This post is the latest installment of our “Monday Reflections” feature, in which a different Just Security editor examines the big stories from the previous week or looks ahead to key developments on the horizon.

Over at Lawfare, Ashley Deeks has offered some thoughtful reflections on the ad bellum “imminence” discussion in Brian Egan’s recent ASIL speech. Jack Goldsmith followed up with a post suggesting that Egan’s articulation of the law of anticipatory strikes in self-defense was virtually equivalent to the Bush Administration’s “preemption” doctrine. Sir Daniel Bethlehem took issue with Jack, insisting that there is a big difference; and Jack responded that the difference appears to be in name only, not in substance. I thought it might be useful, in light of those posts, to unpack the “imminence” question just a bit, and to situate it properly in a broader discussion of at least two important topics in the jus ad bellum—necessity and “first use.” (My previous posts on the Egan speech are here and here.)

Although the Egan speech does correspond to understandings of the jus ad bellum, including “anticipatory” strikes, that have been fairly well-established for a long while (going back at least as far as The Caroline controversy in the 1830s), there is at least one very significant difference between the Egan speech and the Bush Administration “preemption doctrine” of 2001-2003: The Bush Administration argued that international law permits the United States to engage in a “first use” strike, in a nonconsenting state, against a state or nonstate actor that has not already engaged in an armed attack against the United States, before any threat of attack is “fully formed” — indeed, even where the probability of any such future attack is “relatively low.” The Bush Administration used this very aggressive reading of self-defense to argue that international law would permit the use of force, in nonconsenting states, against all terrorist groups of “global reach,” regardless of whether they had developed, let alone engaged in, attacks against the United States, and, at least in theory, that self-defense would justify an invasion of Iraq, despite virtually no evidence that Iraq was likely to attack the United States. That’s what made the Bush “preemption” doctrine so controversial. Neither Brian Egan nor (as far as I know) any other Obama Administration official has endorsed such an untenable understanding of international law.

1. “Imminence.” Before turning to the jus ad bellum, a word about “imminence” generally. The term “imminent” or “imminence” is, of course, used in many diverse legal contexts, and it can have subtly different meanings or applications across various doctrines: as the Supreme Court has recently reminded, it is “concededly a somewhat elastic concept.” Notably — and wholly independent of its use in the jus ad bellum — in most contexts “imminent” does not necessarily, or even primarily, mean “immediate” or “very soon.” Rather, it more commonly means “impending.” Webster’s offers: “ready to take place; especially: hanging threateningly over one’s head,” while a recent edition of Black’s Law Dictionary defined “imminent” as “[n]ear at hand; mediate rather than immediate; close rather than touching; impending; on the point of happening; threatening; menacing; perilous.”

2. Article 51 and necessity in the law of self-defense. In order to make sense of the dispute about Brian Egan’s speech, it’s essential to understand the context in which “imminence” is relevant to the jus ad bellum, in particular. As I explained last week in discussing the speech, the question Egan was discussing is when, if at all, it is lawful for a state to use force unilaterally — that is, without Security Council authorization — in the territory of a nonconsenting state. As Egan explained, most of the U.S.’s uses of force against al Qaeda and ISIL have been with the consent of the host states, and therefore the ad bellum question has been inapposite in those cases. But host-state consent has been absent in other cases—most importantly, the 2011 operation against Osama bin Laden in Pakistan, strikes against the Khorasan Group in Syria in 2014, and the ongoing operations in Syria against ISIL. When, if at all, are those uses of force lawful?

Article 2(4) of the U.N. Charter provides that states “shall refrain in their international relations from the threat or use of force against the territorial integrity or political independence of any state.” Article 51, however, clarifies that neither Article 2 nor any other provision of the Charter “shall impair the inherent right of individual or collective self-defence if an armed attack occurs against a Member of the United Nations.” As I noted in my earlier post, Egan confirms the U.S. view — which virtually no other nations dispute — that this “inherent” right of self-defense applies to threats from nonstate actors — such as ISIL today, or the Canadian rebels in the nineteenth century case of The Caroline — even in situations where the host state is not complicit in, or responsible for, the nonstate actor’s armed attacks upon the acting state.

