The question of the legality of ‘interceptive’, ‘anticipatory’, ‘pre-emptive’ or ‘pre- ventive’ self-defence is one of the most debated topics in international law and has been for over fifty years. Whether armed reprisals or punitive measures are legal is also a well-trodden path in writings since the inception of the United Nations.
Both questions raise a host of issues, including the aim or goal of self-defence (Section 2.5), whether there is a ‘black hole’ (Section 2.1) or the conditionality of an armed attack (Section 2.2.1). The crucial point, however, is how an armed
100 Some lawyers thought that the Court wanted to differentiate between individual response – where forceful actions are allegedly allowed – and collective response, where this is not so. See e.g. Tom J. Farer, Drawing the right line. Appraisals of the ICJ’s decision: Nicaragua v. United States (Merits), 81 American Journal of International Law (1987) 112–116 at 113.
101 Kreò (1995) supra note 41 at 187–195.
102 UN Charter, Preamble (para 1). See the importance attributed to the phrase ‘scourge of war’ by Klein (1964) supra note 28.
Uncertainty in International Law 24
attack needs to be present. It can be argued that if one admits that ‘armed attack’
is a condition for the lawful exercise of self-defence, all the concepts mentioned above refer to illicit forms of response. The only armed attack activating self- defence is a present armed attack. Neither an imminent nor a concluded armed attack suffices, because the condition makes the ending of an armed attack the only valid objective of self-defence under Article 51.103 Therefore, most early proponents of anticipatory self-defence base their support for that concept on a denial of the conditionality of an armed attack, which most, in turn, rest upon the
‘black hole’ theory.
There is an interesting line of thought especially among international lawyers educated in Germany that requires an armed attack, but also argues that self- defence is available against an imminent armed attack.104 It makes no sense to repeat the arguments of Sections 2.1 and 2.2.1 if no further conclusions can be drawn. Hence, we will ask whether the ‘presence’ demanded by Article 51 can be squared with the sufficiency of ‘imminence’. A domestic law analogy may be at the root of this conception. Just like Article 51 UN Charter, Section 32(2) of the German Penal Code only speaks of a present attack: ‘Necessary defence is the defence which is required to avert a present unlawful attack from oneself or another.’105 However, the word ‘gegenwọrtig’ in the German original is inter- preted by domestic criminal law doctrine to mean ‘imminent and present’. This would explain why the ‘presence’ criterion in Article 51 (‘if an armed attack occurs’) is seen in such a light by German scholars, for it can be argued that the terms used there are analogous to the wording of Section 32(2). ‘Occurs’
includes ‘is imminent’ just as much as ‘gegenwọrtig’ includes ‘unmittelbar bev- orstehend’. The municipal criminal codes of two other German-speaking coun- tries expressly include imminent attacks, which points to the conclusion that, for Austrian lawyers at least, ‘present’ does not include imminent: Section 3(1) of the Austrian Penal Code reads: ‘It is not unlawful to use such [means of] defence necessary to repel a present or imminent unlawful attack’.106 The law in Switzerland is similarly worded: ‘If somebody is being attacked or threatened by an unlawful
103 See Brownlie (1962) supra note 20 at 242 for a list of authors believing that the ordinary meaning precludes preventive action.
104 Alexandrov (1996) supra note 20 at 163–165; Dahm (1962) supra note 28; Dinstein (2005) supra note 9 at 190–191; Donner (1995) supra note 28 at 180; Hailbronner (1986) supra note 2 at 81;
Higgins (1962) supra note 6 at 301 (the basis for her support is unclear.); Hanspeter Neuhold, Internationale Konflikte – verbotene und erlaubte Mittel ihrer Austragung (1977) 137–138; (in effect also:) Verdross and Simma (1984) supra note 28 at 287–289; Wengler (1967) supra note 28 at 5–6; Wildhaber (1971) supra note 28 at 153.
105 ‘Notwehr ist die Verteidigung, die erforderlich ist, um einen gegenwọrtigen rechtswidrigen Angriff von sich oder einem anderen abzuwenden.’ § 32 Abs. 2 Strafgesetzbuch vom 15. Mai 1871, RGBl. S. 127 (emphasis added).
106 ‘Nicht rechtswidrig handelt, wer sich nur der Verteidigung bedient, die notwendig ist, um einen gegenwọrtigen oder unmittelbar drohenden rechtswidrigen Angriff . . . abzuwehren.’ Đ 3 Abs 1 Bundesgesetz vom 23. Jọnner 1974 ỹber die mit gerichtlicher Strafe bedrohten Handlungen (Strafgesetzbuch – StGB), BGBl 1974/60 (emphasis added).
