Australia’s Attorney-General, the Hon. George Brandis QC, recently presented his government’s legal position on the use of force in national self-defense. The text of his speech was posted over on EJIL:Talk!, and should be of interest to many Just Security readers as well. Importantly, Brandis’ remarks are intended to

complement, and reflect, a similar approach to that of the Attorney-General of the United Kingdom, the Rt Hon Jeremy Wright QC MP, who on 11 January this year delivered a landmark speech … articulating the United Kingdom’s approach to the law of self-defence.

I offered some critical remarks on Wright’s speech here, here, and here. While Brandis’ speech improves on Wright’s in one respect, many of my criticisms of Wright’s speech apply to Brandis’ as well. Brandis’ speech also contains some new problems all its own.

To frame the discussion, recall that the United Nations Charter prohibits member States from using force against other States, while recognizing their inherent right of self-defense “if an armed attack occurs.” However, according to Brandis, “it is now recognised that customary international law permits self‑defence not only against an armed attack that has occurred but also against one that is imminent.” Brandis then offers his government’s view of the scope of this alleged customary law permission.

Like Wright, Brandis endorses the approach of Sir Daniel Bethlehem, according to which “whether an armed attack is imminent will fall to be assessed by reference to all relevant circumstances,” including:

  • the nature and immediacy of the threat;
  • the probability of an attack;
  • whether the anticipated attack is part of a concerted pattern of continuing armed activity;
  • the likely scale of the attack and the injury, loss, or damage likely to result therefrom in the absence of a mitigating action; and
  • the likelihood that there will be other opportunities to undertake effective action in self‑defence that may be expected to cause less serious collateral injury, loss or damage.

As I’ve explained in an earlier post, these circumstances are a bit of a dog’s breakfast. While they are all potentially legally relevant in some way or another, most have no obvious relevance to whether an armed attack is imminent.

At the same time, Brandis departs from Wright in one important respect. While Wright did not clearly distinguish between the imminence of an armed attack and the immediate need to use defensive force, Brandis is less equivocal: 

In today’s world, the precise timing of a threatened attack is often unclear or unknown. In Australia’s view, then, the appropriate question to ask when it comes to assessing imminence is: “What is the last feasible window of opportunity to act against the threatened armed attack?”

Put another way, a State may act in anticipatory self‑defence against an armed attack when the attacker is clearly committed to launching an armed attack, in circumstances where the victim will lose its last opportunity to effectively defend itself unless it acts.

On this view, States may use force in self-defense not only when an armed attack is about to occur, but also when the last window of opportunity to prevent an armed attack is about to close.

It is not obvious how this “last chance” approach is supposed to relate to Bethlehem’s list of relevant circumstances. My guess is that those circumstances are supposed to give content to the terms “feasible” and “effectively,” but Brandis does not draw this connection explicitly.

Importantly, Brandis attempts to distinguish anticipatory self-defence from pre‑emptive self‑defence, endorsing the former while rejecting the latter. According to Brandis, “[a]nticipatory self‑defence refers to the resort to force in response to an imminent (rather than an actual) armed attack.” In contrast, “States are not permitted to use force to respond to threats which have not yet crystallised but which might materialise in the future.”

Unfortunately, the “last chance” approach blurs the line between anticipation and preemption. There is a fairly clear distinction between an attack that is about to occur and a threat which might materialize in the future. In contrast, a State might easily claim that it will lose its last opportunity to effectively defend itself against a threat that has not yet crystallized unless it acts before the threat materializes in the future. Now, Brandis does indicate that the opposing State or armed group must be “clearly committed to launching an armed attack.” However, he does not explain whether this involves the general intention to attack, the general capacity to attack, active planning of a specific attack, active preparation of a specific attack, or something else entirely.

This ambiguity led Jack Goldsmith to comment that Australia has effectively embraced the Bush preemption doctrine:


Relatedly, Monica Hakimi suggests that Australia’s position effectively reclassifies many cases of preemption as cases of anticipation:


At the very least, Brandis should have drawn the distinction between anticipation and preemption more carefully, spelling out the “last chance” approach in greater detail.

