A coda to Bobby's post below asking about the legal views underlying US operations in Somalia over the past three weeks. Three weeks ago, SEALs attempted a capture operation against a target on the coast of Somalia. The SEAL team withdrew without capturing its target, on account of risks to noncombatants, it was reported. Three weeks later, the US undertook another raid, only this time with a drone missile attack, killing the target. I have no way to know whether these operations are part of a change of policy that is now having actual effects on operations, or whether these two strikes are merely two blips in the policy that the Obama administration has insisted has been in place all along - a preference for capture over kill, where "feasible."
These possibilities have been vigorously analyzed in commentary, including here at Lawfare. But it would be a mistake to assume that attempting a (unsuccessful) capture operation three weeks ago or a drone strike two days ago says much about the policy of "feasibility." There are reasons to believe that the administration means what it says about having a policy preference for capture where feasible, despite having the lawful targeting authority (on the US government's view of the relevant international law) to undertake first resort to lethal force - or at least to remain agnostic as to what actual US government policy might be. The reason is that the more the US attempts capture operations, and the more that becomes widely known and believed as US policy by terrorist targets, the less feasible over time these operations are likely to become.
Feasible means different things in different contexts. It means one thing in the law of human rights. It means another in the law of targeting and its requirement of feasible precautions in attack. (There is also a tendency, I would say, of human rights monitors to conflate those two distinct legal concepts.) In the context of a US targeting policy, however, that at this point is deliberately more restrictive than the law requires (in the US government's view of its lawful targeting authorities), "feasible" here is a policy term, not a narrowly legal one. Like the term "kill or capture," feasibility here is a policy term in establishing a mission's rules of engagement, under the legal view (so far as the US is concerned) that the target is already in the "first resort to lethal force is lawful" category.
As part of a mission's ROE, feasibility takes into account benefits from capturing this particular target, on the one hand - but set against risks to civilians, risks to own forces, and "geopolitical" factors that include, for example, the international political risks of entering into a firefight with civilians in the midst of it that (far more than a drone strike) risks many casualties, on the other. Such a fight in a neutral country might inflame a national population, as the President pointed out in his May 23 speech. Strategic and political planners thus might quite correctly judge that a capture operation carries risks that a drone attack is less likely to bring about. They might (correctly, in my view) reach this conclusion, despite efforts of anti-drone campaigners to insist that drone strikes are uniquely susceptible of political blowback - rather than considering what the blowback of the genuine alternatives for the use of force through drones would be. The plausible alternative is not likely to be no use of force, not over the longer run; it is likely to be either US forces on the ground or else the Pakistani military engaged in aerial bombardment or artillery shelling.
II. The Choice Between Lethal Air Attack and Human Ground Team in the Bin Laden Raid
There are also political risks at the "grand strategy" level - geopolitical consequences reverberating potentially far into the future. Presidents properly take these possibilities into account, even when sometimes they involve estimates, guesses, and bets on contingent events. The bin Laden raid, for example, was meticulously planned and rigorously executed - still, US forces benefited from some luck, as the President quite correctly said in his May 23 speech. Meticulous contingency planning notwithstanding, the US was lucky not to have faced fighting either with Pakistani civilians living near bin Laden's compound, or with Pakistani military in Abbottabad. The result might have been a complete rupture of state-to-state relations, bringing about almost inevitably enormous but unpredictable consequences.
Civilian neighbors living around the bin Laden compound in Abbottabad did not try to challenge the SEALs, for example, not knowing who they were, why they were there - in violation of Pakistani domestic law, after all, given the widely reported claim that the Pakistani government was not informed in advance or its consent sought, whatever the US's view of its international claim to be able to attack the author of 9/11. These neighbors did not try to rush the Abbottabad compound or the SEALs, or even simply insist on walking into the compound unarmed, in a sort of nonviolent protest against foreign invasion, local inhabitants against an armed foreign invader (can there be a non-violent levee en masse)? Whatever one thinks of Zero Dark Thirty, the broad narrative accuracy of the tactical scenes near the film's end hasn't been disputed: the JSOC soldier holding the outer perimeter and the team's translator face the possibility of civilians entering the compound. The translator is able to go forward and calm the gathering crowd of men, but they might instead have started chanting slogans and marching forward, apparently unarmed and yet threatening, and perhaps armed after all. The US soldier would have had to make immediate decisions with lasting and almost certainly bad consequences.
The bin Laden raid was an "attack" using "first resort to lethal force," not a raid specifically designed for capture; it is, however, an example of the risks inherent in a human ground operation. The larger point is that some good luck spared US forces from having to face issues that arise from the fact of having a human team on the ground in the first place, on which, in capture operations over time, the US cannot depend. Moreover, when the US government decides between a drone attack or ground team operation, it has to take into account not only possible casualties among the American team, but the possibility of an American soldier getting captured as well. A particular form of this was sharply debated in the bin Laden raid's planning; according to Mark Bowden's account, US military planners initially assumed that had the SEAL team been surrounded by the Pakistani military, it would hunker down in the compound and let the governments negotiate a solution. But the President rejected that option. He feared the international consequences of a protracted standoff with the Pakistani government - the President's very own hostage crisis, with Navy SEALs as pawns in a larger geopolitical game - where Pakistani authorities would face enormous domestic pressures not to let the SEALs go, after nakedly violating Pakistani sovereignty, without consent or prior warning (irrespective of how the US might view the situation under international law).
