Home EJIL Analysis What Is An Imminent Armed Attack? A Hopefully Helpful Hypo

What Is An Imminent Armed Attack? A Hopefully Helpful Hypo

Published on January 12, 2017        Author: 

Yesterday we had the privilege to publish the speech by the UK Attorney General, Jeremy Wright QC MP, setting out some of the UK Government’s views on the law of self-defence. The speech focused in particular on the criteria for assessing the imminence of an armed attack by a non-state actor, and essentially endorsed some of the principles set out in Daniel Bethlehem’s 2012 AJIL article. Thus, the Attorney stated in particular (following a speech by the US State Department Legal Adviser, Brian Egan, at last year’s ASIL meeting) that:

[Bethlehem’] Principle 8 on imminence, as part of the assessment of necessity, is a helpful encapsulation of the modern law in this area.

Sir Daniel’s proposed list of factors was not exhaustive, but included (at Principle 8), the following:

  • 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.

It is my view, and that of the UK Government, that these are the right factors to consider in asking whether or not an armed attack by non-state actors is imminent and the UK Government follows and endorses that approach.

In each exercise of the use of force in self-defence, the UK asks itself the questions that flow from that articulation. Questions like – how certain is it that an attack will come? How soon do we believe that attack could be? What scale of attack is it likely to be? Could this be our last clear opportunity to take action? And crucially – is there anything else we could credibly do to prevent that attack?

I don’t think the Attorney broke any new ground here, nor do I wish to dispute the accuracy or normative desirability of this analysis. But what struck me most about it is the lack of conceptual clarity, in particular the lack of clear delineation between the concepts of imminence, necessity and proportionality and the legal role that these concepts are playing. (I would highly recommend, in that regard, this piece in the AJIL by Dapo and Thomas Liefländer). For example, what is the work that the idea of imminence does here? Is its main purpose to delineate between permissible anticipatory and prohibited preemptive self-defence, which goes around the Article 51 Charter language ‘if an armed attack occurs‘? Or is imminence an aspect of the broader concept of necessity? And can a word such as imminence encompass non-temporal elements? Conceptual clarity matters because without a common understanding of the words we are using we cannot actually properly debate the soundness or desirability of any given approach. Without it, it is hard to even have a conversation.

So here’s a hypo that I hope might be helpful in this regard. Again, the point of the hypo is not to argue for any particular interpretation of self-defence, whether expansive or restrictive. It is only to help us understand how people use particular words, such as imminence, and for what precise purpose.

Dr. Evil is a very capable terrorist, who has decided to attack the United Kingdom, even though he has never done so before. He manages to get his hands on a mid-range cruise missile with a 150 kt thermonuclear warhead, in perfect working order. He places the missile launching system in a building in a Paris suburb, and uploads a video of himself to YouTube showing him arming a very specific firing mechanism. The missile is aimed at London, and will launch in exactly 30 days; there is no off-switch, code or remote signal that can disarm it. Absent forcible intervention in the causal chain, there is complete certainty that the missile will fire in 30 days and that it will destroy a substantial part of London.

Is this armed attack ‘imminent’ in any legally relevant jus ad bellum sense?

Note that this hypo is specifically designed to eliminate most of the real-world uncertainties about armed attacks – the reliability of the intelligence, the likelihood of the attack, not knowing the exact time, location or scale of the attack. In this hypo, we know everything with absolute certainty. And if you have a problem with the non-state actor nature of the attacker, we can easily turn him into a French state agent. Again, the main point here is that a causal chain has been set in motion which, without some further action, as its certain end has the destruction of London. Does this mean that the attack is ‘imminent’? If so, would it be imminent even if the timer was set to 60 days, 120 days, or 10 years? At what point (if any) is there a switch from an anticipatory to a preemptive scenario? When does imminence end, and necessity begins?

Print Friendly

Related posts


26 Responses

  1. Marko, isn’t your hypothetical an ongoing armed attack, assuming Dr Evil is in the process of initiating the launch?

  2. Jordan J Paust

    Aurel: hasn’t the process of attack actually begun? The trigger has been fired but the bullet has not yet left the chamber. Isn’t this almost like Putin announces that Russian mechanized units will roll into Latvia and they are moving yards before crossing into Latvia, although the hypo has an additional assumption of inevitability (Putin, in contrast, might be able to order the bullet to stop).

