This post by Kelley Vlahos at The American Conservative looks at attempts to assess the number of civilians killed by U.S. “drones” in Pakistan and Yemen. She notes that the Obama administration has not exactly been forthcoming with estimates of the number of civilians killed–either because they aren’t tracking it or aren’t willing to make the assessments public. We’ve also been informed recently by the New York Times, that the administration may have a somewhat unorthodox method for deciding who counts as a “civilian.” (According to Vlahos, a watchdog group, the Bureau for Investigative Journalism, has estimated about 200 civilian casualties.)
This official reticence isn’t new. Although our leaders frequently tout our military’s efforts to minimize “collateral damage,” the U.S. military has long refused to offer body counts of civilian casualties from its wars (see, e.g., this article on efforts to get the Pentagon to release an estimate of civilians killed in the Iraq war). But this raises anew a question I’ve had for a long time: if we don’t know how many civilians have been killed in a particular war, how can we know whether the war is just?
To offer some context, most versions of what is usually referred to as “just war” theory (which is not really a single theory, but a family of theories) contain a principle of non-combatant immunity. This doesn’t mean that non-combatants* can never be killed, but that they may never be directly targeted. It’s only permissible, according to this principle, for non-combatants to be killed when this is an unavoidable side-effect of attacking a legitimate military target. Or, to put it another way, the principle forbids inflicting harm on non-combatants as an end in itself or as a means to another end.
However, this principle is qualified by another principle: the principle of proportionality. The non-combatant immunity principle doesn’t provide a blank check for inflicting harm on non-combatants so long as you’re not intentionally targeting them. It’s constrained by the principle of proportionality, which says that the good effect one aims to achieve must outweigh, or be greater than, the damage to noncombatants that one foresees, but does not intend. Moreover, the principle requires that there not be another available option that would achieve the same benefit with less collateral damage.
So, to summarize, collateral damage (i.e., harm or killing of non-combatants) can be permissible only if (1) it is an unintended (though possibly foreseen) side-effect of attacking a legitimate target and (2) the good to be achieved outweighs the evil of the collateral damage.
What seems to follow from this, however, is that, in order to know whether these conditions have been met, one would have to know, at least with some degree of accuracy, what the extent of the “collateral damage” was. Otherwise, how can we assess whether the good achieved outweighed the bad of the harm to non-combatants? (This assumes that the first condition is met: that non-combatants were not directly targeted; if that isn’t the case, then no amount of proportionate good can “make up” for the violation of the immunity principle, at least according to most versions of JWT.**)
And there’s the rub: if our government isn’t tracking the number of civilian deaths in a particular war, then they can’t know whether conditions (1) and (2) have been met. Or, alternatively, if they do know and they’re just not telling us, then we aren’t in a position to know whether they’ve been met. But without that knowledge, it’s hard to see how we can be assured that the war is being carried out according to widely recognized principles of just war.
——————————————————
*There are legitimate debates about who exactly counts as a non-combatant (e.g., suppliers of weapons), but wherever we end up drawing the line, there are clearly some groups that fall on the “non-combatant” side (e.g., children).
**The first principle acts a kind of deontological constraint on the conduct of war, while the second more closely resembles a consequentialist principle. So a pure consequentialist would likely reject (1) as absolutely binding (Michael Walzer, a noted theorist of just war, rejects (1) for cases of what he calls “supreme emergencies”).
A Thinking Reed 
“It’s constrained by the principle of proportionality, which says that the good effect one aims to achieve must outweigh, or be greater than, the damage to noncombatants that one foresees, but does not intend.”
I think the main interpretation of proportionality relates to the military objective, not a necessarily highly speculative calculation of “good effect” vs “number of non-combatant casualties.” So, at the link you tweeted me earlier from the Stanford Encyclopedia of Philosophy, “Proportionality” under “jus in bello” is defined as follows:
Obviously, there are broad limits, and in some cases the notion that our government should provide statistics might take on great important, but it’s not as though, say, 50,000 non-combatant casualties will be “OK,” but 100,000 turns the same, otherwise just war into an un-just one, especially if actual responsibility for the non-combatant deaths is shared or in dispute. The key question is, in a way, not “how many,” but “how.”
So the key preliminary assumption re the strike campaign against AQ and associates is that military action against them is legitimate. That’s debatable, of course, but, since the theoretical potential damage of an unchecked terror campaign is very high, you might be able to justify a high number of noncombatant casualties if that was your only criterion. You might not accept that Israel’s invasion of Gaza against minimally lethal but otherwise unstoppable rocket attacks was truly just and “proportional,” but, from Israel’s point of view, there was no other way to stop them, and their cumulative threat to the coherence and therefore political viability of the Israeli state justified military action that in the immediate frame of reference appears numerically disproportionate – 1,000 Palestinian deaths vs .mere 10s of Israeli casualties.
The drone assassination program is seen by its proponents very much to be what the professor quoted in the Vlahos article says it is: a “least bad option.” It’s proportional to the military objective and discriminating – and results in far few casualties, especially to non-combatants, than alternative traditional means for achieving the same objective – such as ground invasion or last-century bombing techniques. Disproportionate relative to military necessity would in this case be massive area bombing, weapons of mass destruction, ground invasion and occupation.
