The garbage is burning on the streets of Port-au-Prince. After a decade in which the international community spent well over $2 billion to foster stable institutions in Haiti, cars and trucks crush the uncollected trash into a paste on the city's shattered roads, while citizens burn it down to a smelly, blackened nub along what once were sidewalks. The same people who loathed the dictator Jean-Claude Duvalier now recall that, under ''Baby Doc,'' Port-au-Prince at least had a sanitation service.

I was in Haiti just a few weeks after a small American-led force debarked in Port-au-Prince, this time to restore order after the ouster of the very same Jean-Bertrand Aristide whose restoration to power had been protected by American troops 10 years before -- a shocking reversal of fortune not only for the populist leader but also for the nation-building project itself. What had happened to all those training programs for the police, judges and jailers, to all those World Bank loans for development? Peacekeeping experts argued that the international effort had simply been too halfhearted. I heard something of that in Port-au-Prince, but not much. When I all but invited one of Haiti's leading economists, a former official in the Aristide government, to blame the nation-building effort, he shot back, ''You could have put in three, four, five billion dollars, and you would have had the same results.'' Aristide, he said, would have made sure it was misused.

For Americans, and for the United Nations and the other multilateral bodies that over the past decade have taken up the worldwide burden known as nation-building (in Pentagonese, ''stability operations''), Haiti offers an object lesson -- just as do the ethnic riots that flared up in Kosovo last month, leading to the deaths of 28 Albanians and Serbs. The lesson is not ''nation-building can't work''; the lesson is that it is very hard, and that it demands a great deal of both patience and modesty -- qualities that do not come naturally to American policymakers or, for that matter, to Americans.

Since 2000, I have spent time in four countries with major peacekeeping missions -- Sierra Leone, East Timor, Kosovo and Haiti. Some, like Sierra Leone, consist largely of keeping people from one another's throats. In others, like East Timor, the work involves the laborious building up of what has been systematically wrecked. All these countries, and almost all candidates for peacekeeping, have been profoundly damaged physically, culturally, even psychically, by the time they reach the world's attention. Nation-building is a rehabilitative act, often performed on a patient who does not consider himself sick and is busily trying to tear off the bandages. This, of course, is the situation in which the Bush administration, which blithely ignored the lessons of past missions, has found itself in Iraq. The culture of Iraq often seems so deeply poisoned as to utterly defeat our well-intentioned meddling. Perhaps, in the end, that will prove to be the case.

But what is almost impossible turns out to be indispensable. It has become obvious since 9/11 that we cannot allow collapsed states, or rogue states, to fester; their failures have become our problem. Nation-building is no longer a subject for debate: we will get it right or pay the price.

During the 2000 presidential debates, George W. Bush mocked the idea of nation-building as a dangerous Democratic folly. The function of the American military, he often repeated, was ''to fight and win wars.'' Bush gave the impression that nation-building was something Bill Clinton and his team of woolly-headed multilateralists had dreamed up. But the truth is that while the term is new, the endeavor is not, and the people who know this best are current and former members of the United States military. ''Look at the horse cavalry in the Old West,'' said Col. George Oliver, an infantry officer who served with the Rangers and the Special Forces and headed the Army's Peacekeeping Institute at its War College in Carlisle, Pa. ''They were keeping down the renegade Indians while the settlers established themselves.'' After World War II, Army civil affairs officers ran conquered cities across Europe. Oliver said that his father served in the Constabulary Forces, a gendarmerie that helped police postwar Germany.

Then came the cold war, and peacekeeping began to seem an unaffordable luxury. ''The cold war was the aberration,'' said Scott Feil, a retired Army colonel who has studied postconflict situations. ''If you took a battalion out of Germany, you could predict how much quicker the Soviet forces would get to the Rhine. It was easy to do the risk analysis and say, 'Too risky.''' The civil wars that consumed nations around the globe during the 60's and 70's became proxy battles for the two great powers, making impossible the cooperative effort required for peacekeeping. And the Vietnam War taught the military a different but related lesson: stay away from anything ''political'' -- the task is to fight and win wars, period. Meanwhile, the cold war paralyzed the United Nations Security Council, which had been designed to enforce and preserve peace. The only peacekeeping missions that could be mounted were passive affairs, in Cyprus, the Sinai and elsewhere, where United Nations soldiers stood between two armies.

