Although the Erbil handover was a sham that Prince Potemkin might have admired, it was not easily arranged. The Bush administration had wanted the handover to take place before the U.S. congressional elections in November. But it also wanted an Iraqi flag flown at the ceremony and some acknowledgement that Iraq, not Kurdistan, was in charge. The Kurds were prepared to include a reference to Iraq in the ceremony, but they were adamant that there be no Iraqi flags. It took months to work out a compromise ceremony with no flags at all. Thus the ceremony was followed by a military parade without a single flag — an event so unusual that one observer thought it might merit mention in Ripley's Believe it or Not.

Mowaffak al-Rubaie, the Iraqi national security adviser, attended the ceremony alongside Kurdistan's prime minister, Nechirvan Barzani, but the Iraqi government had no part in supervising the nonexistent handover. While General Mixon, a highly regarded strategist with excellent ties to the Kurds, had no choice but to make the remarks he did, Mowaffak al-Rubaie acknowledged Kurdistan's distinct nature and the right of the Kurds — approximately six million people, or some 20% of Iraq's population — to chart their own course.

On July 12, the White House released a congressionally mandated report on progress in Iraq. As with the sham handover, the report reflected the administration's desperate search for indicators of progress since it began its "surge" by sending five additional combat brigades to the country in February 2007. In recent months the Bush administration and its advocates have been promoting the success of the surge in reducing sectarian killing in Baghdad and achieving a turnaround in Anbar province, where former Sunni insurgents are signing up with local militias to fight al-Qaeda.

Although reliable statistics about Iraq are notoriously hard to come by it does appear that the overall civilian death toll in Baghdad has declined from its pre-surge peak, although it is still at the extremely high levels of the summer of 2006. Moreover, the number of unidentified bodies — usually the victims of Shiite death squads — has risen in May and June to pre-surge levels. How much of the modest decline in civilian deaths in Baghdad is attributable to the surge is not knowable, nor is there any way to know if it will last.

The developments in Anbar are more significant. Tribesmen who had been attacking U.S. troops in support of the insurgency are now taking U.S. weapons to fight al-Qaeda and other Sunni extremists. Unfortunately, the Sunni fundamentalists are not the only enemy of these new U.S.-sponsored militias. The Sunni tribes also regard Iraq's Shiite-led government as an enemy, and the U.S. appears now to be in the business of arming both the Sunni and Shiite factions in what has long since become a civil war.

Against the backdrop of modest progress, much has not changed, or has gotten worse. The Baghdad Green Zone is subject to increasingly accurate mortar attacks and is deemed at greater risk of penetration by suicide bombers. Moqtada al-Sadr, the radical Shiite cleric whose Mahdi Army was a major target of Bush's surge strategy, remains one of Iraq's most powerful political figures. The military activity against his forces seems only to have enhanced his standing with the public.

Even if the surge has had some modest military success, it has failed to accomplish its political objectives. The idea behind Bush's new strategy was to increase temporarily the number of U.S. troops in Baghdad and Anbar. The aim was to provide a breathing space so that Prime Minister Nouri al-Maliki's government might enact a program of national reconciliation that would accommodate enough Sunnis to isolate the insurgents. Meanwhile, Iraqi forces, improved by their close relations with U.S. troops and additional training, would take over security.

The core of the national reconciliation program is a series of legislative and political steps that the government should take to address the concerns of Iraq's Sunnis, who feel left out of the country they dominated until 2003. These steps include an oil revenue–sharing law (to ensure that the oil-poor Sunni regions get their share of revenue); holding provincial elections (the Sunnis boycotted the January 2005 provincial and parliamentary elections leaving them underrepresented even in Sunni-majority provinces); revising Iraq's constitution (the Sunnis want a more centralized state); revising the ban on public sector employment of former Baathists (Sunnis dominated the upper ranks of the Baath Party and of the Saddam-era public service), and a fair distribution of reconstruction funds. Both the administration and Congress have placed great emphasis on the obligation of the Iraqi government to achieve these so-called benchmarks. Congress has, by law, linked US strategy on Iraq and financial support of the Iraqi government to progress on these benchmarks and other steps.

Iraq's government has not met one of the benchmarks, and, with the exception of the revenue-sharing law, most are unlikely to happen. But even if they were all enacted, it would not help. Provincial elections will make Iraq less governable while the process of constitutional revision could break the country apart.

Ryan Crocker, the U.S. ambassador to Baghdad, likes to talk of the disparity between the Iraqi clock and the U.S. clock, suggesting that Iraqis believe they have more time to reach agreement than the American political calendar will tolerate. Crocker is the State Department's foremost Iraq hand but, more generally, American impatience often reflects ignorance. For example, both Congress and the administration have expressed frustration that the ban on public service by ex-Baathists has not been relaxed, since this appears to be a straightforward change, easily accomplished and already promised by Iraq's leaders.

