In a recent speech at NYU, Department of Defense General Counsel Jennifer O’Connor suggested that after an impending revision, the US Law of War Manual would reflect the position that the principle of proportionality applied to “civilians who are proximate to military operations, such as workers in a factory that produces military weapons”. This would correct one of the most glaring deviations of the Manual from established legal practice. Yet, O’Connor qualified this encouraging message with a crucial caveat for the protection of persons in a group reminiscent of the pre-Geneva Conventions category of “quasi-combatant”: “[T]heir choice to support military operations in or on a military objective may be weighed by commanders as a factor in the proportionality analysis.”
As Marty Lederman, Oona Hathaway, and Mike Schmitt noted in their recent Just Security post, neither customary law nor the relevant provisions in API support such a “sliding scale” of civilian protection. No appeal to moral principle, no plea to take account of the humanitarian purpose of Article 57 API is necessary to show that the Protocol has once and for all laid to rest the idea of the “quasi-combatant”. The dichotomy between civilians and combatants as exhaustive and mutually exclusive categories is the core structuring principle of contemporary IHL. If the location or “general use” of a person were legally relevant, the drafters might have chosen a formula for the definition of a combatant echoing that of a military objective. Instead, distinction among persons is based on a priori status. The text is as determinate as law gets when stipulating that a civilian is immune from attack for as long as she does not directly participate in hostilities. In turn, a civilian either counts as one civilian in a proportionality calculus or she is herself a legitimate target of intentional attack.
If IHL leaves no room for a “quasi-combatant” or a “lighter than usual civilian,” why do such eminent scholars as Bothe, Partsch and Solf defend the legal salience of a person, who is not a combatant for the purposes of distinction, but not a civilian for the purposes of proportionality? In their influential Commentary on API, Bothe, Partsch and Solf explain their startling endorsement of such civilians’ weighing “less heavily” in the application of the rule of proportionality with the observation that “being within, or in the vicinity of a military objective, these civilians assume the risk of collateral injury from the effects of attack.” The same reference to risk-taking is at the heart of the never adopted ICRC instrument the Manual currently quotes as authority to justify its exclusion of certain groups of civilians from the proportionality calculus. It observes that “should members of the civilian population … be within or in close proximity to a military objective, they must accept the risks resulting from an attack directed against that objective.”
As a matter of fact, it barely warrants mention that a civilian working in a munitions factory assumes a higher risk of incidental harm than one working in a bakery. Neither of the quoted authorities, however, shows how this is relevant as a matter of law. No amount of indifference to risk magically constitutes a person as a combatant. Neither is direct participation in hostilities a type of risk-taking or an attitude towards incidental harm. The required “belligerent nexus” is more than a mental state – it must be invested into an action for IHL to even register it. While one could argue that being in or on a military objective is an action, the ICRC’s Interpretive Guidance also demands that the action reveals a specific motive as part of the required mental state: namely the design “to directly cause the required threshold of harm”. Even if working in a military objective allowed for the inference that the civilian consciously chose an increased risk of harm, neither that action (working in a military objective) nor the presumed attending mental state (reconciling herself with heightened risk) would at all be indicative of her motives. Working in a munitions factory is not only “not direct participation in hostilities”; it is not a lesser version of it either. How then could it affect the level of protection a civilian is legally owed?
But is there really no difference between the munitions worker and the baker dying as a foreseeable side-effect of an attack on a (nearby) military objective? It is a powerful intuition that what happens to civilians in war should at least to some extent be a result of their own choices. Crucially, this is an intuition about what the law should be, not about what it is. As an intuition about what the law should be, it is deeply misguided. On some accounts, all things being equal, killing an innocent bystander who assumes no risk of harm is worse than killing one who consciously put herself in a position where such an outcome was likely. But all else is not equal! The moral status of civilians and the imprint that their deaths leave in the moral universe depend on their reasons for acting, their motives, their justifications and excuses. As a result, reducing the protection of all civilians who choose to be near or on a military objective does not necessarily mean legally permissible harm is more closely aligned with morally justified harm.
