First off, some personal news – I am very happy to report that I will be taking up a lectureship at the University of Nottingham School of Law starting this September. It’s a truly excellent school, with some wonderful colleagues, and I do look forward to working there. Now, on to business:

The whole Gaza flotilla affair has occupied so much public attention and legal commentary that there has been little response so far to the publication of Prof. Philip Alston’s report to the Human Rights Council on targeted killings. The report is on any view a valuable contribution to the debate. Over at Opinio Juris, Ken Anderson has published a short ‘not-yet-response’ to the report, and I’d like to add a few thoughts of my own, mostly with regard to the relevance of self-defense.

The independent self-defense justification for targeted killings

As readers are aware, Ken has in the past argued for self-defense as an independent justification for (some) targeted killings. He has done so (and I am in full agreement with him on that point) because the justification offered by the Bush administration at the time, that it is engaged in some amorphous, global armed conflict with Al Qaeda and that it can kill combatants in that conflict, cannot justify the whole of the US targeted killings policy even if stretched to its utmost limit. In other words, even if we accept that there is such a thing as a global non-international armed conflict, the US has killed people, and thinks it needs to kill people, who have no connection whatsoever to that conflict. And there, says Ken, self-defense comes in.

Today, of course, Ken is not alone in so arguing – recently, at the last ASIL meeting, the legal adviser of the State Department, Prof. Harold Koh, has offered the same justification.

In a prior post, I argued that it is doubtful that self-defense can do all that Ken wants it to do. In particular, I argued that if the person being targeted has rights under human rights law, self-defense cannot preclude the wrongfulness of the killing. Rather, the killing would have to be justified within the human rights framework. If, on the other hand, human rights treaties did not apply, then there would be little need for self-defense.

Ken has responded to my critique by saying the following:

Meaning, Marko starts from two points – one is extraterritorial application of the ICCPR.  I don’t buy that, the US doesn’t buy it – and I don’t think its position unprincipled or ungrounded.  If one disagrees not just as to the view, but also as to whether it is a principled position or not, it seems hard to have traction in the rest of the discussion, with all admiration and respect to Marko.  Marko’s second point (really the first) went to self-defense as being about the state whose sovereignty is being violated, not the terrorists.  I truly think we – and the US – and Marko are on different, irreconcilable pages here; I can’t imagine the US thinking that the act of self-defense is anything other than aimed at the terrorists, and the violation of sovereignty of the local state is collateral to that.  It is a violation of that state’s sovereignty, but territorial integrity is not everything, as Sofaer said in his 1989 speech and Koh essentially reiterated.  But I think I must not understand Marko well, because I couldn’t understand how he meant self-defense.

Now I’d like to offer a rejoinder – in part because of a real disagreement between us, and in part to clarify what I think is, or may be, a misunderstanding, either linguistic or semantic, or perhaps legal and conceptual.

(Warning! long post — but hopefully not a boring one!)

Four basic scenarios

I’d first like to set out four basic factual scenarios, so that we could get a bit away from the heady heights of abstraction, and thus hopefully facilitate mutual understanding. First, for our present purposes, we need to get IHL out of the picture. In other words, we need to envisage the killing of a suspected terrorist that takes place outside any armed conflict. With that out of the way, the other important contextual element is where the killing takes place, i.e. whether the person being killed was located within or outside the territory of the state which is doing the killing. So:

(1) Imagine an Osama Bin Laden clone, who is equally nefarious, but who had nothing to do with 9/11 and is in no way affiliated with Al Qaeda. Let’s call him Mephistopheles. Old Mephisto has his own terrorist cabal somewhere in Pakistan, and plans an imminent attack on a US city, say with a dirty bomb or something equally horrific. Pakistan’s military is either unwilling or unable to prevent Mephisto from completing the attack, and therefore the US sends a drone which manages to take him out.

(2) For our second scenario, let’s take all of the facts from the first, but with one important difference: while in the first scenario Pakistan does not want to, or cannot prevent Mephisto from conducting his attack, in this scenario Pakistan gives its consent to the US to do its drone strike.

(3) In our third scenario, Mephisto has really outdone himself. Not only is he the mastermind of his own terrorist organization, but he actually operates from a volcanic rock in the middle of the Pacific, a terra nullius which no state claims as its own, and from ships which deliberately do not fly the flag of any state. The US military’s drones still find these ships to be easy targets, and Mephisto’s island base fares no better.

(4) In our fourth and final scenario, Mephisto is a Pakistani national, but he is actually living in the United States, where he has completed doctoral studies in nuclear physics, engineering, or what have you with flying colors, and has maybe even obtained US citizenship. He somehow manages to build a dirty bomb, and decides to put it in a car and drive it all the way to Times Square in New York City, where he intends to detonate it. Fortunately, the US military in Pakistan is informed by Pakistani security services of Mephisto’s plan, and the President authorizes a drone attack (or, perhaps slightly more plausibly, a mere sniper) to take out Mephisto while he’s driving on the New Jersey Turnpike.

