Since the start of the American war in Afghanistan, dozens of mass-casualty incidents involving Afghan civilians have plagued the enterprise. According to a document obtained by The Nation, three incidents in particular led the US command to issue a series of so-called tactical directives, other guidances and what the military calls fragmentary orders (FRAGOs) designed to restrict military operations and reduce civilian casualties.

Yet according to a 2011 Joint Civilian Casualty Study (JCCS) for the military called “Reducing and Mitigating Civilian Casualties,” and backed by interviews conducted by The Nation, those directives were less than ideally effective. The JCCS was written by Sarah Sewall of Harvard and Larry Lewis of the US Joint Forces Command, with a foreword by Gen. David Petraeus, then commander of the International Security Assistance Force (ISAF) in Afghanistan.

The three incidents, which collectively resulted in as many as 300 civilian deaths (and possibly more), were airstrikes in the provinces of Herat (2008), Farah (2009) and Kunduz (2009); the strike in Herat is commonly known as “the Azizabad incident,” after the village in which it occurred. In this article, each one is described in some detail, along with the aftermath and military fallout. As the JCCS report notes:

After the Azizabad incident, both ISAF and CENTCOM [Central Command] released Tactical Directives in September 2008 involving reporting of potential civilian casualties from airstrikes. The ISAF version also specifically called for limiting airstrikes on compounds to avoid civilian casualties when ISAF forces are not in imminent danger. The Farah incident led to a second ISAF Tactical Directive in July 2009 that reiterated the need to limit airstrikes on compounds. And the Kunduz incident led to the redefining of the term “Troops in Contact” to prevent self defense criteria from being applied inappropriately.

Though well-intentioned, each of the tactical directives and other guidances didn’t stop mass-casualty incidents from taking place. They generated controversy within the military’s own ranks, and troops at all levels often chafed under what they felt were overly severe restrictions that handcuffed them in dangerous situations. Whether the directives were taken seriously at all often depended on how much emphasis the senior command put on them; after all, similar directives had been in place since at least 2007 but were often ignored in practice.

According to the JCCS, whose authors interviewed Gen. Stanley McChrystal, the head of ISAF in 2009–10, McChrystal argued strongly that without top-down emphasis on civilian casualty prevention, the directives would have little or no impact.

Still, despite the directives, a steady drumbeat of mass-casualty incidents took place, along with hundreds of smaller-scale civilian killings involving night raids as well as “escalation of force” and ground-fire incidents. Although advocates of the counterinsurgency doctrine, such as McChrystal and Petraeus, recognized that civilian casualties threatened the US effort, it is also true that war cannot be waged without significant civilian casualties.

In what follows, The Nation recounts the incidents in Azizabad, Farah and Kunduz and relates them to subsequent tactical directives and guidances, using the JCCS as a guide.

I. Azizabad (Herat), 2008

The bloody events that transpired in Azizabad, a village in the Shindand district of Herat Province, emerged slowly, unfolding in horror in the days and weeks after Friday, August 22, 2008. At the beginning, it appeared to be just another airstrike in the midst of an escalating war, following an ambush by the Taliban of ISAF soldiers on patrol. After the incident, the coalition “denied killing civilians,” according to Reuters. Instead, ISAF issued a terse statement that read: “Insurgents engaged the soldiers from multiple points within the compound using small-arms and rocket-propelled grenade fire. The joint forces responded with small-arms fire and an airstrike killing 30 militants.”

Not so. As would soon be revealed, as many as ninety-two civilians, including sixty children, died that night.

The airstrike, by an AC-130H gunship and an MQ-9 Reaper UAV drone, began between 1 and 2 am, when US and Afghan troops reportedly pursued a Taliban commander named Mullah Sadiq to Azizabad, got into a firefight and then called in air support. According to Human Rights Watch, citing military reports, “Airstrikes lasted for two to three hours, and reportedly entailed the dropping of a 500-pound bomb from a drone and shellfire from the gunship’s M102 105mm howitzer and 40mm grenades.” Initially, the Afghan Interior Ministry and local officials reported scores of civilian casualties. A local elder in Azizabad told Reuters: “Last night, around 2 a.m., some people were attending a holy Koran recitation in Shindand district when Americans started bombing.” While the Afghan Defense Ministry reported that only five civilians had died, the early count by the Interior Ministry put the death toll at seventy-six. The next day, President Hamid Karzai said the death toll had reached ninety-five. The Afghan government, the US military and the UN all pledged to investigate the slaughter.