The “inherent” right of self-defense against armed attacks, which Article 51 preserves, allows the use of force in a nonconsenting state (such as Syria) only if that force satisfies two fundamental conditions: necessity and proportionality. Thus, for example, if the host state is willing and able to eliminate the threat from the nonstate actor on its territory, then the threatened state may not use force within the host nation’s territory, because it would be unnecessary.[1]

The notion of “imminence,” historically, has been used to flesh out what it means to determine whether the threatened state’s use of force in self-defense would be necessary. To quote the classic formulation agreed to by both the U.S. and England in the Caroline case, the “necessity of self-defence” (rather than the threat) must be “instant, overwhelming, leaving no choice of means.” In other words, what must be “imminent,” in the Caroline sense, is not the threat of armed attack itself, but the need to use force in another state’s territory to suppress that threat. This is where I think many formulations of the “test,” including Egan’s, engender confusion. Egan referred, as many do, to “considering whether an armed attack is imminent”; but the relevant consideration is, instead, whether the need to use force in the territory of the nonconsenting state is “imminent.”

Thus, wholly apart from the fact (noted above) that the term “imminent” does not necessarily mean “immediate,” the ad bellum necessity test does not require proof that the armed attack from the territorial state or (here) from the nonstate actor is just around the corner. As Dapo Akande and Thomas Liefländer have recently explained, “such an independent temporal limitation would mean that where a highly probable and severe threat exists, whose realization is temporally remote, no action could be taken even where no future opportunity will arise to eliminate the threat. The better understanding of the law is that where a threat is sufficiently probable and severe, the mere fact that it is still temporally remote should provide no independent injunction against action where that action is necessary and proportionate.”

This does not mean that the “temporal” question with respect to the anticipated armed attack is legally insignificant — to the contrary. The immediacy and certainty of the threat are very important considerations in the assessment of necessity; the more remote or speculative the threat is, the harder it is to demonstrate the necessity of using force in self-defense to prevent it, in large measure because a lot can happen between Time A and Time B that might diminish or eliminate the threat. I can’t put the point much better than Akande and Liefländer have done:

[D]enying the existence of an independent temporal limitation does not mean that temporal factors are unimportant. They have a heavy impact on the possibility of making accurate predictions about both the likelihood and the gravity of a threat. The shorter a causal chain is, the easier it becomes to predict what will occur. It will be harder to establish that a threat is sufficiently probable and severe if such a threat is still very far away in a temporal sense. In addition, and more importantly, the temporal dimension affects necessity. As noted, necessity allows using force only where no peaceful alternative is available [or, I would add in this context, where the host state itself is willing and able to suppress the threat]. Thus, where a military option will be available for some time because the threat is temporally remote, other options should be tried first. Other scholars, including those involved in the Chatham House principles, have read the imminence requirement similarly, focusing on the last point in time at which an effective responsive action is possible, rather than temporal proximity [of the threatened attack] per se. What is really at stake is whether some sort of self-defense action is demonstrably necessary—without any alternative, including later in time—rather than how temporally remote the threat is.

Seen in this context, Brian Egan’s “factors” make perfect sense, because they are exactly the sorts of things one would be required to assess — in addition to the willingness and ability of the host state to deal with the problem — in considering whether the use of force is necessary, namely, “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 mitigating action; and the likelihood that there will be other opportunities to undertake effective action in self-defense that may be expected to cause less serious collateral injury, loss, or damage.”

I don’t think much of this understanding of the “inherent right of self-defense” is very controversial—not, anyway, amongst states. Moreover, it describes common practice: In the Charter era, there have been very, very few cases of asserted self-defense in a nonconsenting state in situations where the armed attack to be addressed was known to be immediately forthcoming. In almost all such cases, the state invoking self-defense has asserted, or implied, that the attack was certain or almost certain; that it was impossible to assess precisely when the attack would occur; and, ultimately, that there was an immediate need to prevent the attack with a (proportional) use of force, because all other means had been exhausted, and/or because there was significant reason to doubt there’d be another “window of opportunity.” It’s almost unheard of, however, for a state to claim that it waited to act until the very moment before the threatened attack.