All four scenarios involve targeted killings. The first scenario is one of extraterritorial killing, and the territorial state (Pakistan) has not consented to the use of force within its boundaries. The second scenario is also extraterritorial, but the territorial state has in fact consented. The third scenario is likewise extraterritorial, but the killing does not take place on the territory of any state. Finally, the fourth scenario is intraterritorial.

So, what does international law have to say about these scenarios of targeted killing, again all of which take place outside armed conflict?

Self-defense

To my mind, self-defense is relevant ONLY for the first scenario. But before I show what that is the case, I need to define what I exactly mean by self-defense – and this is perhaps where Ken and I might be misunderstanding each other.

Self-defense is a concept of the jus ad bellum, an exception to the prohibition on the use of inter-state force set out in Art. 2(4) of the UN Charter. It is a primary rule, arising from customary law and enshrined in Art. 51 of the Charter, in that it directly regulates when force can be used between states – in other words, the use of inter-state force is prohibited, except when it is in self-defense. It is also a secondary rule, in that it precludes the wrongfulness of any wrongful act merely incidental to the lawful use of force in self-defense, such as the violation of a state’s sovereignty and territorial integrity, and in that capacity it is enshrined in Art. 21 of the ILC Articles on State Responsibility. Because self-defense operates as an exception to the prohibition on the use of force, it logically operates only when that prohibition is itself engaged. (For a more detailed exposition, see here).

Now, self-defense is relevant only for our first scenario because:

- in scenario (2) Pakistan has consented to the use of force;

- in scenario (3) the US attack on Mephisto’s base and ships does not implicate the sovereignty of any other state – in other words, Art. 2(4) of the Charter is not engaged, since the killings took place on terra nullius and ships without nationality;

- in scenario (4) for the same reasons as in scenario (3), since the killing took place on US soil.

Even in scenario (1), many authors would question the applicability of self-defense, for example because Mephisto’s acts were not attributable to Pakistan, or because the purported self-defense was anticipatory in nature. These are I think serious issues, but they do not concern me here – I will postulate the applicability of self-defense.

Now, Ken states that “we – and the US – and Marko are on different, irreconcilable pages here; I can’t imagine the US thinking that the act of self-defense is anything other than aimed at the terrorists, and the violation of sovereignty of the local state is collateral to that.”

Let me say very clearly that I do not actually dispute that. The act of self-defense (a drone attack) is of course directed against the non-state actor. But the only reason why self-defense is needed is to preclude the wrongfulness of what would otherwise be an infringement of Pakistani sovereignty. This is what it does. It is the fact that the US would be using force in Pakistan without Pakistan’s consent that engages Art. 2(4), and self-defense is needed for the US to be able to say to Pakistan, look, it is necessary for us to attack this non-state actor, and this is not a violation of your sovereignty because you are unable or unwilling to prevent this non-state actor from attacking us, and our narrow and confined response is a proportionate one.

Conversely, if Pakistan actually consents, as in our scenario (2), then self-defense becomes perfectly irrelevant so long as the US acts within the boundaries of Pakistani consent. Art. 2(4) is not engaged, and Pakistani sovereignty is not being violated even presumptively.

Finally, in scenarios (3) and (4) the US does not need self-defense because its use of force is simply not prohibited in the first place. There is no rule of international law which in principle prevents states from using force against individuals or non-state actors in areas outside the sovereignty of other states.

Self-defense and the preclusion of wrongfulness

Now, maybe Ken is thinking about some different concept of self-defense than I am. I at least don’t think that any other such concept exists in international law, but I am of course open to contrary argument. However, such an argument would need to explain how self-defense is relevant in situations where the sovereignty of third states is NOT implicated.

If it IS implicated, or if we are in our scenario (1) where it is postulated that it applies, then the issue is what is self-defense as a secondary rule, a circumstance precluding wrongfulness, actually capable of precluding. I have explained in my earlier post why, if human rights law applies, self-defense is as such incapable of precluding the wrongfulness of any violation, just as it is incapable of precluding the wrongfulness of violations of IHL.

This is exactly what Prof. Alston says in his report (paras. 42-44), concluding that:

In sum, even if the use of inter-state force is offered as justification for a targeted killing, it does not dispose of the further question of whether the killing of the particular targeted individual or individuals is lawful. The legality of a specific killing depends on whether it meets the requirements of IHL and human rights law (in the context of armed conflict) or human rights law alone (in all other contexts).

