Three Problems with the “As-If” Thesis of Proportionality

Professor Robert D. Sloane*

This contribution is a brief preview derived from part of a full forthcoming
article that will run in the spring issue of the Harvard National Security Journal.

It is a privilege to contribute the first piece for the Boston University International Law Journal’s online publication series. Tragically, the issue I will discuss is, once again, as timely as it is significant to the evolving law of war in the twenty-first century. For the third time in six years, a brutal armed conflict has afflicted the Gaza Strip, killing or injuring thousands. As in the past, the term proportionality permeates both legal commentary and media coverage of the conflict. And as in the past, the term gets bandied about but rarely defined precisely and often misused, rhetorically and otherwise.[1]

In part, this is because the principle of proportionality in the law of armed conflict (“LOAC”), or jus in bello, is a singularly subjective and indeterminate legal standard. Additional Protocol I of 1977 to the 1949 Geneva Conventions (“API”),[2] prohibits as disproportionate any attack “which may be expected to cause incidental loss of civilian life, injury to civilians, damage to civilian objects, or a combination thereof [collectively, ‘collateral damage’],[3] which would be excessive in relation to the concrete and direct military advantage anticipated.” This canonical definition subsumes diverse considerations, which have distinct legal, ethical, and political dimensions—all of which, ideally, would inform operational judgments but, realistically, seldom can or do. Not only can these dimensions of proportionality be difficult to distinguish in theory (and a fortiori in practice); participants in armed conflicts also may have propagandistic incentives to obscure them. The unfortunate rhetorical use of the term disproportionate contributes to a relatively commonplace but mistaken view: that proportionality as a legal constraint on war is so subjective and indeterminate as to be no real constraint at all.

While plainly hyperbolic, there is undoubtedly some truth to this view: proportionality is subjective and indeterminate—ineluctably so. There are many and diverse reasons for this. But perhaps the chief problem is that the core of the principle is a directive to weigh incommensurables. As Yoram Dinstein writes, in what strikes me as an understatement, proportionality requires “pondering dissimilar considerations—to wit, civilian losses and military advantage—[which] is not an exact science.”[4]

There is an unfortunate tendency, encouraged by the ordinary meaning and connotation of proportionality, to conceptualize the question in pseudo-mathematical terms, as though it were susceptible to determination by an algorithm or tit-for-tat calculation. This leads to exasperated statements such as that of former U.S. Ambassador to the United Nations John Bolton, speaking to the press at the outset of the 2006 war between Israel and Hezbollah:

What Hezbollah has done is kidnap Israeli soldiers and rain rockets and mortar shells on innocent Israeli civilians. What Israel has done in response is act in self-defense. And I don’t quite know what the argument about proportionate force means here. Is Israel entitled only to kidnap two Hezbollah operatives and fire a couple of rockets aimlessly into Lebanon?[5]

LOAC makes clear that proportionality is not a matter of lex talionis.[6] Nor, of course, is there a “precise formula . . . such as that 2.7 enemy lives equal one of ours.”[7] So what guidance can be supplied?

We need not, in my view, throw up our hands in the face of this question. But neither is it susceptible to a formulaic answer. In the balance of this contribution, I want to offer three critiques of perhaps the most prominent and popular of these formulaic answers, what might be termed the “as-if” thesis. I stress at the outset that I do not suggest that this thesis lacks value; only that it does not suffice, and at times, it may affirmatively mislead. The basic idea of the as-if thesis is to turn the subjective nature of proportionality upon itself, to make a virtue of a “vice.” The as-if thesis posits that all civilians—those of the attacker, the attacked, and third parties—must be treated identically. For purposes of proportionality, a civilian is a civilian, period. A military commander should therefore determine the acceptable level of expected collateral damage relative to the anticipated military advantage as if the civilians put at risk by a particular attack were citizens of his own state.

Michael Walzer and Avishai Margalit sparked a continuing debate about the as-if thesis by advancing it in the context of the first (2008-2009) Gaza War.[8] They explained their view in an editorial in the New York Review of Books (“NYRB”).[9] Like Just and Unjust Wars,[10] Walzer’s celebrated work of just war theory prompted by his reflections on the Vietnam War, the editorial spurred an extended debate, first played out in a NYRB exchange of replies and rejoinders with Asa Kasher and Amos Yadlin.