With the end of the cold war, many of the old client states, suddenly left to their own devices, slipped into chaos. Intervention had become possible, and the fear of humanitarian catastrophe made it unavoidable. So the United Nations and the great powers embarked on a new kind of quasi-military operation for which, it quickly turned out, there was no blueprint. In Somalia, in 1993, the Clinton administration agreed for the first time to permit the American military to serve within the structure of a United Nations peacekeeping operation. Somalia began as a humanitarian mission in which troops oversaw the distribution of food -- successfully, by most accounts -- then suddenly became a military adventure when warlords killed a group of Pakistani soldiers. The attempt to capture the chief warlord, Muhammad Farah Aidid, ended, of course, in catastrophe. The death of 18 Rangers in the ''Black Hawk Down'' incident was a mini-Vietnam, with its very own never-again lesson: no ''mission creep.'' If you can't go in big, with all guns blazing, don't go in at all. When a genocidal civil war broke out in Rwanda the following year, the Clinton administration made it clear from the outset that the United States would not intervene.

The Pentagon now termed the range of postconflict activities as ''military operations other than war,'' referred to as ''mootwa.'' Gen. John Shalikashvili, chairman of the Joint Chiefs of Staff from 1993 to 1997, is said to have declared, ''Real men don't do mootwa.'' (Shalikashvili said that he was himself pro-mootwa, but he did attribute the sentiment to other senior officers.) The Clinton administration was, in fact, deeply divided over the merits of peacekeeping, with the State Department typically on one side and the Pentagon on the other. The military itself was divided on the subject, in part along generational lines. Many younger military officials came to feel that peacekeeping and the inevitable political and diplomatic tasks that went with it were not a distraction from the real work, but were the real work itself, or much of it. This generation included Marine Corps leaders like Gen. Anthony Zinni -- who served as an officer or a diplomat in Kurdistan, Somalia (where he was police chief of Mogadishu), Eritrea and the former Soviet republics -- as well as top Army figures like Wesley Clark and the former and current commanders in Iraq, Tommy Franks and John Abizaid.

The lesson of Somalia, for them, was ''Let's get it right.'' Gen. Gordon Sullivan, the Army chief of staff in the early 90's, served not only in Korea and Vietnam but also in postwar settings in Panama and Kurdistan. Sullivan concluded that peacekeeping operations were not aberrations to be avoided; they were the inevitable consequence of a disintegrating, post-cold-war world. ''Saying, 'You can't do peacekeeping,''' he noted, ''is like saying, 'I'm going to leave all these petri dishes with this liquid in them and let viruses breed.''' While American troops were still in Somalia, Sullivan successfully pushed for the creation of a Peacekeeping Institute on the campus of the Army War College. The institute would both teach and help formulate a doctrine of peacekeeping. It offered two elective courses, in peacekeeping operations and negotiation -- both always oversubscribed -- to the 350 lieutenant colonels who went through the War College every year. Meanwhile, the Pentagon created a new Office of Peacekeeping and Peace Enforcement Policy.

Among military people, the fight-and-win-wars view had come to seem like a cold-war artifact; most active-duty officers had, after all, seen most of their action in postconflict settings. Two years ago, the Peace Through Law Education Fund, a Washington-based research organization, convened a group of 33 senior military officials, including service chiefs (and several NATO commanders), to survey their views on peacekeeping. The consensus, according to the final report, was that peacekeeping operations not only do not harm combat-readiness, as had often been claimed, they also ''enhance the great majority of combat skills.'' The report said ''engagement in multilateral peace operations is in our national interests and will be a key ingredient in the war against terrorism'' -- and that the United Nations, while generally ineffective at fielding a fighting force, can play a crucial role in forging a political settlement and helping establish the rule of law.

There is a nomenclatural confusion in the peacekeeping world that mirrors the complexity and variety of the work itself. ''Peacekeeping'' is a United Nations term, which is one reason it has a bad reputation among the hard of head. United Nations officials use terms like ''peace-building'' and ''post-conflict reconstruction'' to distinguish military operations -- designed to preserve stability in a war-torn country -- from the largely civilian work of recreating institutions in order to sustain and fortify the peace. The Pentagon term ''stability operations'' is intended to draw a bright line between preventing people from killing one another and putting a country back together again, but one of the first things American peacekeepers discovered after the cold war is that the distinction was unsustainable on the ground.

Late in the fall of 1995, NATO sent a force of 60,000 into Bosnia to enforce the terms of the Dayton accords. Maj. Gen. William Nash, who had gotten a taste of peacekeeping as the governor, effectively, of southern Iraq in the immediate aftermath of the gulf war, was the first commander of the American division in Bosnia. For the first four months, Nash said, he policed the new Dayton-mandated borders, confiscated and stored weapons and kept the ethnic factions separated. Then he realized that his real problem was idle young local men. So Nash put them to work building things. Then he and his troops had to secure several polling areas for the first national election. That sounds like mission creep; but Nash had discovered that he couldn't separate peacekeeping from nation-building. ''The first rule of nation-building is that everything is related to everything,'' Nash said, ''and it's all political.'' Everything, that is, impinges on somebody's power, and in order to establish stable democratic institutions you have to deal with, and often confront, the political structures that provoked the conflict in the first place.