2.2.4 Self-defence 25
attack, the attacked and any other person is entitled to repel the attack by means appropriate to the situation.’107 However, as noted in the introduction to this chapter, the existence of a right of self-defence within one legal order does not necessitate the existence or a certain scope for self-defence in another legal order.108
How is the ‘slight expansion’ argument made? Frequently it is an assertion to the effect that one ought to admit self-defence in certain narrowly defined cases where prevention is necessary.109 This is sometimes ‘buttressed’ by an argumentum ab inconvenienti110 to the effect that ‘[i]n conditions of modern warfare it is unreasonable for a state to have to wait for an armed attack’.111 This is a political, not a legal argument and as such this kind of argument will be excluded, because this does not provide an insight into how the law is shaped. Law does not change because its application may be perceived as inconvenient.
If one assumes – as the majority of scholars do – that it is certain that an armed attack is a necessary condition of the exercise of self-defence under Article 51 UN Charter (Section 2.2.1), the key issue regarding the temporal scope of self-defence becomes what counts as ‘armed attack occurs’ and the key term becomes ‘occurs’.
The question, then, ‘which should be posed is not when is anticipatory action justified but, when has an attack occurred?’112 A typological classification of the precise moment (as distinct from a determination of the factual circumstances sufficient to constitute an ‘occurrence’) results in three logical possibilities: an armed attack occurs (1) when an armed attack begins (an attack is launched), (2) when it takes effect (the harm at the intended target starts occurring) or, even, (3) when it ends (the attack is consummated).
One of the features of the ‘sitting duck’ argument – states do not have to present themselves as stationary targets and may therefore take action before the attacker does – is that its proponents use a straw man. They impute that their opponents restrict the right of self-defence to a great degree and thus try to show that these doctrines are absurd. The view that ‘[s]ome authorities . . . have interpreted “if an armed attack occurs” to mean “after an armed attack has occurred” ’,113 does not reflect the other side’s view correctly. Indeed, if self-defence
107 ‘Wird jemand ohne Recht angegriffen oder unmittelbar mit einem Angriffe bedroht, so ist der Angegriffene und jeder andere berechtigt, den Angriff in einer den Umstọnden angemessenen Weise abzuwehren.’ Art. 15 Schweizerisches Strafgesetzbuch vom 21. Dezember 1937, AS 54 757 (emphasis added).
108 Josef Kunz acknowledges that an imminent attack suffices in municipal laws, but he is adamant as far as the Charter is concerned: ‘The “imminent” armed attack does not suffice under Art. 51.’
Kunz (1947) supra note 19 at 878.
109 Donner (1995) supra note 28 at 180; Hailbronner (1986) supra note 2 at 81; Wengler (1967) supra note 28 at 5–6: ‘Art. 51 is not to be taken quite literally’ ‘Art. 51 [ist] nicht ganz wửrtlich zu nehmen’; Wildhaber supra note 28 at (1971) 153.
110 Brownlie (1962) supra note 20 at 243.
111 Jennings and Watts (1992) supra note 22 at 422.
112 Brownlie (1962) supra note 20 at 258.
113 Waldock (1953) supra note 12 at 497.
Uncertainty in International Law 26
could only be employed after an attack it would defy the logic of self-defence: to repel an attack. A right of this sort would be legalised revenge.
Those scholars who admit action against imminent attacks as well as those who do not admit imminent action regard the beginning of an armed attack as the relevant moment for self-defence. The question thus becomes when an armed attack can be said to have begun. The approach taken by the ‘slight expansion’
doctrine is to interpret the attack to have begun earlier than orthodoxy would have done. Unlike the proponents for a wide right of self-defence, the ‘slight expansionists’ do not claim that self-defence is legal if an armed attack has not (yet) begun. It is thus a reinterpretation of the same point of reference, not its denial. This reinterpretation occurs by two means: by altering the meaning of the word ‘beginning’ and by a change in the way facts are interpreted.
There are four different, but frequently intertwined grounds on which the slight expansion is based. First, it is claimed that the presence of an animus belligerendi, the will to make war, constitutes the beginning of an armed attack.114 The reason for such an approach is the fact that frequently there is very little physical evidence of force in the time between the decision to attack and the attack; therefore, these scholars make the forming of a mens rea (the decision, not its execution) the relevant point of reference. The most expressive and convincing example is where the putative attacker gives notice of his intention to attack the future victim;
even restrictionists would tend to admit preventive actions in such a case.115 The second, difficult to distinguish, approach is a shift to the epistemological plane.