Abdul Aziz al-Hakim leads the Supreme Islamic Iraqi Council (SIIC, previously known as SCIRI), which is Iraq's leading Shiite party and a critical component of Prime Minister al-Maliki's coalition. He is the sole survivor of eight brothers. During Saddam's rule Baathists executed six of them. On August 29, 2003, a suicide bomber, possibly linked to the Baathists, blew up his last surviving brother, and predecessor as SCIRI leader, at the shrine of Ali in Najaf. Moqtada al-Sadr, Hakim's main rival, comes from Iraq's other prominent Shiite religious family. Saddam's Baath regime murdered his father and two brothers in 1999. Earlier, in April 1980, the regime had arrested Moqtada's father-in-law and the father-in-law's sister — the Grand Ayatollah Baqir al-Sadr and Bint al-Huda. While the ayatollah watched, the Baath security men raped and killed his sister. They then set fire to the ayatollah's beard before driving nails into his head. De-Baathification is an intensely personal issue for Iraq's two most powerful Shiite political leaders, as it is to hundreds of thousands of their followers who suffered similar atrocities.

Iraq's Shiite leaders are reluctant to spend reconstruction money in Sunni areas because they believe, not without reason, that such funds support the Sunni side in the civil war. In a speech in late June on the Senate floor Indiana Republican Richard Lugar reported that Iraq's Shiite-led government has gone "out of its way to bottle up money budgeted for Sunni provinces" and that the "strident intervention" of the U.S. embassy was required in order to get food rations delivered to Sunni towns.

Iraq's mainstream Shiite leaders resist holding new provincial elections because they know what such elections are likely to bring. Because the Sunnis boycotted the January 2005 elections, they do not control the northern governorate, or province, of Nineveh, in which there is a Sunni majority, and they are not represented in governorates with mixed populations, such as Diyala province, northeast of Baghdad. New elections would, it is argued, give Sunnis a greater voice in the places where they live, and the Shiites say they do not have a problem with this, although just how they would treat the militant Sunnis who would be elected is far from clear. The Kurds reluctantly accept new elections in the Sunni governorates even though it means they will lose control of Nineveh and have a much-reduced presence in Diyala.

The American benchmark of holding provincial elections would also require new elections in southern Iraq and Baghdad. If they were held, al-Hakim's Shiite party, the SIIC, which now controls seven of the nine southern governorates, would certainly lose ground to Moqtada al-Sadr. His main base is in Baghdad and new elections would almost certainly leave his followers in control of Baghdad Governorate, with one quarter of Iraq's population. Iraq's decentralized constitution gives the governorates enormous powers and significant shares of the national budget, if they choose to exercise these powers. New local elections are not required until 2009 and it is hard to see how early elections strengthening al-Sadr, who is hostile to the U.S. and appears to have close ties to Iran, serve American interests. But this is precisely what the Bush administration is pushing for and Congress seems to want.

Constitutional revision is the most significant benchmark and it could break Iraq apart. Iraq's constitution, approved by 79% of voters in an October 2005 referendum, is the product of a Kurdish–Shiite deal: the Kurds supported the establishment of a Shiite-led government in exchange for Shiite support for a confederal arrangement in which Kurdistan and other regions like the one SIIC hopes to set up in the south, are virtually independent.

Since there is no common ground among the Shiites, Kurds, and Sunnis on any significant constitutional changes in favor of the Sunnis, such changes must come at the expense of the Kurds or Shiites. Since voters in these communities have a veto on any constitutional amendments, they are certain to fail in a referendum. A revised constitution has no chance of being enacted but its failure will exacerbate tensions among Iraq's three groups.

Constitutionally, Iraq's central government has almost no power, and the Bush administration is partially to blame for this. When the constitution was being drafted in 2005, the United Nations came up with a series of proposals that would have made for more workable sharing of power between regions and the central government. The U.S. embassy stopped the UN from presenting these proposals because it hoped for a final document as centralized as (and textually close to) the interim constitution written by the Americans.

When the constitution finally emerged in its present form, then U.S. Ambassador Zalmay Khalilzad brokered a deal with several Sunni leaders whereby, in exchange for Sunni support for ratification, there would be a fast-track process to revise the constitution in the months following ratification to meet Sunni concerns. Like the Bush administration, the Sunnis want a more centralized state. While the U.S. insists that constitutional revision is a moral obligation, the Sunnis actually never lived up to their end of the bargain. Almost unanimously, they voted against ratification of the current constitution.