Of course, the DoD’s rationale for clinging to the existence of a “quasi-combatant” or “almost-civilian” may not be to bring IHL closer in line with moral principles. But it is the notion that civilians’ moral status and hence their protection-worthiness diverge that lends this differentiation a false moral appeal. After all, IHL’s division of persons into combatants, directly participating civilians and immune civilians cannot hope to properly align the protection of individuals with their moral status or in accordance with the relative moral evil of harming them. It is worth bearing in mind though that IHL draws these bright and imperfect lines for good moral reasons. Nothing less is at stake than its very ability to make a difference. As Lederman, Hathaway, and Schmitt point out; a sliding scale of civilian protection bears the risk of making IHL too hard to apply.
The DoD’s particular approach to the “less than proper civilian” might have consequences even worse than IHL missing the chance of guiding the conduct of a conscientious attacker who does not know how to operationalize an overly complicated law. O’Connor indicates that a commander “may” (or may not?) “weigh the choices” of the civilians on the other side. This amounts to suggesting that the allocation of legal protections should follow a commander’s sitting in judgement of his enemy’s choices. Is a fair and measured assessment really a reasonable expectation in the uniquely stressful situation of military combat? Rather than better tracking moral principle, IHL could end-up licensing additional unjustified violence, if it made the protection of civilians conditional on their moral status as perceived by the attacking commander.
The double detriment of hinging legal protections and permissions in war on unverifiable and necessarily often mistaken assumptions about civilians’ moral status as supposedly evidenced in what we frame as their choices should be kept in mind when we revisit the thorny issue of human shields. O’Connor again encouragingly suggests that the revised Manual “will provide greater clarity on the DoD’s legal view of human shields, noting that whether voluntary or involuntary, civilian “human shields” would not be considered to be directly participating in hostilities and would not thereby lose their protections from attack.” But does a “voluntary human shield” count as one “full” civilian? Elsewhere O’Connor refers to voluntary human shields as civilians who “willfully place themselves on or near military objectives,” which arguably puts them in the category of civilians who make a “choice to support military operations in or on a military objective” though the speech is not clear whether this is the intended implication. If it is, then commanders would be able to discount the lives of voluntary human shields in the proportionality analysis, it seems, according to the DoD’s analysis.
The very differentiation between voluntary and involuntary human shields, found nowhere in contemporary treaty law, arguably draws its enduring appeal from the same misguided intuition about what IHL should be: a means to allocate harm and violence in war in accordance with civilians’ individual moral status. But a law that distinguished between voluntary and involuntary human shields would not necessarily be fairer or adequately track morally relevant categories if we simply reduced the protection of all voluntary human shields. Consider the following types of voluntary human shields: Voluntary shield A receives a warning that his house will be bombed because a belligerent has stored rockets in the basement. He does not leave the house in the hope that he can prevent the destruction of his life’s work and the resulting homelessness of his family. He invites a family friend (voluntary shield B) to come over to bring up the numbers of civilians in the house. Voluntary shield C hears of an impending attack against a house used as an IED workshop on the outskirts of his town. In an effort to help his troops he drives there and climbs on the roof. All three shields intend to prevent an attack on a legitimate military objective, but their actions are morally different.
For someone to count as a voluntary human shield according to O’Connor’s definition, their primary motive does not even have to be shielding an object from attack. Indeed, the term is often used simply for civilians who through their non-coerced conduct knowingly make attacking a military objective less likely. Shield D does not intend to shield his house from attack, but he does not want to subject an elderly relative to the stress of displacement. He in effect shields his house but this is a side-effect of his intention to protect vulnerable members of his family. Yet, he looks exactly like shield A, who intends to prevent an attack on his house and not all that much different from Shield B, who in turn looks a lot like Shield C. If law tried to track these morally relevant categories by differentiating among voluntary human shields, it would be too complicated to be useful to the conscientious attacker. It would also likely lead to a general erosion of civilian protection as intent is only sometimes, motive very rarely and justification almost never visible in conduct.
Complicating, indeed just changing, IHL is always a gamble with its action guidance and compliance pull. Some changes may be necessary to allow law to adjust to the evolving character of war. Some complications may be worth it to reduce the attenuation between legally permissible and morally justified harm in war. The resurrection of the long dead quasi-combatant and the differentiation between voluntary and involuntary human shields are neither necessary nor worth it. While there is often a trade-off between IHL’s practical efficacy and moral appeal, inviting commanders to differentiate among civilians’ protection-worthiness based on evaluations of their perceived choices, in light of poorly understood and almost certainly unverifiable moral criteria, has the rare distinction of making IHL less useful in both military and moral terms.