Still denying civilian casualties even as more and more evidence emerged that innocents had been killed, ISAF released the following statement: “Coalition forces are aware of allegations that the engagement in the Shindand District of Herat Province Friday may have resulted in civilian casualties. Allegations of civilian casualties are taken very seriously. Coalition forces make every effort to prevent the injury or loss of innocent lives. An investigation has been directed.”

Attempting to piece together what happened, the New York Times reported that a team of Afghan army commandos and US Special Forces called in an airstrike that was based on poor or faulty intelligence or deliberate misinformation. Said the Times:

A [Karzai] presidential aide who declined to be identified said that the Interior Ministry and the Afghan intelligence agency had reported from the region that there were no Taliban present in the village that night. The Afghan National Army, whose commandos called in the airstrike along with American Special Forces trainers, were unable to clarify their original claim, he said….

A tribal elder from the region who helped bury the dead, Haji Tor Jan Noorzai, said people in the village were gathered in memory of a man who was anti-Taliban and had been killed last year, and that tribal enemies of the family had given out false information.

“It is quite obvious, the Americans bombed the area due to wrong information,” he said by telephone. “I am 100 percent confident that someone gave the information due to a tribal dispute. The Americans are foreigners and they do not understand. These people they killed were enemies of the Taliban.”

A human rights team from the United Nations Assistance Mission in Afghanistan was sent to the scene to investigate. Sifting through the wreckage, UNAMA carefully compiled information on every civilian who perished. It reported:

[The] team met with the District Governor and local elders yesterday. They also interviewed people from a number of households in Nawabad village who confirmed to us that at around midnight on the 21st August, foreign and Afghan military personnel entered the village of Nawabad in the Azizabad area of Shindand district. Military operations lasted several hours during which air strikes were called in. The destruction from aerial bombardment was clearly evident with some 7-8 houses having been totally destroyed and serious damage to many others. Local residents were able to confirm the number of casualties, including names, age and gender of the victims.

Investigations by UNAMA found convincing evidence, based on the testimony of eyewitnesses, and others, that some 90 civilians were killed, including 60 children, 15 women and 15 men. 15 other villagers were wounded or otherwise injured.

Kai Eide, the UN special representative for Afghanistan at the time, was furious. “This is a matter of grave concern to the United Nations,” he said. “It is vital that the international and Afghan military forces thoroughly review the conduct of this operation in order to prevent a repeat of this tragic incident.”

Several days later, the military was still disputing the claims of civilian casualties. On August 27, the New York Times’s Eric Schmitt reported: “In the face of an investigation by Afghan officials and a report by a United Nations team that support the high number of deaths, United States officials maintain that 25 militants and 5 civilians were killed in airstrikes called in after Afghan and American commandos came under heavy fire during a raid on the compound of a top Taliban commander.” While refusing to comment publicly, a US military official told the Times there was “no evidence to support the higher civilian death toll [nor] any evidence of a large number of recently dug graves or large number of injuries reported in local hospitals.”

But the Azizabad story wouldn’t go away. On September 7, the Times published the results of its own investigation. Its description was horrific:

At the battle scene, shell craters dotted the courtyards and shrapnel had gouged holes in the walls. Rooms had collapsed and mud bricks and torn clothing lay in uneven mounds where people had been digging. In two places blood was splattered on a ceiling and a wall. An old woman pushed forward with a cauldron full of jagged metal bomb fragments, and a youth presented cellphone video he said was shot on the day of the bombing…. The smell of bodies lingered in one compound, causing villagers to start digging with spades. They found the body of a baby, caked in dust, in the corner of a bombed-out room.

