10 Mar Mea Culpa Regarding Israel’s Attacks on Hezbollah in 2006
In a number of posts (see, for example, here and here), I have claimed that the League of Arab States (LAS) formally rejected the “unwilling or unable” test in the context of Israel’s 2006 attacks on Hezbollah in Lebanon. Thanks to comments by Ori and Tom Ruys on the most recent post, I now realize I have been guilty of the same kind of methodological sloppiness that characterizes most scholarly work in defence of the test. If you read the statement by the LAS — you can find it here — there is no way to determine whether the it denounced Israel’s attack because it rejected the “unwilling or unable” test or — and this actually seems more likely — because it simply rejected Israel’s claim that it was acting in self-defence. (I disagree with Ori that the statement can be read as an indictment of Israel solely for using disproportionate force in self-defence.) And if we cannot determine the precise reason why LAS rejected Israel’s self-defence claim, that rejection obviously cannot provide opinio juris against the “unwilling or unable” test.
That said, loathe though I am to disagree with Tom, I don’t see the international response to Israel’s attacks on Hezbollah in Lebanon as supporting the “unwilling or unable” test. Most obviously, Israel claimed that Hezbollah’s actions were attributable to Lebanon — it did not invoke the test at all. Moreover, no state specifically invoked “unwilling or unable” during the Security Council debate over Israel’s actions — some expressed concern over Lebanon’s failure to exercise effective control over the entirety its territory, but a number of those states attributed that failure to Israel’s occupation of southern Lebanon, not to Hezbollah’s actions. So I agree with Olivier Corten that “these standpoints are highly ambiguous and so it seems a very difficult business to deduce from them any opinio juris.”
My thanks to Ori and Tom for weighing in — and to Ori for providing links to the relevant documents. Apologies to readers for being so sloppy. I just hope my lack of care will not distract from my basic point, which is that scholars who claim that the “unwilling or unable” test represents customary international law have failed to identify (anywhere near) sufficient significant state practice or opinio juris in defense of their position.
Thanks for that post, Kevin. Still, now that I understand your argument — namely, that the “inherent right of individual or collective self-defence if an armed attack occurs against a Member of the United Nations” incorporated into Article 51 simply does not include self-defense against a nonstate attacker unless the host state is complicit (in a strong sense) in the attack — I honestly don’t see what the basis for it might be. That was exactly the scenario in the Caroline case itself, in which the U.S. and Britain agreed that Britain had such a right to attack the rebels *if* the U.S. could not but an end to the rebel attacks. As Ashley demonstrates, there are plenty of pre-Charter historical precedents — presumably incorporated in the Art. 51 reference — and more examples still after the Charter. There is no obvious reason why any states would have ever agreed that they must stand by helpless if attacked by a group based in a state that is unable and unwilling to stop the attacks. And I’m not aware of any states — let alone sufficient states to establish a new prohibition — articulating the official view that a state may… Read more »
Marty,
I quoted and linked to Christian Tams’ article in response to your last comment, in which he explains (and his article provides plenty of cites) why, prior to 9/11, the overwhelming position of states (and, of course, the ICJ in Nicaragua) was that there was no right of self-defence against a non-state actor whose attacks were not in some way attributable to a state.
As for Caroline, I’ll simply repeat what the International Law Commission said in its ASR Commentary (p. 196): ‘The Caroline incident of 1837, though frequently referred to as an instance of self-defence, really involved the plea of necessity at a time when the law concerning the use of force had a quite different basis than it now has.”
Another good article discussing the traditional attribution requirement can be found here. These are the key paragraphs — and note that the author reaches the same position as Tams, which is that attribution to a state is still required, but the standard of attribution articulated by the ICJ (effective control) has been relaxed post-9/11: Currently, the most significant issue in the law of self-defence is whether or not an armed attack must be attributable to the State that is the prospective target of an exercise of self-defence, and if so, what the test for that attribution requirement might be. Three primary views have emerged. First, there is the view that the attribution test set out by the ICJ in the Nicaragua case must be satisfied: the State subjected to an exercise of self-defence must have had ‘effective control’ over the conduct of the individuals who carried out the attacks. Second, some argue that no attribution is required. That is, States may exercise self-defence against any State and within any State’s borders, regardless of whether or not that State is in any way responsible for the attacks. Third, there is the view that, whilst attribution is required, the test for attribution… Read more »
Marty: re: The Caroline, the UK claim was that it was entitled to engage in measures of self-defense against NSA armed attacks that were ongoing, just as if rebels had erected a shore battery on U.S. soil and were firing across the border, killing persons in Canada — not “if” the U.S. was unable. And Kevin, the ILC was wrong (perhaps in ways that you have addressed here and re: the so-called unwilling/unable limitation on the inherent right of, what did the UK constantly say, “self-defense”!!
