Why Iraq Has No Army

An orderly exit from Iraq depends on the development of a viable Iraqi security force, but the Iraqis aren't even close. The Bush administration doesn't take the problem seriously—and it never has

By James Fallows
The Atlantic, December 2005

When Saddam Hussein fell, the Iraqi people gained freedom. What they didn't get was public order. Looting began immediately, and by the time it abated, signs of an insurgency had appeared. Four months after the invasion the first bomb that killed more than one person went off; two years later, through this past summer, multiple-fatality bombings occurred on average once a day. The targets were not just U.S. troops but Iraqi civilians and, more important, Iraqis who would bring order to the country. The first major attack on Iraq's own policemen occurred in October of 2003, when a car bomb killed ten people at a Baghdad police station. This summer an average of ten Iraqi policemen or soldiers were killed each day. It is true, as U.S. officials often point out, that the violence is confined mainly to four of Iraq's eighteen provinces. But these four provinces contain the nation's capital and just under half its people.



The crucial need to improve security and order in Iraq puts the United States in an impossible position. It can't honorably leave Iraq—as opposed to simply evacuating Saigon-style—so long as its military must provide most of the manpower, weaponry, intelligence systems, and strategies being used against the insurgency. But it can't sensibly stay when the very presence of its troops is a worsening irritant to the Iraqi public and a rallying point for nationalist opponents—to say nothing of the growing pressure in the United States for withdrawal.

Therefore one question now trumps others in America's Iraq policy: whether the United States can foster the development of viable Iraqi security forces, both military and police units, to preserve order in a new Iraqi state.

The Bush administration's policy toward Iraq is based on the premise that this job can be done—and done soon enough to relieve the pressures created by the large-scale U.S. presence in Iraq. These include strains on the U.S. military from its long overseas assignments, mounting political resistance in America because of the cost and casualties of the war, and resentment in Iraq about the open-ended presence of foreign occupation troops. This is why President Bush and other officials say so often, "As Iraqis stand up, we will stand down." American maximalists who want to transform Iraq into a democracy, American minimalists who want chiefly to get U.S. troops out as soon as possible, and everyone in between share an interest in the successful creation of Iraq's own military.

If the United States can foster the development of a sufficiently stable political system in Iraq, and if it can help train, equip, and support military and police forces to defend that system, then American policy has a chance of succeeding. The United States can pull its own troops out of Iraq, knowing that it has left something sustainable behind. But if neither of those goals is realistic—if Iraqi politics remains chaotic and the Iraqi military remains overwhelmed by the insurgent threat—then the American strategy as a whole is doomed.

As Iraqi politicians struggle over terms of a new constitution, Americans need to understand the military half of the long-term U.S. strategy: when and whether Iraqi forces can "stand up."

Early in the occupation American officials acted as if the emergence of an Iraqi force would be a natural process. "In less than six months we have gone from zero Iraqis providing security to their country to close to a hundred thousand Iraqis," Donald Rumsfeld said in October of 2003. "Indeed, the progress has been so swift that ... it will not be long before [Iraqi security forces] will be the largest and outnumber the U.S. forces, and it shouldn't be too long thereafter that they will outnumber all coalition forces combined." By the end of this year the count of Iraqi security forces should indeed surpass the total of American, British, and other coalition troops in Iraq. Police officers, controlled by Iraq's Ministry of the Interior, should number some 145,000. An additional 85,000 members of Iraq's army, plus tiny contingents in its navy and air force, should be ready for duty, under the control of Iraq's Ministry of Defense. Since early this year Iraqi units have fought more and more frequently alongside U.S. troops.

But most assessments from outside the administration have been far more downbeat than Rumsfeld's. Time and again since the training effort began, inspection teams from Congress, the Government Accountability Office (GAO), think tanks, and the military itself have visited Iraq and come to the same conclusion: the readiness of many Iraqi units is low, their loyalty and morale are questionable, regional and ethnic divisions are sharp, their reported numbers overstate their real effectiveness.

The numbers are at best imperfect measures. Early this year the American-led training command shifted its emphasis from simple head counts of Iraqi troops to an assessment of unit readiness based on a four-part classification scheme. Level 1, the highest, was for "fully capable" units—those that could plan, execute, and maintain counterinsurgency operations with no help whatsoever. Last summer Pentagon officials said that three Iraqi units, out of a total of 115 police and army battalions, had reached this level. In September the U.S. military commander in Iraq, Army General George Casey, lowered that estimate to one.

Level 2 was for "capable" units, which can fight against insurgents as long as the United States provides operational assistance (air support, logistics, communications, and so on). Marine General Peter Pace, who is now the chairman of the Joint Chiefs of Staff, said last summer that just under one third of Iraqi army units had reached this level. A few more had by fall. Level 3, for "partially capable" units, included those that could provide extra manpower in efforts planned, led, supplied, and sustained by Americans. The remaining two thirds of Iraqi army units, and half the police, were in this category. Level 4, "incapable" units, were those that were of no help whatsoever in fighting the insurgency. Half of all police units were so classified.

In short, if American troops disappeared tomorrow, Iraq would have essentially no independent security force. Half its policemen would be considered worthless, and the other half would depend on external help for organization, direction, support. Two thirds of the army would be in the same dependent position, and even the better-prepared one third would suffer significant limitations without foreign help.

The moment when Iraqis can lift much of the burden from American troops is not yet in sight. Understanding whether this situation might improve requires understanding what the problems have been so far.

Over the summer and fall I asked a large number of people why Iraq in effect still had no army, and what, realistically, the United States could expect in the future. Most were Americans, but I also spoke with experts from Iraq, Britain, Israel, France, and other countries. Most had served in the military; a large number had recently been posted in Iraq, and a sizable contingent had fought in Vietnam. Almost all those still on active duty insisted that I not use their names. The Army's press office did arrange for me to speak with Lieutenant General Dave Petraeus, who was just completing his year's assignment as commander of the training effort in Iraq, before being replaced by Martin Dempsey, another three-star Army general. But it declined requests for interviews with Petraeus's predecessor, Major General Paul Eaton, or others who had been involved in training programs during the first months of the occupation, or with lower-ranking officers and enlisted men. Many of them wanted to talk or correspond anyway.