  3. Jordan J Paust

    p.s. an imminent armed attack would be one that has not begun.
    Another question: given tbe express and unavoidable limitation in Article 51, “if an armed attack occurs,” would use of force not justified under articles 2(4), 39-42-48, 51, or 52-53 (e.g., “regional action” as re Kosovo) be “manifestly” unlawful?

  4. Jordan J Paust

    Yet another, assume that North Korea has created missiles capable of hitting Ann Arbor, Michigan and Notre Dame, Indiana; the missiles are known to have newly created nuclear warheads; the missiles are known to be preped for immediate launch; and the dictaor has declared that they will be launched in three days.
    Has the process of armed attack begun?

  5. Marty Lederman

    I think you’re headed in exactly the right place, Marko. As I’ve explained here (–borrowing liberally from Dapo and Liefländer–where many have gone astray is in assuming that the *attack* must be “imminent.” That’s not right. The question, instead, is whether the use of force in another (unconsenting) state would be *necessary*–and the notion of “imminence,” historically, has been used as a factor in determining whether that necessity requirement is met. To quote the classic formulation agreed to by both the U.S. and England in the Caroline case, it’s the “necessity of self-defence” (rather than the threatened attack) that must be “instant, overwhelming, leaving no choice of means.” In other words, what must be “imminent,” in the Caroline sense, is not the 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 and Wright’s, engender confusion. They refer, 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.” (Dapo & Liefländer: “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.”

    So, as for your terrific hypo:

    The UK could use force on the suburban French building if it’s necessary to do so, regardless of whether the missile will be launched in 30 minutes, 30 days, or 30 months. If it’s going to launch in 30 days or months, however–as opposed to 30 minutes– the UK is obliged to figure out whether the French will either (i) consent to the UK strike (in which case the 2(4) problem would disappear), or (ii) disable the threat themselves.

  6. Jordan J Paust

    I meant not only prepped and will be launched, but will be launched against cities in Michigan and Indiana.

  7. LTC Steven Gariepy

    I enjoyed your post and will shamelessly swipe your hypothetical for future exam use. In my humble opinion, the distinction that you (along with the folks remarking to your post) are making is remarkably similar to a Military Commander’s decision whether or not to use military force. Just because the use of military force is legal does not necessarily mean that the use of military force is the correct response to a problem. The fundamental difference between “can I do something” and “should I do something” is what distinguishes judgment and discretion. On a related note, in my experiences working with NATO, a particular lawyer’s definition of what is or is not “imminent” generally stems from their own domestic legal systems interpretation of that particularly loaded term. Take care and hope to see you back at West Point in the near future.

  8. Kriangsak Kittichaisaree

    Rule 15 (Imminence and immediacy of self-defence) of the Tallinn Manual on the International Law Applicable to Cyber Warfare: “The right to use force in self-defense arises if a cyber armed attack occurs or is imminent. It is further subject to a requirement of immediacy”.

    Rationale: in a cyber armed attack, just one click on your laptop etc. and the destruction is accomplished at the other end. Waiting for a cyber armed attack to actually occur before you may exercise the right of self-defence would be too late. This seems to make sense in the cyber context. Q: Whether it also makes sense outside the cyber context depends on the circumstances at hand?

  9. Jordan J Paust

    KK: but how can it be consistent with “if an armed attack occurs”? In any event, is the “click” the only factor in determining when an attack is occuring? Re the pistol analogy, can it be, in view of other features of context, when the pitol is drawn and pointed?

  10. Kriangsak Kittichaisaree

    Dear Jordan,

    Re the pistol analogy, when the pistol is drawn and pointed, the other guy can exercise his right of self-defence under domestic criminal law. Self-defence under IL is different from the one recognized under domestic criminal law, though (cf. the different thresholds under Art. 2(4) and Art. 51, UN Charter and the customary IL of self-defence).