3. First use and the Bush Doctrine. This brings us to the area of real controversy, and to the major distinction between President Bush’s “doctrine of prevention” and the Obama Administration’s uses of force in self-defense — namely, the question of first use.

In a very real respect, virtually all instances of self-defense are “anticipatory,” or “preemptive,” in the sense that they are designed to prevent future (rather than ongoing) armed attacks. As I’ve explained above, that, in and of itself, is not terribly controversial. Yes, it is common ground between the Bush and Obama Administrations — but it was also fairly well-uncontroverted internationally for a very long time before 2001. (I’m generally trying to minimize use of the common adjectives “anticipatory,” “preemptive” and “preventive” here, because they have been so promiscuously employed and have come to mean very different things to different people. But the type of self-defense I’m describing here is often termed “preemptive,” whereas the Bush doctrine, described below, is often characterized as a form of “preventive” self-defense.)

Notably, however, in the Charter era, such uses of force in self-defense against future attacks have almost always occurred after the targeted party has already engaged in an armed attack, and thereby demonstrated its design and capability of threatening future attacks: very few, if any, cases of ad bellum self-defense in the Charter era have been against states or nonstate actors before they have engaged in any armed attacks against the acting state. (The U.S. invasion of Iraq in 2003 is often cited as a counterexample; importantly, however, the United States conspicuously declined to invoke self-defense in justification of that use of force.) That dearth of what we might call “purely” anticipatory self-defense is, in no small measure, a function of international law, because, as I explain below, it is not clear whether and to what extent Article 51 and the “necessity” requirement of the jus ad bellum permit first use of force, even in self-defense, in the territory of a nonconsenting state.

The Obama Administration has said little, if anything, about such “purely anticipatory” first use of force in self-defense. The two obvious cases in which self-defense has been a necessary predicate of U.S. action during this Administration have been the 2011 operation against bin Laden in Pakistan, and the strikes against the Khorasan Group in Syria in 2014. Bin Laden was the head of al Qaeda and, according to the President, the Khorasan Group consists of “elements of al-Qai’da.” Al Qaeda, of course, has already engaged in large-scale attacks on the United States, and is dedicated to as many further such attacks as its capabilities will allow; as David Luban has put the point, al Qaeda has already sent an “unmistakable signal that [it] has crossed the line from diplomacy to force.” The U.S. might not have known exactly when al al Qaeda would next strike (or attempt to do so) — no more than the British knew exactly when the Canadian rebels would next strike when they set fire to The Caroline; but in both cases there was virtual certainty that the organized armed group would attack again if afforded the opportunity, because it had already demonstrated the requisite intent, design and capability, and there was no reason to believe that it had abandoned its campaign against the threatened state.

Of course, the fact that al Qaeda will almost certainly continue to try to attack the United States, in and of itself, does not satisfy the “necessity” test for the U.S.’s use of force in a nonconsenting state — the U.S. still could not have lawfully struck the Khorasan Group in Syria, for instance, unless such intervention were “necessary to defend the United States and our partners and allies against the threat posed by these [al Qaeda] elements,” as the President claimed they were. So why were such strikes deemed necessary? A “senior administration official” explained, at the time of the attacks, that the Khorasan Group consisted of “al Qaeda veterans who have established a safe haven in Syria to develop and plan external attacks in addition to construct and test improvised explosive devises and to recruit Westerners for external operations.” The President’s order to strike against them reportedly was based on intelligence that “these senior Syria-based al Qaeda operatives were nearing the execution phase for an attack in Europe or the homeland,” and that “their plotting was reaching an advanced stage.”

