An important function of the 2015 DoD Law of War Manual has been to address longstanding confusion concerning the U.S. relationship to the 1977 Additional Protocol I to the Geneva Conventions of 1949 (AP I). The U.S. is not a party to AP I and therefore is not bound to its rules on the conduct of hostilities as a matter of treaty law. But it had long been clear the U.S. regards much of AP I to reflect customary international law binding on parties and non-parties to AP I alike. The devil, however, had been in the details. Aside from very general observations offered by sporadic statements, the U.S. view on many important AP I rules had for decades been notoriously difficult to discern. Thus, the Manual represents a significant effort to fill persistent gaps in the U.S. view on AP I that have dogged law-of-war practitioners for years.
A second, related function of the Manual has been to present a vision of the general, customary law of war that exists outside AP I, especially that applicable to targeting operations. While an overwhelming majority of States has ratified AP I, a still-significant number of States has not. Many of these non-party States are especially significant in their military capabilities and combat experience. While there are, doubtless, independent views among non-AP I parties as to what the law of targeting outside AP I looks like, U.S. views on the subject, when they have been voiced, have been especially influential.
One of the leading respects in which the U.S. has maintained that AP I does not reflect customary international law has been its view on AP I, Article 57(3). That article provides:
‘When a choice is possible between several military objectives for obtaining a similar military advantage, the objective to be selected shall be that the attack on which may be expected to cause the least danger to civilian lives and to civilian objects.’
Consistent with decades of U.S. practice, the original 2015 edition of the Manual reproduced Article 57(3) and simply observed:
‘5.11.5 AP I Provision on Choice Possible Between Several Military Objectives. . . . The United States has expressed the view that this rule is not a requirement of customary international law.’
A footnote to this provision reinforced the view, citing a 1991 U.S. comment to an International Committee of the Red Cross (ICRC) memorandum on the law of war in the First Gulf War. The U.S. comment rejected the customary status of Article 57(3) and further indicated the provision is not always mandatory, as some sources allege. Thus, even if Article 57(3) reflected custom, the case might be made that the U.S. has achieved persistent objector status to it. And, for what it’s worth, the Rome Statute of the International Criminal Court does not include violation of AP I, Article 57(3) in the subject matter jurisdiction of the court.
But in a subtle yet surprising new section, yesterday’s update of the Manual clarifies the original Manual‘s guidance on Article 57(3) with the following:
5.11.7 Selecting Military Objectives. ‘. . . [W]hen attempting to achieve a particular military advantage through an attack, a commander may confront a choice among several military objectives for achieving that advantage. When facing such a choice, provided that all other factors are equal, the object to be selected for attack shall be the object the attack on which may be expected to cause the least danger to civilian lives and to civilian objects.’
The new passage’s similarity to Article 57(3) is striking. Like AP I, the updated provision assumes that equal military advantage can occasionally be achieved by attacking different or various military objectives. Also like AP I, the new provision anticipates a commander’s duty to identify and analyze multiple options with respect to objective choice. And like AP I, the updated section employs the imperative “shall” regarding selection of the objective that will result in least danger to civilians. Most importantly, the passage resembles Article 57(3) of AP I in requiring that selection of a military objective result not merely in acceptable or proportionate danger to civilians, but rather that selection of a military objective must result in the least danger to civilians.
So, has the U.S. now abandoned its assiduous rejection of Article 57(3)?
I confess that on my initial reading of the updated passage cited above, I worried the answer was yes. In fact, the immediately succeeding passage of the new provision reinforced my impression. It provides,
‘For example, in seeking to deny an adversary the ability to use a railroad network, it may be possible to disable the railroad network just as effectively by striking the railroad lines away from inhabited areas as by striking the railroad station located near civilians.’
Military lawyers and law-of-war specialists will no doubt recognize the example. It is widely used in law-of-war instruction. In scenario-based instruction, the rail system is often graphically depicted as presenting a central hub or rail station surrounded somewhat densely by civilian objects. From the station, spokes of rails emanate outward with comparatively fewer civilians and civilian objects along the rail spokes. Students are asked whether a commander may lawfully attack the hub or whether he must attack the various rail lines.