Likewise, even in scenario (2), where Pakistan, the territorial state, consents to the use of force, thereby obviating any need for reliance on self-defense, that consent cannot preclude the wrongfulness of violations of IHL and IHRL (if applicable). As Alston puts it (para. 37):

The proposition that a State may consent to the use of force on its territory by another State is not legally controversial. But while consent may permit the use of force, it does not absolve either of the concerned States from their obligations to abide by human rights law and IHL with respect to the use of lethal force against a specific person. The consenting State’s responsibility to protect those on its territory from arbitrary deprivation of the right to life applies at all times. A consenting State may only lawfully authorize a killing by the targeting State to the extent that the killing is carried out in accordance with applicable IHL or human rights law.

Note, finally, that even though an invocation of self-defense is unnecessary in scenarios (2)-(4), it would in fact be necessary in scenario (1) even if the person killed was undoubtedly a combatant in an armed conflict. A killing may be perfectly lawful as a matter of the jus in bello, but still be unlawful as a matter of the jus ad bellum. Assuming, for example, that Osama Bin Laden is a combatant in such a conflict, and that he is found somewhere in Germany, the United States could not just immediately send a drone to kill him – it would at the very least need to show that Germany was unable or unwilling to do the killing (or the capture) itself.

These conclusions are not, I think, particularly controversial. I’d even go so far to say that they are black letter law. The whole point of the separation of the jus ad bellum from the jus in bello is precisely  that lawfulness under the former cannot affect lawfulness under the latter. And, in my view, and in that of the ILC as well as Prof. Alston, that same reasoning extends to human rights law, if applicable.

Extraterritorial application of human rights treaties

This is why both the Bush and the Obama administrations cannot just say that self-defense precludes the wrongfulness of any human rights violation. Rather, they have to argue that human rights treaties do not apply at all – in other words, that the individual killing is as such not prohibited – and this is in fact what they have done. In our scenarios (1)-(3), this lack of application would be due to the fact that the killing is extraterritorial (while in scenario (4) such reasoning of course could not apply); in the absence of human rights, because in (1) self-defense applies, in (2) we have consent, and in (3) we need neither, the targeting killing would be lawful, as it would not be prohibited by any applicable norm of international law.

This is precisely what Ken does when he says that “I don’t buy that [the extraterritorial application of HR treaties], the US doesn’t buy it – and I don’t think its position unprincipled or ungrounded.  If one disagrees not just as to the view, but also as to whether it is a principled position or not, it seems hard to have traction in the rest of the discussion.”

Let me make two points. First, substantively, the argument of my previous post was precisely that self-defense was not enough IF human rights law applies – that the US (and Ken) need to do more if it does. One thing to do, of course, is to deny extraterritorial applicability and be done with that. But if that argument turns out to be wrong, then self-defense just doesn’t cut it. Perhaps Ken and I don’t even disagree on this point, and scenario (4) would of course be a good place to test that.

Second, I of course did not mean to suggest that those who argue that human rights treaties do not apply extraterritorially are themselves unprincipled. Far from it – and apologies if I have given the wrong impression. But, but, I still maintain that there is no reason of principle to deny that human rights treaties apply altogether.

By ‘principle’ I simply mean a normative justification as to why one category of human beings, those within the state’s territory, are entitled to be protected from the arbitrary exercise of the state’s power, while those outside the state’s territory are not so entitled. If, as all human rights treaties explicitly say, their normative basis is in universality, in the idea that ALL human beings are by virtue of their humanity entitled to some minimum of rights and worthy of protection, no matter how otherwise bad or evil, then I don’t see a principled justification for a wholesale denial of rights.

Having said so, I don’t mean to deny (1) that the US has a good, even better, purely textual argument that the ICCPR should not apply extraterritorially; (2) that this has been a consistent, long-standing position of the US government; and (3) that even under ECHR case law, most notably Bankovic, drone attacks in a territory outside a state’s control (eg. US in Pakistan) would not be covered by the ECHR.

Nonetheless, those in the US who argue against the ICCPR’s extraterritorial application should likewise realize that (1) text is as such not absolutely paramount, as it must be tempered with object and purpose; (2) that the text is more ambiguous than they make it seem (even the US Supreme Court has frequently read an ‘and’ to mean an ‘or’ in other contexts), and so is the drafting history; (3) and that though the US position is consistent, it may well be that it is consistently wrong.

As for Bankovic and the ECHR, I can only say that this unfortunate case has been steadily eroded  – but we’ll see how far the European Court itself will go in Al-Skeini in a matter of months. I would just like to note that the issue of extraterritorial application is not relevant only to these ‘big’ cases of drones and wars on terror or what have you. For example, currently pending before the Court is the application filed against Russia by the family of Alexander Litvinenko, famously assassinated in London through poisoning by radioactive sushi, supposedly by a KGB agent or with the KGB’s collusion.

