Gen. Tommy Franks, the U.S. commander in the Iraq war last year, spelled it out before the invasion began. "We don't do body counts," he said, referring to the Iraqis who might be killed in the forthcoming conflict. His deputies were left to explain why a careful toll of American dead was kept but Iraqi deaths went unrecorded. "It just is not worth trying to characterize by numbers," Brig. Gen. Vince Brooks, the deputy director of operations at U.S. Central Command, said just days before the fall of Baghdad.
"And, frankly, if we are going to be honorable about our warfare, we are not out there trying to count up bodies. This is not the appropriate way for us to go."
Occasionally the generals have not been able to resist. After the assault on Fallujah last month commanders said at least 1,200 rebels were killed. It was a claim impossible to verify. Even now the city is cordoned off by U.S. troops, the roads leading there are still extremely dangerous, and the Iraqi Red Crescent, the only aid agency operating inside the city, has had to pull out. When it has been possible to investigate the claims they have been difficult to verify.
Last December the military said it had killed 54 insurgents in ambushes in Samarra, a town north of Baghdad. Hospital officials, however, put the toll at only eight or nine. There is a tempting propaganda value for U.S. commanders to cite tolls of enemy dead -- during the Vietnam War "success" was often measured in the number of Vietnamese killed, and those numbers were routinely inflated to mask the painful realities on the ground. But the Americans have a broader policy in Iraq, and over the past 20 months the military has made no effort to record the thousands of Iraqi civilians who have been killed.
Wednesday British Prime Minister Tony Blair rejected a call from more than 40 diplomats, peers, scientists and religious leaders who pressed for an independent inquiry for a civilian death toll. "Figures from the Iraqi Ministry of Health, which are a survey from the hospitals there, are in our view the most accurate survey there is," he told Parliament.
The Health Ministry has produced a figure of 3,853 civilians killed between April and October of this year. But it is not clear whether those figures cover the entire country, how they were confirmed or what the causes of death were. No figures have been produced for the first year since the invasion. The task of trying to assess the civilian toll has been taken up by small groups working in a difficult and dangerous environment. With security in Iraq ever more fragile, and with nearly every Western charity out of the country, it has become harder to keep an accurate record.
Straight after the war the Guardian examined Baghdad hospital ledgers and gained an insight into the difficulty of keeping accurate figures. One large hospital, al-Kindi, officially recorded 192 civilians killed between the invasion and the fall of the regime on April 9. The Yarmouk hospital had 99 civilian deaths before it closed on April 6. But even then staff were cautious, saying many civilians were buried without being taken to hospital and that some may have been fighters out of uniform.
The Associated Press surveyed hospitals and found 3,420 civilian deaths in the war. Another survey by the Campaign for Innocent Victims in Conflict, which checked each reported death against several sources, came up with just under 2,000. It was slow and often inaccurate work.
Since then much of the recording has been done by Iraq Body Count, independent researchers based in Britain who use press reports and other public sources to establish a death toll not only for last year's war but also the occupation that has followed. Wednesday night its overall Iraqi civilian toll stood at between 14,619 and 16,804. The broad range is evidence in itself of how hard an accurate count is.
In June a report by the left-wing Institute for Policy Studies and Foreign Policy in Focus estimated 11,317 Iraqi civilians and 6,370 Iraqi soldiers or insurgents had died. Another survey by U.S. academics published in the Lancet in October was based on a survey of 33 randomly chosen districts in Iraq and came up with some 100,000 probable civilian deaths so far.
By comparison, the U.S. and British military keeps a careful record of servicemen and women killed in Iraq: As of Wednesday night, 1,276 Americans and 71 Britons since the invasion. In Iraq the relatives of those civilians killed often seem less concerned about the figures.
Most speak instead of a search for recognition: an acknowledgment of what happened, an investigation, an apology and compensation. For some it has come. For many others it has not, and that has left the bitter and dangerous taste of frustration.