What I heard amounted to this: The United States has recently figured out a better approach to training Iraqi troops. Early this year it began putting more money, and more of its best people, on the job. As a result, more Iraqi units are operating effectively, and fewer are collapsing or deserting under pressure. In 2004, during major battles in Fallujah, Mosul, and elsewhere, large percentages of the Iraqi soldiers and policemen supposedly fighting alongside U.S. forces simply fled when the shooting began. But since the Iraqi elections last January "there has not been a single case of Iraqi security forces melting away or going out the back door of the police station," Petraeus told me. Iraqi recruits keep showing up at police and military enlistment stations, even as service in police and military units has become more dangerous.

But as the training and numbers are getting somewhat better, the problems created by the insurgency are getting worse—and getting worse faster than the Iraqi forces are improving. Measured against what it would take to leave Iraqis fully in charge of their own security, the United States and the Iraqi government are losing ground. Absent a dramatic change—in the insurgency, in American efforts, in resolving political differences in Iraq—America's options will grow worse, not better, as time goes on.

Here is a sampling of worried voices:

"The current situation will NEVER allow for an effective ISF [Iraqi Security Force] to be created," a young Marine officer who will not let me use his name wrote in an e-mail after he returned from Iraq this summer. "We simply do not have enough people to train forces. If we shift personnel from security duties to training, we release newly trained ISF into ever-worsening environs."

"A growing number of U.S. military officers in Iraq and those who have returned from the region are voicing concern that the nascent Iraqi army will fall apart if American forces are drawn down in the foreseeable future," Elaine Grossman, of the well-connected newsletter Inside the Pentagon, reported in September.

"U.S. trainers have made a heroic effort and have achieved some success with some units," Ahmed Hashim, of the Naval War College, told me in an e-mail. "But the Iraqi Security Forces are almost like a black hole. You put a lot in and little comes back out."

"I have to tell you that corruption is eating the guts of this counter-insurgency effort," a civilian wrote in an e-mail from Baghdad. Money meant to train new troops was leaking out to terrorists, he said. He empathized with "Iraqi officers here who see and yet are powerless to stop it because of the corrupt ministers and their aides."

"On the current course we will have two options," I was told by a Marine lieutenant colonel who had recently served in Iraq and who prefers to remain anonymous. "We can lose in Iraq and destroy our army, or we can just lose."

The officer went on to say that of course neither option was acceptable, which is why he thought it so urgent to change course. By "destroy our army" he meant that it would take years for the U.S. military to recover from the strain on manpower, equipment, and—most of all—morale that staying in Iraq would put on it. (Retired Army General Barry McCaffrey had this danger in mind when he told Time magazine last winter that "the Army's wheels are going to come off in the next twenty-four months" if it remained in Iraq.) "Losing" in Iraq would mean failing to overcome the violent insurgency. A continuing insurgency would, in the view of the officer I spoke with, sooner or later mean the country's fracture in a bloody civil war. That, in turn, would mean the emergence of a central "Sunni-stan" more actively hostile to the United States than Saddam Hussein's Iraq ever was, which could in the next decade be what the Taliban of Afghanistan was in the 1990s: a haven for al-Qaeda and related terrorists. "In Vietnam we just lost," the officer said. "This would be losing with consequences."

How the Iraq story turns out will not be known for years, but based on what is now knowable, the bleak prospect today is the culmination of a drama's first three acts. The first act involves neglect and delusion. Americans—and Iraqis—will spend years recovering from decisions made or avoided during the days before and after combat began, and through the first year of the occupation. The second act involves a tentative approach to a rapidly worsening challenge during the occupation's second year. We are now in the third act, in which Americans and Iraqis are correcting earlier mistakes but too slowly and too late.

As for the fourth act, it must resolve the tensions created in the previous three.

I. Autumn 2002-Autumn 2003: Taken by Surprise

"It was clear what might happen in a highly militarized society once the regime fell," Anthony Cordesman wrote recently. Cordesman, of the Center for Strategic and International Studies, in Washington, has produced an authoritative series of reports of the new Iraqi military, available at the CSIS Web site. "The U.S. chose to largely ignore these indicators."

In explaining the early failures that plagued the occupation, Cordesman cited factors that have become familiar: an unrealistic expectation of how long Iraqis would welcome a foreign force; a deliberate decision to hold down the size of the invading army; too little preparation for postwar complications; and so on. Before the invasion Saddam Hussein had employed at least half a million soldiers and policemen to keep the lid on Iraq. The United States went in with less than a third that many troops, and because virtually none of them spoke Arabic, they could rarely detect changes in the Iraqi mood or exert influence except by force.

But the explanation of early training problems also leads in some less familiar directions.

One view about why things went so wrong so fast is espoused by Ahmed Chalabi, onetime leader of the Iraqi National Congress, and American supporters of the war such as James Woolsey, a former CIA director, and Richard Perle, a former chairman of the Defense Policy Board. "My view is pretty straightforward," Woolsey told me. "We lost five years, thanks to the State Department and the CIA." The years in question were from 1998, when Congress passed and Bill Clinton signed the Iraq Liberation Act, advocating regime change in Iraq, to 2003, when U.S. troops moved on Baghdad. The act provided $97 million for arms and for training expatriate Iraqi forces. "All we had to do was use some of that money to train mainly Kurdish and Shia units to fight with us, like the Free French in 1944," Woolsey said. The main counterargument is that a Kurdish-Shiite invading army would have made it even harder to deal with Sunnis after Saddam fell.

A different view is strongly held by others among the war's early advocates within the Bush administration. In discussions with former members of the administration I was told they felt truly bad about only one intelligence failure. It did not concern WMD stockpiles in Iraq; the world's other intelligence agencies all made the same mistake, my informants said, and Saddam Hussein would have kept trying to build them anyway. What bothered them was that they did not grasp that he was planning all along to have his army melt away and re-emerge as a guerrilla force once the Americans took over. In this view the war against Saddam's "bitter enders" is still going on, and the new Iraqi forces are developing as fast or as slowly as anyone could expect.

