I remember, just a few years back, Wes Clark was feeling pretty good about his war. This was in July 1999, just after he'd commanded the victorious NATO fight for Kosovo, just before he was fired by the Pentagon, and way before anyone thought about him as a presidential candidate. I had been in Belgrade when the allied air forces were attacking, and General Clark wanted to hear, firsthand, what it was like. "Pretty low-key for a bombed city, wasn't it?" he suggested over coffee.

"Well, yeah," I said, trying to explain just how strange the sensation had been. "You know, people came to have a lot of faith, in fact, in the accuracy of NATO bombs."

"Yeah," chortled Clark. "I know."

If Clark was feeling cocky, it was because thanks to the Air Force (and despite his own frustrated desire to send in ground troops), he'd just fought the cleanest, most efficient war in history, without a single allied casualty. The bombing campaign went on for 78 days, 38,000 sorties, yet by the end, people in Belgrade's sidewalk cafes barely looked up from their espressos when the air-raid sirens wailed. One of the city's main streets was known as the Boulevard of Ashes, because the government buildings on it had been destroyed systematically and repeatedly. But the buildings right next to them were mostly unscathed.

Gosh, that seems like a long time ago. It was before 9/11, of course. Before Afghanistan. Before Iraq. It was at a time when the entire West was mobilized to fight a war on moral grounds, to stop the suffering of the Muslims in Kosovo at the hands of the Serbian tyrant Slobodan Milosevic. The issue wasn't muddled by fear-mongering. Nobody claimed Milosevic was about to attack America with weapons of mass destruction. And it was at a time when the lives of American soldiers were held very dear, as were those of the people they were attacking.

Yet even then, the war wasn't as clean as Clark and other U.S. officials would claim. The Pentagon acknowledged only 20 or 30 incidents of "collateral damage" in the whole conflict. Human Rights Watch, after meticulous research, counted "about 500 civilians" killed in "90 separate incidents as a result of NATO bombing."

When the United States and Britain invaded Iraq last March, some of the lessons learned in Kosovo were applied, and some weren't. The bombing was amazingly meticulous. Even Human Rights Watch investigator Marc Garlasco gives the Air Force credit for "an outstanding job." But this time the U.S. Army and Marines did go into action, using weaponry that included multiple-launch rockets that rain cluster bombs on the enemy--and it appears that most of the civilians who died were killed by infantry.

How many civilians would that be? The Pentagon piled up statistics on just about everything, except for the number of innocents killed in Operation Iraqi Freedom and its aftermath. When asked why, the answer is always that the job is too difficult, that this was the Iraqis' business, that families buried their dead without ever reporting them. "We don't do body counts," Gen. Tommy Franks declared last year after the big offensives in Afghanistan.

What that means in Iraq is that while the Bush administration may spend $1 billion to find those elusive weapons of mass destruction, it just doesn't want to know about the people whose lives the war destroyed, much less pay them compensation.

Even as the administration talks about building democracy and preaches the rule of law, the U.S. government turns its back on the claims of civilians who lost fathers, mothers, brothers, sisters, sons and daughters in this war. Under American law, which is the only one that counts in this case, civilians have no redress if they are killed in combat--and in practice, any time an American soldier fires his weapons and an Iraqi, any Iraqi, gets killed, that's combat.

In August, when American military officials adjudicating Iraqi claims for financial compensation were asked about civilian casualties, they couldn't cite a single instance of wrongful death. They'd paid out a lot of money for accidental property damage caused by the occupation, but if money was paid for someone who'd been killed, which they doubted, it wasn't enough to make a blip on their screen. "The value of a life in Iraq--and I hate to say it--is probably a lot less in the United States, or Britain, for that matter," one officer explained.

Since then, the United States did give $11,000 to a woman who was eight months pregnant when American soldiers gunned down her husband and three of her children in Baghdad, but that was "for sympathy," she was told, and not an admission of error. The Reuters cameraman who was shot, the Iraqi police in Baghdad who were killed--they died under "the rules of engagement" and have no redress. The only case still under investigation is of the eight U.S.-trained Iraqi policemen blown away by U.S. troops in the rebellious city of Fallujah.

It's fallen to the press and to human-rights activists to try to count the Iraqi dead, and to one woman, a 26-year-old Californian named Marla Ruzicka, to try to help the victims' families.

In June, the Associated Press surveyed Iraqi hospitals and counted 3,420 confirmed civilian deaths from March 20 to April 20, at the height of major combat operations. But that's a very, very conservative number. Because the AP did not want to risk counting any Iraqi soldier as civilian, and Basra hospitals did not make clear distinctions, it counted no civilian deaths at all in the nation's second-largest city, where there was a lot of bloody urban combat.

A volunteer organization in Britain, IraqBodyCount.net, compiles a number based entirely on press reports. The current figure is between 7,376 and 9,178, but because this is an openly left-wing antiwar group, many people find its reporting suspect. Cofounder Hamit Dardagan takes the position that even if the numbers aren't totally accurate, "if the war was avoidable, so was all the collateral damage."

To tell you the truth, I think this emphasis on numbers, as such, is deeply ambiguous. I felt that way during Kosovo, and I feel that way about Iraq. It is people who are mourned, not statistics. Beyond a certain point (perhaps a few hundred) the scale becomes impossible to comprehend, a matter of curiosity and sterile debate. Just ask yourself, was the experience of September 11 eased when we found "only" 3,000 people had died in the World Trade Center and not 20,000 as we thought at first? Did those accounting clarifications help the grieving families?

That's why other groups deserve more attention, I think, than the simple body-counters. Human Rights Watch, which came up with the best number for civilian dead in Kosovo, has also sent researchers to Iraq. But they now put less emphasis on numbers than trends and legal issues, looking at the way civilian casualties can be reduced in future conflicts.

Ruzicka, who did similar work with civilian casualties in Afghanistan, has organized a group called the Campaign for Innocent Victims in Conflict (CIVIC) which has sent survey teams all over Iraq "to find cases," as she puts it, not just numbers. The idea is to push the American government to help the survivors, which you'd think would be a smart move politically as well as morally. So far, Ruzicka's teams have identified the cases of 1,860 civilians killed and 3,560 injured. "Why does the [Bush] administration say they go to great lengths to prevent these deaths," asks Ruzicka, "but when they occur, I am the one to go to the families and say 'sorry'?"

Good question. If the United States is going to fight wars to help people, it should learn to help, first, the ones that it hurts.