With input from the United Nations (belatedly brought back into the process last year), the Iraqi Parliament's mainly Arab Constitutional Review Committee (CRC) is considering amendments that would strip Kurdistan of many of its powers, including its right to cancel federal laws, to decide on taxes applicable in its own territory, and to control its own oil and water. The Sunni Arabs would also like Iraq declared an Arab state, a measure the non-Arab Kurds consider racist and exclusionary.

Thanks to Khalilzad's expedited procedures, constitutional revision may be the final wedge between Kurdistan and Arab Iraq. If approved by the CRC, the constitutional amendments will be subject to a vote in the parliament as a single package and then to a nationwide referendum. Kurdistan's voters are certain to reject the proposed package (or any package affecting Kurdistan's powers), and this could push tense Sunni–Kurdish relations into open conflict. Kurdish NGOs, who ran a 2005 independence referendum, are poised to make a "NO" campaign on constitutional revision a "No to Iraq" vote. In its July 12 report to Congress, the White House graded the CRC's work as "satisfactory," an evaluation that was either grossly dishonest, or, more likely, out of touch with Iraqi reality.

For the most part, Iraq's leaders are not personally stubborn or uncooperative. They find it impossible to reach agreement on the benchmarks because their constituents don't agree on any common vision for Iraq. The Shiites voted twice in 2005 for parties that seek to define Iraq as a Shiite state. By their boycotts and votes the Sunni Arabs have almost unanimously rejected the Shiite vision of Iraq's future, including the new constitution. The Kurds' envisage an Iraq that does not include them. In the 2005 parliamentary elections, 99% of them voted for Kurdish nationalist parties, and in the January 2005 referendum, 98% voted for an independent Kurdistan.

But even if Iraq's politicians could agree to the benchmarks, this wouldn't end the insurgency or the civil war. Sunni insurgents object to Iraq being run by Shiite religious parties, which they see as installed by the Americans, loyal to Iran, and wanting to define Iraq in a way that excludes the Sunnis. Sunni fundamentalists consider the Shiites apostates who deserve death, not power. The Shiites believe that their democratic majority and their historical suffering under the Baathist dictatorship entitle them to rule. They are not inclined to compromise with Sunnis, whom they see as their longstanding oppressors, especially when they believe most Iraqi Sunnis are sympathetic to the suicide bombers that have killed thousands of ordinary Shiites. The differences are fundamental and cannot be papered over by sharing oil revenues, reemploying ex-Baathists, or revising the constitution. The war is not about those things.

2.

The Iraq war is lost. Of course, neither the President nor the war's intellectual architects are prepared to admit this. Nonetheless, the specter of defeat shapes their thinking in telling ways.

The case for the war is no longer defined by the benefits of winning — a stable Iraq, democracy on the march in the Middle East, the collapse of the evil Iranian and Syrian regimes — but by the consequences of defeat. As President Bush put it, "The consequences of failure in Iraq would be death and destruction in the Middle East and here in America."

Tellingly, the Iraq war's intellectual boosters, while insisting the surge is working, are moving to assign blame for defeat. And they have already picked their target: the American people. In The Weekly Standard, Tom Donnelly, a fellow at the neoconservative American Enterprise Institute, wrote, "Those who believe the war is already lost — call it the Clinton-Lugar axis — are mounting a surge of their own. Ground won in Iraq becomes ground lost at home." Lugar provoked Donnelly's anger by noting that the American people had lost confidence in Bush's Iraq strategy as demonstrated by the Democratic takeover of both houses of Congress. (This "blame the American people" approach has, through repetition, almost become the accepted explanation for the outcome in Vietnam, attributing defeat to a loss of public support and not to fifteen years of military failure.)

Indeed, Vietnam is the image many Americans have of defeat in Iraq. Al-Qaeda would overrun the Green Zone and the last Americans would evacuate from the rooftop of the still unfinished largest embassy in the world. President Bush feeds on this imagery. In his May 5, 2007, radio address to the nation, he explained:

If radicals and terrorists emerge from this battle with control of Iraq, they would have control of a nation with massive oil reserves, which they could use to fund their dangerous ambitions and spread their influence. The al Qaeda terrorists who behead captives or order suicide bombings would not be satisfied to see America defeated and gone from Iraq. They would be emboldened by their victory, protected by their new sanctuary, eager to impose their hateful vision on surrounding countries, and eager to harm Americans.