But here is the view generally accepted in the military: the war's planners, military and civilian, took the postwar transition too much for granted; then they made a grievous error in suddenly dismissing all members of the Iraqi army; and then they were too busy with other emergencies and routines to think seriously about the new Iraqi army.

"Should we have had training teams ready to go the day we crossed the border?" asks Lieutenant General Jim Mattis, who commanded the 1st Marine Division during the assault on Iraq (and whom Harrison Ford is scheduled to play in a film about the battle of Fallujah). "Of course! The military has one duty in a situation like this, and that is to provide security for the indigenous people. It's the windbreak behind which everything else can happen." Mattis argued before the war that teams of civic advisers should have been ready to flood in: mayors from North America and Europe to work with Iraqi mayors, police chiefs with police chiefs, all with the goal of preparing the locals to provide public order. "But we didn't do it, and the bottom line was the loss of security."

Many other people suggest many other sins of omission in preparations for the war. But at least one aspect of the transition was apparently given careful thought: how to handle the Iraqi military once it had surrendered or been defeated. Unfortunately, that careful thought was ignored or overruled.

After years of misuse under Saddam Hussein, the Iraqi military had severe problems, including bad morale, corrupt leadership, shoddy equipment, and a reputation for brutality. But the regular army numbered some 400,000 members, and if any of them could be put to use, there would be less work for Americans.

By late 2002, after Congress voted to authorize war if necessary, Jay Garner, a retired three-star Army general, was thinking about how he might use some of these soldiers if the war took place and he became the first viceroy of Iraq. Garner, who had supervised Kurdish areas after the Gulf War, argued for incorporating much of the military rank-and-file into America's occupation force. Stripping off the top leadership would be more complicated than with, say, the Japanese or German army after World War II, because Iraq's army had more than 10,000 generals. (The U.S. Army, with about the same number of troops, has around 300 generals.) But, I was told by a former senior official who was closely involved in making the plans, "the idea was that on balance it was much better to keep them in place and try to put them to work, public works—style, on reconstruction, than not to." He continued, "The advantages of using them were: They had organization. They had equipment, especially organic transport [jeeps, trucks], which let them get themselves from place to place. They had a structure. But it was a narrow call, because of all the disadvantages." Garner intended to put this plan into action when he arrived in Baghdad, in April of 2003. He told me recently that there were few signs of the previous army when he first arrived. "But we sent out feelers, and by the first week in May we were getting a lot of responses back. We had a couple of Iraqi officers come to me and say, 'We could bring this division back, that division.' We began to have dialogues and negotiations."

Then, on May 23, came a decision that is likely to be debated for years: Coalition Provisional Authority Order Number 2, to disband the Iraqi military and simply send its members home without pay.

"I always begin with the proposition that this argument is entirely irrelevant," Walter Slocombe, the man usually given credit or blame for initiating the decision, told me in the summer. During the Clinton administration Slocombe was undersecretary for policy at the Pentagon, the job later made famous by Douglas Feith. A month after the fall of Baghdad, Slocombe went to the Green Zone as a security adviser to Paul Bremer, who had just replaced Garner as the ranking American civilian.

On arrival Slocombe advocated that the Coalition Provisional Authority, or CPA, should face the reality that the previous Iraqi army had disappeared. "There was no intact Iraqi force to 'disband,'" Slocombe said. "There was no practical way to reconstitute an Iraqi force based on the old army any more rapidly than has happened. The facilities were just destroyed, and the conscripts were gone and not coming back." The Bush administration officials who had previously instructed Garner to reconstitute the military endorsed Slocombe's view: the negative aspects of consorting with a corrupt, brutal force were still there, and the positives seemed to be gone. "All the advantages they had ran away with the soldiers," the senior official involved in the plans said. "The organization, the discipline, the organic transport. The facts had changed."

The arguments about the decision are bitter, and they turn on two points: whether the Iraqi army had in fact irreversibly "disbanded itself," as Slocombe contends, and whether the American authorities could have found some way to avoid turning the hundreds of thousands of discharged soldiers into an armed and resentful opposition group. "I don't buy the argument that there was no army to cashier," says Barak Salmoni, of the Marine Corps Training and Education Command. "It may have not been showing up to work, but I can assure you that they would have if there had been dollars on the table. And even if the Iraqi army did disband, we didn't have to alienate them"—mainly by stopping their pay. Several weeks later the Americans announced that they would resume some army stipends, but by then the damage had been done.

Garner was taken by surprise by the decision, and has made it clear that he considers it a mistake. I asked him about the frequently voiced argument that there was no place to house the army because the barracks had been wrecked. "We could have put people in hangars," he said. "That is where our troops were."

The most damaging criticism of the way the decision was made comes from Paul Hughes, who was then an Army colonel on Garner's staff. "Neither Jay Garner nor I had been asked about the wisdom of this decree," Hughes recalls. He was the only person from Garner's administration then talking with Iraqi military representatives about the terms of their re-engagement. On the eve of the order to disband, he says, more than 100,000 Iraqi soldiers had submitted forms to receive a one-time $20 emergency payment, from funds seized from Saddam's personal accounts, which they would show up to collect.

"My effort was not intended to re-activate the Iraqi military," Hughes says. "Whenever the Iraqi officers asked if they could re-form their units, I was quite direct with them that if they did, they would be attacked and destroyed. What we wanted to do was arrange the process by which these hundred thousand soldiers would register with [the occupation authorities], tell us what they knew, draw their pay, and then report to selected sites. CPA Order Number Two simply stopped any effort to move forward, as if the Iraqi military had ceased to exist. [Walt Slocombe's] statement about the twenty dollars still sticks in my brain: 'We don't pay armies we defeated.' My Iraqi friends tell me that this decision was what really spurred the nationalists to join the infant insurgency. We had advertised ourselves as liberators and turned on these people without so much as a second thought."

The argument will go on. But about what happened next there is little dispute. Having eliminated the main existing security force, and having arrived with fewer troops than past experience in the Balkans, Germany, and Japan would suggest for so large a territory, American officials essentially wasted the next six months. By the time they thought seriously about reconstituting Iraq's military and police forces, the insurgency was under way and the challenge of pacifying Iraq had magnified.