The current operations against ISIL in Syria complicates the picture, but only by a bit. Recall that the United States’s primary and formal justification for using force against that group in Syria is the collective defense of Iraq, a nation that ISIL has repeatedly attacked. Self-defense of the United States therefore has not been a necessary justification for our actions in Syria. Even so, Brian Egan indicated that the U.S. believes its own self-defense would be an apparent alternative justification under Article 51, even though in recent years ISIL has not yet engaged in an armed attack specifically against the United States: “[I]n Syria,” said Egan, “U.S. operations against ISIL are conducted in individual self-defense and the collective self-defense of Iraq and other States.” Similarly, Great Britain has formally invoked both collective and individual self-defense as justifications for its use of force against ISIL in Syria, even though ISIL has (probably) not yet attacked that nation.[2]

These statements by the U.S. and the U.K. are undoubtedly the most forward-leaning and potentially controversial of those that Western nations have made in connection with the jus ad bellum and ISIL, because they at least imply that the U.S. and U.K. could use force in Syria against ISIL even before ISIL has operationalized or planned any armed attacks on those two nations. Recall that Article 51 preserves “the inherent right of individual or collective self-defence if an armed attack occurs against a Member of the United Nations.” Read literally, this language would appear to exclude all “purely anticipatory” uses of force in self-defense, i.e., any “first uses” of force against actors that have not yet attacked the threatened state — in which case the British and American statements would be in tension with the treaty language.

That has not been the predominant understanding of Article 51, however: virtually all states and commentators acknowledge that a state must be allowed to employ force — even a “first use” — to prevent, at a minimum, a certain and immediately impending attack that it is about to suffer (such as the proverbial case where tanks are massing on its border and planning to attack in the morning). And, in any event, the U.S. and U.K. statements about a hypothetical “first use” of force in self-defense against ISIL are hardly surprising or alarming, because ISIL has already demonstrated, beyond peradventure, that it is ready, willing and able to attack every member of the Western coalition arrayed against it, just as it has already attacked, e.g., France, Egypt, Iraq, Tunisia, Belgium, etc. Moreover, ISIL has brutally executed U.S. nationals who it has captured, and it killed dozens of British tourists in Tunisia. The primary rationale for limiting the permissibility of “first uses of force” in self-defense is the uncertainty of whether such anticipatory action is necessary: unless and until the nation or group in question has demonstrated a design and capability of making such attacks, it is often difficult to know whether it would actually strike. That concern is virtually absent here: There’s little doubt that, left to its own devises, ISIL would attempt to attack the U.S. and the U.K. Therefore, the British and American suggestions that the Charter would permit “first use” of force against ISIL in Syria, although provocative, are not terribly ground-breaking. (Once again, such “first use” in Syria would only be lawful if it were necessary to prevent the future ISIL attacks, and proportional to that objective. Those assessments remain largely academic, however, because the U.S.’s primary rationale for use of force in Syria — collective defense of Iraq — is compelling.)

Contrast all of this with what was so controversial about the United States’s “doctrine of preemption” in 2001-2003. The Bush Administration not only argued that “first use” strikes in self-defense of the U.S. could be permissible, but also provocatively suggested that such first strikes could be launched against groups or states that had not yet demonstrated that they would, or could, engage in armed attacks against the United States.

This much more robust — and much more controversial — notion of the “inherent right of self-defense” emerged in two very high-profile contexts. First, on September 20, 2001, President Bush famously declared that “[o]ur war on terror begins with Al Qaeda, but it does not end there. It will not end until every terrorist group of global reach has been found, stopped and defeated.” This statement certainly appeared to suggest that the United States could (and would) use force against any terrorist group by virtue of its “global reach,” regardless of whether it had already attacked the United States or had concrete and operational plans to do so.

Then, in September 2002, the President’s National Security Strategy (NSS) went much further still, indicating a right under international law to strike at state and nonstate actors before they even have the capability of attacking the United States (and presumably before they had expressed or operationalized any serious designs to do so): According to the NSS, the United States must “stop rogue states and their terrorist clients before they are able to threaten or use weapons of mass destruction against the United States and our allies and friends,” which justifies using force “against such emerging threats before they are fully formed.”