But there will be no Saigon moment in Iraq. Iraq's Shiite-led government is in no danger of losing the civil war to al-Qaeda, or a more inclusive Sunni front. Iraq's Shiites are three times as numerous as Iraq's Sunni Arabs; they dominate Iraq's military and police and have a powerful ally in neighboring Iran. The Arab states that might support the Sunnis are small, far away (vast deserts separate the inhabited parts of Jordan and Saudi Arabia from the main Iraqi population centers), and can only provide money, something the insurgency has in great amounts already.

Iraq after an American defeat will look very much like Iraq today — a land divided along ethnic lines into Arab and Kurdish states with a civil war being fought within its Arab part. Defeat is defined by America's failure to accomplish its objective of a self-sustaining, democratic, and unified Iraq. And that failure has already taken place, along with the increase of Iranian power in the region.

Iraq's Kurdish leaders and Iraq's dwindling band of secular Arab democrats fear that a complete U.S. withdrawal will leave all of Iraq under Iranian influence. Senator Hillary Clinton, Foreign Relations Committee Chairman Joe Biden, and former UN Ambassador Richard Holbrooke are among the prominent Democrats who have called for the U.S. to protect Kurdistan militarily should there be a withdrawal from Iraq. The argument for so doing is straightforward: it secures the one part of Iraq that has emerged as stable, democratic, and pro-Western; it discharges a moral debt to our Kurdish allies; it deters both Turkish intervention and a potentially destabilizing Turkish–Kurdish war; it provides U.S. forces a secure base that can be used to strike at al-Qaeda in adjacent Sunni territories; and it limits Iran's gains.

In laying out his dark vision of an American failure, President Bush never discusses Iran's domination of Iraq even though this is a far more likely consequence of American defeat than an al-Qaeda victory. Bush's reticence is understandable since it was his miscalculations and incompetent management of the postwar occupation that gave Iran its opportunity. While opposing talks with Iran, the neoconservatives also prefer not to discuss its current powerful influence over Iraq's central government and southern region, persisting in the fantasy — notwithstanding all evidence to the contrary — that Iran is deeply unpopular among Iraq's Shiites and clerics. (At the same time, U.S. officials accuse Iran of supplying Iraqi Shiite militias with particularly lethal roadside bombs.)

3.

On June 25, without giving the press or White House any advance notice, Richard Lugar, the most respected Republican voice on foreign affairs in Congress, spoke in the Senate about "connecting our Iraq strategy to our vital interests." On the face of it, the idea is as sensible and conservative as the senator delivering the speech. He observed that political fragmentation in Iraq, the stress suffered by the U.S. military, and growing antiwar sentiment at home "make it almost impossible for the United States to engineer a stable, multi-sectarian government in Iraq in a reasonable time frame." Lugar noted that agreements reached with Iraqi leaders are most often not implemented, partly, as Lugar observed, because the leaders do not control their followers but also because Iraqi leaders have also discovered that telling the Bush administration what it wants to hear is a fully acceptable substitute for action.

Lugar is blunt in his description of the situation in Iraq:

Few Iraqis have demonstrated that they want to be Iraqis.... In this context, the possibility that the United States can set meaningful benchmarks that would provide an indication of impending success or failure is remote. Perhaps some benchmarks or agreements will be initially achieved, but most can be undermined or reversed by a contrary edict of the Iraqi government, a decision by a faction to ignore agreements, or the next terrorist attack or wave of sectarian killings. American manpower cannot keep the lid on indefinitely. The anticipation that our training operations could produce an effective Iraqi army loyal to a cohesive central government is still just a hopeful plan for the future.

Lugar concluded his speech by urging that we "refocus our policy in Iraq on realistic assessments of what can be achieved, and on a sober review of our vital interests in the Middle East." After four years of a war driven more by wishful thinking than strategy, this is hardly a radical idea, but it has produced a barrage of covert criticism of Lugar from the administration and overt attack from the neoconservatives.

Lugar's focus on the achievable runs against main currents of opinion in a nation increasingly polarized between the growing number who want to withdraw from Iraq and the die-hard defenders of a failure. We need to recognize, as Lugar implicitly does, that Iraq no longer exists as a unified country. In the parts where we can accomplish nothing, we should withdraw. But there are still three missions that may be achievable — disrupting al-Qaeda, preserving Kurdistan's democracy, and limiting Iran's increasing domination. These can all be served by a modest U.S. presence in Kurdistan. We need an Iraq policy with sufficient nuance to protect American interests. Unfortunately, we probably won't get it.

Peter W. Galbraith, a former US Ambassador to Croatia, is Senior Diplomatic Fellow at the Center for Arms Control and a principal at the Windham Resources Group, a firm that negotiates on behalf of its clients in post-conflict societies, including Iraq. His The End of Iraq: How American Incompetence Created a War Without End is now out in paperback