Just before the war in Afghanistan, Noam Chomsky accused the United States of preparing the slaughter of three million to four million innocent people, whether directly or from starvation. The actual figure: between 500 and 1200 civilian deaths from combat. A massive influx of Western aid after the war averted the famine that was threatening the country before.

And now we return to Iraq. In the runup to the present conflict, the United Nations forecast civilian casualties in the range of 500,000. A little over two weeks into the war, the Iraq Body Count project puts the number at between 876 and 1049. But since the project's count is based on news reports, themselves based on claims by the Iraqi regime, the real number is probably a fraction of that.

At any rate, in all these recent conflicts, we are dealing with numbers of civilian deaths for which there are no historic comparisons. The "Chaco War" (1932-35) between Paraguay and Bolivia, to take a deliberately obscure example, resulted in 70,000 civilian deaths. The body count in the Vietnam War numbered in excess of one million. Even the Six Day War is thought to have killed as many as 50,000. The final stages of this war may be bloodier than the first, but on current trends it would be astonishing if the civilian death toll exceeded that of the first Gulf War, even in the face of the Iraqi regime's unprecedented efforts to cause as many civilian casualties as possible.

What's going on? Obviously technology is a part of it: the well-known preponderance of "smart bombs" in this war as opposed to the much smaller proportion, perhaps 10%, in Gulf War I, which ushered in the age of precision weaponry. But technology could as easily have been turned to producing the maximum number of fatalities, rather than the minimum, as indeed it was in the age of Mutually Assured Destruction.

A choice has been made, rather, to adhere to techniques that kill as few civilians as possible, even at some risk to the coalition forces themselves (though here again the losses are infinitesimal by historic standards).

It is quite clear that the coalition is going to the most extraordinary lengths to limit civilian casualties, beyond anything required under the Geneva Convention: avoiding "dual-use" targets such as bridges or the electrical supply, allowing the narrowest possible margins for error around a target, and so on. Indeed, one theme of the criticism directed at coalition military strategy has been a suggestion that it was being too dainty about civilian lives. "The prolongation of war caused by half measures kills many more people," writes Edward Luttwak in The Daily Telegraph, "than the restricted collateral damage of precision bombing ever could." This is very much in line with traditional military thinking. War's not a pretty business. Listen to the professionals, use maximum force, and leave politics out of it. As the Duke of Marlborough advised three centuries ago, "The pursuit of victory without slaughter is likely to lead to slaughter without victory." But the course the coalition is pursuing is driven by more than purely ethical concerns. It is a conscious act of policy, reflecting the very latest strategic thinking -- though in truth it relies on another centuries- old military doctrine, which is that you can't leave politics out of it.

When Clausewitz wrote that war was the continuation of politics by other means, he was not simply observing the truism that wars have political aims, but that aims and means are one: that military strategy must be informed by political considerations, through and through.

The armchair generals grouse that Donald Rumsfeld, the Defence Secretary, has tailored the campaign to fit his own doctrine of military "transformation." So he should. As Eliot Cohen writes in Supreme Command, a book that is known to have influenced administration thinking, only an active and engaged political leadership can take full account of the wider "theatre" in which any war inevitably takes place.

That is doubly true of this war. Simply put, it is an essential part of the coalition's war aims that there be as few civilian casualties as humanly possible. This is not only a matter of limiting the political fallout in the region. Rather, in the 21st century, this is the only kind of warfare that a democratic public is willing to wage, at least so long as its own territory is not being invaded. Fortunately, we have the technological capability to indulge our scruples, enough to compensate for the Baathists' conspicuous lack of same (their own tactical advantage.) For it is a fact that cannot be ignored.

A military that was unable or unwilling to adapt to this reality, that deployed its firepower according to more traditional notions of acceptable civilian losses, would be musclebound. If the only way the United States could put its massive military might into the field was in ways that inflicted thousands upon thousands of civilian deaths, it would cease to have any practical likelihood of being used, and as such would no longer act as a deterrent -- as is arguably the case with its nuclear arsenal. But a military that can topple dictatorships thousands of miles away at a cost in civilian lives numbering in the hundreds, now that is something truly to be feared.