    Now, re your Q about R 15 of TM, the accompanying Commentary explains as follows:
    – “Textually”, Art. 51 refers to a situation in which ‘an armed attack occurs’, but the majority of the International Group of Experts (IGEs)writing TM agreed that “a State need not wait idly as the enemy prepares to attack. Instead, a State may defend itself once the attack is ‘imminent’. Such action is labelled ‘anticipatory self-defence’. This position is based on the standard of imminence articulated in the C19 by Webster following the Caroline incident……” Preventive strikes (ie., the Bush doctrine), on the other hand, are not justifiable under IL.
    – “Consider a situation in which the intelligence service of State A receives incontrovertible info that State B is preparing to launch a cyber attack that will destroy A’s primary oil pipeline within the next 2 weeks. The attack involves causing the microcontrollers along the pipeline to increase the pressure of the pipeline, resulting in a series of explosions. Intel services have no info on the specific vulnerability to be exploited, thereby preventing effective cyber defence of the microcontrollers. However, they do have info that those involved in conducting the attack will be gathered at a particular location and time. State A would be justified in concluding that the necessity of self-defence against those individuals would be lawful as proportionate anticipatory self-defence should lesser means be inadequate.” (Q: Does this look like what Marty has described above?)

    Public Disclosure: I was a member of the IGEs that wrote the TM 2.0 on the IL Applicable to Cyber Operations (Cambridge UP, Feb. 2017) and had nothing to do with TM (1.0) and its Rule 15. But, as a matter of courtesy/etiquette, I should not say whether IGEs 1.0 were right or wrong. I also still earn my living as a full-time diplomat; hence, I should not say whether the UK position as explained by its AG is OK or not. Having said that I have joined the discussion here to draw your attention to the possibility mentioned in R15/TM. It seems that the speed of the Internet is the main reason for IGEs 1.0 to adopt the view stated above. Q: Are there other similar circumstances that would justify this same IL position?

    The debate between those supporting only ‘if an armed attack occurs’ will do versus those Ok’ing’anticipatory self-defence’ has a long and unsettled history, e.g. Prof. Ian Brownlie (the former view) v. Prof. Derek Bowett (the latter view). States adopting either position must find State practice plus opinion juris to support their action ‘in self-defence’.

  11. Marko Milanovic Marko Milanovic

    Many thanks to everyone for your great comments, and apologies for the late reply. Let me try to develop this a bit further:

    (1) On one view (Aurel), the armed attack in my hypo may already have started and be ongoing, and that’s why it’s not imminent. This obviously begs the question of how we define an armed attack in the first place, and how far up the causal chain the scope of this definition extends.

    So, let’s try another hypo: early in the morning of Saturday the 21st, President Trump gets totally fed up with all this golden shower/Russian stooge business and decides to set the record straight by nuking Moscow. He opens the nuclear football, pushes whatever button and inputs whatever code he needs to, and the order to strike Moscow is given.

    At what point does the armed attack against Russia begin:(i) when Trump gives the order; (ii) as the ICBM is being prepped; (iii) as the ICBM launches; (iv) at some point of no return at which the launch or the missile’s flight can’t be aborted; (v) when the missile leaves US airspace; (vi) when the missile penetrates Russian airspace; (vii) when it detonates over the Kremlin (or the Ritz Carlton on Tverskaya St).

    Again, the only question this hypo is meant to address is when does the armed attack occur. Logically the attack could be imminent only before that point of occurrence.

    (2) I think that Jordan is quite right that my hypo is analogous with the invading army example, as the tanks are rolling towards the border.

    (3) Aurel, if the timer in my original example was set to 10 years, would you still be happy to say that the armed attack has already started?

    (4) Marty, I think your position is entirely logically coherent – you are basically saying (like Dapo and Thomas) that imminence is a matter for the necessity analysis – do I have to act to defend myself, or do I have the time to wait? But isn’t then the fundamental problem with your view not only that it renders nugatory the language of Article 51 of the Charter ‘if an armed attack occurs’, but also the commonly held view that anticipatory self-defence in response to an imminent attack is fine whereas a preemptive or preventive one wouldn’t be? You are basically saying that an ‘imminent armed attack’ is misnomer, or at least not a legally relevant standalone concept, aren’t you? An imminent attack is a future attack that creates necessity to defend against it today.