The Afghanistan Independent Human Rights Commission investigated the bloodbath, interviewing witnesses and collecting data, two days after it occurred. One witness, a “local worker for a mine clearance company,” told the AIHRC:

I heard the noise of helicopters and stepped out to the courtyard. Then I got injured in my head and left arm. I ran into a drain and stayed there up to 8:00 in the morning. When I came back to my house, I saw my wife, my two daughters, and my son had all died. Beside them, my brother, his wife, his two sons and two daughters had also died. My house and my two brothers’ houses were destroyed as well.

Another witness, a 75-year-old man, told the AIHRC:

It was about 2:00 am. We were asleep when a heavy explosion woke me up. Following that I heard noise of helicopters, started firing rockets, shelling and bombings, and lasted for 6 hours.

And Soraya Sarhaddi Nelson reported for NPR that many of those killed were visitors gathered for a memorial for a prominent local villager:

Abdul Rashid treads gently across the rubble that was once his uncle’s home in Azizabad, Afghanistan. He spots a little girl’s shoe, caked in dried blood. He picks it up and waves it angrily in a visitor’s face.

“Does this look like it fits a Taliban fighter?” Rashid says. Still clinging to the tiny shoe, Rashid takes a few more steps and picks up a torn woman’s veil.

“Does this look like something the Taliban would wear? Can you believe it? This is what Afghanistan has become,” Rashid says, his rage dissolving into tears….

Rashid says the latest victim, a 6-year-old girl, was buried the day before.

Afghan officials say Rashid’s neighborhood, which was the target of the military operation, was packed with visitors the morning of the attack. They were there to attend a memorial ceremony honoring his late brother [Timor Shah], a local strongman.

Rashid says vats of meat, rice and potatoes were being prepared for the scores of mourners when the soldiers attacked.

Although US Special Forces conducted what the Times described as “an initial battlefield review, including a building by building search,” followed four days later by a US military team’s visit to the “vicinity” of the attack, the Pentagon wasn’t backing off its early assessment. Finally, on September 7, more than two weeks later, Gen. David McKiernan, the US commander, ordered a review and investigation. The inspection team interviewed villagers, reviewed cellphone videos of the aftermath, visited at least six burial sites and determined that more than thirty civilians had died. Those conclusions, which would generate controversy, were included in an official report issued by Gen. Michael Callan.

Among other things, the partial admission of civilian casualties in the Callan report was tempered by the military’s assertion that the target was still a legitimate one. Speaking anonymously to the Times, a military official attributed the initial assessment to the fact that the troops were operating in a hostile environment: “We were wrong on the number of civilian casualties partly because the initial review was operating under real limitations. [We] were definitely not welcome there.” Yet the military didn’t explain why, without having conducted a thorough inquiry, it was so confident in its early assessment that no civilians had been killed.

In Washington, Gen. James Conway, the Marine Corps commandant, noted that sometimes, in situations such as the attack in Azizabad, the Taliban deliberately use civilians in residential compounds as shields. “Sometimes we think there’s been overt efforts on the part of the Taliban, in particular, to surround themselves with civilians so as to, at a minimum, reap an I.O. [information operations] advantage if civilians are killed. You want to strike the precise building that you’re targeting, but sometimes there are other people there. And you don’t know that, because you’re not on the inside looking out.”

The Azizabad incident led both the coalition and the US military to order specific changes in policy, two of several directives designed to reduce civilian casualties, build local support for the US counterinsurgency effort, and ease political pressure on ISAF from the Afghan government. According to the JCCS, a tactical directive instructed foreign forces to avoid airstrikes against Afghan compounds. “After the Azizabad incident, both ISAF and CENTCOM released Tactical Directives in September 2008 involving reporting of potential civilian casualties from airstrikes,” the report said. “The ISAF version also specifically called for limiting airstrikes on compounds to avoid civilian casualties when ISAF forces are not in imminent danger.”