There is no single comprehensive explanation for what went wrong. After the tension leading up to the war and the brilliant, brief victory, political and even military leaders seemed to lose interest, or at least intensity. "Once Baghdad was taken, Tommy Franks checked out," Victor O'Reilly, who has written extensively about the U.S. military, told me. "He seemed to be thinking mainly about his book." Several people I spoke with volunteered this view of Franks, who was the centcom commander during the war. (Franks did not respond to interview requests, including those sent through his commercially minded Web site, TommyFranks.com.) In retrospect the looting was the most significant act of the first six months after the war. It degraded daily life, especially in Baghdad, and it made the task of restoring order all the more difficult for the U.S. or Iraqi forces that would eventually undertake it. But at the time neither political nor military leaders treated it as urgent. Weeks went by before U.S. troops effectively intervened.

In June of 2003, as the looting was dying down but the first signs of insurgent violence were appearing, the CSIS sent a team of experts who had worked in past occupations. They were alarmed by what they saw. "There is a general sense of steady deterioration in the security situation, in Baghdad, Mosul, and elsewhere," they reported. "Virtually every Iraqi and most CPA and coalition military officials as well as most contractors we spoke to cited the lack of public safety as their number one concern." At that time, the team pointed out, some 5,000 U.S. troops were tied down guarding buildings in Baghdad, with two and a half battalions, representing well over a thousand troops, guarding the American headquarters alone.

Anthony Cordesman, of the CSIS, says there was never a conscious decision to delay or ignore training, but at any given moment in the occupation's first months some other goal always seemed more urgent or more interesting. Through the first six months of the occupation capturing Saddam Hussein seemed to be the most important step toward ending the resistance. His two sons were killed in July; he himself was captured in December; and the insurgency only grew. Along the way the manhunt relied on detention, interrogation, and break-down-doors-at-night techniques that hastened resentment of the U.S. presence. "The search for Saddam colored everything," Victor O'Reilly told me. "It is my belief that the insurgency was substantially created by the tactics used by the occupying force, who were initially the saviors, in their search for Saddam. Ambitious generals, who should have known better, created a very aggressive do-what-is-necessary culture. Frustrated troops, with no familiarity with the language or culture, naturally make mistakes. And in a tribal society if you shoot one person it spreads right through the system."

The hunt for WMD troves, conducted in the same way as the search for Saddam and by troops with the same inability to understand what was being said around them, had a similar embittering effect. The junior-level soldiers and Marines I interviewed consistently emphasized how debilitating the language barrier was. Having too few interpreters, they were left to communicate their instructions with gestures and sign language. The result was that American troops were blind and deaf to much of what was going on around them, and the Iraqis were often terrified.

General Mattis had stressed to his troops the importance of not frightening civilians, so as not to turn those civilians into enemies. He, too, emphasizes the distractions in the first year that diminished the attention paid to building an Iraqi security force. "There was always something," he told me. "Instead of focusing on security, we were trying to get oil pipelines patched, electrical grids back into position, figure out who the engineers were we could trust, since some of them hated us so much they would do sabotage work. It was going to take a while."

When Americans did think about a new Iraqi army, they often began with fears that it might become too strong too fast. "Everybody assumed that within Iraq it would be peaceful," says T. X. Hammes, the author of The Sling and the Stone, who was then in Iraq as a Marine Corps colonel. "So the biggest concern was reassuring all of Iraq's neighbors that Iraq would not be a threat. One of the ways you do that is by building a motor infantry force with no logistics"—that is, an army that can't sustain any large-scale offensive operation. Such an army might assuage concerns in Syria and Iran, but it would do little to provide internal security, and would not be prepared for domestic counterinsurgency work. (This tension has not been resolved: to this day the Iraqi government complains that the United States will not help it get adequate tanks, armored vehicles, and artillery.) Corrupt use of U.S. aid and domestic Iraqi resources was a constant and destructive factor. Last August the Knight Ridder newspapers revealed that Iraq's Board of Supreme Audit had surveyed arms contracts worth $1.3 billion and concluded that about $500 million had simply disappeared in payoffs, kickbacks, and fraud.

Training the police would be as big a challenge as training the army. "There was no image of a non-corrupt police force anywhere in the country," Mattis says. And to make matters more difficult, the effort began just as the police were coming under attack from insurgents' bombs and grenades.

Throughout the occupation, but most of all in these early months, training suffered from a "B Team" problem. Before the fighting there was a huge glamour gap in the Pentagon between people working on so-called Phase III—the "kinetic" stage, the currently fashionable term for what used to be called "combat"—and those consigned to thinking about Phase IV, postwar reconstruction. The gap persisted after Baghdad fell. Nearly every military official I spoke with said that formal and informal incentives within the military made training Iraqi forces seem like second-tier work.

There were exceptions. The Green Berets and other elite units of the Special Forces have long prided themselves on being able to turn ragtag foreign armies into effective fighting units. But there weren't enough Special Forces units to go around, and the mainstream Army and Marine Corps were far less enthusiastic about training assignments. Especially at the start, training missions were filled mostly by people who couldn't get combat postings, and by members of the Reserves and the National Guard.

Walter Slocombe told me that there could have been a larger structural attempt to deal with the B-Team issue. "If we knew then what we know now," he said—that is, if people in charge had understood that public order would be the biggest postwar problem, and that Iraqis would soon resent the presence of foreigners trying to impose that order—"we would have done things differently. It would have made sense to have had an American military unit assigned this way from the beginning. They would be told, 'You guys aren't going to fight this war. You're not going to get Medals of Honor. But you will get due recognition. Your job is to run the occupation and train the Iraqis.' And we'd configure for that mission."

But of course that didn't happen. "I couldn't believe that we weren't ready for the occupation," Terence Daly, a retired Army colonel who learned the tactics of counterinsurgency in Vietnam, told me. "I was horrified when I saw the looting and the American inaction afterward. If I were an Iraqi, it would have shown me these people are not serious."

