Self-defense in response to armed attacks by non-state actors is undoubtedly one of the most interesting – and controversial – issues in modern international law. It is of great practical relevance, as for instance, with the ongoing use of drones for the targeted killings of suspected terrorists (a question I’ve discussed here), and has attracted a great deal of scholarly attention. Lindsay Moir has just published a book with Hart/Oxford that I’m sure will provide a strong contribution to the field. I would particularly like to draw our readers’ attention to the recent discussion in the EJIL provoked by Christian Tams’ excellent article on the use of force against terrorists, with responses by Federico Sperotto and Kimberley Trapp, and a rejoinder by Christian.
In this post, I would like to add a few thoughts on how the indeterminacy of state practice and opinio juris has caused an indeterminacy in the law, which I don’t think can be denied or removed by any legal analysis, no matter how exhaustive and competent.
Consider, first, that we are (I think) in broad agreement that the prohibition on the use of force in international law, as set out in Article 2(4) of the UN Charter, operates exclusively between states. If, instead of from Afghanistan, Al-Qaeda operated from and launched the 9/11 attacks against the United States from Antarctica, the high seas or some Dr. Evil-style terra nullius volcanic rock in the middle of nowhere (or the Moon even), the jus ad bellum would not in any way limit the US response to the armed attack. It is only if in responding to the attack the US has to encroach on the sovereignty of some other state that Article 2(4) is engaged. (See, similarly, our discussion regarding Israel’s intervention in Gaza). Of course, that in such a situation the jus ad bellum would be inapplicable does not mean that the use of force might not be regulated by some other body of law, such as humanitarian law or human rights law – but that is for now beside the point.
Consequently, it is only if Article 2(4) is engaged that the Article 51 right of self-defence comes into play, since it is an exception to the prohibition on the use of inter-state force. Sure, one could generally say that in responding to an attack by a non-state actor which does not operate from the territory of another state the attacked state is acting in self-defence. But such a characterization would be merely descriptive – this would not be the Article 51 notion of self-defence, which is relevant only when the attacked state responds by violating the sovereignty of some other state. This was, I believe, the case with the Israel’s incursion into Gaza.
However, and this is a crucial point, that Article 51 operates as an exception to the ban on the use of inter-state force, does not logically and conceptually entail that the ‘armed attack’ within the meaning of Article 51 must be attributable to a state. Such a conclusion may follow from an examination of state practice, but it simply does not follow from the text of the Charter or from some broader legal principle, nor from the fact that Article 2(4) itself is inter-state in operation.
So, with this in mind let’s turn to the 9/11 attacks as the paradigmatic example of a non-state actor attacking one state while operating from the territory of another state. The international community has been practically unanimous that the US invasion of Afghanistan was a lawful exercise of self-defence, though some have expressed certain doubts as to the proportionality of the regime-toppling intervention. If the US invasion is accepted as lawful, as I think it must be, there are two, and only two, possible explanations for its lawfulness:
(1) Article 51 requires that the ‘armed attack’ be attributable to a state, thereby engaging its responsibility. Therefore, the 9/11 attacks must have been attributable to the state of Afghanistan. However, the general rules of attribution of acts of non-state actors to states, as articulated by the ILC in its Articles on State Responsibility and by the ICJ in the Nicaragua and Genocide cases, do not allow for a reasonable interpretation that would attribute the 9/11 attacks to Afghanistan, because they require proof that Afghanistan either (a) had complete control over Al-Qaeda, rendering it a de facto state organ; or (b) that Afghanistan had effective control over Al-Qaeda’s conduct in question, i.e. the 9/11 attacks. Since there is no proof of either – indeed, far from it – the 9/11 attacks cannot be attributed to Afghanistan under the general rules. But, because the US invasion was a lawful exercise of self-defence (so we all agree), the attacks MUST have been attributable to Afghanistan. Therefore, the general rules of attribution have either changed, or lex specialis rules of attribution have emerged, whether confined to ‘terrorist’ armed attacks or to the jus ad bellum more broadly, to allow for looser standard of attribution, such as harbouring terrorists or complicity in their actions.
(2) Article 51 does NOT require the attribution of the armed attack by a non-state actor to a state. Rather, for the attacked state to respond against the non-state actor which is operating in another state, the conduct of this latter state must be such to justify the ensuing violation of its sovereignty. Various authors have proposed different standards, but three scenarios seem generally possible, on a scale from the most to the least stringent: (a) the territorial state was complicit or was actively supporting the non-state actor in its armed attack; (b) the territorial state failed to exercise due diligence, i.e. it did not do all that it could reasonably have done to prevent the non-state actor from using its territory to mount an armed attack against another state, or is not doing all it can to prevent further attacks; (c) the territorial state may have exercised due diligence, but it was nonetheless unable to prevent the attack, or to prevent further attacks. The US post-9/11 invasion of Afghanistan could be quite comfortably justified under these standards.
One of these two options MUST be true if the US invasion was a lawful exercise of self-defense. No third option exists, as far as I can see. True, the two options may not lead to different ultimate results in the same factual situations. They are, however, conceptually significantly different and have other, broader implications. The first option couples the jus ad bellum and the law of state responsibility, while the second does not. If the first option is true, then the law of state responsibility must change whenever the jus ad bellum needs to accommodate a more lax standard of self-defense in response to armed attacks by non-state actors. And this is precisely what Christian argues in his initial EJIL piece – that a less stringent rule of attribution which is specific to the jus ad bellum context has evolved. Similarly, see Steve Ratner’s article in 96 AJIL 905 (2002).