That report said that the September 2008 directive issued by General McKiernan focused on “raids, use of air-to-ground and indirect fire, escalation of force, and being first with the truth.” It “called for forces to use these fires only when there was no other option to protect the force, removing the option to use that type of force in order to accomplish the mission.” According to the JCCS, a parallel directive applied not just to ISAF but to US forces specifically.

The report added that McKiernan’s successor, Gen. Stanley McChrystal, expressed unhappiness in an interview about having to apologize to the Afghans when the US military got civilian casualty information wrong. Said the report: “He recounted that he had been troubled by civilian casualty incidents early in his command where he had defended US force accounts against Afghan allegations, only to discover later that those initial US assessments had been wrong. He was then compelled to apologize for his wrong information.”

However, in a January 2009 letter to then–Defense Secretary Robert Gates, Brad Adams, the director of the Asia division for Human Rights Watch, was highly critical of the Defense Department and the military regarding Azizabad. Adams did acknowledge that the US military changed its procedure, including the new tactical directive. However, in the letter to Gates, Human Rights Watch blasted the Callan report for its deficiencies:

Instead of being an exemplary US investigation derived from a new operational mandate, the Callan Report Summary appears to be little more than a return to the discredited inquiries of recent years. It simply and summarily dismisses the methodology used in the investigations by the United Nations, the government of Afghanistan, and the Afghan Independent Human Rights Commission (AIHRC); rejects information provided by villagers by arbitrarily calling into question their motivations; effectively places responsibility for preventable civilian deaths on Taliban forces; and exonerates US forces of any wrongdoing.

Human Rights Watch was also critical of the report’s conclusion that the military’s level of force and tactics in Azizabad were “necessary” and “proportional.” In particular, it noted with dismay that the military failed to explain how it concluded that many of the dead were Taliban-linked insurgents. Apparently, according to Human Rights Watch, the military simply concluded that because some of the dead were armed military-age males, they were probably insurgents — even though at least some of those killed were apparently working for a pro-American private security firm. According to the AIHRC, among those killed were thirteen armed men “[who] were engaged in combat with the forces when they entered the village.” But it’s unclear how the military sorted out who was who. Said Human Rights Watch:

As the US investigation admits, several of the dead men were employed by the British company Armor Group, working as security guards for the US military in Shindand. Government officials and villagers have alleged that the operation was the result of misinformation resulting from tribal rivalries in the area. The Callan Summary dismisses these allegations, but military sources have told Human Rights Watch that they are credible.

The AIHRC’s conclusion:

AIHRC questions the way that international military authorities handled the aftermath of the incident — first denying any civilian casualties by forces involved, then admitting only 5 to 7 casualties, and then only 33 without releasing any further details of the investigation, despite repeated reports of as many as 90 casualties by various independent monitors and public bodies. This type of public foot-dragging in accepting responsibility increases the anger and resentment that Afghan communities feel. The failure to compensate or apologize to those civilians whose deaths were recognized only angers the community more.

II. Farah, 2009

The tiny village of Granai, deep in western Afghanistan, is far from the epicenter of the battle against the insurgency. Located in the Bala Baluk district of Farah Province, south of Herat, it’s the scene of a mass-casualty debacle that is referred to as the “Bala Baluk incident” and the “Granai incident.” On May 4, 2009 — eight months after the September 2, 2008, tactical directive aimed at restraining airstrikes that might kill civilians — somewhere between twenty-six and 140 Afghan civilians perished in the blink of an eye. Just after 8 pm, several waves of aircraft attacked three targets in Granai, including three F-18 fighter jets, which dropped a total of five laser-and satellite-guided munitions, and a B-1 bomber, which dropped “three 500-pound satellite-guided bombs on a tree grove, four 500-pound and 2,000-pound satellite-guided bombs on one building, and one 2,000-pound satellite-guided bomb on a second building,” according to the New York Times. Writing ten days after the attack, Carlotta Gall of the Times reported:

The bombs were so powerful that people were ripped to shreds. Survivors said they collected only pieces of bodies. Several villagers said that they could not distinguish all of the dead and that they never found some of their relatives….