please see http://ssrn.com/abstract=2459649 and http://ssrn.com/abstract=2402414 (also re: when does an armed attack begin).
Thanks, Kevin. I look forward to reading the Tams and Michael articles thoroughly when I get a chance. But on a very quick first glance, the Tams hardly supports your characterization of “the overwhelming position of states [in 1989] that there was no right of self-defence against a non-state actor whose attacks were not in some way attributable to a state.” He appears to rely principally on the UN’s condemnations of the US action in Libya in 1986 and the 1985 Israeli raid on Tunis — but he does not say whether those condemnations were based upon the view that such attacks on nonstate actors are always prohibited (as opposed to, e.g., the lack of the prerequisite armed attack by a nonstate group, or the failure of Israel and/or the U.S. to comply with necessity and proportionality, etc.). The excerpt from Michael that you quote appears to be concerned with the distinct question of when a state can be made “the prospective *target* of an exercise of self-defence,” such as when the attacked state uses force against the host state’s forces or military targets — see, e.g., the U.S. actions against the Taliban circa 2001. And I agree — who… Read more »
Marty,
If Tams does not provide enough state practice to convince you, I suggest you read chapter 5 of Tom Ruys’ book, which reaches the same conclusion as Tams regarding the state of the jus ad bellum prior to 9/11.
As for your second comment, the idea that targeting a non-state actor does not involve targeting the state on which the non-state actor is located is largely an invention of scholars trying to expand the right of self-defence. Article 2(4) protects the “territorial integrity” of states; that integrity is violated regardless of whether the state acting in “self-defence” does not believe the territorial state is involved in the non-state actor’s attacks.
Finally, a question for you: why would states in the global south have ever endorsed a conception of self-defence that would allow their territorial sovereignty to be violated whenever a state in the global north decided they were “unwilling” or “unable” to prevent terrorist attacks? After all, it is their territory, not the territory of states in the global north, that will always be the object of such “self-defence.”
Just checked the two UN resolutions that Tams principally relies upon. SC Res. 573, involving the 1985 Israeli attack on Tunis, did not offer an explanation for its condemnation, and, in particular, did not even mention, let alone repudiate, any self-defense claim, and did not discuss whether self-defense against nonstate actors is ever legitimate. GA Res. 41/38, respecting the U.S. attack on Libya in 1986, is even less apposite, on several grounds: It involved an attack on a state, not on a nonstate actor. And again, the resolution did not address self-defense. Reportedly, the objection by some nations was based on the absence of sufficient proof that the nightclub bombings were an armed attack, and/or that Libya was responsible for them. Finally, of course, many nations thought the U.S.’s response was lawful. Neither resolution, then, even discusses whether there is a “right of self-defence against a non-state actor whose attacks were not in some way attributable to a state,” let alone establishes an “overwhelming position of states [in 1989]” that there is no such right. So, I’m still looking for evidence that *any* state publicly adopted such a view — after which, we can consider whether those statements (if any)… Read more »
Kevin: OK, since Tams likely does not offer *any* state practice to support your claim, I’ll gladly read Ruys next.
On the 2(4) question, let me be clear: The use of force against a nonstate group that has taken sanctuary in State X certainly *would be* a 2(4) violation against State X, even if not directed at the state as such, if there were no legitimate self-defense claim. My point is simply that in deciding whether the use of force is a necessary and proportional response in self-defense to an attack by a nonstate group, the force ought to be targeted at the nonstate group. The force presumptively may not be directed at State X’s own forces or military targets absent evidence that the attack by the nonstate group was in some fashion attributable to State X.