But President Obama of course had the domestic political implications of a protracted, quasi-hostage standoff with the Pakistani government in mind - potentially disastrous consequences for him in the election - and there is nothing wrong with political leadership considering those consequences, too. He therefore insisted, according to Bowden, on a plan in which the SEALs would shoot their way out if necessary - backed by the full weight of the US military crossing the Pakistani border, shooting down that "ally's" aircraft and engaging its forces as necessary in order to extract the team. Had it occurred, the proper name for such an encounter, let's be clear, is "interstate war." It was a gutsy decision and, in my view, clearly the right one. But it was surely impelled as much by the President's domestic political calculations as by geopolitical or international ones for the United States.
That these very negative scenarios did not play out is a blessing for the America. But from the standpoint of "feasibility" as policy, such risks are perfectly okay to take into account, and of course must be, in the bin Laden raid as in any other situation. For a target other than bin Laden, the President implied, however, a decision to enter Pakistan in force with US military personnel in a capture operation is (highly) likely to be infeasible on political, strategic, and geopolitical grounds. But these are considerations of feasibility quite apart from tactical feasibility in the conduct of any particular operation.
III. A Policy of More Capture Attempts Will Likely Make Them Less Feasible
But there is a further consideration in choosing between lethal drone attack or ground team capture operation that is essentially operational and tactical. Capture missions are likely to become less feasible the more of them the US attempts. Today, people who think they might be targets of drone strikes take some precautions, of course - but they do so at the front end, so to speak, before the drone is able to put them in its sights, ready to fire a missile. They seek to hide so that they can't be surveilled, and to stay out of drone surveillance by staying under cover, indoors, in the company of civilians who can't be targeted, etc. But short of strapping a baby to one's back, it is very difficult for anyone who is operational, and whose location is generally known, to remain isolated from a drone strike forever. Short of using civilians as shields (which, in the case of a drone strike using precise weaponry combined with heavy surveillance, means having women or children highly physically, tightly proximate to you), a terrorist target has few if any ways of defending against the missile strike itself. It strikes without warning, and unless you have civilians very, very closely surrounding you at all times, it will be difficult to avoid becoming a target. So while you might try to conceal your location and stay under cover as much as possible, and stay with children generally around you, you can't really do anything to protect against the actual drone attack itself. It's too unpredictable and you don't have protections against a missile.
However, if it is announced and becomes widely believed that the US has changed its policy to favor capture operations instead, the incentives change. In addition to trying to hide, protect your location, stay out of surveillance, etc., a capture operation requires a human team launching an attack that takes time - time to arrive, time to make the attack and capture, time to withdraw. It's not a blink of an eye, like a drone missile strike. US special operators are very smooth and fast, but any human attack requires time - an eternity, by comparison, to the missile strike. That time gives the target plenty of reasons and opportunities to make that assault costly in terms of civilians. And where time is longer than a second, and is measured in many minutes at least, it isn't necessary to strap the baby to your back to gain protection. It's enough to have civilians in the area - women and children loosely in the area who can be pulled closer to protect the target as the commando assault is launched. Within the same group of civilians, moreover, fighters can hide and then kick off precisely the sort of firefight that the President warned, in his May 23 NDU address, a drone strike could, and would be used to, avoid. Which is why, presumably, the SEALs withdrew three weeks ago without making their capture.
The more convinced a terrorist target is that the US will attempt a capture operation, then, the greater the incentive he has to surround himself with civilians, and to prepare a gauntlet that a US special forces team will have to run to carry out the attack. As terrorist targets of US special operations refine their tactics - as they will - the likelihood of civilian and special forces casualties increases and the likelihood of capturing the target decreases - indeed, killing the target might itself become difficult or infeasible. The more likely targets believe that the favored US policy is capture, the more they will prepare against it - increasingly certain that they have something tangible to gain by it, and increasingly correct in that assumption. Their preparations against a human attack can be made effective and costly in ways not available to them against a drone strike.
The irony is that capture operations become most likely to succeed, and most feasible, when they are few and far between, unpredictable, and sufficiently rare that they do not invite the targets to attempt preparations to prepare against an attack aimed at capture. The more you attempt capture, by contrast, over time the less feasible it will be.
(Ben and I discuss this and other problems of feasibility, capture and kill, drone strikes and special forces raids, in Chapter 3 of our book, Speaking the Law, which has just been posted online by the Hoover Institution. Chapter 3 is a an analysis of President Obama's May 23 NDU speech, and the section on targeting and drone warfare takes up these questions of feasibility. Update: Some grammar and syntax cleaned up.)