“There was someone’s legs, someone’s shoulders, someone’s hands,” said Said Jamal, an old, white-bearded man with rheumy eyes, who lost two sons and a daughter. “The dead were so many.”

The carnage followed clashes in or near Granai between a contingent of some 400 Taliban fighters and Afghan forces. According to subsequent investigations, most if not all of the Taliban had left Granai when the bombs started falling, and those killed were villagers, including entire families, who’d taken cover in several buildings. An investigation by the Afghan Defense Ministry found that at least 140 people died in the attack, including ninety-three children. But two weeks after the bombing, a preliminary investigation by the US military concluded that perhaps twenty to thirty civilians had been killed, along with sixty to sixty-five “Taliban extremists,” according to the Times.

Human Rights Watch visited the area and put together a detailed timeline of the events, reporting: “Villagers and local officials have told Human Rights Watch that many villagers took shelter from the bombing in the houses of local religious and tribal leaders, including the homes of Sayyed Naim and Mualem Rahmadi. Dozens of civilians are reported to have been killed when each of these compounds was bombed.”

Human Rights Watch emphasized that US procedures, including those adopted after the Azizabad killings, were clearly inadequate and that “further reform is required.” The UN said the toll included at least sixty-four civilians, adding that while the US military “acknowledged that it had failed to comply with internal military guidelines, a proper assessment prior to the attack could have determined whether civilians would have been disproportionately harmed and whether it was appropriate to use air strikes in a residential area.”

Later, when reports surfaced that WikiLeaks had obtained damaging video of the Farah massacre and, to the consternation of the Pentagon, was planning to release it, Daniel Ellsberg, who leaked the Pentagon Papers record of the Vietnam War in 1971, called on President Obama to release the video officially:

I’d call for President Obama to post that videotape online. Let’s see whether it confirms what his officials…said about it earlier, or what the truth is. Has he seen it himself? He certainly should. He has access to it. And if he does, what excuse would he have for not revealing it? So why is he waiting for WikiLeaks to use its sources to decrypt that, when he can just easily release it, as he should have some time ago?

According to The Guardian, in 2009 General Petraeus admitted that the military had video of the Farah incident and said he would release it. But he didn’t.

Still, the increasing toll of civilian killings was putting pressure on the military to change its ways, especially in light of the emerging counterinsurgency doctrine that the new commander, General McChrystal, was putting in place. And the pressure was coming from the top. Two weeks after Farah, Adm. Mike Mullen, then chairman of the Joint Chiefs of Staff, told a gathering at the Brookings Institution: “We cannot succeed in Afghanistan — or anywhere else, but let’s talk specifically about Afghanistan — by killing Afghan civilians…. We can’t keep going through incidents like this and expect the strategy to work. [But] we can’t tie our troops’ hands behind their backs.”

That tension would increasingly vex US strategy. At its core, it underscored the enormous difficulty — perhaps the impossibility — of waging a humane counterinsurgency.

A month after the Farah massacre, a Pentagon investigation by Brig. Gen. Raymond Thomas showed that there were “problems” with the military’s actions that day. Geoff Morrell, a spokesman for the Defense Department, said: “There were some problems with tactics, techniques and procedures, the way in which close air support was supposed to have been executed in this case.” He said that a B-1 bomber pilot had erred: “That plane, because of how it takes its bombing routes, had to break away from positive identification of their target at one point to make its elongated approach,” and he called it “the fundamental complaint that was rendered I believe from this investigation.”

The Joint Civilian Casualty Study by Sewall and Lewis reports that General McChrystal issued another tactical directive to ISAF in July 2009, following the killings in Farah, that instructed the military to “look for all tactical alternatives, including withdrawal, when considering airstrikes on compounds,” adding that airstrikes against compounds would be permitted only in self-defense. After pointing out that “air incidents had produced the highest number of civilian casualties over the past few years, the JCCS says:

Units were directed to look for all tactical alternatives, including withdrawal, when considering airstrikes on compounds, and airstrikes were only to be considered on compounds in self-defense when forces were receiving effective fire and had no alternative to save the lives of Coalition forces. The tactical directive also directed that BDA [Battle Damage Assessment] would be done for all airstrikes and uses of indirect fire, in order to assess the effects on civilians from the engagement. This BDA was to be done for all engagements, not just engagements involving compounds.