In 1995, the Clinton administration began devising a ''generic political-military plan'' to apply to future missions. Len Hawley, the N.S.C. official who drew up the plan, said that over the ensuing years, as he and others learned more about the dynamics of peacekeeping and other interventions, this blueprint grew from 7 to 60 pages. A vast new literature of peacekeeping ''lessons learned'' began to proliferate around the same time. For example, the 21,000-man American-led multinational force that arrived in Haiti in 1994 was widely credited with quickly restoring order in a volatile situation. The moral: go in big. But the military made no serious effort to disarm rival factions, which ensured that violence would flourish as soon as the troops left. And the troops left quickly; with the memory of Somalia still fresh, the Americans turned over peacekeeping operations to the United Nations after six months. The United Nations, at the direction of the Security Council, withdrew almost all its forces over the next two years. In order to prove that Haiti was ready for self-government, the international community agreed to hold legislative elections in the summer of 1995. But amid allegations of blatant poll-rigging by Aristide's Lavalas Party, the opposition boycotted the vote, creating a crisis of legitimacy that was never resolved. The moral: nation-building takes a long time, and elections should not be held too soon. And don't forget to take away the guns.

''Peacekeeping is not an act, it's a process,'' said Jock Covey, who served as deputy to the first civilian head of the Kosovo operation. ''There's not an exit from it. The intervention peaks when you go in there, and you have to always find ways to build it down without ever talking about an exit.'' This building-down process involves a delicate navigation from a phase of neocolonial dominance to the full return of sovereignty, as well as from an essentially military operation to an essentially civilian one. The initial phase must be very dominant. According to James Dobbins, a former diplomat who played a central role in virtually every peacekeeping mission during the Clinton years: ''There's a clear inverse proportion between the size of the force and the number of casualties inflicted. You need about 20 soldiers per 1,000 inhabitants to stabilize an unsettled population.'' That's a breathtakingly large number; in Iraq, it would work out to almost half a million troops. And yet this overwhelming military force must be coupled with a nuanced awareness of local conditions. ''These places tend to be chaotic, dynamic,'' said Frederick Barton, an analyst with the Center for Strategic and International Studies and a veteran of many peacekeeping operations. ''Our own institutions tend to be static. You have to head things in the right direction rather than controlling them.'' One cannot easily find a peacekeeping mission that exemplifies this peculiar mix of characteristics.

If there is such a thing as state-of-the-art peacekeeping, it would be the NATO-United Nations operation in Kosovo, which began with the retreat of Slobodan Milosevic's forces from the province in June 1999. NATO went in big -- 40,000 soldiers in a territory of two million people. And it worked. The Serbian forces stayed on their side of the border, and the 800,000 Albanians who had fled into the freezing mountains, hounded by rampaging Serbian paramilitaries, returned to their towns and villages and began to rebuild their lives. On the civilian side, administrators streamlined the unwieldy structure they had imposed on Bosnia, ensuring that the various agencies reported to a single official, the special representative of the United Nations secretary general. A RAND Corporation study described Kosovo as ''the best-managed of the U.S. post-cold-war ventures in nation-building.''

When I arrived in Kosovo early this year, I found officials at Unmik and KFOR -- the United Nations and NATO operations, respectively -- quite hopeful, and the Kosovars themselves rather sullen, if grudgingly optimistic. The Kosovars were sullen because they were sick and tired of international tutelage. The war had ended without resolving Kosovo's status, and the major powers had done little to press the rump Yugoslav state to grant Kosovo its sovereignty. But only the month before, the provisional government of Kosovo had agreed with the United Nations administrator to work together to achieve a series of standards -- on ethnic reconciliation and integration, privatization and the like -- in exchange for which the international community promised to begin discussions on future status in mid-2005. So there was optimism, grudging and otherwise.

And something like peace prevailed. Soon after I arrived, I paid a visit to Mitrovica, the notorious city that, in the first months of the international occupation, had become a site of pitched battles between the Albanians who lived south of the Ibar River and the Serbs to the north of it. Despite a few incidents, the Serbs and the Albanians seemed to have reached an uneasy truce. The shutters were pulled at the infamous Dolce Vita Cafe, where Serbian ''bridgewatchers'' had once lounged in thuggish glory. The situation had become so placid that the French NATO troops charged with guarding the region had relinquished control of the bridge to the local United Nations-trained police force, which operated out of a trailer parked in the center of the bridge. This was precisely the way the mission was supposed to ''build down.'' I walked onto the bridge to visit the officers of the Kosovo Police Service and asked the squad leader, Branko Nikolic, if anything had happened recently. He thought about it. ''The other day,'' he volunteered, ''a Serbian girl came and said, 'Is it O.K. to go to the bank?''' -- the bank, that is, on the Albanian side. ''We said, 'It's O.K.'''

