When soldiers commit atrocities, we must ask why. The question is being raised once again by reports that a handful of U.S. soldiers in Afghanistan carried out premeditated killings—murders—of Afghan civilians. The soldiers allegedly took photographs of themselves posing with corpses and body parts, including fingers and heads.

The alleged ringleader is Sergeant Calvin Gibbs. In an interrogation video leaked to CNN, Specialist Adam Winfield, a member of Gibbs's platoon, said that Gibbs "likes to kill things. He is pretty much evil incarnate. I mean, I have never met a man who can go from one minute joking around, then mindless killings."

Military officials invariably blame these sorts of atrocities on "bad apples." That was the phrase that former Secretary of Defense Donald Rumsfeld used to describe American guards accused of abusive behavior at Abu Ghraib prison in 2004.

Some evidence supports the bad-apples theory of atrocities. In a previous post, I cited a report by two psychiatrists that after 60 days of continuous combat, 98 percent of World War II infantrymen suffered from psychiatric illness, with some succumbing to a near-catatonic "vegetative phase." But 2 percent of the soldiers, far from being traumatized by intense, prolonged combat, enjoyed it. The psychiatrists diagnosed these soldiers with "aggressive psychopathic personalities."

Dave Grossman, a former professor of psychology at West Point and Army lieutenant colonel, acknowledged in his 1995 book On Killing (Little, Brown, 1995) that a small number of men—whom he called "the two percent who like it"—can "kill without regret or remorse." According to Grossman, these men may be excellent soldiers when competently trained and supervised, but they are also more likely than other men to use excessive force and commit atrocities.

The description above of someone joking one minute and killing the next sounds like textbook psychopathy. In "Inside the Mind of  a Psychopath," in the September/October issue of Scientific American MIND, the neuroscientists Kent Kiehl and Joshua Buckholtz stated that psychopaths "are guilty of the most erratic and irresponsible, sometimes destructive and violent behavior," for which they "feel no compunction or regrets." Psychopaths, who comprise as much as 35 percent of U.S. prisoners, seem incorrigible; they may behave worse after treatments such as group psychotherapy, Kiehl and Buckholtz said, because "insights into others' vulnerabilities become opportunities to hone their manipulation skills."

Today, some psychiatrists prefer the terms "sociopathy" or "antisocial personality disorder" to psychopathy. According to the Diagnostic and Statistical Manual of the American Psychiatric Association, antisocial personality disorder is characterized by extreme aggression, lack of empathy for others, lack of remorse for one's actions—and, not surprisingly, a propensity for violent crime. The manual estimates that 3 percent of all males have the disorder, which is suggestively close to the "two percent who like it." The disorder is much less common among women.

A British study of school-age twins published in 2005 [pdf] suggests that psychopathy has a strong genetic component. Teacher surveys revealed psychopathic tendencies (including antisocial behavior and "callous-unemotional traits") in 234 children—all less than 10 years old—out of a total of 3,687 pairs of twins, or roughly 3 percent. If one identical twin was psychopathic, the other was much more likely to be so; the concordance between fraternal twins was smaller.

In Final Solutions (Cornell, 2005), the political scientist Benjamin Valentino asserted that small percentages of men caused much of the slaughter of the 20th century. "The impetus for mass killing usually originates from a relatively small group of powerful leaders and is often carried out without the active support of broader society," Valentino stated. This pattern was true of mass killings in the Soviet Union, China, Cambodia, the Balkans, Guatemala and elsewhere.

Similarly, the biologist Barbara Oakley argued in Evil Genes (Prometheus, 2007) that Hitler, Stalin, Mao Zedong and other notorious tyrants displayed symptoms of psychopathy. Oakley concluded that they were "born to be bad." But in her 1963 essay on the Nuremberg trials, Hannah Arendt noted that psychiatric evaluations of Nazi mass murderers such as Adolf Eichmann suggested that they were "neither perverted nor sadistic, that they were, and still are, terribly and terrifyingly normal." Arendt attributed Eichmann's crimes to "circumstances that make it well-nigh impossible for him to know or feel that he is doing wrong." Arendt was rejecting the bad-apple theory and blaming "circumstances" for the Holocaust.