Long afterward, in 2011, the US Joint Center for Operational Analysis used the slaughter in Farah as a prime example of the military’s continuing failure to adapt its policies and procedures under wartime conditions in a way designed to minimize or eliminate civilian casualties — particularly in regard to training and instruction before troops are deployed. Written by Larry Lewis and Lt. Col. Michael “Strike” Messer and based on more than 200 interviews in and out of Afghanistan, the document, called “Adaptive Learning for Afghanistan,” presents a withering criticism of that failure, though couched in military-speak:

In Afghanistan, lessons learned organizations did not appear to make a significant contribution to in-theater adaptation. The in-theater organizations frequently did not align their efforts or share lessons effectively. In addition, these organizations often lacked the necessary capability and capacity, in terms of resourcing and training, to quickly analyze information and get it back out to the force….

ISAF and the ISAF Joint Command (IJC) also had lessons learned organizations, but they were not resourced adequately with sufficient personnel, and the personnel typically had no experience with lessons or analysis….

There were multiple barriers to effective sharing of information, knowledge, and lessons.

The Farah incident was used by multiple training sources in scenario and vignette-based training…. Since the Farah incident in May 2009, the joint and Service communities have implemented many changes to integrate CIVCAS [civilian casualty] reduction and mitigation into the “road to deployment” for Afghanistan…. [But] in-theater lessons did not always inform longer-term…change.

Yet another study, released in June 2012 and called “Afghanistan Civilian Casualty Prevention,” seems more concerned about the propaganda importance of mass civilian casualties than it does about the casualties themselves: “The unfortunate circumstances of the Gerani [sic] Village incident are a reminder of the complexities of the strategic information battle.” In summarizing the events there, the study says that Taliban forces in the area that day “validated the lawful military nature of the air strikes,” but added, “The inability to discern the presence of civilians and assess the potential collateral damage of those strikes is inconsistent with the U.S. government’s objective of providing security and safety for the Afghan people.” And it insists that “USCENTCOM’s investigation report estimates that at least 78 Taliban fighters were killed,” adding, “No one will ever be able conclusively to determine the number of CIVCAS that occurred on 4 May 2009. The USCENTCOM investigation does not discount the possibility that more than 26 civilians were killed in this engagement.”

III. Kunduz, 2009

In the 2010 Joint Civilian Casualty Study by Sewall and Lewis, a third mass-casualty incident, the bombing of two fuel tankers in the Ali Abad district of Kunduz Province in September 2009, is cited as having provoked yet another series of US military directives, guidances and FRAGOs. That incident, in which as many as ninety-one people — including many children — died, revealed a fatal flaw in the military’s directives to avoid bombing Afghan buildings where civilians might be present; in this case, the target was a pair of tanker trucks that had been surrounded by a crowd of civilians seeking fuel. As the JCCS succinctly notes:

The majority of targets in CIVCAS incidents after mid-2009 were typically individuals and vehicles, not compounds. However, a few incidents demonstrated that many people can be killed even when the target does not involve a compound. For example, in the Kunduz incident in September 2009, masses of people gathered around tanker trucks taking fuel, and the subsequent engagement resulted in about 30 civilians killed.

Yet according to the JCCS, it took the military until August 1, 2010, to issue another tactical directive to address the issue of airstrikes directed against other targets, in addition to buildings and residential compounds, that could result in mass casualties.

The Kunduz attack unfolded at about 2 am on September 4. From the scene, the New York Times reported two days later that the tankers had stalled after being hijacked by insurgents, who then invited local villagers to come siphon the fuel. The tankers had entered the country from the north, via Tajikistan. Said the Times:

Some people wounded by the strike said that they had gone to the scene with jerrycans after other people had run through their villages saying that free fuel was available. “They were just telling us, ‘Come and get the fuel,'” Wazir Gul, a 23-year-old farmer, said at the hospital, where he was treated for serious burns on his back. He estimated that hundreds of people from surrounding villages went to siphon fuel from the trucks before the airstrike. Mr. Gul said his older brother Amir was among the villagers incinerated in the blast. “When the tanker exploded and burned, I knew he was dead,” Mr. Gul said.