Yet two months later in this same river, a Serb reportedly would force two Albanians into the water, where they drowned -- and on the Albanian side of the river mobs set upon Serbs, burning down their homes and killing eight people. Both the police and the troops were overwhelmed by the size and the sudden violence of the crowds; had the French not regrouped to block the bridge, the Albanians would have rushed onto the north bank to wreak yet worse mayhem. Ethnic reconciliation was the core of the mission, and ethnic reconciliation had failed. Negotiations on Kosovo's final status have been derailed.

A senior United Nations official I contacted from New York insisted that the violence was the work of ''nefarious elements'' rather than the expression of widespread hatred, and he added that although the event was catastrophic, ''there have been many elements of the situation that have been promising.'' I'm not sure about the first, but the second is plainly true. I was reminded of Branko Nikolic, the police officer. He was a Serb, commanding a detachment in the most volatile spot in Kosovo. The squad also included an Ashkali man -- the Ashkali are a small, Albanian-speaking ethnic minority -- and, almost more remarkably, a woman. They were all members of the Kosovo Police Service, easily the most successful institution built by the international presence in Kosovo. The training of a local police force had been the single most impressive achievement of the international presence in Haiti -- until it was undermined by graft and largely conscripted into the pursuit of President Aristide's private political objectives. In Kosovo, an American ex-cop, Steven Bennett, arrived almost immediately behind the KFOR troops to begin a police training program.

By the time I arrived in Kosovo, Bennett had been sent to Amman to try to do for Iraq what he had done for Kosovo. When I visited the police academy in Vucitrn, north of Pristina, it was being run by Dave Cummings, a retired police sergeant from Akron, Ohio, who had been in Kosovo almost from the beginning of the mission. Cummings explained that the hunger for work of any kind had allowed trainers to be extremely choosy about recruits: only 20 percent of those who took a written exam were even considered, and then an oral exam winnowed out another 90 percent. ''They're told up front the first day what the expectations are,'' Cummings said. ''There are Serbs in every class. There are women in every class. We have zero tolerance for violations of human rights.'' And the Serbian officers I talked to confirmed that the Kosovo Police Service was, in fact, one of the very few truly multiethnic institutions in Kosovo.

The Kosovo Police Service is an emblem of the united Kosovo that the international community seeks. It is, in this sense, a message to Kosovars. Veton Surroi, publisher of Koha Ditore, Kosovo's best-known newspaper, told me that at the Macedonian border his papers had been checked by a Serbian policeman who spoke to him in Albanian -- an unusual act of respect. It was, Surroi said, a significant exception to ''the enclavization of the mind'' that otherwise dominates Kosovar experience. Nevertheless, the Kosovo Police Service has not provoked the Albanian-dominated government to promote reconciliation on its own or to integrate the upper ranks of the civil service. And Serbs still think of the police as an Albanian institution. Several of the Serbian civilians I spoke to told me that they would entrust themselves first to KFOR, second to the United Nations police and last -- that is, not at all -- to the new local police. Even the optimistic Cummings said that it may take a generation before attitudes change. Cummings also granted that the Kosovo Police Service is far from ready to take over all police duties in Kosovo; the international police will continue to field a large contingent.

The quick training of the local police was part of a larger recognition that grew out of previous operations, especially in Bosnia: you cannot achieve the stability needed to promote democratic institutions simply by ending organized hostilities. A state with little or no political legitimacy or tradition of legitimate authority is almost bound to descend into chaos absent an intense focus on the rule of law. This is precisely what happened in the initial months in Kosovo. The power vacuum created by the swift withdrawal of Serbian paramilitaries was immediately filled by the rebels of the Kosovo Liberation Army. Albanian mobs took vengeance on suddenly defenseless Serbs, while KFOR troops, who had come to protect Albanians from Serbs, usually just stood by. And the international police force assembled by the United Nations arrived so slowly that it was of little use in helping maintain order. By the end of 1999, an unholy alliance of rebel leaders, political operators and organized criminals had seized control of many of Kosovo's institutions, including much of its municipal government.

Thanks to its exploits in the months leading up to the NATO war, the Kosovo Liberation Army enjoyed heroic status among ordinary Kosovars even as its leaders came to be accused of war crimes as well as trafficking in prostitutes, guns and drugs. The organization posed a serious threat to Kosovo's future and to KFOR and Unmik's supremacy. ''The K.L.A. was poised to go into opposition,'' a former United Nations official said. ''We could have whacked them, but NATO wouldn't have held together.''