Finally, I have not addressed at all the very hard question, stretching back at least as far as the Caroline, as to *who decides* that the host state is U&U, and what the evidentiary predicate must be. Those are, indeed, the hard questions, which is why they are the focus of Ashley’s excellent article.
I don’t understand your interpretation of SC Res. 573. In response to the PLO hijacking a yacht and murdering three Israelis (an armed attack), Israel attacked the headquarters of the PLO, a non-state actor, in Tunisia. There was no claim by Israel or any other state that the PLO’s actions were attributable to Tunisia. The SC rejected the idea that Israel had the right to act in self-defence, describing its actions as an “act of armed aggression perpetrated by Israel against Tunisian territory in flagrant violation of the Charter of the United Nations.” How is that not an affirmation of the attribution requirement? I suppose you could argue that the SC rejected Israel’s argument because it did not believe the PLO had engaged in an armed attack in Israel, but I don’t think such an interpretation is particularly compelling.
Re: The Caroline: “a claimed [British] right of “self-defense” and “self-preservation” against prior and ongoing armed attacks by insurgents against British rule in Canada and complicitous conduct of the U.S. vessel Caroline that had also already occurred. The United States had claimed that there was a very strict limitation on particular methods of responsive force as opposed to when the right of self-defense pertains. No one disagreed that non-state actor armed attacks trigger the right to engage in certain measures of self-defense and that armed attacks had already occurred, but the U.S. claimed that use of a particular means of self-defense when the right of self-defense had been triggered should only be permissible when the “‘necessity of that self-defense is instant, overwhelming, and leaving no choice of means, and no moment for deliberation;’” and when it can be shown that the authorities responding to an attack “did nothing unreasonable or excessive; since the act, justified by the necessity of self-defense must be limited by that necessity, and kept clearly within it.” The U.S. had claimed that under the circumstances the British could have waited until daylight and seized the vessel Caroline when it re-entered Canadian waters and, therefore, that the… Read more »
I don’t really know why the Security Council voted that way, Kevin. The resolution itself does not say. (The U.S. abstained and apparently expressed the view that self-defense could be appropriate.) A quick look at the literature suggests that the other voting members might have relied upon at least three other grounds:
— that Israel lacked proof of the PLO’s culpability
— that the Israeli response was designed to be retaliatory rather than to prevent future attacks
and
— according to Jochen Frowein (apparently; I haven’t read the original source), the Security Council was of the view that “an armed attack cannot consist of a terrorist action against citizens on foreign territory,” as was the case with the PLO attack in Cyprus.
Marty, Regarding your distinction between targeting a non-state actor and targeting the state that is “unwilling or unable” to prevent that non-state actor group from using its territory to launch armed attacks, here is what James Green has to say in his excellent book The International Court of Justice and Self-Defence in International Law (p. 49): It is worth noting here that an interpretation of the Court’s jurisprudence on this issue has been advanced that makes a distinction between, on the one hand, a response taken by the victim state against a host state when only the non-state actors operating within that state are targeted and, on the other hand, actions in self-defence when the government or forces of the host state itself are additionally targeted. It has been argued that in Nicaragua, Israeli Wall and DRC v Uganda, the Court has held that there must be a level of collaboration on the part of the host state in the latter situation but not in the former. In cases in which the response is directed solely at, say, a terrorist military base, all that is required is that the ‘host’ state was unable or unwilling to act. This has a… Read more »
My understanding is that, at least since the Caroline case, international law has not prohibited a state from acting in self-defense against a non-state actor that launches an armed attack from another state, *assuming* that the self-defensive actions otherwise satisfy the requirements of necessity and proportionality — which in some cases may well prohibit actions against the host state forces as such, requiring that the force to targeted at the nonstate actor, and which will presumably prohibit the use of force where the host state is willing and able to prevent future attacks — as well as other legal constraints, such as the jus in bello and any applicable human rights limitations. You still have not pointed to a single state — much less an “overwhelming position of states” — that has expressly rejected this view.