In short, of those who argue that there is a principled reason – rather than one of political expediency – to deny the extraterritorial application of human rights treaties, I would ask the following two questions:

(1) Why is it okay for a government to kill, or be complicit in the killing, of one of its political opponents, merely because the killing took place in London rather than Moscow? Note that, the application of IHL aside, to my mind nothing legally or morally distinguishes killing by drone-fired missile from killing by polonium-laced sushi.

(2) Would Auschwitz count as a human rights violation? Yes, that’s my question, absolutely bizarre though it may seem. Imagine if somehow the Third Reich was a party to the ICCPR. Would it really not have been a violation of this treaty to set up the death camp, just because it might be located in occupied Poland, rather than in Germany proper? And is this really a result that its drafters would have found acceptable, legally or morally?

Targeted killing within a human rights framework

Thus, in order to justify the practice of targeted killing, the US and other states engaging in the practice would be well advised not to put all their eggs in one basket, that of denying the extraterritorial application of human rights, but to articulate an argument within a human rights framework, even if it is only in the alternative. This, of course, is something that the Israeli Supreme Court has already done in its Targeted Killings judgment.

Within that framework, as within IHL, self-defense is as such perfectly irrelevant. It is an ad bellum consideration which has no place in IHL analysis, and if it should have a place within a human rights analysis (I make no judgment on this point), that would be become the human rights framework itself would mandate this result (thus, for example, one could argue that killings done during a use of force that is unlawful under the jus ad bellum could never be justified under human rights law).

The reason why states are reluctant to articulate such arguments is simple. They are afraid that human rights law is too rigid and inflexible, and that it would render a policy of targeted killings almost invariably unlawful. I personally don’t think that this is the case. States can, and they do, lawfully take lives under human rights law if they show the necessity for doing so. In that regard, I am somewhat concerned with some of Alston’s finding in his report, such as that (para. 85):

Outside the context of armed conflict, the use of drones for targeted killing is almost never likely to be legal. A targeted drone killing in a State’s own territory, over which the State has control, would be very unlikely to meet human rights law limitations on the use of lethal force.

This, I think, greatly overstates the case against drones. Coming back to our scenario (4), a drone may well be an appropriate way to perform a killing. Whether it is or it isn’t depends on concrete facts which by their nature cannot be taken into account in the abstract. Likewise, Alston remarks that (para. 86):

Outside its own territory (or in territory over which it lacked control) and where the
situation on the ground did not rise to the level of armed conflict in which IHL would apply, a State could theoretically seek to justify the use of drones by invoking the right to anticipatory self-defence against a non-state actor. It could also theoretically claim that human rights law’s requirement of first employing less-than-lethal means would not be possible if the State has no means of capturing or causing the other State to capture the target. As a practical matter, there are very few situations outside the context of active hostilities in which the test for anticipatory self-defence – necessity that is “instant, overwhelming, and leaving no choice of means, and no moment of deliberation” – would be met. This hypothetical presents the same danger as the “ticking-time bomb” scenario does in the context of the use of torture and coercion during interrogations: a thought experiment that posits a rare emergency exception to an absolute prohibition can effectively institutionalize that exception. Applying such a scenario to targeted killings  threatens to eviscerate the human rights law prohibition against the arbitrary deprivation of life. In addition, drone killing of anyone other than the target (family members or others in the vicinity, for example) would be an arbitrary deprivation of life under human rights law and could result in State responsibility and individual criminal liability.

Alston is justifiably concerned with the integrity of the human rights regime, which might be compromised by carving out exceptions from the right to life, and even by injecting IHL into it. But, at the same time, his concern leads to a lack of flexibility that may in turn simply lead states to think that the application of the human rights regime to such situations would be hopelessly utopian.

In other words, for the extraterritorial application of human rights treaties to become a practical reality, rather than empty words on paper, a price must be paid, and that price may involve watering down human rights in order to be able to give full regard to extraordinary circumstances – but human rights must not be watered down too much. Thus, for example, I find Alston’s position that the killing in a drone attack of any bystander would necessarily be arbitrary to be far too rigid.

Without concrete facts and cases, it is impossible to devise a balance in the abstract between these two competing considerations of flexibility and regime integrity. But that balance can be struck. To my mind, at least, it would be perfectly justifiable within a purely human rights framework for the US to kill Osama Bin Laden by a drone, so long as it could show that capture was entertained, but that it was wholly impracticable. The same human rights principles that apply, for instance, to the killing of a hostage-taker or a prison fugitive or during a riot can also be effectively applied here, again, so long as due regard is given to the exceptional circumstances. A mere invocation of self-defense, however, would simply not suffice.