The strike, carried out by American F-15E fighters, was ordered by German officers, whose 4,000 troops were responsible for the area. A 500-pound bomb struck each tanker. The German weekly Der Spiegel, which concluded that the airstrike was a war crime, quoted an eyewitness:

At first, there was a loud droning, like what you hear when a generator short-circuits. Then there was a bright flash. I just let myself fall forward and went down underwater. Even from there, I could feel the shock wave. For a few seconds, it was as bright as day. Even the water was heating up. When I came out of the water, the whole area around the tanker trucks was on fire. It looked like the ground was spitting up fire, though it was just the fuel from the trucks. It was unbearably hot. There were bodies lying everywhere; they were completely carbonized…. I believe there were about 120 there before the bombing; only a handful survived.

The UN’s civilian protection unit in Afghanistan reported:

It is not disputed that some Taliban were at the site but it should have been apparent that many civilians were also in the vicinity of the trucks. According to UNAMA HR’s investigations, 74 civilians, including many children, were killed. Despite several requests, by UNAMA HR, to the Civilian Casualty Tracking Cell, ISAF did not release the unclassified version of its report nor show video footage as requested. As a result of the air strike, several high ranking German officials resigned after it came to light that they had withheld information that civilians had been killed and injured.

Initially, German military officials claimed that no civilians had been killed in the bombing, and that aerial surveillance had detected no civilians present. That, of course, turned out to be false. Eventually, horrific video of the bombing was released by the German newspaper Bild. Both the German Army chief of staff and the defense minister resigned over the Kunduz strike and subsequent cover-up.

General McChrystal, recognizing that the fallout from yet another civilian mass-casualty incident could be devastating, personally visited the scene of the atrocity, wading through the river to inspect the fire-blackened tankers as a NATO investigative team flew over the scene. He also visited a local hospital to comfort the wounded.

IV. Conclusion

Questions swirl around the US-led counterinsurgency effort and the series of tactical directives and other orders designed to protect civilians. The Joint Civilian Casualty Study, written to shine a spotlight on civilian protection issues, has at least started the Pentagon thinking along these lines. Sarah Sewall, one of the authors, said in an interview: “It’s safe to say that the JCCS was the basis for a number of changes that the army has made to its doctrine and training.”

Once the military starts down that path, she says, it tends to put more and more effort into it. “One of the things about the military is, once they set about trying to do X, then they are constantly trying to do X better. And once they’ve decided to understand the civilian security piece as part of their assessment method, that became something that improved.”

But Sewall raises the $64,000 question: As the Afghan War winds down and the military inevitably shifts its war-fighting strategy away from counterinsurgency, is the Pentagon going to remember the lessons learned in Afghanistan? As she puts it, “I think the right question to be asking from the outside is, [given] the current focus on the return to full-spectrum combat operations and the real desire on the part of everybody to put these long, extended counterinsurgencies behind them, whether that might accidentally mean that people [won’t be] paying enough attention in future engagements to these issues, because they mistakenly think that they’re COIN-specific, as opposed to having broad applicability to the use of violence.”

Part of the problem is that, as the Joint Civilian Casualty Study says, “No organization or person within DOD [Department of Defense] is responsible for assessing the issue comprehensively.” Despite its importance, it’s just not a priority for the bureaucracy.