The “last chance” approach also invites the question: How sure must you be—that is, how strong must your evidence be—that an armed attack will occur unless you use defensive force now? With respect to the applicable evidentiary standard, Brandis largely follows Wright:

Australia’s longstanding view is that there must be a reasonable and objective basis for determining that an attack is imminent. And this view can only be formed on the basis of all available evidence when the assessment is made.


Nevertheless, it is an assessment which must be carried out in “good faith and on the basis of sound evidence”, the standard reflected in the Chatham House Principles. Nothing less will establish a sufficient level of confidence to justify the use of force in self‑defence against an imminent armed attack.

In my view, it is far from obvious that a “reasonable and objective basis” and “good faith and … sound evidence” constitute “a sufficient level of confidence to justify the use of force.” After all, an erroneous decision to use force will violate the UN Charter as well as the sovereign rights of another State, and typically kill or injure human beings, including civilians. Arguably, the use of force would not be justified absent overwhelming evidence, or at least a preponderance of reliable evidence, that all applicable substantive rules are satisfied.

Finally, like Wright, Brandis says virtually nothing about seeking the consent of a State before using force against armed groups on its territory. In my view, it is not lawful to use defensive force without consent unless it is necessary both to use defensive force and to do so without consent. In some cases, there may be no time to seek consent. In others, the involvement of a State with an armed group may render it pointless to seek consent. However, in many cases, the only way to demonstrate the necessity to use defensive force without consent will be to seek consent in good faith and see what happens.

Now that we understand the content of the Australian position, we can inquire into its legal basis. Here we quickly encounter a methodological problem. As we saw at the outset, the UN Charter generally prohibits the use of force while narrowly permitting self-defense “if an armed attack occurs.” In contrast, Brandis claims that customary international law permits self-defense if an armed attack has not occurred but will occur unless defensive force is taken now.

One problem with this claim is that customary international law is created and modified by general state practice accepted as law. Yet Brandis cites the legal positions of only two states—the United Kingdom and the United States—in support of his claim that the “last chance” approach reflects customary law. Brandis does cite a UN High Level Panel Report for the proposition that international law permits defensive force against an imminent armed attack. However, the Report seems to understand an imminent armed attack as one that is about to occur. Brandis also cites the letters of Canada and Turkey to the Security Council regarding their use of force in Syria. However, neither letter clearly endorses the “last chance” approach.

Another problem with this claim is that is not clear when this customary law permission arose. For the most part, Brandis presents this as a post-Charter evolutionary development that is still underway. If so, then the first States to apply the “last chance” approach may have violated the UN Charter, since their initial practice presumably would not have instantly changed customary law. Brandis also invokes the Caroline incident of 1837, suggesting that the customary law right of anticipatory self-defense preexists and is codified by the Charter. While analysis of the Caroline incident would take us too far afield, my own view is that it tells us very little about the contemporary law of force and self-defense (see, e.g., here).

Finally, Brandis does not explain how this alleged development in customary international law affects the scope of lawful self-defense under the UN Charter. The Charter is, after all, a treaty, which we should presumptively interpret in accordance with the ordinary meaning of its terms, in their context, and in light of the treaty’s object and purpose. Brandis does not attempt to defend the “last chance” approach using traditional methods of treaty interpretation, perhaps because such an attempt is unlikely to succeed.

No doubt, the interpretation of a treaty should take subsequent practice into account, but only to the extent that such subsequent practice establishes the agreement of the parties to the treaty regarding its interpretation. Obviously, the subsequent practice of a handful of States hardly establishes the agreement of the parties to the UN Charter, that is, the agreement of nearly all States.

Perhaps Brandis has in mind the principle that the interpretation of a treaty should take into account other relevant rules of international law. This principle, though generally sound, should be applied with great care to the UN Charter. The prohibition on the use of force is the general rule, and self-defense is the limited exception. Where the scope of the exception is in doubt, the prohibition should prevail.

Image: Getty