This conclusion was corroborated by famous experiments carried out in the 1960s by the psychologist Stanley Milgram. The son of Jewish immigrants, Milgram devised his experiments in part as a reaction to the Nuremberg trials, which left him wondering about the motivation of Eichmann and other Nazis. In Milgram's experiments—the details of which are still chilling—subjects were told that they were participating in a test of another person's learning ability. The "learners" were actually actors in cahoots with Milgram.

The subject read pairs of words to the learner—who was in an adjoining room and could be heard but not seen by the subject—and then tested his ability to remember the pairings. Each time the learner failed to remember a pairing, the scientist, who was in the same room as the subject, ordered him to give the learner a stronger electric shock. As the shocks increased, the learner reacted with audible distress, crying out in pain, banging on the wall or even claiming that he was about to have a heart attack. After a certain point, the learner would fall silent.

Of course, the learner was pretending to be shocked. If the subject hesitated to deliver stronger shocks, the scientist insisted that the subject continue, adding that he would not be held responsible for anything that happened to the learner. Only if the subject resisted four successive commands from the scientist was the experiment stopped. Otherwise the experiment continued until the subject had administered a "shock" of 450 volts to the learner.

Before the experiment, Milgram asked several dozen psychiatrists to predict the results; the average guess was that only 1 percent of the subjects, those with sadistic tendencies, would deliver the strongest shock. But in Milgram's initial experiment 26 out of 40 subjects, or almost two-thirds, administered what they believed to be the strongest, life-threatening shock. Only one subject refused to continue the experiment before reaching the 300-volt level. Versions of Milgram's experiment have been repeated in the U.S. and elsewhere with similar results.

In 1974 Milgram spelled out the implications of his research: "The extreme willingness of adults to go to almost any lengths on the command of an authority constitutes the chief finding of the study and the fact most urgently demanding explanation. Ordinary people, simply doing their jobs, and without any particular hostility on their part, can become agents in a terrible destructive process. Moreover, even when the destructive effects of their work become patently clear, and they are asked to carry out actions incompatible with fundamental standards of morality, relatively few people have the resources needed to resist authority."

In 1971 the psychologist Philip Zimbardo, Milgram's former classmate at a New York City high school, carried out the so-called Stanford Prison Experiment, which has become almost as renowned as Milgram's work. Zimbardo created a mock prison in which Stanford students played the roles of either prisoners or guards. The "guards" quickly became so abusive—and the "prisoners" so distressed—that Zimbardo had to discontinue the experiment. Some of the guards' abusive acts—which included forcing prisoners to strip and to engage in simulated homosexual intercourse—were strikingly similar to acts perpetrated more than three decades later by American soldiers against Iraqis held in the Abu Ghraib prison.

This research suggests that human aggression and cruelty stem less from the "disposition" of individuals than from their environment, or "situation," Zimbardo argued in The Lucifer Effect (Random House, 2008). Studies of modern suicide bombers, torturers and war criminals, Zimbardo wrote, have revealed that many are, in Arendt's words, "terrifyingly normal." People behave badly not because they are bad apples but because they are in "bad barrels," situations that encourage brutality. War is the ultimate bad barrel. "In all wars, at all times, in every country, wars transform ordinary, even good men into killers," Zimbardo stated.

In War of the World: History's Age of Hatred (Penguin, 2007), the British historian Niall Ferguson described how combatants in World Wars I and II became consumed with hatred for their opponents. As a result, even Americans and British soldiers, the putative good guys, engaged in escalating atrocities, including bombing civilians, torturing and killing prisoners and mutilating the dead. This emotion-fueled descent into brutality is an inevitable consequence of the bad barrel of war.

If genocide, war crimes and other atrocities were all perpetrated by a few bad apples born with bad genes, we could perhaps identify them through genetic testing and sequester them from the rest of us good, decent, peaceful folk. If only things were that easy.

Photograph of 1968 massacre in My Lai, Vietnam, courtesy of Wiki Commons