In 1953, after the armistice ending the Korean War, South Korea lay in ruins. President Eisenhower was eager to put an end to hostilities that had left his predecessor deeply unpopular, and the war ended in an uneasy stalemate. But the United States had a strong interest in regional stability, and some worrisome enemies to keep in check. So Eisenhower decided to leave tens of thousands of troops behind, and signed a treaty with the U.S.-backed government to formalize their presence. Thirty-five years later, South Korea emerged as a stable democracy. 

The situation in Iraq today bears some intriguing similarities. The reduction of American forces in Iraq to 50,000 is thus good news—but not because it is a step closer to complete withdrawal. In the coming year, the Iraqi government (once it is formed) is likely to ask the U.S. to keep some significant number of troops in the country after the pullout date of summer 2011. If so, President Obama may well agree, because it is just about the only way to avoid a resurgence of civil war and continue Iraq's tenuous progress toward consolidating democracy. As in South Korea—where nearly 30,000 U.S. troops remain today, almost 60 years after the war ended—patience may pay off. Then there is the ethical side of the issue: If the elected Iraqi government asks for help, the U.S. owes it to them to continue its commitment.

Iraq faces a raft of difficulties if it is to become an effective, self-governing nation, and all of them point to the need for a continuing U.S. role in security and beyond. Start with the basic question of who or what is an Iraqi. In the last three years, Iraq's citizens pulled back from the brink of civil war. But they did so because of the surge of U.S. troops, not because they had forged a national consensus on living together democratically.

Iraqis' primary identities are still of religious denomination or ethnicity, not of Iraqi nationhood—and that may remain the case indefinitely. Iraqi national identity under Saddam Hussein never truly incorporated Shiites or Kurds. Sunnis, who identified most closely with the Iraqi nation, remain in some ways disenfranchised relative to the other groups, or at least they perceive themselves that way. Having run the country since its birth under the British, many Sunnis experience the rise of the Shiites and Kurds as a defeat for them in the zero-sum game of domination.

A new Iraqi identity might conceivably emerge on the basis of common interest—the way Flemish and French-speaking Belgians have, despite historic distrust, so far stayed uneasily together out of inertia, convenience and economic advantage. But that is a far cry from the sort of commonality needed to sustain a state under difficult conditions or the threat of dissolution. Civic citizenship on the basis of pride in democratic institutions is a more desirable endpoint—but to get there, those institutions first must do a good job of functioning.

Then there is the danger that Iraq's fissiparous character could drive it back into civil war. Civil war began to break out in 2004 and 2005 because, in the absence of a national military or an effective occupying force, ordinary people did not believe the government was capable of protecting them. That forced everyone to find someone who might be capable of doing the job—and for most people, the best candidate was a militia. Sunnis squared off against Shiites. Meanwhile Iraq's Kurds cast a baleful eye over the whole situation, and spoke in robust terms about regional autonomy that sounded very much like independence.

Faced with the collapse of Iraq into something like Lebanon—or worse, Somalia—the Bush administration opted for a new counterinsurgency strategy. Violence was reduced because, for the first time since the U.S. invaded Iraq in 2003, Iraqis felt that there was a force capable of dominating the situation and ensuring basic order. The surge worked because it managed to convince the vast majority of Iraqis that they had more to gain by pursuing power peacefully than through violence.

But if war is politics by other means, peaceful politics in Iraq is still not far removed from war. A perceived (or real) security vacuum like the one that sparked the civil war will put us right back where we started. The Iraqi security forces have grown enormously, to more than 660,000 split roughly in half between police and military. They have also improved markedly in quality (a low bar considering how weak they were for most of the last seven years). But they are still some distance from being able to ensure safety for ordinary citizens. Their ranks could still break apart into their constituent ethnic and denominational groups. And they are nowhere near able to protect Iraq's borders from serious threats that might arise in, for example, a regional war involving Iran.

Iraqi politics is still unsettled and unsettling. No party won a clear majority in the March 2010 elections, and since then, no government has been formed. The two largest Shiite parties, which split before the election, have since reunited, giving them together nearly half the seats in the assembly; but they would still need to make a deal with the Kurdish parties to form a government, and that has not yet happened. A Shiite-Kurdish coalition would parallel the structure of elected governments in Iraq since the U.S. occupation formally ended. But it would further alienate Sunnis, who after previous boycotts finally participated fully in an election, supported the winning party (led by Ayad Allawi)—and yet may find themselves out of power once more.