  12. Marko Milanovic Marko Milanovic

    (5) Kriangsak, many thanks for this and the very helpful references to the Tallinn Manual. As I read the discussion in the TM, the majority of experts rejected imminence as a strict temporal threshold, but saw the concept in essentially causal terms – once a causal chain that leads to an attack is initiated, that attack becomes imminent at the point in time (with some wiggle-room) in which the attacked state would lose the ability to defends itself effectively, and is thus forced to act. To my eyes this very much looks like necessity, doesn’t it, much like in Marty’s comment? The concept of imminence doesn’t seem to do any work by itself, except to say that the armed attack is yet to occur. Thus when there is necessity to defend against a future attack the attack is said to be imminent and the self-defence is labelled as anticipatory, but when there is no such necessity (e.g. there are plenty of other options to be exhausted) the attack is said to be non-imminent and the self-defence is labelled as pre-emptive.

  13. Marko, I appreciate the hypothetical nature of the scenario, but you could have given the President-elect at least a week in office, you know. Anyway, I would be content with declaring an armed attack has occured on point (iii) under virtually all circumstances. I might also be content with (i) and (ii), depending on the context and other factors. Questions to ask would include: “How often have ICBM’s been prepped before”? “Does an order to fire and active preparation to do so mean with near certainty that the missiles will be fired?” “Did the President issue the order on TWITTER?”

    On your second question, I think imminence is not exclusively a question of proximity in time, but that time does have an effect on the assessment of necessity. Ten years down the line is hardly imminent, but the attack could still be inevitable, in principle. The critical question that would have to be answered therefore is whether the attack is in fact inevitable, despite the fact that it will only take place in 10 years. I will not speculate, as it’s a hypothetical, but would suggest that the answer could go both ways, depending on the facts.

  14. Marty Lederman

    Marko: You have, of course, put your finger right on the most inscrutable question–the effect of Article 51’s “if an armed attack *occurs*” language. Two things about this for starters:

    First, by this point virtually no one (no state, anyway) believes that it ought to be read literally–even in the “Egyptian troops massing on the border for attack tomorrow” case. So there must be an implied term. I’ll suggest this one: “occurs *or is virtually certain*.” (Of course, that’s simply to articulate the necessity requirement already built into the “inherent right of self-defense.”) The point of the phrase, on this view, is to prohibit the use of force in self-defense unless and until there is a high degree of certitude that the nation or NSA in question truly has the capability, design and intent–commitment, even–to strike.

    Second, in the vast mine run of cases that have actually transpired, this has been a nonissue, because the group or nation has *already* engaged in an armed attack against the acting state (“an armed attack occur[red]”), and thus demonstrated capability and design (and, for good measure, has satisfied the literal condition of Art. 51). And when the group/state does so repeatedly, e.g., in an armed conflict, there is no longer much question that it will do so again, unless it has been vanquished. That describes the Caroline, by the way: The rebels had already engaged in “armed attacks,” and there was no reason to think they would not again, even though the time and place were unknown.

    The interesting thing about your hypo is that you have (deliberately) designed it to be the outlier, test case: (i) to *remove* the past armed attack, which is what usually establishes the evidence of a threat of certain future attack, but (ii) to establish virtual certitude of future attack, anyway. Egypt massing on the border for a certain (first) attack a month from now, in a sense.

    What’s especially interesting now is that the UK has apparently acted upon that theory, with the Khan strike. (This assumes the UK would not retroactively say that that strike was also part of the collective self-defense of Iraq, which ISIL had already attacked.) That’s fairly groundbreaking.

    In my view, because in that rare case–where the target group/state/individual has not yet attacked–there is more uncertainty than the usual case about whether the actor will, in fact, engage in an armed attack, the evidentiary burden to show “necessity” is, and ought to be, much greater, given the possibility that the person/group is bluffing, or will not have the guts to go through with it, etc.–something that we know *not* to be the case where they have already attacked.

    I discuss all this in greater detail in the “First Use” section of that post to which I linked in my earlier comment. I’ll include an excerpt in a follow-up comment; I’d be curious whether you and Dapo agree with it:

  15. Marty Lederman

    Excerpt from

    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.

    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 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 [against AQ]: the 2011 operation against bin Laden in Pakistan, and the strikes against the Khorasan Group in Syria in 2014. 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. . . .

    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.[FN: 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.]

    [The British defense of the Khan strike is] undoubtedly the most forward-leaning and potentially controversial that a Western nation has made in connection with the jus ad bellum and ISIL, because [it at least implies] that the U.K. could use force in Syria against ISIL even before ISIL has operationalized or planned any armed attacks on it. 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 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.K. statement about a hypothetical “first use” of force in self-defense against ISIL is 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.