James Dobbins, who during the Clinton administration oversaw the demobilization of the Haitian Army, said that he came up with the idea of dissolving the Kosovo Liberation Army and establishing in its stead the Kosovo Protection Corps, a civilian emergency response unit to which ex-Kosovo Liberation Army fighters would be invited to apply. (He's not the only one to take credit for this inspiration.) It was a tremendously controversial idea, since the United Nations would be rewarding a fair number of extremists and thugs with jobs and allowing the Kosovo Liberation Army to reconstitute itself as an army-in-waiting. But, as Dobbins said, ''it channeled a lot of restless and potentially very destructive energies in a positive direction.'' The Kosovo Protection Corps has been very far from an unmitigated success, but it still seems a vastly better solution to the problem of armed malcontents than what the United States has pursued in Iraq, for example, where the army was simply dissolved.

Conscripting a small fraction of ex-rebels into the Kosovo Protection Corps scarcely brought an end to Kosovo's burgeoning culture of impunity. In those early months, first NATO and then the international police began to arrest malefactors. But who would mete out justice? Could you depend on the judiciary in a place so violently divided against itself? ''At the very beginning, we had a debate about maybe bringing in international judges,'' said a former high-ranking Unmik official. ''We felt that we already had Kosovar judges, and they could do it. We were just wrong.'' Fearing allegations of neocolonialism, Unmik restored Albanian judges and prosecutors to their old jobs. But the judicial system only formalized the ongoing national vendetta, exonerating Albanians and handing down stiff sentences against Serbs.

In 2000, after a rash of ethnic violence, Unmik imported international judges and prosecutors to serve on the most important and sensitive cases. At first, one international judge served on each five-judge panel. Then, when the Kosovar judges simply outvoted the imported ones, the United Nations changed the rules: in cases involving alleged war crimes, or politically powerful defendants who could intimidate local judges, the internationals would control the prosecution and the courtroom. Here was a frank admission that sensitivity to local feelings had to give way before the need to maintain order. Michael Hartmann, one of the most experienced of the international prosecutors, later wrote that peacekeeping missions must be willing to initially brush aside concerns about neocolonialism and ''condition their initial deployment in the judiciary upon a worst-case scenario.''

It is now obvious that all these institutional and cultural changes did not penetrate deeply enough into the fiber of Kosovo to soothe the fear and outrage provoked by the violence before, during and after the war. Is that so very surprising? My Greek KFOR escort pointed out to me that Greeks and Turks have only begun to stop hating one another 80 years after the pillage of Smyrna. Peacekeeping can't change hearts. Nor can it simply transfuse the spirit of democracy and impartiality. Agnieszka Klonowiecka-Milart, the first international judge to arrive in Kosovo, admitted to me that even after years of training and mentoring, her Kosovar colleagues were not ready to handle politically sensitive cases on their own and would not be safe trying to do it. ''You don't internalize these imposed values without seeing the obvious benefit of it,'' she said. ''You must be accorded respect for independence before your own professional class and political class in order to be immune from pressures from society and the media.'' She said she believed that it will be a long time before people like her can go home.

No handle matters, in the end, if one cannot grasp the handle of politics. Kosovo actually has a thoroughly admirable prime minister in Bajram Rexhepi, a former physician, but Rexhepi can exercise only those powers that Unmik chooses to share with him; and so independent political institutions have been infantilized. As Kosovars have waited and waited for their final status to be determined, gratitude has curdled into resentment. When I asked a Rexhepi aide if he could think of something good the United Nations had done, he said grandly, ''In the last five years, we have seen not a single act that shows Kosovo what is democracy, what is rule of law, what is justice.'' This sense of growing embitterment not only poisons Unmik's relationship with Albanians in Kosovo; it also fuels ethnic violence, for Albanians increasingly fear ending up under some form of Serbian control.

East Timor provides an obvious contrast. By the time an Australian-led multinational force waded ashore in September 1999, the Indonesian military and local militias had gutted East Timor even more thoroughly than the Serbs had Kosovo; most of the country's buildings were burned to the ground. When I arrived the following January, the United Nations was running everything: United Nations officials stamped your passport, passed the laws, ran the courts -- one lesson learned from Kosovo -- and arrested common criminals. The head of the mission, Sergio Vieira de Mello, who would later die in Baghdad, occupied the old Portuguese governor's office and ruled as an all-powerful proconsul. There were complaints, even then, that the United Nations was ignoring local capacity and that the highly paid internationals were distorting the local economy. There were angry demonstrations and a few spontaneous brawls. Yet by mid-2000, Vieira de Mello was able to begin transferring authority to an interim Timorese government; a year later, that government staged multiparty elections to a constituent assembly. By 2002, East Timor was an independent country.