To convince Iraqis that their government is not about to collapse and fail to protect them, a guarantor is needed. Right now, only the U.S. can offer a credible guarantee. This is the reason that many observers, including Ryan Crocker, former U.S. ambassador to Iraq, expect Iraq's elected leaders to ask for troops past the planned pull-out target. The political class has the most to lose if the shaky edifice it has erected begins to teeter. Its members have also grown accustomed to working with the U.S., often at high levels. The cost to Iraqi politicians of asking the former occupier to stick around is likely to be offset by the tremendous gains in public confidence associated with a prolonged American commitment—especially if they ask early in their own election cycle.

A request in the next year, even after midterm elections in the U.S., will still feel like bad timing for the Obama administration, which would prefer to go into its own run for a second term with Iraq completely off the political agenda. But the risks of rebuffing a serious Iraqi request for troops would be enormous. A negative reply would probably speed up disaster within Iraq, since it would send a message of no-confidence that would frighten Iraqis and reactivate militias. Renewed civil war before a U.S. election would allow a Republican challenger to argue that Obama had taken his eye off the ball in Iraq. By contrast, leaving behind some tens of thousands of troops would carry little domestic cost, at least so long as Iraq remains relatively quiet. Al Qaeda may use the troops' presence as proof of an American plot to occupy Arab lands, but this charge can be balanced by the fact that the request will have come from a democratically elected Iraqi government.

What's more, the U.S. has serious, long-term interests in keeping Iraq stable and on the road to democracy—and in maintaining a role in regional security. The reasons go well beyond Iraq's oil or its proximity to Iran, which remains one of the handful of countries that actively considers itself an enemy of the U.S. An Iraq that devolved back into civil war would heighten regional tensions between Sunnis and Shiites—a state of affairs that was becoming increasingly dangerous when the civil war in Iraq was smoldering before the surge.

Civil war would also increase the risks of Kurdish declaration of independence. This would inflame Turkey, a country that, concerned with the progress of its European Union application, is already looking to Syria and other Middle Eastern states as potential allies more than it has in the past. Then there is the tremendous benefit that a reversal of Iraq's progress would bring to al Qaeda, both as a matter of propaganda and also because they could take advantage of the resulting chaos.

As for democracy, Iraq is only at the beginning of a lengthy, tenuous and risky process of consolidating the rough set of political norms that have been put in place. The first and most important element for making democracy grow is time. Repetition of successful elections and successful government operation is the single best mechanism for convincing both political elites and ordinary people that democracy can work for them. The practical core of democracy, defined functionally, is the peaceful exchange of power between different groups of powerful political players arranged in parties. Each party must believe that, when it is out of power, it is worth waiting for its own turn to govern and share the spoils—or else it would have an interest in breaking the system and reaching for power by nondemocratic means. The more times power changes hands, the greater the odds that the players will come to believe that their chance will indeed come.

The other crucial element of democratic consolidation is the spread of belief in its core ideals: voting, equality and liberty. For these values to be more than mere slogans, there must be security and stability. Order precedes law, not the other way around. Once order is in place, government officials need to respect the decisions of electoral majorities and administer justice fairly and effectively. The presence of U.S. civilians can help provide assurances that the government is operating the way it is supposed to do—and those civilians need the protection of U.S. forces if they are to operate safely in a still-dangerous environment.

Over time, however, South Korea grew economically, and political development came in tandem. In 1987, real elections took place, and since then, South Korea has blossomed into a free and functioning democracy. The presence of U.S. forces provided a background security guarantee throughout that process, one that Koreans have used to good effect.

Today, many South Koreans would like to see U.S. troops leave their country. If the U.S. were to find itself a generation from now in a similar situation in Iraq, that would be cause for celebration. In the meantime, if Iraqis seek continued help, we would be well advised to agree. We broke Iraq, let us remember. And if we have never quite bought it, we have a basic responsibility to help put it back together.

—Noah Feldman is a professor at Harvard Law School and a former senior adviser to the Coalition Provisional Authority in Iraq.