting with those they believe are aiding their enemies in the civil war. This process is already underway in Iraq; the 2005 hotel bombings in Amman, Jordan were organized from Iraqi territory. Iraq-based groups are also inspiring others to emulate their targets and tactics. As they regularly do in Iraq, jihadist terrorists have tried to strike Saudi Arabia's oil infrastructure, a switch from the jihadists' past avoidance of oil targets. Moreover, their Improvised Explosive Device technologies are showing up in Afghanistan.¹ Suicide bombing, heretofore largely unknown in Afghanistan, is also now a regular occurrence, with the Iraq struggle providing a model to jihadists in Al-Qaeda's former home.

In turn, an ongoing civil war can contribute to the radicalization of populations in neighboring countries. Already, the war has heightened Shi‘a-Sunni tension throughout the Middle East. In March 2006, after Sunni jihadists bombed the Shi‘i Askariya Shrine in the northern Iraqi city of Samarra, over 100,000 Bahraini Shi‘a took to the streets in anger. Bahraini Shi‘a are simultaneously horrified at the suffering of their co-religionists in Iraq and emboldened by their political successes. As one Bahraini Shi‘a politician noted, "Whenever things in Iraq go haywire, it reflects here."²

And as Iraq descends into further violence, the numbers of refugees will likely increase exponentially. Iraq has already generated roughly two million refugees with another one million internally displaced. These represent large groupings of embittered people who can serve as a ready recruiting pool for armed groups still waging the civil war. And as the wars in Africa's Great Lakes region show, foreign countries where refugees find shelter can become caught up in the civil war. At times, the refugees simply bring the war with them: The fighters mingle with non-combatant refugees and launch attacks back in their home countries, while those who drove them out continue the fight against the refugees in their new bases. Neighboring governments may try to defend refugees on their soil from attacks by their enemies or at times exploit the refugees as a proxy for the governments' own ambitions. Moreover, large refugee flows can overstrain the economies and even change the demographic balances of small or weak neighboring states, upsetting what is often a delicate political balance.

Then there is the "demonstration effect" caused when a civil war involves one group seeking separation or independence as the solution to its problems. Sometimes, other groups in similar circumstances (either in the state in civil war or in neighboring countries) follow suit if the first group appears to have achieved some degree of success. Thus Slovenia's secession from Yugoslavia started the first of those civil wars, but it also provoked Croatia to declare its independence, which forced Bosnia to follow suit-and in both of the latter cases Serb enclaves within both countries sought to secede from the seceding state and join with Serbia.

All the problems created by these and other forms of spillover often provoke neighboring states to intervene-to stop terrorism as Israel tried repeatedly in Lebanon, to halt the flow of refugees as the Europeans tried in Yugoslavia and the Iranians in Afghanistan, or to end (or respond to) the radicalization of its own population as Syria did in Lebanon. These interventions usually turn out badly for all involved. Iraq is already seeing both actual intervention and threats of intervention. Iran has hundreds, or perhaps thousands, of intelligence and paramilitary personnel in Iraq and is arming an array of Iraqi groups. Leaders of Turkey, Saudi Arabia and Jordan have all threatened interventions of their own, both to mollify domestic sentiment and to counter what they see as unchecked Iranian gains from Tehran's intervention.

What Can We Do-or Avoid Doing?

IF IRAQ spirals into an all-out civil war, the United States will have its work cut out attempting to prevent spillover from destabilizing the region and threatening key governments, particularly Saudi Arabia. Not being prepared to quickly fall back to a containment posture will lead to an ad hoc approach that will involve many avoidable mistakes and missed opportunities.

One of the most difficult challenges for the United States will be simply to not make a bad situation worse. Many of the policy options being discussed for Iraq, however, have the potential not only to fail, but to further undermine U.S. interests.

In this respect, Washington should not try to pick "winners." The temptation for the United States to try to aid one Iraqi faction against another in an effort to manage such a civil war from within is enormous. In Washington, such management may seem like playing a clever game of chess. Unfortunately, the historical reality often proves very different. Proxies frequently fail in their assigned tasks or turn against their masters. As a result, such efforts rarely succeed, and in the specific circumstances of Iraq, such an effort appears particularly dubious.

Once an internal conflict has metastasized into all-out civil war, military leadership often proves to be a crucial variable in determining which faction prevails (sooner or later). However, it is extremely difficult to know a priori who the great military commanders are. We know about Moqtada Sadr and Abdul Aziz al-Hakim but very little about the field commanders of either the Mahdi Army or the Badr Organization, to name only the two best-known Iraqi militias. And in some cases we don't even know the relevant militias, let alone their leaders.

Moreover, many communities are divided, fighting against one another more than against their supposed enemies. Commentators often speak of "the Shi‘a" or "the Sunnis" as if they were discussing the Confederates or the Roundheads. In fact, Iraq's Shi‘a population is fragmented among dozens of militias, many of which hate and fight one another as much as they hate and fight the Sunnis. It is an important element in the chaos of the country today-and one attested to by recent battles in Amarah, where Mahdi Army forces squared off against the Badr Brigade, and Diwaniyah, where Mahdi Army forces fought Iraq's Shi‘a-dominated security services, as well as the nightly bloodshed in Basra-all of which is Shi‘a-on-Shi‘a. Thus Iraq's Shi‘a may go the way of the Palestinians or the various Lebanese factions, who generally killed more of their own than of their declared enemies. What is true for the Shi‘a is just as true for the Sunnis.

A second specific problem for the United States in trying to pick (or create) a winner in an Iraqi civil war is the question of how America would support its choice. Say we choose the Shi‘a: All of the Shi‘a militias are strongly anti-American and some are closely tied to Iran, and none of Iraq's Sunni neighbors (Saudi Arabia, Kuwait, Jordan or Turkey) would help us to engineer a Shi‘a militia conquest of Iraq. The Sunni neighbors would be glad to help us support a Sunni militia gain control of the country, but most of these militias are closely aligned with Al-Qaeda in Mesopotamia and other salafi jihadi groups-the principal target of the U.S. War on Terror-certainly an unpalatable choice.

And whichever group the United States chooses to support would have to slaughter large numbers of people to prevail and establish control over the country-especially true in the case of the Sunni Arabs, who make up no more than 20 percent of the population.

This is why some argue that the solution to civil war is partition-as Senator Joe Biden (D-DE) and John Owen IV, among others, have argued in these pages. The basic problem with pursuing any version of partition today in Iraq is that it is probably impossible to do so without either causing an all-out civil war or deploying the hundreds of thousands of American and other first-world troops whose absence has been the first-order problem preventing reconstruction from succeeding. Other than the Kurds, few Iraqis-whether political leaders, militia commanders or ordinary citizens-want their country divided. And many of those who are fleeing their homes are not merely peacefully resettling in a more ethnically homogeneous region, but are joining vicious sectarian militias like the Mahdi Army in hope of regaining their homes, or at least extracting revenge on whoever drove them out.