sorry, in that last comment “to” should be “be” — “force be targeted”
I will leave it to readers to decide whether your position is consistent with the evolution of the jus ad bellum since the Caroline. I will simply end my contribution to this debate by noting my agreement with the conclusion of Chapter 5 to Tom Ruys’ magisterial book, which is based on a close textual analysis of essentially every case of self-defence in the post-Charter era: Having examined the relevant customary practice and ICJ case law, we must again confront our key question: to what extent do cross-border attacks by non-State actors amount to ‘armed attacks’, warranting a defensive response against the non-State presence abroad and possibly against the infrastructure of the State from whose territory the attacks were prepared, directed and/or launched? In the past, as we have seen, this question was answered in a restrictive manner. A general right of hot pursuit was never recognized, and – despite the open-ended phrasing of Article 51 – purely ‘private’ attacks were excluded from its scope. Rather, it was agreed that Article 51 only applied when close links existed between the non-State actor and the latter State, which in fact employed it as an instrument of proxy warfare. Article 3(g) of… Read more »
You still have not pointed to a single state — much less an “overwhelming position of states” — that has expressly rejected this view.
The Definition of Aggression to which Ruys refers in his conclusion, which explicitly requires attribution (and thereby rejects “unwilling or unable”), is UNGA Res. 3314, adopted by consensus in 1974. UNGA resolutions count, of course, as state practice.
The Ruys excerpt speaks volumes, Kevin:
Let’s see . . .
On the one hand, we have consistent state practice, including a “considerable number of interventions,” plus “numerous security doctrines and official statements” of numerous states.
And on the other hand, we have . . . a curious, repeated use of the passive voice (“this question was answered in a restrictive manner’; “a general right of hot pursuit was never recognized”; “it was agreed”), with no references to any state practice or state opnio juris.
(Art. 3(g) of Res. 3314 merely indicates that State A is guilty of an act of aggression if armed bands, groups, irregulars or mercenaries carry out certain acts of armed force against State B when they were “sent by,” or “acting on behalf of,” State A. It doesn’t say a thing about whether and in what circumstances IL prohibits State B from attacking the armed group in the territory of State B.)
As I said, I will leave it to readers to decide whether Marty or the ICJ and all those scholars who have reached precisely the same conclusion about the attribution requirement — including Ruys, who cites numerous examples of state practice, contrary to Marty’s claims — have the better of the argument.
Thanks for another informative post, I suggested in the past in an article http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1744190 that there some other examples to the unable and unwilling test, (including the 2006 war) such as the attack by French forces in kolwezi congo
Ilan,
You write in your article (p. 22): “The Zairian despot Mobutu Sese-Seko requested foreign assistance from Belgium, France and the United States. The French Foreign Legion with Belgian forces mounted a rescue operation which involved a large military force that operated on Zairian territory alongside local military forces.” If Zaire (now the DRC) consented to the operation, the rescue operation was not self-defence under Art. 51. The example thus provides no support for the “unwilling or unable” test.
UN GA Resolutions are NOT relevant “state practice”! They are at best reflections of patterns of opinio juris. The “practice” would involve conduct (acts and omissions) supportive or thwarting of the norm (e.g., the prohibition of genocide). Moreover, “states” have never been the only actors with formal participatory roles in international law. It is professionally irresponsible to focus merely on the practice of “states”!
Marty is correct re: the GA Dec. on Aggression, which reqs. a state actor violation of the Charter and does not address self-defense as such.
Marty is correct re: “reprisal” actions re: PLO/Tunisia (not self-defense)
I always wondered, why persist to speak of “unwilling or unable” as if both situations carry the same implications? Does the general principle of necessity in self-defence not require the victim State to offer its assistance to the territorial State when the latter is unable – yet willing – to deal with non-State actors operating from its territory?
A State which is merely unable (e.g. due to largely collapsed governmental infrastructure) should be offered an opportunity to remedy the situation with outside help. Acting in self-defence before offering assistance consented to by the territorial State, seems unnecessary.
Only if assistance is neither sought nor accepted, then the territorial State becomes unwilling.
Remy: That is exactly the sort of question we ought to be asking. And it is exactly the sort of question that Ashley Deeks considers in her article, which I commend to you:
http://www.vjil.org/assets/pdfs/vol52/issue3/Deeks_Post_Production.pdf
Remy: I completely agree with Marty that Deeks does a very good job exploring what the unwilling/unable test might look like in practice. But, of course, she does not even remotely establish her claim that the test reflects customary international law. On the contrary, as I have pointed out, she openly admits she “found no cases in which states clearly assert that they follow the test out of a sense of legal obligation.” (The UK and US have made such assertions since she published her essay.)