II. Autumn 2003-Autumn 2004: Overwhelmed

By late 2003 the United States had lost time and had changed identity, from liberator to occupier. But in its public pronouncements and its internal guidance the administration resisted admitting, even to itself, that it now faced a genuine insurgency—one that might grow in strength—rather than merely facing the dregs of the old regime, whose power would naturally wane as its leaders were caught and killed. On June 16 Army General John Abizaid, newly installed as centcom commander, was the first senior American official to say that in fact the United States now faced a "classical guerrilla-type campaign." Two days later, in congressional testimony, Paul Wolfowitz, the deputy secretary of defense, seemed to accept the definition, saying, "There is a guerrilla war there, but ... we can win it." On June 30 Rumsfeld corrected both of them, saying that the evidence from Iraq "doesn't make it anything like a guerrilla war or an organized resistance." Two days after that President Bush said at a White House ceremony that some people felt that circumstances in Iraq were "such that they can attack us there. My answer is, Bring them on." Meanwhile, the insurgency in Iraq grew worse and worse.

Improving the training of Iraqis suddenly moved up the list of concerns. Karl Eikenberry, an Army general who had trained Afghan forces after the fall of the Taliban, was sent to Iraq to see what was wrong. Pentagon briefers referred more and more frequently to the effort to create a new Iraqi military. By early 2004 the administration had decided to spend more money on troop training, and to make it more explicitly part of the U.S. mission in Iraq. It was then that a grim reality hit: how hard this process would actually be.

"Training is not just learning how to fire a gun," I was told by a congressional staff member who has traveled frequently to Afghanistan since 9/11. "That's a part of it, but only a small part." Indeed, basic familiarity with guns is an area in which Iraqis outdo Americans. Walter Slocombe says that the CPA tried to enforce a gun-control law—only one AK-47 per household—in the face of a widespread Iraqi belief that many families needed two, one for the house and one for traveling.

Everyone I interviewed about military training stressed that it was only trivially about teaching specific skills. The real goal was to transform a civilian into a soldier. The process runs from the individual level, to the small groups that must trust one another with their lives, to the combined units that must work in coordination rather than confusedly firing at one another, to the concept of what makes an army or a police force different from a gang of thugs.

"The simple part is individual training," Jay Garner says. "The difficult part is collective training. Even if you do a good job of all that, the really difficult thing is all the complex processes it takes to run an army. You have to equip it. You have to equip all units at one time. You have to pay them on time. They need three meals a day and a place to sleep. Fuel. Ammunition. These sound simple, but they're incredibly difficult. And if you don't have them, that's what makes armies not work."

In countless ways the trainers on site faced an enormous challenge. The legacy of Saddam Hussein was a big problem. It had encouraged a military culture in which officers were privileged parasites, enlisted soldiers were cannon fodder, and noncommissioned officers—the sergeants who make the U.S. military function—were barely known. "We are trying to create a professional NCO corps," Army Major Bob Bateman told me. "Such a thing has never existed anywhere in the region. Not in regular units, not in police forces, not in the military."

The ethnic and tribal fissures in Iraq were another big problem. Half a dozen times in my interviews I heard variants on this Arab saying: "Me and my brother against my cousin; me and my cousin against my village; me and my village against a stranger." "The thing that holds a military unit together is trust," T. X. Hammes says. "That's a society not based on trust." A young Marine officer wrote in an e-mail, "Due to the fact that Saddam murdered, tortured, raped, etc. at will, there is a limited pool of 18-35-year-old males for service that are physically or mentally qualified for service. Those that are fit for service, for the most part, have a DEEP hatred for those not of the same ethnic or religious affiliation."

The Iraqi culture of guns was, oddly, not an advantage but another problem. It had created gangs, not organized troops. "It's easy to be a gunman and hard to be a soldier," one expert told me. "If you're a gunman, it doesn't matter if your gun shoots straight. You can shoot it in the general direction of people, and they'll run." Many American trainers refer to an Iraqi habit of "Inshallah firing," also called "death blossom" marksmanship. "That is when they pick it up and start shooting," an officer now on duty in Baghdad told me by phone. "Death just blossoms around them."

The constant attacks from insurgents were a huge problem, and not just in the obvious ways. The U.S. military tries hard to separate training from combat. Combat is the acid test, but over time it can, strangely, erode proficiency. Under combat pressure troops cut corners and do whatever it takes to survive. That is why when units return from combat, the Pentagon officially classifies them as "unready" until they have rested and been retrained in standard procedures. In principle the training of Iraqi soldiers and policemen should take place away from the battlefield, but they are under attack from the moment they sign up. The pressure is increased because of public hostility to the foreign occupiers. "I know an Iraqi brigade commander who has to take off his uniform when he goes home, so nobody knows what he's done," Barak Salmoni told me. The Iraqi commander said to him, "It really tugs at our minds that we have to worry about our families' dying in the insurgency when we're fighting the insurgency somewhere else." The GAO found that in these circumstances security units from "troubled townships" often deserted en masse.

The United States, too, brought its own range of problems. One was legislative. Because U.S. forces had helped prop up foreign dictators, Congress in the 1970s prohibited most forms of American aid to police forces—as distinct from armies—in other countries. For the purposes of containing the insurgency in Iraq the distinction was meaningless. But administration officials used up time and energy through 2004 figuring out an answer to this technical-sounding yet important problem.

Language remained a profound and constant problem. One of the surprises in asking about training Iraqi troops was how often it led to comparisons with Vietnam. Probably because everything about the Vietnam War took longer to develop, "Vietnamization" was a more thought-through, developed strategy than "Iraqization" has had a chance to be. A notable difference is that Americans chosen for training assignments in Vietnam were often given four to six months of language instruction. That was too little to produce any real competence, but enough to provide useful rudiments that most Americans in Iraq don't have.

The career patterns of the U.S. military were a problem. For family reasons, and to keep moving up in rank, American soldiers rotate out of Iraq at the end of a year. They may be sent back to Iraq, but probably on a different assignment in a different part of the country. The adviser who has been building contacts in a village or with a police unit is gone, and a fresh, non-Arabic-speaking face shows up. "All the relationships an adviser has established, all the knowledge he has built up, goes right with him," Terence Daly, the counterinsurgency specialist from the Vietnam War, says. Every manual on counterinsurgency emphasizes the need for long-term personal relations. "We should put out a call for however many officers and NCOs we need," Daly says, "and give them six months of basic Arabic. In the course of this training we could find the ones suited to serve there for five years. Instead we treat them like widgets."