How was East Timor able to make such a relatively smooth transition to sovereignty? First, because there were few bad guys; the militias largely melted away and left an ethnically homogeneous society. Second, because even the most impatient Timorese knew that independence was in the offing. And third, because of local political leadership. The rebel leader, Jose Alexandre Gusm-o, preached a doctrine of reconciliation and almost always treated the United Nations and Vieira de Mello respectfully. I watched the two men work together as equals, even though they weren't. One lesson drawn by a recent study of peacekeeping in East Timor is that it is extremely difficult for U.N. missions to be effective in situations where influential sections of the host country's population do not support the U.N.'s presence.

In Haiti, by contrast, international administrators could exercise almost no effective control over an increasingly dictatorial ruler. Take police training as an example. Alix Fils Aime, who served as chief security aide to the prime minister in one of Aristide's governments, told me that the training from French, American and Canadian officials was quite thorough. For several years, the police, long a feared arm of the military (Haiti had no civilian police), enjoyed an almost respectable reputation. Then, in preparation for the 2000 election that would return him to power, Aristide began to turn the police into a private instrument of power. One high-ranking police official told me that, in 2002, he was removed from his post after he refused an order to provide no protection to peaceful anti-Aristide demonstrators, in order to give free rein to the chimeres, his personal squads of goons. And in the final weeks, according to Fils Aime, ''Aristide began giving police uniforms to the chimeres. By the end, there were all these guys from his personal security unit running command posts all over the country.''

This degenerative pattern seemed to obtain in every area. The French established an ecole de magistrature, but Aristide sent his half-educated proteges to the school so they would emerge as judges and refused to give jobs to many of the independent figures who graduated from the program. The Haitian government was decently bankrolled with foreign aid, but Aristide demanded that his ministers fork over millions of dollars to the National Palace, where much of it went to finance the organisations populaires that served as his base of support. The Haitian professionals and analysts I spoke to were quick to blame not only Aristide himself but Haiti's ''winner take all'' political culture. I was constantly struck by the contrast with Kosovo. The Kosovars, having never been masters of their own fate, blamed everything on the United Nations; the Haitians, after 200 years of self-government, blamed everything on themselves.

The RAND study concludes that what distinguishes successful nation-building efforts -- Germany and Japan as well as Bosnia and Kosovo -- from failed ones is not the beneficiary's prior level of development but ''the levels of effort the international community has put into their democratic transformations.'' This is a fine polemic to aim at the Bush administration, but it seems overdrawn. The beneficiary must cooperate in the project of democratic transformation. East Timor was more successful than Haiti or Kosovo principally because it offered a more hospitable climate for nation-building, and in part for that reason, the international community put in more of an effort. At the same time, it's clear that the swift withdrawal of American forces and the failure to disarm rival factions, and then the willingness of international actors to throw up their hands in disgust at Aristide's intractability, all but ensured failure in Haiti.

Is there any hope for Haiti? A big contingent from the United Nations Department of Peacekeeping Operations was making an assessment while I was there. No one knows what will happen yet, but the international community is clearly preparing for Round 2.

Despite the increasingly widespread acceptance of peacekeeping in the ranks of the military, the conservative thinkers who staffed the upper reaches of the Bush foreign-policy apparatus adhered to the ''fight and win wars'' philosophy. They came in, as James Dobbins puts it, with ''an ideological blinder'': nation-building didn't work; nation-building represented the triumph of the nanny state on an international scale. Defense Department officials initially planned to abolish the Pentagon's Office of Peacekeeping and Humanitarian Affairs; in the end, however, they contented themselves with changing its name to the far more manly Office of Stability Operations, and then ignoring it as much as possible. Even after 9/11, President Bush said, ''We're not into nation-building; we're focused on justice.''

The administration made its aversion unmistakable when the Army announced in May 2002 that the Peacekeeping Institute would be closed. The Army explained the closing as a budgetary decision, but very few of the current or retired soldiers and civilian experts who constitute Washington's peacekeeping community were convinced. ''I think the decision was made at the highest level,'' said Beth C. DeGrasse, who conducted the 2002 survey of leading military commanders. ''People in the O.S.D.'' -- the Office of the Secretary of Defense -- ''will tell you that peacekeeping was a dirty word there.''