Yes, but I’d bet every state would acknowledge that it is obligated to adhere to the requirement of necessity under the jus ad bellum — and the U/U precondition is merely one aspect of that necessity requirement.
Nope — U/U is an aspect of whether there is an “armed attack” within the meaning of Art. 51, which is a condition precedent to any consideration of necessity or proportionality.
“U/U is an aspect of whether there is an “armed attack” within the meaning of Art. 51, which is a condition precedent to any consideration of necessity or proportionality.”
This is simply wrong. The *armed attack* by the nonstate actor is the armed attack. (The question is *not* whether the host state has attacked, or is responsible for it. The hypo in question assumes that the state had nothing to do with it.) Whether the host State is U/U to prevent further attacks is merely a question that is relevant to whether it is permissible to strike the nonstate actor on the territory of the host state.
It seems that U/U preference is only relevant as a theoretical construct if one claims that Article 51 does not restrict self-defense to what it expressly declares (“if an armed attack occurs”). Once an armed attack “occurs” the “necessity” circ. has been met. So this entire discussion would only be relevant if we changed the text of the UN Charter to allow either anticipatory self-defense (when an armed attack has not begun but the attack is imminent) or GW Bush’s “preemptive” self-defense (which is even more restrained than Obama’s unacceptable “imminent threat” test b/c an “imminent” “threat” is not yet a real threat, much less an imminent attack).
To make this more vivid, assume the following:
Nonstate armed group A attacks State X from State Y (or on the land of State X itself), at a time when A has *nothing to do* with State Z.
Group A then, in order to avoid being attacked by State X, and to be able to plan and conduct its future attacks on State X, repairs to an “ungoverned and ungovernable” area of State Z, without any assistance from State Z, and much to its chagrin.
State A wishes to strike Group A within State Z, in order to prevent further attacks. State Z has no way of dealing with the threat from Group A, but also does not consent to State A’s use of force.
IL does not prohibit State A from acting (in a proportionate way, consistent with the jus in bello, etc.) in this case, even though State Z was not in any way responsible for the past attacks.
p.s., once an “armed attack occurs,” the state from which the NSA armed attacks emanate would be decidedly “unable” with respect to the armed attack and/or continual armed attacks. http://ssrn.com/abstract=2402414
Marty: State X can respond in self-defense to continual NSA armed attack (use a movie camera instead of a snapshot to become aware of the fact that attacks are continuous — http://ssrn.com/abstract=2402414 )
State A, however, has not been subject to a NSA armed attack and has no right under the express language in Article 51 to attack the NSA in Sate Z.
sorry Marty, the name should have been “Jordan”
And State A could engage Group A in State Z if State A has “consent” from State X to participate in collective self-defense (as does the U.S. etc. from Iraq).
Marty, thank you for the article. My point was merely, if there is any change in jus ad bellum (any additional justification or ground for SD) it should be named appropriately. Mere inability is meaningless; it is unwillingness that matters. To speak of a “U/U-test” carries the suggestion that both situations have equal implications for the existence of SD. More generally, “even though State Z was not in any way responsible for the past attacks.” Isn’t that exactly the crucial point of disagreement? Even if one accepts that there is a U/U-rule or principle, then one might still disagree on how fits in with more tradtional and established notions of state responsibility and jus ad bellum. Is U/U is a relaxation of the traditional jus ad bellum in the sense that only States can commit armed attacks, yet with U/U being an extension of the concept of “armed attack”: the State commits an armed attack as a result of its unwillingness. (Debatable whether this is actually a relaxation of the procedural rules of attribution [U/U instead of overall/effective control], or more properly a reappraisal of what materially constitutes an armed attack [originally understood to encompass only action, now perhaps also… Read more »
Remy: international law has never been merely state-to-state (“traditional” or otherwise). Armed attacks by other actors has been part of the “traditional” jus ad bellum. Consider the formal roles, even “wars” with, actors such as nations, peoples, tribes, “belligerents” (http://ssrn.com/abstract=1701992 ) Nothing in UN Article 51 requires that an armed attack occur by a “state.”
“Armed attack” means any attack with arms — kind of like “Exchange established by the state” means only exchanges established by a state.