In the original editorial, Walzer and Margalit invite readers to imagine that before the 2006 conflict between Hezbollah and Israel, a Hezbollah militia captures Manara, a kibbutz “in northern Israel adjacent to the Lebanese border.”[11] The authors then posit four possible variations: In the first, Hezbollah holds the Israeli kibbutz citizens as hostages, “[mingling with them] so as to be shielded by them from any counterattack.”[12] In the second, noncombatant, non-Israeli-citizen volunteers occupy Manara at the moment of capture, and Hezbollah uses them as human shields. In the third, when Hezbollah captures Manara, the militia finds it occupied not by “well-wishing,” noncombatant volunteers but by a group of foreign protesters from abroad who came to the region to raise their voices against Israel’s policies toward Lebanon. In the final variation, Manara had been evacuated before capture, but Hezbollah forcibly brings in South Lebanese villagers to occupy the kibbutz and to use them as human shields.

With these variations in mind, the authors further assume that the IDF decides to retake Manara, which would be a legitimate, ad bellum proportionate, military response.[13] How, they ask, should a military commander acting in good faith carry out this military operation consistent with in bello proportionality?

Walzer and Margalit deliberately avoid the devils lurking in the details of their four-variation hypothetical. Rather than wade into this morass, they propose that the answers reached by a reasonable military commander, acting in good faith based on their view of what in bello proportionality requires may not depend on either the nature of Hezbollah’s conduct (for example, that it may have deliberately created greater risks to civilians in order to exploit LOAC to its advantage) or the identity of the civilians put at risk by the effort to recapture Manara (for example, whether they are nationals or foreigners). In other words, “[w]hatever Israel deems acceptable as ‘collateral damage’ when its own captured citizens are at risk—that should be the moral limit in the other cases too.”[14] The authors thus conclude succinctly: “This is the guideline we advocate: Conduct your war in the presence of noncombatants on the other side with the same care as if your citizens were the noncombatants.”[15]

Walzer and Margalit argue that failure to abide by the as-if thesis would “erode the distinction between combatants and noncombatants, which is critical to the theory of justice in war (jus in bello).”[16] The reason for this erosion is hardly recondite. To say that force must be proportionate naturally invites the question: “proportionate to what?” Were the answer pitched at an unduly high level of abstraction, that is, in terms of one side’s ultimate ad bellum military objective (“victory”), then almost any attack, no matter how many civilians it would foreseeably injure or kill, would arguably be in bello proportionate. That is why in bello proportionality must refer to atomized rather than ultimate military advantage—although, just as clearly, that advantage “need not be limited to [neutralizing] an individual soldier, tank or aircraft.”[17] The nebulous nature of victory in many modern wars[18]—of which the most recent Gaza War is a tragic but paradigmatic example[19]—makes it even more vital to define and circumscribe “anticipated military advantage” in the law of in bello proportionality.

But does the as-if thesis help? Its apparent merit is twofold. First, by definition, it respects the dualistic axiom. If all civilians must be treated identically, there can be no danger of injecting perceptions of which party is the just or lawful belligerent into proportionality decisions. Second, the as-if thesis seems to offer a uniform way to handle the multiple subjectivities and indeterminacies that afflict the principle of proportionality. It does so by investing each arguably subjective or indeterminate component of the proportionality definition with substantive content by relying on the attacker’s own intuitions. It tells military elites to weigh the lives of civilians—their own nationals, enemy civilians, or those of a third party—equally to decide “what a reasonable military commander would [deem] an acceptable loss of civilian life for the military advantage sought.”[20]

Now, the as-if thesis may well inform judgments of in bello proportionality in practice. Perhaps it can serve in this regard as a salutary “reality check” on operational proportionality judgments.[21] But further reflection on the as-if thesis suggests that it is misguided or problematic in at least three ways: first, as a matter of law, it begs the question or, at best, supplies an excessively variable standard; second, as a matter of ethics, it conflicts with broadly shared intuitions; and third, and most damningly, as a matter of practice, it is simply unrealistic: no military has ever conducted, or even aspired to conduct, an armed conflict in compliance with the as-if standard.