Thanks for commenting, CK.
It’s true that proportionality is frequently discussed in terms of the fitting of military means to ends, but it is also crucial in most JW accounts on limiting the permissible amount of “collateral damage.” Otherwise, the principle of double-effect could be seen to license an unlimited amount of civilian carnage so long as it was “unintended.” (More on double effect: http://plato.stanford.edu/entries/double-effect/)
The point I’m trying to make, though, is that if we agree that there should be limits on civilian casualties (and almost everyone agrees there should), how can we assess whether we’ve reached an “unacceptable” threshold if we don’t actually know how many people are dying? Or is it enough to say that we’ve done our due diligence by using the most discriminating means available? That’s not a rhetorical question–I’m not actually sure what the right answer is.
I agree that the “drone war” may well be the “least bad” option if we assume that we have to wage some kind of “war” on AQ and that it is more “proportionate” in the sense you describe than, e.g., widespread aerial bombing. Though, “least bad” doesn’t necessarily mean “just,” and the tradition that informs JW theorizing places limits on what means we can justly pursue even in the pursuit of otherwise legitimate ends. (It’s possible, of course, to reject or modify the JW framework.)
Still seems to me that there are different arguments and types of argument that get balled up together in what is more a Just War “discourse” than a single coherent and comprehensive, universally applicable “theory.”
On the level of your original question, taken separately: To know whether the US government owes us better numbers, I’d still like to know how those numbers were actually supposed to be used, or how they might be expected to change a moral calculation regarding a particular conflict.
We are also in a peculiar place regarding war, perhaps reflecting how much our conduct of war has changed, and also the luxury of being the overwhelmingly militarily superior power. In a “real” war, there may be little opportunity to get a remotely accurate count of non-combatant casualties. Alternatively, the “good effect” might be something more like the Hiroshima calculation in its simplified form: Better a large number of casualties all at once today, all of them, today, then much, much larger numbers of combatant and non-combatants tomorrow.
Finally, to pick up on Derek Olsen’s last point below, it is a peculiar aspect of Terror War that, even at much lower overall numbers of casualties (at least so far!), like Total War it already represents a failure of the conventions of state-to-state war, including the identification of combatants and non-combatants, something AQ has specifically emphasized. It is little understood, but widely intuited, that we are all, from Manhattan to rural Yemen, already “combatants” or “potential combatants” – or conscriptable sacrifices – for the unlimited total/biopolitical state.
One of the issues that I’m having with all of this remains the kind of war. This is no longer a war between nation-states that involves capture of territory, occupation of said territory, and replacing a governmental structure. The central difference, it seems to me, is that the major objectives to be gained are ideological/psychological in nature and that the central goals are to occupy perspective and sympathy.
That is, the typical purpose of a terrorist strike is not the disruption of the *material* capacity to wage war. Rather, it seeks to destabilize a populace and government by undermining their psychological capacity–their will–to engage in certain behaviors perceived as war-making (like carrying on a Western-style market economy if the 9/11 justifiers are to be believed…). The psychological attack is, in effect, an attack directed at non-combatants designed to raise the level of fear around everyday activities and implant thoughts like: “if my government keeps doing this, I could be randomly killed by these terrorist people who show no discrimination in killing.”
The purpose of a drone strike seems to have one foot in conventional war (in that most of these are about disrupting the material capacity to conduct war through dismantling the leadership structure) and one foot in psychological war. I’d assess the psychological attack here to be a much more directed one that speaks directly to *combatants* saying: “If you accept a leadership position in this organization, we will target you and you will die.”
A central differences between the two is that drone attacks are designed to establish a psychological premise by minimizing fear of non-combatant casualties—to both Americans and those in the theater of war (i.e., Afghanis and Pakistanis). One of the key counter-attacks is to disrupt this premise and suggest that, no, drone attacks are just as indiscriminant as “conventional” terrorist attacks. As a result, a media campaign (using that word *quite* deliberately) showing children killed by a drone attack is precisely an attempt to subvert the largely utilitarian calculus of most just war theorization and the sway opinion (both here and there) with a strong deontological message.
At stake is not simply porportionality but, more important for this kind of conflict, the *perception* of porportionality—and a well-orchestrated campaign can skew the perception quite seriously.
Having said that, should the US be keeping statistics? To the degree that they can, yes. However, in a campaign like this one, the labels of combatant and non-combatant are notoriously flexible. Of course, children should not be considered combatants. Anyone over, say, fourteen, however, gets a lot fuzzier.
Derek, thanks for the thoughtful comments. I think it’s indisputable that drone attacks are more discriminating than most conventional attacks (and this post wasn’t really intended to be about drones per se, but I guess I invited that interpretation by hanging it on that particular hook). On the other hand, I wonder if the actual psychological effect of the attacks really is to minimize fear: if I thought flying robots could rain fiery death upon us anywhere, anytime (not just in a designated “war zone”), I’m not sure how much comfort I’d take in their ability to discriminate. Things get even murkier when we consider that the countries where this is taking place–Pakistan and Yemen–aren’t countries we’re actually at war with. Any utilitarian calculus that tries to take this all into account will be fuzzy indeed.