    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.”

    As Marko Milanovic has noted, this self-defense analysis by OLC so obviously did not reflect the customary jus ad bellum, 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.

  16. Jordan J Paust

    Aurel & Marty: once we conclude that an attack has begun there it should not be required that its completion is virtually certain. Marko has stretched out the choice points nicely with respect to whether an attack is occurring. And we should be looking at the process of attack when making such a choice. See re process of attack and the need for choice.
    If the attack is occurring, its intended consequences can be denied by a responsive use of force in self-defense, like use of a missile “shield” to destroy incoming rockets.

  17. Jordan J Paust

    For example, re the gun and bullet hypo, if we conclude that an attack is occurring when the gun is loaded, drawn, and aimed, there is no need to wait until it is fired or to conclude with “certainty” that it will be fired.
    By the way, the oft only partially quoted phrase from The Caroline incident and necessity actually focused on the necessity of the means used, not necessity to engage (and the attacks were already occurring). For example, in the early 19th Century, the means used were, from the US point of view, not necessary because the UK could have waited for the vessel to return to Canadian waters and to do so in daylight. On the UK side it was noted, however, that if a NSA shore battery was firing across the border self-defense would be proper and as well as the means used, destruction of the shore battery. See ssrn artcle.

  18. Marko Milanovic Marko Milanovic

    Marty, thanks – I agree with most of what you say, in particular the point about most cases of self-defence being anticipatory in the sense that an attack has already happened and action is now being taken to prevent future attacks.

  19. Marty, I disagree on the Khan strike. I don’t think it is analogous to Marko’s example at all.

    There is a whole lot of background to the Khan strike, namely the ongoing NIAC to which the UK and ISIL were opposing parties. Until December 2015, the British Parliament did not authorise the Government to engage targets in Syria. However, this domestic limitation did not take away from the fact that, from the perspective of LOAC, a lawful ISIL target was a lawful ISIL target for the UK whether located in Iraq or in Syria. To the extent that Reyaad Khan was a member of ISIL’s armed forces, his position and targeting under LOAC was straightforward. The difficulty lied elsewhere. First, to rely on self-defence, the UK clearly had to embrace the unable or unwilling standard. Second, since Parliament did not permit the deployment of British forces in Syria as part of the coalition effort to support Iraq, the Government could only act in Syria pursuant to individual self-defence. That is exactly the justification that David Cameron gave for the strike in Parliament in September 2015. However, internationally, nothing prevented the UK from invoking both the individual and the collective right of self-defence to justify the strike, which it did in its letter to the Security Council. (The apparent discrepancy was then seized upon by the opposition.)

    The point in all of this is that by the time of the Khan strike, the UK was already deep in ad bellum and in bello territory. Though as far as the Khan strike is concerned, not, I might add, in ECHR territory.

  20. Marty Lederman

    Marko: Thanks. The other way of dealing with the Art. 51 language is, of course, to point to the French version–“est l’objet d’une agression armée,” which, I thinks, fits more comfortably with the practice and understanding, and does not tie us into pretzels parsing “occurs.”

    Aurel: I agree entirely that the Khan strike is not as groundbreaking to the extent it was justified on the ground that the UK was already in a NIAC with ISIL (which it has done after-the-fact). Even then, however — and as I discussed in my post w/r/t the U.S. and the Khoresan faction of AQ — for the UK to breach Syrian sovereignty *within* that NIAC, it would still have to be necessary for the UK to use force *in Syria* (and perhaps against Khan, in particular, since that’s all it was doing at the time in Syria). That showing can, of course, be met by virtue of the fact that ISIL was using Syria as a base for attacks against Iraq (and France), but for the reasons you explained, the British didn’t invoke that rationale under later.