First, the as-if thesis is question-begging—or, to state the objection less categorically given its value as one way to inform judgments of in bello proportionality—excessively variable. It makes the answers to in bello proportionality questions depend on intuitions that vary widely from state to state and commander to commander.[22] Doubtless some degree of variability in any legal standard, especially that of in bello proportionality, is both inevitable and unobjectionable. But recall that one of the putative virtues of the as-if thesis is precisely that it purports to reduce (and ideally eliminate) the degree of that variability. It is unclear that the as-if thesis may justifiably claim this virtue.

Consider a hyperbolic example, which nonetheless starkly illustrates the problem. If war on the Korean Peninsula were to resume, North Korea’s current regime would almost certainly not hesitate to impose very high risks of serious harm on its civilians in the service of the state’s military objectives. Is the in bello proportionality constraint for North Korea therefore far less demanding than that for liberal states that place a much higher value on the welfare of their own civilians? Of course not. Yet, strictly speaking, the as-if standard—“[c]onduct your war in the presence of noncombatants on the other side with the same care as if your citizens were the noncombatants”[23]—would imply that it is. No one would seriously argue that North Korea respects proportionality just in case it imposes risks of the same nature and scope on the enemy’s civilian population as it would be prepared to impose on its own noncombatants. I do not doubt that Walzer and Margalit, too, would reject this notion. The risks themselves, the expected collateral damage, presumably should be equally and illegally disproportionate even if North Korea’s political or military elite were prepared to impose those risks on its civilians.

Because this example is admittedly hyperbolic, one might object that in most cases—in particular, where a LOAC-abiding state considers the as-if thesis in good faith relative to its own civilians—it produces a laudable result. Not so. As explained below, the truth is that even the armed forces of states that strive to respect LOAC do not treat enemy civilians and their own equally. What the example of North Korea shows hyperbolically is true of all states to some extent. Outer-boundary cardinal constraints on collateral damage, even if they cannot be specified precisely, must be part of the proportionality standard. Yet the as-if thesis cannot supply them. It only seems to do so because we tend to assume that other peoples and military cultures would make the relevant judgments roughly as we do.

Second, as a matter of ethics, the as-if thesis is in considerable tension with broadly shared intuitions. The utilitarian ideal of giving equal concern to the welfare of “our” and “their” civilians, which the as-if thesis cleverly polices, conflicts with the common conviction that states, like people, may—and perhaps must—assign greater weight to the welfare of those with whom they have a morally, socially, or politically salient relationship.[24]

Thomas Hurka advances this argument clearly with an analogy to interpersonal relations: It is surely not unethical for a father, faced with the tragic choice, to save his daughter’s life at the cost of the lives of several strangers—and while “relations among citizens of a nation are not as close as between parents and children,” normative ethics similarly authorize “some partiality toward fellow citizens.”[25] The same goes for governments and their militaries relative to their civilians.[26] A state therefore may value the welfare of its own civilians to some degree over the welfare of enemy, or third-party, civilians.

The typical response to this argument, as Hurka points out, is to stress the distinction between attacking and defending: hence “[a] father may and even should prefer saving his daughter’s life to saving five strangers, but he may not kill those strangers in order to save his daughter.”[27] A soldier, so this argument runs, occupies a moral position comparable to that of a criminal assailant: like the assailant, who would be equally culpable for attacking any innocent person (whatever his relationship or lack thereof to the victim), the soldier may not distinguish between civilians and expected collateral damage to them based on nationality—at least, not insofar as he is attacking. All civilians, after all, are equally innocent in the relevant ethical and legal sense of being “currently harmless.”[28] Not surprisingly, Walzer and Margalit adopt this view.[29]

But, as often, the apparent analogy between war and crime turns out to be misguided. A thug’s attack in peacetime is a crime. A soldier’s attack in wartime, assuming it otherwise respects LOAC, is not—even if that soldier fights for an unjust belligerent. The dualistic axiom prohibits distinctions in the conduct of war based on judgments about the justice or legality of each side’s casus belli. And legally, a soldier must take precautions to minimize harm to enemy civilians.[30] Ethically, however, there is force to the argument that he may nonetheless prefer the welfare of the civilians of the state on behalf of which he fights to that of both enemy and third-party civilians—at least to a certain extent. “The fact that [a soldier] is killing rather than failing to save is not irrelevant; it still plays a significant moral role and in particular reduces the degree of partiality he may show below what would be permitted if he were merely distributing benefits.”[31] But the identity of the civilians placed at collateral risk by a military attack is also not morally irrelevant. Insofar as the as-if thesis suggests otherwise, it is unlikely to be accepted by states as a means to give more precise legal content to in bello proportionality.