The 9/11 terrorist attacks made both war and postwar operations unavoidable; but they didn't change the fundamental calculus of leading officials. After the swift and efficient battle to dislodge the Taliban, the administration was praised for inviting the United Nations to adjudicate the process of selecting a new Afghan government led by Hamid Karzai. But just as Afghanistan was supposed to represent a new kind of war, with fewer and more mobile troops, so its postwar governance was intended as an alternative paradigm to East Timor-style nation-building. Matthew Vaccaro, the head of peace operations in the Pentagon in the last months of the Clinton administration and the first years of Bush, characterized the administration position this way: Liberate the people, install good, indigenous leadership and provide those leaders with a national army. ''They take the view that state-building is the responsibility of the indigenous people,'' Vaccaro said. ''It's the process of the indigenous actors building the state that really creates the practices and the norms that are necessary for the state to sustain itself.'' (Douglas Feith, the Defense Department's undersecretary for policy and one of the principal administration architects of this view, declined a request for an interview.)

The obvious shortcomings of nation-building make this bootstrap alternative appealing, but Afghanistan has not exactly vindicated it. The country has a government and a constitution. But Karzai's writ scarcely runs past Kabul, fear of violence has brought much development activity to a halt and warlords control most of the countryside. Between the 5,700-man NATO contingent and the 13,000 American troops, Afghanistan has less than one peacekeeper per thousand citizens. The RAND nation-building study describes the Afghanistan mission as ''a clear regression.'' Only now has the American military seemed to acknowledge the depth of the problem: American troops have begun using classic nation-building tactics in order to win over a hostile population. ''These people have seen military forces for 30 years,'' a captain said in a recent Times article. ''Nothing is going to change until they see schools built and wells dug.''

We now know, thanks in part to reporting in this magazine, that the administration's aversion to nation-building made a fiasco of postwar planning for Iraq. But what I was struck by, in the course of talking to peacekeeping veterans, is just how blatantly administration planners ignored the lessons of earlier operations, as if the liberation of Iraq were so sui generis that the rules simply did not apply. Certainly there was no lack of sound advice. Between September and December 2002, the Center for Strategic and International Studies, a leading research center on postconflict issues, convened a series of panels that included leading officials from across the government. The recommendations contained in the final report sum up the wisdom abstracted from past experience: build a transitional security force to arrive immediately behind combat troops and to focus on classic constabulary duties; prepare to demobilize and then selectively reintegrate Iraqi soldiers; establish a United Nations-led transitional administration; recruit standby teams of judicial officials, as well as of international police officers; and so on.

Past experience also showed that the postwar is harder than the war and takes more planning. And yet Jay Garner, the retired lieutenant general who was chosen to lead postwar reconstruction, had all of eight weeks between the time of his appointment and the beginning of the war to organize the government of a nation of 25 million. Garner's administration ignored virtually all the lessons taught by experience. No international police force was readied, no international judges were contacted and the planned reintegration of the Iraqi Army was abruptly abandoned.

Michael Bailey, a retired military officer with long peacekeeping experience, said that he was sent to Iraq on contract with the Pentagon to implement plans for the Iraqi army. Hours before beginning a pilot program to find jobs for disarmed soldiers, Bailey said he was told to stop. And when L. Paul Bremer III replaced Garner in May, one of his very first acts was to cancel any plans for reintegration. ''You put 400,000 breadwinners in the unemployment line,'' Bailey said. ''They possess unique skills, and they have to feed their family.'' Some unknown number of ex-soldiers have joined the resistance to American occupation.

No matter what approach the United States had taken, Baathists and fundamentalists would still be doing their best to kill American soldiers and the Iraqis who work with them, as well as the private contractors who play an increasingly large role in the occupation. But a peacekeeping effort planned as meticulously and as pragmatically as the military campaign could have persuaded far more Iraqis that the occupying forces had their well-being at heart, and thus that their interests lay in cooperation -- in which case, fewer American soldiers might have died. Gen. Anthony Zinni, who in the months before the war testified before the Senate Foreign Relations Committee about the immense difficulties of postcombat stabilization, grows furious when he talks about the mess in Iraq and said of the senior Pentagon figures he holds responsible for it: ''What is remarkable to me is that they have failed the president, and no one is held accountable.''

It is easy to blame ideologues in the administration for getting Iraq all wrong. But it's important to recall that peacekeeping is extraordinarily hard, that nobody can be said to have gotten it ''right'' before and that even if you accept the need for ambitious and expensive efforts, we lack any institutional means of mounting such efforts. Frederick Barton, the co-director of the Center for Strategic and International Studies study on postwar Iraq and himself a former U.S. Agency for International Development and United Nations official with long experience in postconflict settings, said, ''You can't help but come back to the systemic problem: There really hasn't been a lead person, office, place within the U.S. government'' where postwar responsibility lies. ''Every one of these efforts seems to be a surprise to the system,'' he continued. ''If you'd been in any of these rooms and said, 'This is going to be consuming our national-security apparatus in a year,' you'd be run out of there in a straitjacket.''