Less than five weeks later, the Office of Legal Counsel demonstrated the import and vast breadth of this novel idea: OLC opined that going to war against Iraq would be a permissible exercise of a so-called “reformulated test for using force in anticipatory self-defense” because Iraq might develop chemical and biological weapons and then might one day use them against the United States. OLC concluded that the President could legitimately use military force against Iraq even in the absence of “information regarding whether the use of force against Iraq at a particular time would be necessary to take advantage of a window of opportunity to prevent the threat of a WMD attack from materializing,” and, most strikingly of all, “even if the probability that Iraq itself would attack the United States with WMD, or would transfer such weapons to terrorists for their use against the United States, were relatively low.”[3]

As Marko Milanovic has noted, this self-defense analysis by OLC so obviously did not reflect the customary jus ad bellum,[4] and was “so outlandish that it was in the end not formally relied on by the United States to justify its use of force against Iraq” in 2003. It did, however, reflect the remarkable breath of the Bush Administration’s efforts to broaden the concept of “necessity” for purposes of anticipatory self-defense under the jus ad bellum, beyond what any reasonable reading of customary law would sanction.

Neither Brian Egan’s speech, nor any other statement or action by the Obama Administration, has come anywhere close to suggesting an embrace of this, the most controversial aspect of the Bush doctrine as articulated in the President’s speeches, the 2002 NSS, and the OLC Iraq opinion. Indeed, Egan’s speech — unlike the most controversial instances of the Bush Administration’s invocations of “preemptive use of force” — does not even address question of first use, at least not expressly. And that’s because the Obama Administration has not engaged in any such practice, nor had any occasion to discuss whether and when such first use of force would be lawful.


[1] Fionnuala Ní Aoláin thus is mistaken in describing the “unable or unwilling” assessment as “random” or as having “more than a hint of ‘legal lego-land’ to it.” There is nothing new or controversial about requiring the threatened state to assess whether the territorial state is unwilling or unable to suppress the threat. That is a simple and straightforward precondition to satisfying the requirement of “necessity,” without which the victim state cannot act in self-defense on the territory of the nonconsenting host state. I would be surprised if any state disagrees with this basic requirement. Fionnuala appears to be concerned that “unable and unwilling” would be viewed as a sufficient condition for the use of force without the host state’s consent. But that is not correct, and Egan didn’t imply any such thing: Of course, in addition to satisfying the condition of necessity, the threat itself has to be of the sort that justifies the “inherent right of self-defense,” and the use of force must be proportionate to preventing the threatened future armed attacks.

[2] Thirty of the 38 victims of the attack at the Sousse Hotel in Tunisia last year were British tourists; but it is not clear whether that was an ISIL-designed attack on Great Britain, in particular.

[3] That October 2002 OLC opinion also offered broad and unconvincing accounts of the scope of the customary law of self-defense and of the President’s constitutional authority to start a full-scale war without congressional approval. All and all, it might have been the most far-reaching and misguided of all the Bush Administration OLC’s opinions.

[4] OLC made virtually no effort to show that its analysis actually reflected customary international law. Indeed, its historical survey uncovered only one instance in which a single nation relied upon such a theory — Israel invoked it to defend its 1981 attack on the Osirak nuclear reactor under construction in Iraq — and in that case, as OLC conceded, “the international community, including the United States, condemned the Israeli attack.” The Security Council unanimously adopted a resolution “[s]trongly condemn[ing]” the Israeli strikes as a “clear violation” of the U.N. Charter, and the U.S. Legal Adviser in the first Bush Administration, Abe Sofaer, later confirmed that there was an “absence of any evidence that Iraq had launched or was planning to launch an attack that could justify Israel’s use of force.” “While Israel’s anxiety concerning Iraq’s intentions may have been reasonable,” wrote Sofaer, “the presence in a State of the military capacity to injure or even to destroy another State cannot itself be considered a sufficient basis for the defensive use of force.” 126 Mil. L. Rev. 89, 109 (1989).

Indeed, OLC went far beyond even what Grotius opined would be lawful almost 500 years ago: Grotius reasoned that although “an injury not yet inflicted, which menaces either person[] or property,” can be a justifiable basis for war if it is “immediate and certain, not when it is merely assumed,” it would be “untenable” for a state to attack a “growing power which, if it became too great, may be a source of danger.”