'Kill team' trial: Are atrocities inevitable in war?
Why do soldiers kill innocent civilians in wartime?
The question is an old one, animated again by the "kill team" prosecution of US soldiers from the US Army's 5th Stryker Brigade and the court martial now under way of the alleged ringleader of the atrocities, Sgt Calvin Gibbs.
Several civilians were alleged to have been wantonly killed in Afghanistan by five soldiers in 2010.
It was a troubling episode, hardly typical of the conduct of American military men and women in the wars in Afghanistan and Iraq.
Yet, atrocities large and small recur with disturbing frequency.
The habitual response of the military is to describe such rampages as an "aberration" - bad actors among generally well-behaved soldiers.
An alternative account is that atrocities result from structural reasons: poor training, aggressive commanders, a permissive military.
And there are the conditions of war itself.
The fog of war now involves a blurry distinction between enemy fighters and civilians, and daily operations - house-to-house searches, roadblocks and vulnerable convoys, among others - in which soldiers anxiously encounter civilians.
Which explanation accounts for the "kill team" and the many other incidents of murders in Iraq and Afghanistan?
The answer is likely to involve all three.
The "bad actor" theory is most convenient for the military as it implicitly exonerates the military and the war policy.
Without a doubt, however, there are bad actors.
In the case of the "kill team", drugs and alcohol fuelled some of their deadly escapades. Prosecutors say Sgt Gibbs in particular goaded the others into the series of crimes.
But other known atrocities did not have these elements.
In the massacre of more than 100 unarmed civilians at No Gun Ri in Korea in 1950, only revealed 50 years later, no such bad actors were apparently at work.
Even at My Lai, where nearly 400 Vietnamese were murdered in 1968, the perpetrators did not manifest aberrant behaviour before that day.
The structural causes are more plausible as a general explanation.
Training on the rules of war is clearly inadequate, while training to kill is intensive.
A survey of US troops in Iraq at the height of the violence there found one-third of marines and one-quarter of soldiers saying that their leaders failed to tell them not to mistreat civilians.
Another army survey in 2007 found that "only 38% of marines and 47% of soldiers said non-combatants should be treated with dignity and respect."
More than one-third said torture of civilians was permissible to get information, and 17% viewed all civilians as insurgents.
Attitudes among officers during the Vietnam War were similar.
A survey of officer candidates in 1967 found half willing to use torture to get information, and a very large survey in 1970 found that 15% of officers and enlisted men did not comprehend the rules of war.
One-third of them gave incorrect answers to more than half of the questions about those rules, which are meant to protect civilians.
At the same time the training of soldiers to be killers is powerful.
A study of US soldiers in World War II found that only 15% to 20% ever discharged their weapons, even in the midst of combat.
Training changed that to 55% in the Korean War and 90% in Vietnam.
Some scholars say this shows a natural reluctance to kill other humans, a behaviour that must be hammered out of new soldiers in training.
The attitude of field officers is another major problem.
Officers were under pressure to produce high body counts of enemy in Vietnam and those pressures persist.
Killing civilians requires an investigation, so officers are prone to err on the side of describing all fatalities as those of the enemy.
In a survey, less than half of troops in Iraq told researchers that they report civilian casualties while 10% admitted they had harmed civilians.
Nowadays, too, soldiers do several tours of duty in these difficult circumstances.
Sgt Gibbs himself was on his third tour - the first a difficult one in Iraq. This form of battle fatigue creates in many young men mental health stresses that have brutal consequences.
We don't know precisely how many civilians are dying in Iraq and Afghanistan. Even less data is available from earlier wars.
But the estimates for Iraq from household surveys suggest it is in the hundreds of thousands. Afghanistan, which is less violent, is likely in the tens of thousands.
Much of this is attributable to insurgents. But the coalition forces have their share of responsibility.
How to change this dreadful outcome is not an easy question.
Better training, rules of engagement that emphasise civilian protection, fostering attitudes of respect toward the populations we are there to protect - these measures could help but they conflict with many other pressures and practices.
These practices and the sheer violence of war have tragic consequences.
When journalist Seymour Hersh was probing the My Lai massacre he spoke with the mother of one of the accused soldiers on her Indiana farm.
"I sent them a good boy," she told him, "and they sent me back a murderer."
John Tirman is Executive Director and principal research scientist of the MIT Center for International Studies, and the Author of The Deaths of Others: The Fate of Civilians in America's Wars