Finally, the as-if thesis, as a legal standard rather than only an ethical ideal, is at odds with state practice. Even belligerents that have made a serious effort to respect the law of war have historically, and unsurprisingly, privileged the lives of their own civilians over those of enemy civilians.[32] No state military force that is or has ever been treats enemy civilians and its own civilians equally. So whatever may be said for the as-if thesis as an ethical ideal, it is implausible as law. It would render the vast majority of strikes that cause collateral damage supposedly disproportionate, even in circumstances in which the attacker sought in good faith to respect LOAC.

In Kosovo, for example, NATO decided it would not introduce ground troops because its constituent states, including the United States, believed (with good reason) that given NATO’s avowed military objective (to prevent war crimes against Albanian Kosovars), their domestic political constituencies would not tolerate combatant, let alone civilian, casualties. Of course, the perceived political acceptability of a military operation is not the measure of its legality. Yet had the relevant risk of collateral damage been to the citizens of the United States and the other western European states that contributed to the NATO force, rather than to the Serb and Albanian civilians who were placed at risk by the aerial assault, it is nearly impossible to imagine that NATO’s generals would have been authorized to undertake the campaign. They would almost surely have sought to minimize the risks to their civilians by an assault with ground troops,[33] an alternative that many analysts suggest would have posed a lesser risk to Serb and Albanian civilians.[34]

At any rate, the relevant question is whether this counterfactual renders NATO’s conduct disproportionate. According to the ICTY Kosovo Report, at least, it does not. The Report found, to the contrary, that military elites could reasonably disagree about the proportionality of the aerial strikes in 1999.[35] No clear violations of proportionality, still less violations at the level of war crimes, could be established. Yet application of the as-if standard would apparently yield the contrary conclusion. A similar analysis applies to the first Persian Gulf War. It is regarded as one of the most legally sanitized armed conflicts in history; figurative armies of lawyers advised literal armies before almost every strike.[36] But despite the coalition’s efforts, the U.S.-led forces did not meet (nor did they try to meet) the as-if standard.[37] Did the 1991 coalition therefore systematically violate in bello proportionality?

In the final analysis, the as-if thesis offers scant guidance to military elites seeking to abide by in bello proportionality, and at times its application may be affirmatively misguided. It is especially noteworthy that the as-if thesis proves unhelpful in modern conflicts characterized as “asymmetric.” Walzer and Margalit argue that if “soldiers . . . take fire from the rooftop of a building, they should not pull back and call for artillery or air strikes that may destroy most or all of the people in or near the building; they should try to get close enough to the building to find out who is inside or to aim directly at the fighters on the roof.”[38] In an interstate war, however, this scenario, which yields the relatively uncontroversial conclusion just stated, is the exception. By contrast, in an asymmetric conflict with a non-state belligerent (“NSB”), it may at times be, not only the rule, but the modus operandi of the NSB. For if a NSB knows that it will be outgunned on the battlefield, it will naturally seek to shift the literal and figurative terrain in an effort to neutralize the enemy’s technological, geographic, and other advantages. Robin Geiss notes, for example, that “direct attacks against civilians, hostage-taking and the use of human shields—practices that have long been outlawed in armed conflicts—have seen a revival in recent conflicts in which the far weaker party has often sought to gain a comparative advantage over the militarily superior enemy by resorting to such practices as a matter of strategy.”[39]

This is not a new phenomenon; it characterized the Vietnam War, for example. But because of technological advances in warfare, it is much more widespread than it was a century—or even a few decades—ago. Bearing in mind that the only (virtual) consensus legal principle of in bello proportionality to date is API’s, as informed by the nebulous idea of the “reasonable military commander,” one of the most difficult questions for the law of war in the twenty-first century is to what extent it is reasonable (or realistic) to demand that armies committed to LOAC systematically, not just in periodic incidents, risk their soldiers’ lives to minimize collateral damage. That is a question I cannot pursue here, but one that will, I believe, be critical to the resilience of the principle of proportionality in future, especially asymmetric, armed conflicts.