All indications from the home front were that training Iraqis had become a boring issue. Opponents of the war rarely talked about it. Supporters reeled off encouraging but hollow statistics as part of a checklist of successes the press failed to report. President Bush placed no emphasis on it in his speeches. Donald Rumsfeld, according to those around him, was bored by Iraq in general and this tedious process in particular, neither of which could match the challenge of transforming America's military establishment.

The lack of urgency showed up in such mundane ways as equipment shortages. In the spring of 2004 investigators from the GAO found that the Iraqi police had only 41 percent of the patrol vehicles they needed, 21 percent of the hand-held radios, and nine percent of the protective vests. The Iraqi Civil Defense Corps, a branch of the military, had received no protective vests at all. According to the GAO report, "A multinational force assessment noted that Iraqis within the Iraqi Civil Defense Corps felt the multinational force never took them seriously, as exhibited by what they perceived as the broken promises and the lack of trust of the multinational force."

Although most people I spoke with said they had warm relations with many of their Iraqi counterparts, the lack of trust applied on the U.S. side as well. American trainers wondered how many of the skills they were imparting would eventually be used against them, by infiltrators or by soldiers who later changed sides. Iraq's Ministry of Defense has complained that the United States is supplying simpler equipment, such as AK-47s rather than the more powerful M4 rifles, and pickup trucks rather than tanks. Such materiel may, as U.S. officials stress, be far better suited to Iraq's current needs. It would also be less troublesome if Iraq and the United States came to be no longer on friendly terms.

And the biggest problem of all was the kind of war this new Iraqi army had to fight.

"Promoting disorder is a legitimate objective for the insurgent," a classic book about insurgency says: "It helps to disrupt the economy, hence to produce discontent; it serves to undermine the strength and the authority of the counterinsurgent [that is, government forces]. Moreover, disorder ... is cheap to create and very costly to prevent. The insurgent blows up a bridge, so every bridge has to be guarded; he throws a grenade in a movie theater, so every person entering a public place has to be searched."

The military and political fronts are so closely connected, the book concludes, that progress on one is impossible without progress on the other: "Every military move has to be weighed with regard to its political effects, and vice versa."

This is not a book about Iraq. The book is Counterinsurgency Warfare: Theory and Practice, which was published nearly forty years ago by a French soldier and military analyst, David Galula, and is based on his country's experiences in Algeria and Vietnam.

Counterinsurgency scholarship has boomed among military intellectuals in the 2000s, as it did in the 1960s, and for the same reason: insurgents are the enemy we have to fight. "I've been reading a lot of T. E. Lawrence, especially through the tough times," Dave Petraeus said when I asked where he had looked for guidance during his year of supervising training efforts. An influential book on counterinsurgency by John Nagl, an Army lieutenant colonel who commanded a tank unit in Iraq, is called Learning to Eat Soup With a Knife. That was Lawrence's metaphor for the skills needed to fight Arab insurgents.

"No modern army using conventional tactics has ever defeated an insurgency," Terence Daly told me. Conventional tactics boil down to killing the enemy. At this the U.S. military, with unmatchable firepower and precision, excels. "Classic counterinsurgency, however, is not primarily about killing insurgents; it is about controlling the population and creating a secure environment in which to gain popular support," Daly says.

From the vast and growing literature of counterinsurgency come two central points. One, of course, is the intertwining of political and military objectives: in the long run this makes local forces like the Iraqi army more potent than any foreigners; they know the language, they pick up subtle signals, they have a long-term stake. The other is that defeating an insurgency is the very hardest kind of warfare. The United States cannot win this battle in Iraq. It hopes the Iraqis can.

Through the second year of occupation most of the indications were dark. An internal Pentagon report found, "The first Iraqi Army infantry battalions finished basic training in early 2004 and were immediately required in combat without complete equipment ... Absent-without-leave rates among regular army units were in double digits and remained so for the rest of the year.

I asked Robert Pape about the AWOL and desertion problems that had plagued Iraqi forces in Mosul, Fallujah, and elsewhere. Pape, of the University of Chicago, is the author of Dying to Win, a recent book about suicide terrorism. "Really, it was not surprising that this would happen," he said. "You were taking a force that had barely been stood up and asking it to do one of the most demanding missions possible: an offensive mission against a city. Even with a highly loyal force you were basically asking them to sacrifice themselves. Search and destroy would be one of the last things you would want them to do."

A GAO report showed the extent of the collapse. Fifty percent of the Iraqi Civil Defense Corps in the areas around Baghdad deserted in the first half of April. So did 30 percent of those in the northeastern area around Tikrit and the southeast near al-Kut. And so did 80 percent of the forces around Fallujah.

This was how things still stood on the eve of America's presidential election and the beginning of a new approach in Baghdad.

III. Autumn 2004-Autumn 2005: Progress but no Urgency

At the end of June 2004 Ambassador Bremer went home. His Coalition Provisional Authority ceased to exist, and an interim Iraqi government, under a prime minister selected by the Americans, began planning for the first nationwide elections, which were held in January of this year. The first U.S. ambassador to postwar Iraq, John D. Negroponte, was sworn in as Bremer left. And a new American Army general arrived to supervise the training of Iraqis: Dave Petraeus, who had just received his third star.

The appointment was noticed throughout the military. Petraeus, who holds a Ph.D. from Princeton, had led the 101st Airborne during its drive on Mosul in 2003 and is one of the military's golden boys. What I heard about him from other soldiers reminded me of what reporters used to hear about Richard Holbrooke from other diplomats: many people marveled at his ambition; few doubted his skills. Petraeus's new assignment suggested that training Iraqis had become a sexier and more important job. By all accounts Petraeus and Negroponte did a lot to make up for lost time in the training program.

Under Petraeus the training command abandoned an often ridiculed way of measuring progress. At first Americans had counted all Iraqis who were simply "on duty"—a total that swelled to more than 200,000 by March of 2004. Petraeus introduced an assessment of "unit readiness," as noted above. Training had been underfunded in mid-2004, but more money and equipment started to arrive.

