Over the past week, many thoughtful posts have appeared, here and elsewhere, reacting to the US military’s report on the 2015 airstrike of a Médicins Sans Frontières (MSF) hospital in Kunduz, Afghanistan by US forces. The attack killed 42 civilians and wounded dozens of others. Many have criticized the military’s conclusion that no war crimes were committed. I share their concerns, for reasons that I will discuss later.

First, I want to underscore the military’s finding that US forces violated the law of armed conflict. It is no small thing for a state to admit that its armed forces unlawfully attacked a hospital and unlawfully killed 42 people. This candid admission should be welcomed, even by those who believe that the attack was not merely unlawful but also criminal.

As Just Security readers know, I am very concerned that the US Defense Department’s recently released Law of War Manual misrepresents the law of armed conflict in several important ways, all of which favor attackers and endanger civilians (for a more systematic treatment, see here). I am relieved to say that, as I read it, the military report seems to follow the law of armed conflict rather than the DOD manual. 

For example, the DOD manual rejects the provisions of Additional Protocol I to the Geneva Conventions which require attackers to presume that persons and objects are civilian rather than military in cases of doubt. According to the manual,

Under customary international law, no legal presumption of civilian status exists for persons or objects, nor is there any rule inhibiting commanders or other military personnel from acting based on the information available to him or her in doubtful cases.

Importantly, since the United States is not a party to Protocol I, it is bound only by customary international law. According to the manual, customary law only prohibits attacks “based on merely hypothetical or speculative considerations regarding their possible current status as a military objective.” Were this low standard applied to the attack in Kunduz, it seems entirely possible that the intelligence relayed to the crew, though clearly unreliable, would be found neither “hypothetical” nor “speculative.”

In contrast, the report states that

The GFC [ground forces commander] and the Aircraft Commander failed to identify the MSF Trauma Center as a lawful target. Therefore, it should have been presumed to be a civilian compound. …

The commander’s intent coupled with the defense of others ROE [rules of engagement] should have created doubt in their mind that the target was a lawful target. (pp. 59, 61)

It seems that, contrary to the manual, the report accepts that attackers should presume that persons and objects are civilian in case of doubt.

Importantly, under Protocol I, attackers must “do everything feasible to verify that the objectives to be attacked are neither civilians nor civilian objects and are not subject to special protection but are military objectives … .” The International Committee of the Red Cross (ICRC) concluded that this provision reflects customary law as well.

Strangely, the 1,200 page DOD manual does not quote this provision even once, neither to accept the provision as reflecting customary law, nor to reject the provision as reflecting only treaty law. Instead, the manual asserts that

In assessing whether a person or object that normally does not have any military purpose or use is a military objective, commanders and other decision-makers must make the decision in good faith based on the information available to them in light of the circumstances ruling at the time.

While Protocol I seems both active (“do everything feasible”) and objective (“to verify”), the manual seems both passive (“information available to them”) and subjective (“deci[de] in good faith”). Whether the DOD manual accepts or rejects the Protocol I standard, it should say so clearly.

By way of contrast, the US Army’s 1956 Manual on the Law of Land Warfare provides that attackers “must take all reasonable steps to ensure … that the objectives are identified as military.” As we have seen, the standard adopted in the DOD manual contains no such active and objective language.

In any event, the report does not claim that US forces did not act in good faith, but instead that their “actions were not reasonable under the circumstances” (p. 60). Indeed, the law of armed conflict requires not only subjective good faith but also objective reasonableness.

In addition, the report does not claim that US forces did not base their decision on the information available to them, but instead that “[t]here were no legitimate circumstances requiring the crew members to make decisions to engage without clarifying or requesting more information” (p. 59).

This was not a time-sensitive target. The aircraft knew the location of the [convoy] and could assess that there was no threat posed against the convoy. The aircraft had time to confirm the target through multiple commands. The aircrew had time to execute the deliberate targeting process prior to engagement. The aircraft was not low on fuel, as it had recently refueled prior to the engagement. (p. 61)

According to the report, US forces were required either to seek additional information, sufficient to confirm that their target was a lawful target, or to refrain from attack given the strong reasons for doubt. Their failure to do so was unlawful, irrespective of their presumed good faith.

How much doubt regarding the legal status of a target is sufficient to preclude lawful attack? When has an attacker done everything feasible to verify that a target is military and not civilian? Although experts formulate the relevant standards in somewhat different ways (see here, here, and here), I know of no expert who would consider the Kunduz attack lawful.

My own formulation is that attackers must have decisive evidence that a target is military and not civilian and, in addition, the risk of mistakenly attacking civilians must be outweighed by the risks of restraint (to attacking forces or to other civilians). In Kunduz, US forces had very weak evidence that the building and the people inside were lawful targets. As the report concludes,

The aircraft observed nine people behaving normally. They were told that, “all PAX are hostile.” This was an unreasonable reliance on the GFC’s assessment, to the exclusion of all other readily apparent information … . (p. 61)

The weakness of the evidence alone should have precluded attack. In addition, the risks of restraint were very low, since there was no immediate threat and adequate time to investigate further, while the risk of mistakenly attacking civilians was very great. Accordingly, the attack was doubly unreasonable.