Joseph Felter, a retired colonel who served as a senior aide to both McChrystal and Petraeus, says that all the directives issued between 2007 and 2011 left “some gray area,” adding, “Life doesn’t lend itself to black-and-white: ‘Here’s a checklist. I can shoot back; OK, I can’t.’ There are a lot of subjective assessments that have to be made by a 20-year-old squad leader under fire.” Felter, who co-authored a study by the National Bureau of Economic Research on how civilian casualties make an insurgency worse [see Dreyfuss, “Creating Insurgents,”], points out that there was “some variance in how commanders would interpret directives from Command ISAF. I was disappointed in what I saw. You’d see the candid-comments reports: ‘Hey, we’re getting our hands tied’ or ‘Our people are put under great risk now.'” He stresses that, whatever the directives say, it’s up to the commanders all the way down the line to make it clear to the privates in the field what the policy is. Felter added:

At the private level, this is a leadership issue. If your platoon leader or company commander or battalion commander isn’t educating their subordinates fully on COMISAF’s intent, there’s a leadership breakdown further down the chain. But 19-year-olds will be 19-year-olds, and I can empathize with them too. They see a very narrow view of the battlefield, and they’re the ones on the pointy end of the spear, that see the extreme downside. It’s their friends who are getting killed, up close and personal, and to think strategically is sometimes a bigger challenge for a private, but I can absolutely say that reducing CIVCAS was the only way to make progress.

According to the JCCS, the first tactical directive regarding civilian casualties in Afghanistan was issued in 2007 by Gen. Dan McNeill, the ISAF commander at the time, before the three mass-casualty incidents described above and before the counterinsurgency doctrine took hold in Afghanistan. Still, that directive said, “Whenever our actions in battle cause injury or death to civilians or property damage or destruction, we diminish our effectiveness.” The JCCS added:

The 2007 Tactical Directive focused on three areas: raids, pre-assault or preparatory fires, and air-to-ground or indirect fires. Regarding raids, forces were to be instructed to not go into homes uninvited and to use soft-knock techniques when possible. Exceptions were to be reported to ISAF headquarters detailing why the exception occurred. Also, pre-assault or preparatory fires were to be treated as deliberate targeting operations which required preapproval and formal CDE [collateral damage estimates]. Finally, air-to-ground or indirect fires were to be used only when forces were taking fire from the compound or there was an imminent threat from the compound, and when there were no other options available to the ground force commander to protect the force and accomplish the mission. The Tactical Directive specifically called for use of small arms instead of air-to-ground or indirect fire when tactically feasible.

In 2008, after the Azizabad incident, another tactical directive was issued by General McKiernan. Says the JCCS:

The content of this Tactical Directive was similar to the 2007 Directive, focusing on raids, use of air-to-ground and indirect fire, escalation of force, and being first with the truth. The section on raids was similar to the 2007 Directive with the addition of a section on religious and sensitive sites and the need for respectful behavior towards Afghans. The section on air-to-ground and indirect fire was similar but the new Directive called for forces to use these fires only when there was no other option to protect the force, removing the option to use that type of force in order to accomplish the mission. The Tactical Directive emphasized the need to reduce civilian casualties during escalation of force procedures although it did not change those procedures. Finally, the Directive called for acknowledgement of civilian casualties or property damage at all levels, from community level to national level.

In July 2009, after the Farah incident, another tactical directive by General McChrystal, issued soon after he arrived in Afghanistan, sought to renew ISAF’s focus on civilian casualties. In particular, as the JCCS notes, McChrystal emphasized the restriction on bombing Afghan buildings and compounds where civilians might be present.

The JCCS outlines a series of other, later directives and guidances dealing with warning shots, road encounters while driving, night raids and so on. But it also points out that “the evolution of written directives, FRAGOs and other documents relating to the use of force suggests that guidance has essentially ‘chased’ lessons from specific CIVCAS incidents, often following the occurrence of events with large numbers of casualties and/or with high media visibility.”

Still, lessons were learned. As a platoon sergeant quoted in the Joint Civilian Casualty Study says, “You need to determine if the greater good is to engage [i.e., fire] or if the greater good is not to engage.” Those are wise words, and ones that could be used not just for a single encounter in the field, but for the entire long war in Afghanistan.

Click here to view the interactive database of America’s Afghan Victims.

This article was reported in partnership with The Investigative Fund at The Nation Institute, now known as Type Investigations.