The training strategy also changed. More emphasis was put on embedding U.S. advisers with Iraqi units. Teams of Iraqi foot soldiers were matched with U.S. units that could provide the air cover and other advanced services they needed. To save money and reduce the chance of a coup, Saddam Hussein's soldiers had only rarely, or never, fired live ammunition during training. According to an unpublished study from the U.S. Army War College, even the elite units of the Baghdad Republican Guard were allowed to fire only about ten rounds of ammunition per soldier in the year before the war, versus about 2,500 rounds for the typical U.S. infantry soldier. To the amazement of Iraqi army veterans, Petraeus introduced live-fire exercises for new Iraqi recruits.

At the end of last year, as the Iraqi national elections drew near, Negroponte used his discretion to shift $2 billion from other reconstruction projects to the training effort. "That will be seen as quite a courageous move, and one that paid big dividends," Petraeus told me. "It enabled the purchase of a lot of additional equipment, extra training, and more rebuilding of infrastructure, which helped us get more Iraqi forces out in the field by the January 30 elections."

The successful staging of the elections marked a turning point—at least for the training effort. Political optimism faded with the subsequent deadlocks over the constitution, but "we never lost momentum on the security front," Petraeus told me. During the elections more than 130,000 Iraqi troops guarded more than 5,700 polling stations; there were some attacks, but the elections went forward. "We have transitioned six or seven bases to Iraqi control," he continued, listing a variety of other duties Iraqi forces had assumed. "The enemy recognizes that if Iraqi security forces ever really get traction, they are in trouble. So all of this is done in the most challenging environment imaginable."

Had the training units avoided the "B Team" taint? By e-mail I asked an officer on the training staff about the "loser" image traditionally attached to such jobs within the military. He wrote back that although training slots had long been seen as "career killers," the importance of the effort in Iraq was changing all that. From others not involved in training I heard a more guarded view: If an Iraqi army emerges, the image of training will improve; if it doesn't, the careers of Petraeus and his successor Dempsey will suffer.

Time is the problem. As prospects have brightened inside the training program, they have darkened across the country. From generals to privates, every soldier I spoke with stressed that the military campaign would ultimately fail without political progress. If an army has no stable government to defend, even the best-trained troops will devolve into regional militias and warlord gangs. "I always call myself a qualified optimist, but the qualification is Iraqi leaders muddling through," one senior officer told me. "Certain activities are beyond Americans' control."

Ethnic tensions divide Iraq, and they divide the new army. "Thinking that we could go in and produce a unified Iraqi army is like thinking you could go into the South after the Civil War and create an army of blacks and whites fighting side by side," Robert Pape, of the University of Chicago, told me. "You can pay people to go through basic training and take moderate risks. But unless they're really loyal to a government, as the risks go up, they will run." Almost every study of the new Iraqi military raises doubts about how loyally "Iraqi" it is, as opposed to Kurdish, Shiite, or Sunni. The most impressive successes by "Iraqi" forces have in fact been by units that were really Kurdish peshmurga or Shiite militias.

"There is still no sense of urgency," T. X. Hammes says. In August, he pointed out, the administration announced with pride that it had bought 200 new armored vehicles for use in Iraq. "Two-plus years into the war, and we're proud! Can you imagine if in March of 1944 we had proudly announced two hundred new vehicles?" By 1944 American factories had been retooled to produce 100,000 warplanes. "From the president on down there is no urgency at all."

Since last June, President Bush has often repeated his "As Iraqi forces stand up ..." formula, but he rarely says anything more specific about American exit plans. When he welcomed Iraq's president, Jalal Talabani, to the White House in September, his total comment on the training issue in a substantial welcoming speech was "Our objective is to defeat the enemies of a free Iraq, and we're working to prepare more Iraqi forces to join the fight." This was followed by the stand up/stand down slogan. Vice President Cheney sounds similarly dutiful. ("Our mission in Iraq is clear," he says in his typical speech. "On the military side we are hunting down the terrorists and training Iraqi security forces so they can take over responsibility for defending their own country." He usually follows with the slogan but with no further details or thoughts.)

Donald Rumsfeld has the same distant tone. Condoleezza Rice and Paul Wolfowitz have moved on to different things. At various times since 9/11 members of the administration have acted as if catching Osama bin Laden, or changing Social Security, or saving Terri Schiavo, or coping with Hurricane Katrina, mattered more than any possible other cause. Creating an Iraqi military actually matters more than almost anything else. But the people who were intent on the war have lost interest in the only way out.

A Marine lieutenant colonel said, "You tell me who in the White House devotes full time to winning this war." The answer seems to be Meghan O'Sullivan, a former Brookings scholar who is now the president's special assistant for Iraq. As best I can tell from Nexis, other online news sources, and the White House Web site, since taking the job, late last year, she has made no public speeches or statements about the war.

IV. How to Leave With Honor

Listening to the Americans who have tried their best to create an Iraqi military can be heartening. They send e-mails or call late at night Iraq time to report successes. A Web magazine published by the training command, called The Advisor, carries photos of American mentors working side by side with their Iraqi students, and articles about new training techniques. The Americans can sound inspired when they talk about an Iraqi soldier or policeman who has shown bravery and devotion in the truest way—by running toward battle rather than away from it, or rushing to surround a suicide bomber and reduce the number of civilians who will be killed.

But listening to these soldiers and advisers is also deeply discouraging—in part because so much of what they report is discouraging in itself, but even more because the conversations head to a predictable dead end. Sooner or later the question is What do we do now? or What is the way out? And the answer is that there is no good answer.

Let me suggest a standard for judging endgame strategies in Iraq, given the commitment the United States has already made. It begins with the recognition that even if it were possible to rebuild and fully democratize Iraq, as a matter of political reality the United States will not stay to see it through. (In Japan, Germany, and South Korea we did see it through. But while there were postwar difficulties in all those countries, none had an insurgency aimed at Americans.) But perhaps we could stay long enough to meet a more modest standard.

