Question 2: Does the right of self-defence relate only to an attack from another state, or does it also relate to attacks from non-state actors, e.g. a terrorist group, and if so, under what conditions?
When the focus is directed at the response to an actual armed attack, there seems no reason to limit the right of self-defence to an attack by another state (presumably this is what the question means, i.e. it looks to the person of the attacker rather than the geographical origin of the attack). There is nothing in the text of Article 51 to demand, or even to suggest, such a reading, and logic would be decisively against it. Granted a similar ‘attack’, why should a state’s legal capacity to protect itself depend on the identity of the attacker? To the extent that the ICJ may be thought to have suggested something different in the Wall Advisory Opinion, this should be disapproved. The criteria which emerged following the Caroline34 incident were enunciated in the context of a marauding armed band, not orthodox state-to-state conflict. The necessity and proportionality criteria are perfectly capable of adapting themselves to the foreseeable variety of possible cases; other limitations are covered under question 7 below.
There is no particular reason why a state, confronted with a genuine question of imminence (see under question 4 below), should first have to enter into an investigation of the extent to which the particular attack was ‘attributable’ (e.g. in the state responsibility sense) to the state from the territory of which the attack emanated, or represented a conscious policy decision at the highest level, etc., before its entitlement to an immediate protective response comes into play. Questions of that kind are by no means negligible, and in some cases may take a primary position; in others, they may be more relevant to the second move – or to a longer-term strategy – than to the immediate response.
In the light of the Wall Advisory Opinion, there is some doubt as to whether the right of self-defence relates to attacks by non-state actors. Although the Court sought to limit the scope of its analysis to the Israel-Palestine situation, it is evident from the Separate Opinions of Judges Kooijmans and Higgins that the Court’s thinking went beyond the case before them and challenged the appreciation which informed Security Council Resolutions 1368 (2001) and 1373 (2001).
It is evident, however, from the Nicaragua Judgment, that the Court acknowledges the existence of a right of self-defence against non-state actors if they receive state support. This is consistent with the Security Council’s appreciation in respect of Afghanistan.
In my view, if the law is to be credible, a right of self-defence must be acknowledged against non-state actors conducting themselves from foreign territory in circumstances in which they use or threaten force illegally and the State on whose territory they are based (a) actively supports the group, or (b) takes no effective action to forestall the use or threat of force by the group, or (c) is unable to take effective action to forestall the use or threat of force by the group. Once again, the appropriate principles to limit the scale of any response to an armed attack are the principles of necessity and proportionality.
Only if one disregards UN Security Council resolutions 1368 and 1373, as well as the extensive support for the approach taken by the US in response to the attacks of 11 September 2001.
In my view, it can definitely stem from a terrorist group or other non-state actor. To the extent that the ICJ suggested the contrary in the Wall Advisory Opinion, it was just wrong and its approach is manifestly at odds with state practice in the aftermath of the attacks of 11 September 2001. The text of Article 51 does not contain anything to suggest that an armed attack must emanate from a state. The Caroline incident was all about attacks by non-state actors. Most laymen would think international lawyers were mad if they believed that there was no right of self-defence against terrorist attack. Nor am I in the least persuaded by the argument based on the French text. True the French text of Article 51 uses the term “aggression armée”, and “aggression” is also the term used in Article 39 but (a) the French Government apparently accepted during the debates on the definition of aggression that aggression in Article 39 was not the same concept as aggression armée in Article 51; and (b) the English, Chinese and Spanish texts of the Charter use different terms for Articles 39 and 51.
In my opinion, if the use of force by a terrorist group reaches the level of intensity needed for it to be classed as an armed attack if it had been carried out by a state, then it is to be treated as an armed attack for Article 51 purposes. That was the almost universal reaction to the attacks of 11 September 2001 amongst governments and international bodies35.
Of course, the above analysis does not imply that there is then a right for the state attacked to use force against another state or in the territory of another state. I thought it was lawful to do so in Afghanistan in 2001 because of the scale of Afghan support for Al-Qaeda but such action would only be justified in exceptional circumstances.
The right of self-defence is a right to use force to avert an attack. The source of the attack, whether a state or a non-state actor, is irrelevant to the existence of the right. No-one is obliged by international law passively to accept an attack. The character of the source does, however, affect the measures that can be taken in response. Broadly, if the threat of the attack is made by a state, a response against that state within the bounds of proportionality, as set out following the Caroline incident, is lawful. If the threat emanates from a non-state actor, no forcible action in self-defence is lawful if the state in which the actor is located is able and willing to take reasonable measures to nullify the threat.
Much of the concern with the responsibility for attacks seems to me to stem from the mistaken belief that force can only be used outside the territory of a state against an attacker if the attack emanates from another state that is thereby in breach of international law. That seems to me to be wrong. Force may be used to avert a threat because no-one, and no state, is obliged by law passively to suffer the delivery of an attack. That is what it means to say that the right is ‘inherent’. Defensive force is in no sense dependent upon the attack being a violation of international law.
If the attack does come from another state, the principles of necessity and proportionality as defined following the Caroline incident will clearly apply. I think that they apply also if the attack comes from someone other than a state. In the latter case, however, it would be an unjustified violation of the sovereignty of the state from which the attack emanates if defensive force were used in circumstances where that state was able and willing to take effective action that would neutralise the attack.
Sir Adam Roberts
Although there are hazards in doing so, there is a strong case for recognising openly that a major terrorist attack, and/or sustained terrorist campaign, can constitute an “armed attack”.