U.S. instruction on this scenario frequently advised that AP I, Article 57(3) would likely require the commander to attack the several spokes of the rail system because their destruction would produce less civilian damage than an attack on the hub of the system near which civilians and civilian objects were located. However, under the U.S. view, the hub might remain a lawful target so long as the hub or station itself was a military objective, the attack was discriminate, was not calculated to cause unnecessary suffering to combatants, and the incidental damage to civilians and civilian objects was not excessive in relation to the expected military advantage. In other words, under the U.S. view, if an attack on the station or hub of the rail system was independently lawful under law of war targeting principles, the availability of a “more lawful” option—that is, an attack option that resulted in less collateral damage—did not render the attack on the hub unlawful. Moreover, under the U.S. view, the commander and his staff needed to satisfy themselves only that the attack on the hub was necessary, discriminate and did not result in damage to civilians that was excessive in relation to the military advantage of destroying the rail system. The commander and staff were not required under the law of war to investigate, recon, analyze, consider, or adopt other military objective options, although policy or other non-legal considerations might often commend such efforts.
The railway scenario has also presented the opportunity for military lawyers to examine whether the hub and spoke attack options are truly comparable in terms of military advantage. At first blush, each attack option disables the rail network thus producing the same military advantage to the attacker, namely enemy loss of an important mode of military transport. However, on closer examination, if the spoke option requires more air sorties, more ammunition or ordnance, involves more exposure to enemy fire, requires more reconnaissance, results in less thorough destruction, or presents greater risk to friendly forces than selection of the hub as a military objective, then one might question whether the scenario indeed presents a choice between truly equal options. In such a case, both an AP I party and a non-party might conclude the rule concerning objective selection as an attack precaution is inapplicable. In this respect, a critique might conclude that Article 57(3) imposes significant collection and analytical duties only to address circumstances that would rarely, if ever, arise in combat.
To its credit, the remainder of the new provision 5.11.7 attempts to state something resembling this point. It provides,
‘When the choice of military objectives involves different risks and benefits potentially yielding different military advantages, this rule does not require that the object that may be expected to cause the least danger to civilian lives and to civilian objects be chosen for attack.’
Moreover, at provision 184.108.40.206, the updated Manual reiterates that AP I, Article 57(3) still does not necessarily reflect customary international law. To further confirm this understanding, in a conversation attributable to Department of Defense officials, an official verified that provision 5.11.7 does not reflect a change in the Department’s understanding of the law of war but rather a clarification.
The most important point to appreciate about the clarification may be that while AP I, Article 57(3) operates in cases where various attack objectives offer “similar military advantage” (emphasis added), the U.S. precaution respecting selection of military objectives operates only in cases where various attack objectives offer “equal” military advantage (emphasis added). Thus, I take the update at its word, meaning that unless “all other factors are equal” (emphasis added) between the military advantage offered by various military objectives—including risk to friendly forces, destructive effect, opportunity costs, and expenditure of combat power—the least civilian damage rule does not operate.
Still, it is possible that these subtle nuances will be lost on many readers. Worse, it is possible that jurists sympathetic to Article 57(3) as an expression of custom will seize on provision 5.11.7 as a U.S. capitulation. The striking similarities between early passages of the clarification and Article 57(3) certainly present the opportunity for mistakes or even selective citation and misrepresentation. It seems the original Manual‘s treatment of Article 57(3) presented far less risk in this respect than the updated provision. Understanding the Manual‘s function of clarifying U.S. views on its relationship to AP I, one is left to wonder whether the U.S. view on AP I, Article 57(3) really required clarification. Did it?
Additionally, the Manual‘s original stance on Article 57(3) seemed to greatly reduce the intelligence collection duties and staff work required of attackers. That is, so long as reasonable collection and analysis indicated attack on a military objective would comply with the law of war, no further work or speculation concerning alternative objectives was required. The clarification seems to impose an additional requirement of ensuring that no other potential objectives offer the same military advantage as the objective originally considered for attack. Although the additional burden might be supportable as a matter of humanitarian logic, in a complex battlespace the burden of considering and exhausting endless operational alternatives to ensure compliance with the least civilian damage rule could become quite onerous. Worse, requiring exhaustive consideration of alternative objectives without an identifiable limit could thrust commanders and their subordinates into a minefield of contemplating war crimes or UCMJ litigation.
Ultimately, the prospects of U.S. ratification of AP I are extraordinarily dim. In its waning days, the present U.S. administration appears satisfied with offering policy-based and minor legal clarifications to its understandings of the law of war such as Executive Order 13732 and the White House Report on National Security Legal Frameworks. It is my view the U.S. should continue this course. The U.S. should not ratify AP I and should actively contest many of its provisions’ status as custom. AP I strikes an unsustainable balance between military necessity and humanity. To prevail in combat, even present AP I State parties may be tempted to discount some of its restraints, to contort the plain meaning of its provisions, or even to surreptitiously leave their “dirty work” to an allied non-AP I party. The good news is that the Manual‘s update does not abandon the U.S. longstanding objection to AP I, Article 57(3). However, whether all readers will fully appreciate or honor the update’s clarification as such remains to be seen.