In my view, the report reflects a better understanding of the law of armed conflict than one finds in the DOD manual. Let us hope that the military’s understanding prevails, and guides US targeting practice in the future. The lives of civilians may one day depend on it.

*          *          *

Turning now to the issue of accountability, I join many others in rejecting the Army’s conclusion that no war crime was committed simply because “[t]he label ‘war crimes’ is typically reserved for intentional acts — intentionally targeting civilians or intentionally targeting protected objects” and “none of the personnel knew that they were striking a medical facility.”

It is not entirely clear whether the Army meant that no war crime was committed as a matter of international law or that there was no legal basis for prosecution under US military law. I’ll briefly address both possibilities.

With respect to customary international law, the International Criminal Tribunal for the former Yugoslavia (ICTY) has repeatedly held that it is a war crime to ‘willfully’ attack civilians, where “the notion of ‘wilfully’ incorporates the concept of recklessness, whilst excluding mere negligence.” Accordingly, “[t]he perpetrator who recklessly attacks civilians acts ‘wilfully’.” (See also here and here.) If the Army rejects the ICTY’s understanding of customary international law then it should explain why.

As for domestic law, the United States has court-martialed US soldiers for recklessly killing civilians. Under the Uniform Code of Military Justice (UCMJ), it is murder to kill another person through an inherently dangerous act that shows wanton disregard of human life. In addition, under the UCMJ it is manslaughter to kill another person through a negligent act accompanied by a culpable disregard for the foreseeable consequences to others. The report does not explain the decision not to pursue criminal charges under these provisions of the UCMJ.

In his recent post, Alex Whiting correctly notes that the Rome Statute of the International Criminal Court (ICC) grants the ICC subject matter jurisdiction over attacks that are intentionally directed against civilians or civilian objects, but not over attacks that are recklessly directed against civilians or civilian objects. Whiting is careful to say that “[t]he Rome Statute does not necessarily limit the scope of customary international law.” Unfortunately, based on some conversations that I have had elsewhere, I fear that Whiting’s post may have inadvertently left the impression on some readers that the Rome Statute casts doubt on the ICTY’s rulings, or perhaps reflects a narrowing of customary international law. Let me suggest that is not the case.

For one thing, article 10 of the Rome Statute explicitly disavows any such purpose or effect, stating that “[n]othing in this Part shall be interpreted as limiting or prejudicing in any way existing or developing rules of international law for purposes other than this Statute.” Moreover, the ICTY’s rulings that recklessly attacking civilians is a war crime under customary law were all handed down years after the Rome Statute was adopted and entered into force. Evidently, the judges of the ICTY did not think that the Rome Statute codified or crystallized customary law on this point. Moreover, if the 124 state parties to the Rome Statute believed that they had recently codified or crystallized customary law on this point — an exceptionally important point, one would think — then one would expect many of them to forcefully object to the ICTY’s repeated rulings to the contrary (they have not).

In my view, the ICTY’s rulings should stand or fall on their own merits. However, the Rome Statute casts no doubt upon them.

Whiting also asks whether the crew’s commander could be criminally liable for recklessly ordering the attack. That is an interesting question, but it is not the only one worth asking. After all, under international law, ordering is just one mode of liability among others. For example, military commanders may be criminally responsible for recklessly (indeed, even negligently) failing to prevent war crimes committed by their forces. Accordingly, a military commander who knew or should have known that her forces were about to recklessly attack civilians and failed to prevent them from doing so may be criminally responsible for that attack as well.

Finally, Whiting asks whether the commander and crew were in fact reckless, that is, whether they were aware of a substantial risk that the building and its occupants were civilian rather than military. Given that members of the crew voiced their own doubts about the status of their targets, it seems hard to deny that they were aware of a substantial risk that their target was civilian rather than military. And it seems hard to deny that the commander should have known that the crew was about to launch a reckless attack.

As far as I can tell, there is only one reason to doubt that the commander and crew acted recklessly. On one view, criminal recklessness requires only conscious disregard of a substantial risk that a reasonable person (in this context, a reasonable attacker) would not have taken. However, on an alternative view, criminal recklessness also requires an attitude of culpable indifference toward whether the risk materializes or not. The latter view has some basis in the civil law concept of dolus eventualis as well as in the common law notion that reckless murder must show ‘wanton disregard’ of human life. As we have seen, the UCMJ follows the common law approach. Both views have some basis in the case law of the ad hoc tribunals although, as far as I am aware, neither has definitively prevailed over the other.

Importantly, it is possible that the commander and crew would be found reckless under the former view but would not be found reckless under the latter view. If the report explicitly endorsed the latter view — and made specific findings regarding the relevant attitudes of the commander and crew — then the ultimate decision not to pursue criminal charges might have some basis in law and fact. Unfortunately, by resting its decision on the dubious ground that recklessly attacking civilians is not a war crime at all, the report avoids the legally and ethically relevant issue.

Are the commander and the crew culpable wrongdoers deserving of moral blame and legal punishment? Or are they merely unfit soldiers deserving of demotion, reassignment, or discharge? In my view, these are the questions that the report should have squarely addressed, openly discussed, and fairly resolved. Its failure to do so is a disservice to all involved, most of all to the surviving victims and to the families of those killed.