Not anymore, of course; Iraq has solved that problem. In July 2003, the Army announced that the Peacekeeping Institute was going to be revived and expanded. Peacekeeping is receiving more scrutiny than ever before. A bipartisan study carried out jointly by the Center for Strategic and International Studies and the Association of the United States Army, an advocacy group, proposed that the president appoint a director of reconstruction in each postconflict setting and also establish a new agency to support these directors and to maintain standby lists of judges, police officers, human rights monitors, constitutional authorities and so on who could deploy rapidly to the field. Late last year, a policy advisory group convened by the Senate Foreign Relations Committee brought administration officials together with experts inside and outside government to find a better way to handle the civilian administration of postwar situations. The resulting bipartisan bill would create a directorate within the National Security Council -- and a new office in the State Department -- to run future postconflict activities. It's not clear whether legislators can find the political will to move this law through Congress.

It is now widely understood that the military, too, must adapt. The ''transformation'' that Donald Rumsfeld has championed toward a lighter, faster, more lethal force worked very effectively in combat situations in Afghanistan and Iraq but operated at cross purposes with the needs of peacekeeping; the smaller force that Rumsfeld insisted upon in Iraq proved unequal to the giant policing task it faced after the opposing army melted away. Military leaders have begun thinking about a very different kind of transformation. Gen. Peter J. Schoomaker, the Army chief of staff, plans to close several air-defense and artillery batteries and shift tens of thousands of soldiers into jobs as military police personnel, engineers and civil-affairs officers, thus reversing the process of shedding noncombat forces that began in the aftermath of Vietnam. Schoomaker has also suggested that rather than spending a summer training with the active military, West Point cadets live with a family abroad and learn a foreign language and culture.

The authors of a National Defense University study make the intriguing point that the technological transformation of warfare requires a new way of thinking about postwar operations, since hostilities now cease long before the adversary has been exhausted or obliterated. ''The very rapid defeat of the enemy military,'' they write, ''means the U.S. must be ready to field the resources needed to secure stability and begin the reconstruction process promptly -- ideally concurrently -- with the end of major combat.'' This is more or less what has happened in Iraq and Afghanistan, where soldiers fix schools one day and hunt terrorists the next, but the authors propose that this become a matter of conscious planning. The core of the report is a proposal that the military create two division-size units, one on active duty and the other on reserve, to specialize in postconflict reconstruction and to be deployed with combat forces. The study, which includes trenchant criticism of the planning and structure of the postconflict operation in Iraq, was ordered up by the Office of Force Transformation in the Secretary of Defense's own office.

Clearly, Iraq's postwar order can no longer be left to hazard. Once it became apparent that Saddam Hussein was not harboring weapons of mass destruction, the Bush administration's rationale for war shifted -- from decapitating a regime that threatened our security to fostering political transformation, first in Iraq, then throughout the Middle East. This is, of course, a call to the most ambitious kind of nation-building. The grandiosity of that goal only points up the intellectual poverty, and the ideological self-absorption, of the postwar planning process in Iraq. The insistence on transferring sovereignty this June 30, while Iraqi institutions are so frail and while deadly violence and chaos still rage, gives the impression that domestic politics counts for a good deal more than the tedious work of nation-building. The means seem wholly inadequate to the ends.

Nevertheless, a new day may be dawning. The swift replacement of Jay Garner last spring by Paul Bremer, a seasoned State Department official, offered an implicit acknowledgment that our role in Iraq could not be merely technocratic and humanitarian, but was profoundly and unavoidably political. Bremer has spent much of his time trying to implant democratic institutions and democratic habits in Iraq's arid soil, fostering the creation of a nationwide network of city councils and hectoring the Iraqi Governing Council to reach a compromise on an interim constitution. And an administration that once planned to assign the United Nations a marginal role in Iraqi reconstruction now recognizes that it needs the United Nations to legitimize, and help accomplish, the transfer of sovereignty.

And once we restore Iraq to the Iraqis -- what then? Will the Iraqis somehow find a serious-minded leader acceptable to all factions -- a Karzai of their own? Will they accept the inroads on their autonomy that come with the nation-building project? And even if they do, will the United States have the patience to stick around while the Iraqis make their mistakes? Simon Chesterman, an authority on peacekeeping at New York University School of Law, recently wrote that ''the importance of domestic politics in the exercise of U.S. power means that it has an exceptionally short attention span'' -- far too short for the slow rebuilding of brutalized countries. Not since the reconstruction of Germany and Japan have we had a stake in nation-building as great as we now have in Iraq. It is hardly clear that our current government -- or the country -- understands that.

James Traub, a contributing writer for the magazine, is author of ''The Devil's Playground: A Century of Pleasure and Profit in Times Square.''

Copyright 2004 The New York Times Company. Reprinted from The New York Times Sunday Magazine of April 11, 2004.