Thanks very much to John Merriam for his very thoughtful and insightful post responding to my concerns about the Law of War Manual‘s treatment of how the principle of proportionality applies to persons who are knowingly in the vicinity of lawful military targets. John is absolutely right that I should have done more to distinguish the cases of protected military personnel, including military medical and religious personnel, from cases involving protected civilians in the vicinity of lawful targets. (I didn’t actually argue that the former must be considered in a proportionality analysis; but I also didn’t explain that customary law might treat them differently from civilians, either.)
From the little I know about this particular topic, the customary law of proportionality (and AP I, too, for that matter) has rarely considered whether and how military personnel who are protected — in the sense that they may not be targeted — must be accounted for in any proportionality analysis. (This category includes not only military religious and medical personnel, but also, e.g., wounded and sick forces, and perhaps some parlementaires, too.) I suspect that custom and treaty have not confronted this question because it has not been much of a problem: the vast majority of commanders — all but the most callous — undoubtedly do, for example, take anticipated harms to medical military personnel and the wounded into account in making targeting decisions, whether or not customary law would require them to do so.
There is probably some support for the notion that a proportionality analysis must take into account expected harm to protected military personnel, including medical personnel. According to the ICRC survey, for example, the Manuals in Australia, Canada, and New Zealand, and France’s Code of Defence, all insist that (in the words of the Australian Manual) “[p]roportionality requires a commander to weigh the military value arising from the success of the operation against the possible harmful effects to protected persons and objects.” And, at a minimum, it certainly would not be a stretch to say that the “dictates of conscience” (cf. the Martens Clause) would prohibit any commander from being completely indifferent to such collateral harms.
Even so, John is probably correct that there is not (yet) sufficient evidence of an established custom that requires commanders weigh the foreseeable casualties to protected military personnel in a proportionality assessment, even if there might be a compelling logical or normative case to be made that custom should develop in that direction. After all, even Article 57 of AP I does not impose a proportionality requirement for protected military (as opposed to civilian) persons and objects.
The reason I included the sections about protected military personnel in my list of troubling provisions in the Manual, however, was not to suggest that the proportionality principle applies to such military personnel exactly as it applies to protected civilians, but instead to flag that the rationale expressed in those provisions as to why the individuals in question are not subject to the proportionality principle is wrong — and it’s the same rationale the Manual uses to conclude that protected civilians, too, are not protected by the proportionality principle.
The Manual repeatedly asserts (see sections 7.3.3.1, 7.8.2.1, 7.10.1.1) that the protected persons in question “need not be considered as incidental harm in assessing proportionality in conducting attacks” because they “are deemed to have accepted the risk of death or further injury due to proximity to military operations.” That same exact reasoning is then applied in the Manual to justify the (certainly mistaken) view that protected civilians in proximity to military objectives — such as involuntary human shields, workers in munitions factories, civilians accompanying forces, and nonmilitary medical personnel — don’t have to be considered in the proportionality calculus, either (see, e.g., sections 5.12.3.2, 5.12.3.3, 7.8.2.1, 7.10.1.1).
What should be done about the Manual, then, when it comes to protected military personnel and the proportionality principle? I think it’s probably unnecessary, and unwise, for the Manual to continue to state that protected military personnel do not have to be considered in a proportionality analysis, if only because such statements might undermine what John rightly describes as the strong policy-based U.S. rule that “our collateral damage estimation process requires us to consider not only civilians, but also other specially protected persons or objects.”
If, however, DoD insists that the Manual should expressly state that there is no customary requirement to take such expected harms into account, surely it should amend the Manual to offer the correct reasons why that is so. The principal reason why there is no such requirement is simply that most articulations of customary law have not (yet) extended the principle of proportionality beyond civilians to cover these other categories of persons protected from targeted attack (although DoD would be advised to explain why the Australian, Canadian, and New Zealand manuals appear to conclude otherwise). And if DoD thinks it is useful to offer further explanation of why the customary law has not developed to cover such cases, it would be much better if the Manual were to articulate the pragmatic reasons John offers, rather than to fall back on a troubling and inaccurate “assumption of risk” rationale, which ought to be scrubbed from the Manual wherever it is said to affect the application of the proportionality principle.
In sum, I don’t think John and I disagree on very much; but I’m grateful he’s pressed me to clarify my post and to better explain what is so troubling about the provisions I flagged.