Personally, I find the second option to be significantly more preferable to the first, because it maintains the distinction between primary and secondary rules that was famously introduced by the ILC project on state responsibility. That distinction serves several purposes. It helps us solidify general rules of state responsibility while avoiding politically sensitive questions, such as initially rules on treatment of aliens, or now self-defense against non-state actors. It relegates the discussion of these politically thorny questions to the body of primary rules that regulate them – here to the jus ad bellum. Thus, for instance, Kimberley has persuasively argued in her ICLQ article that it is the customary necessity condition for the lawfulness of self-defense that governs an attacked state’s response to an armed attack by a non-state actor operating from another state. We may disagree in precisely what circumstances on the (a) – (c) spectrum that I have outlined above such necessity would exist, but this is precisely the area where such debates should be had, instead of tinkering with the law of state responsibility. Finally, there is an inherent value in having general, coherent secondary rules of attribution which apply irrespective of the primary rules that are being applied, as this enhances the systemic character of international law, and helps it resist fragmentationist impulses.
In other words, it is as a general matter undesirable to have rules of attribution which are primary rule-dependent. Of course, that this is undesirable doesn’t mean it can’t happen. Article 55 of the ILC ASR expressly allows for secondary rules which are lex specialis. But the rules of attribution are precisely the core concept of the law of state responsibility, where lex specialis should not be accepted lightly. This is especially so because it is hard to justify in principle, say, why there should be special rules of attribution when it comes to the use of force by non-state actors pure and simple, but no such rules when it comes to how that force is used, e.g. in regard of genocide or crimes against humanity, as in the Bosnian Genocide case.
Be that as it may, my main point is this: even though I strongly prefer the second option, I still have to say that both options are on the table. For the time being at least, it is objectively impossible to tell which option is valid from an examination of state practice, opinio juris, and the case law. The ICJ has expressly reserved its position on whether Article 51 requires attribution of the armed attack to a state in Congo v. Uganda, despite its earlier pronouncements in Nicaragua and in the Wall case (the latter in my view being better explained by the fact that the armed attacks in question emanated from non-state actors operating from Israeli-occupied territory, thereby disengaging Article 2(4), and consequently Article 51). As for state practice, the official pronouncements of states in the relevant cases are far too ambiguous to say one way or the other whether states consider that Article 51 requires attribution, but that lex specialis rules of attribution have emerged, or instead that no attribution requirement exists.
Take for example the 2006 Lebanon war. Israel was attacked by Hezbollah, a non-state actor operating from within Lebanon. Israel responded by invading Lebanon, thereby engaging Article 2(4), and requiring a justification under Article 51. The reaction by states and other actors has generally been that Israel was entitled in principle to act in self-defense, but that its actions were disproportionate (e.g. because it did not exclusively target Hezbollah). But neither the Israeli official position on its use of force nor the international reaction to it were clear enough to conclusively decide between options (1) and (2).
For example, on 12 July 2006 Israeli Prime Minister Ehud Olmert made the following statement:
This morning, actions were carried out against IDF soldiers in the north. At this time, the security forces are operating in Lebanese territory. The cabinet will convene this evening in order to approve the continuation of the activity.
I want to make it clear: This morning’s events were not a terrorist attack, but the action of a sovereign state that attacked Israel for no reason and without provocation. The Lebanese government, of which Hizbullah is a member, is trying to undermine regional stability. Lebanon is responsible and Lebanon will bear the consequences of its actions.
Here we obviously have a claim by Israel that the armed attacks against it from Lebanese territory were attributable to Lebanon as a matter of state responsibility. But only a few days later, on 16 July 2006, after the Lebanese government denied that it was responsible for the attack, the Israeli position was different:
Prime Minister Olmert emphasized that Israel is not fighting Lebanon but the terrorist element there, led by Nasrallah and his cohorts, who have made Lebanon a hostage and created Syrian- and Iranian-sponsored terrorist enclaves of murder.
Here Lebanon was no longer the author of attacks and responsible for them, but the ‘hostage’ of a non-state actor. This to me sounds a lot like option (2), but there is nonetheless no way to say this with certainty. And the reactions by other states are even more ambiguous (see, e.g. the Security Council debates in S.PV/5489), though to me it seems that many states have accepted Israel’s claim to self-defense without saying that Lebanon was responsible for the attack.
In other words, not even the state using force, Israel, actually articulated a clear position either in favour of option (1) or option (2). Nor did for, that matter, the United States when it invaded Afghanistan. And when the states responding in self-defense refrain from articulating a clear position – probably because they believe that there is some benefit in the ambiguity – then the reactions by other states are almost by definition useless in choosing between the two options. Sure, these states might tell us whether they thought that the use of force was lawful, but they would not explain the theory on which they base their conclusion.
To me, therefore, the conceptual questions of self-defense against non-state actors remain utterly indeterminate while state practice remains as it is. We as lawyers may certainly choose between the two options, but we do so solely – solely – on policy grounds. There’s nothing wrong with that, mind you, and all of the reasons I’ve given above favouring option (2) are no more, and no less, than reasons of policy. And if this is so, there’s also nothing wrong in openly acknowledging the law’s indeterminacy.