What is needed for an honorable departure is, at a minimum, a country that will not go to war with itself, and citizens who will not turn to large-scale murder. This requires Iraqi security forces that are working on a couple of levels: a national army strong enough to deter militias from any region and loyal enough to the new Iraq to resist becoming the tool of any faction; policemen who are sufficiently competent, brave, and honest to keep civilians safe. If the United States leaves Iraq knowing that non-American forces are sufficient to keep order, it can leave with a clear conscience—no matter what might happen a year or two later.

In the end the United States may not be able to leave honorably. The pressure to get out could become too great. But if we were serious about reconstituting an Iraqi military as quickly as possible, what would we do? Based on these interviews, I have come to this sobering conclusion: the United States can best train Iraqis, and therefore best help itself leave Iraq, only by making certain very long-term commitments to stay.

Some of the changes that soldiers and analysts recommend involve greater urgency of effort, reflecting the greater importance of making the training succeed. Despite brave words from the Americans on the training detail, the larger military culture has not changed to validate what they do. "I would make advising an Iraqi battalion more career-enhancing than commanding an American battalion," one retired Marine officer told me. "If we were serious, we'd be gutting every military headquarters in the world, instead of just telling units coming into the country they have to give up twenty percent of their officers as trainers."

The U.S. military does everything in Iraq worse and slower than it could if it solved its language problems. It is unbelievable that American fighting ranks have so little help. Soon after Pearl Harbor the U.S. military launched major Japanese-language training institutes at universities and was screening draftees to find the most promising students. America has made no comparable effort to teach Arabic. Nearly three years after the invasion of Iraq the typical company of 150 or so U.S. soldiers gets by with one or two Arabic-speakers. T. X. Hammes says that U.S. forces and trainers in Iraq should have about 22,000 interpreters, but they have nowhere near that many. Some 600,000 Americans can speak Arabic. Hammes has proposed offering huge cash bonuses to attract the needed numbers to Iraq.

In many other ways the flow of dollars and effort shows that the military does not yet take Iraq—let alone the training effort there—seriously. The Pentagon's main weapons-building programs are the same now that they were five years ago, before the United States had suffered one attack and begun two wars. From the Pentagon's policy statements, and even more from its budgetary choices, one would never guess that insurgency was our military's main challenge, and that its main strategic hope lay in the inglorious work of training foreign troops. Planners at the White House and the Pentagon barely imagined before the war that large numbers of U.S. troops would be in Iraq three years later. So most initiatives for Iraq have been stopgap—not part of a systematic effort to build the right equipment, the right skills, the right strategies, for a long-term campaign.

Some other recommended changes involve more-explicit long-range commitments. When officers talk about the risk of "using up" or "burning out" the military, they mean that too many arduous postings, renewed too frequently, will drive career soldiers out of the military. The recruitment problems of the National Guard are well known. Less familiar to the public but of great concern in the military is the "third tour" phenomenon: A young officer will go for his first year-long tour in Iraq or Afghanistan, and then his second. Facing the prospect of his third, he may bail out while he still has time to start another, less stressful career.

For the military's sake soldiers need to go to Iraq less often, and for shorter periods. But success in training Iraqis will require some Americans to stay there much longer. Every book or article about counterinsurgency stresses that it is an intimate, subjective, human business. Establishing trust across different cultures takes time. After 9/11 everyone huffed about the shocking loss of "human intelligence" at America's spy agencies. But modern American culture—technological, fluid, transient—discourages the creation of the slow-growing, subtle bonds necessary for both good spy work and good military liaison. The British had their India and East Asia hands, who were effective because they spent years in the field cultivating contacts. The American military has done something similar with its Green Berets. For the training effort to have a chance, many, many more regular soldiers will need to commit to long service in Iraq.

The United States will have to agree to stay in Iraq in another significant way. When U.S. policy changed from counting every Iraqi in uniform to judging how many whole units were ready to function, a triage decision was made. The Iraqis would not be trained anytime soon for the whole range of military functions; they would start with the most basic combat and security duties. The idea, as a former high-ranking administration official put it, was "We're building a spearhead, not the whole spear."

The rest of the spear consists of the specialized, often technically advanced functions that multiply the combat units' strength. These are as simple as logistics—getting food, fuel, ammunition, spare parts, where they are needed—and as complex as battlefield surgical units, satellite-based spy services, and air support from helicopters and fighter planes.

The United States is not helping Iraq develop many of these other functions. Sharp as the Iraqi spearhead may become, on its own it will be relatively weak. The Iraqis know their own territory and culture, and they will be fighting an insurgency, not a heavily equipped land army. But if they can't count on the Americans to keep providing air support, intelligence and communications networks, and other advanced systems, they will never emerge as an effective force. So the United States will have to continue to provide all this. The situation is ironic. Before the war insiders argued that sooner or later it would be necessary to attack, because the U.S. Air Force was being "strained" by its daily sorties over Iraq's no-fly zones. Now that the war is over, the United States has taken on a much greater open-ended obligation.

In sum, if the United States is serious about getting out of Iraq, it will need to re-consider its defense spending and operations rather than leaving them to a combination of inertia, Rumsfeld-led plans for "transformation," and emergency stopgaps. It will need to spend money for interpreters. It will need to create large new training facilities for American troops, as happened within a few months of Pearl Harbor, and enroll talented people as trainees. It will need to make majors and colonels sit through language classes. It will need to broaden the Special Forces ethic to much more of the military, and make clear that longer tours will be the norm in Iraq. It will need to commit air, logistics, medical, and intelligence services to Iraq—and understand that this is a commitment for years, not a temporary measure. It will need to decide that there are weapons systems it does not require and commitments it cannot afford if it is to support the ones that are crucial. And it will need to make these decisions in a matter of months, not years—before it is too late.

America's hopes today for an orderly exit from Iraq depend completely on the emergence of a viable Iraqi security force. There is no indication that such a force is about to emerge. As a matter of unavoidable logic, the United States must therefore choose one of two difficult alternatives: It can make the serious changes—including certain commitments to remain in Iraq for many years—that would be necessary to bring an Iraqi army to maturity. Or it can face the stark fact that it has no orderly way out of Iraq, and prepare accordingly.

*James Fallows is a national correspondent for The Atlantic.