Instead of relying upon the fact of an attack occurring, ‘evidence’ of a future attack is demanded:
Where there is convincing evidence not merely of threats and potential danger but of an attack being actually mounted, then an armed attack may be said to have begun to occur, though it has not passed the frontier.116
However, what is in evidence here is not the armed attack itself, but the prediction of a future attack. A probability, not a certainty is evidenced – its perception or, rather, the epistemological viewpoint the law demands. That is the topic of Section 2.3 and will be discussed there.
A third and enticing variation is to interpret an armed attack as occurring as soon as there is an irreversible course towards the attack. Yoram Dinstein’s ‘inter- ceptive’ self-defence could be subsumed under this heading. Dinstein argues that
‘[i]nterceptive [unlike anticipatory] self-defence . . . takes place after the other side has committed itself to an armed attack in an ostensibly irrevocable way.’117
114 Verdross and Simma (1984) supra note 28 at 288–289.
115 Brownlie (1962) supra note 20 at 259 (not if the declaration is not accompanied by action;
equivocal acts, however, may become univocal); Randelzhofer (2002b) supra note 28 at 803–804 (MN 39).
116 Waldock (1953) supra note 12 at 498; similarly: O’Connell (2002) supra note 24 at 8–9.
117 Dinstein (2005) supra note 9 at 191.
2.2.4 Self-defence 27
This argument approaches the point of launching of an attack. Stanimir Alexandrov argues that this is the case when ‘the aggressor State “pulled the trigger”, . . . when there is no possibility of the aggressor State changing its mind’.118 He refers to imminent harm resulting from forceful action, rather than imminent attack,119 thus changing the nature of the ‘imminence’ required. Dinstein’s view, on the other hand, is broader. He does not require the firing of the first shot, but only a commitment by the other side – not the actions themselves, but the deci- sion. The difference is obvious in the application of the theories to historic facts:
whereas for Dinstein the United States could legally have destroyed the Japanese Fleet even before Japan launched its aeroplanes on Pearl Harbor (provided that this event were transposed to the post-Charter world), and whereas for him the Egyptian behaviour in 1967 signified that ‘Egypt was bent on an armed attack’,120 an application of Alexandrov’s concept would have meant that the United States could only have attacked the planes after they had taken off for their targets and the Israeli Defence Force could only have defended against Egyptian forma- tions moving to attack, even though they might not have passed the border. Ian Brownlie restricts the right even further than Dinstein and Alexandrov. In his opinion only rockets in flight may legally be intercepted in foreign airspace.
Brownlie does not allow this exception to be extended to aeroplanes,121 because his standard hinges on the last human decision versus the automatism of machines.
He demands a virtually unstoppable mechanism, the last human interaction for an armed attack to have occurred when the territorial inviolability of the target is not yet violated, e.g. ICBMs being launched or aeroplanes firing missiles or dropping bombs.
A fourth argument bases its claim partially on the fact that, because in certain cases counter-force would not be effective if one waited until the enemy had entered the territorial domain, the right of self-defence exists from an earlier moment.122 This, again, is a political argument. If, for example, a municipal statute fails to give effective protection against some danger, this in itself does not invalidate or change the statute. The same logic applies in international law: if the correct application of international law leads to a result deemed unsatisfactory, it is not the law’s validity that is in question (a normative problem), but the wisdom of its terms (hence a problem of legal politics).
We are once again reminded by this discussion that there is no definitive clarifi- cation of the issue, neither by the terms of Article 51 nor by its preparatory works.123 There are some difficulties with the various arguments presented above:
the solutions employing a mens rea of the state must grapple with the immense
118 Alexandrov (1996) supra note 20 at 164 (emphasis added).
119 Alexandrov (1996) supra note 20 at 165.
120 Dinstein (2005) supra note 9 at 192.
121 Brownlie (1962) supra note 20 at 259.
122 Brownlie (1962) supra note 20 at 259; Donner (1995) supra note 28 at 180; unclear: Jennings and Watts (1992) supra note 22 at 421–422.
123 Verdross and Simma (1984) supra note 28 at 288.
Uncertainty in International Law 28
difficulty of ascertaining the intentions of governments124 for such a linchpin concept of international law (similar, but more complex than the establishment of opinio iuris).125 Also, ‘imminence’ cannot be assessed by objective criteria and the law126 – if it were shaped thus – would have to take the claims of defending states at face value. Third, ‘there may in fact be no “last irrevocable act” ’127 which could give the protection the proponents of the ‘slight expansion’ seek. Fourth, to some extent all these constructions, all extensions or re-interpretations of the moment an armed attack is said to occur, involve the assumption that armed attacks can occur ‘constructively’,128 because the infringement is not yet fully manifested in physical form. However, there is a wide degree of latitude, because the text of Article 51 sets limited textual boundaries and because the factual uncertainty makes the doctrine intimately connected to particular cases and thus makes all constructions highly casuistic (Section 2.2.2).