    So it’s not so much the *strike* against Khan that was so groundbreaking legally, as much as it was the rationale the UK articulated, which conspicuously did not depend on an existing NIAC. Which doesn’t mean the rationale was dubious, particularly in light of these representations that Cameron made to the House of Commons:

    Junaid Hussain and Reyaad Khan were British nationals based in Syria
    and were involved in actively recruiting ISIL sympathisers and seeking to orchestrate specific and barbaric attacks against the west, *including directing a number of planned terrorist attacks right here in Britain, such as plots to attack high-profile public commemorations, including those taking place this summer.*

    We took this action because there was no alternative. In this area, there is no
    Government we can work with; we have no military on the ground to detain those
    preparing plots; and there was nothing to suggest that Reyaad Khan would ever leave Syria or desist from his desire to murder us at home, so we had no way
    of preventing his planned attacks on our country without taking direct action.

  21. Federica Paddeu

    Hi Marko,

    Interesting hypothetical – and equally interesting discussion of it!

    I’ll just leave a brief note from a generalist’s perspective. The plea of state of necessity (art 25 ARSIWA) also contains a requirement of imminence. Interestingly both the ICJ in Gabcikovo and the ILC have understood imminence as a concept relates not just to time but also to certainty. That is, ab imminent threat is not only one that is about to happen but also one about which, although at some distance in the future, there exists some degree of certainty (what degree of certainty is don course a very complex issue – the future is always uncertain, after all).

    This is of course as a matter of general international law, and it is not the case that imminence in the context of the plea of necessity is identical to imminence in the jus ad bellum. Nonetheless, perhaps worth bearing it in mind in case jus ad bellum possesses no answers to the question.


  22. Kriangsak Kittichaisaree

    Dr. Paddeu,

    Lots of thanks for your voice of reason from across the Mathematical Bridge. It makes me think harder re: Marko’s hypo(s).

    In jus in bello/IHL, the standard used (e.g., by the ICTY) re: military “necessity”, proportionality and so forth is the subjective standard of the commander/soldier in the field at the relevant time and place. This may be because we are talking about the mental element of persons charged with war crimes.

    In jus ad bellum, the ICJ (e.g., in Nicaragua v. USA and subsequent cases) uses the objective standard in assessing whether self-defence is OK in the circumstances, whereas in reality States do make “subjective” decisions whether to resort to the right of self-defence. Of course, more militarily powerful States may abuse their military superiority by interpreting the right too broadly (cf. the Bush doctrine). But less militarily powerful States are also under extreme pressures from their citizens to do something to protect the sovereignty, territorial integrity, etc. Using all the hypos given to us by Marko, the powers that be in a State would assess from their “subjective” perspective whether any of the hypos deserve to be responded by armed force in “self-defence”. “Some degree of certainty”, “imminence” =? It is always subjective, isn’t it? Who are there to provide an “objective assessment” of the situation to guide the putative victim States? The UNSC (but beware of the veto!)? Regional IOs (but may be too late)? When matters come before, say, the UNSC, reactions in the UNSC are always “subjective” depending heavily on the international political stances of each of the UNSC members. Only when matters come before the ICJ does the objective assessment come in, although the ICJ itself needs to take into account State practice and opinio juris which reflect the subjectivity of States in responding to such matters.

    Of course, the “subjectivity” problem might not exist if, as Jordan insists, self-defence is only OK “if an armed attack occurs” (esp. if this means “has already occurred” and physical damage is felt in the victim State).

    This is the messy world we live in. IL is supposed to be the law of peace and hopefully the international community can find a way to respect IL and make the world a more peaceful, and beautiful one, like the River Cam in springtime.

  23. Marko Milanovic Marko Milanovic

    Thanks again for the comments. Federica, with regard to your very helpful point, my growing impression of how many (most?) people use the word ‘imminence’ in the jus ad bellum context is that it is not really a temporal criterion at all, but a certainty/likelihood causal criterion. In that sense, for example, we are CERTAIN that on 12 August 2026 there will be a total solar eclipse visible in Europe. We are also CERTAIN that in 5 billion years (give or take) the Sun will have exhausted its core hydrogen, exit its main sequence and start turning into a red giant, engulfing (and destroying) the inner planets. For both of these events we are as certain that they will occur as humans can be certain about anything. In the sense that many lawyers seem to be using the word ‘imminence’, these two events would be regarded as ‘imminent,’ even if they will take place decades or even billions of years from now. The only relevance of time is in the assessment of whether it is NECESSARY to act now, when there is ample time to consider other options. Or, the word ‘imminence’ is used in an entirely circular way (an attack is imminent if its necessary to act now to repel it, but it is not imminent if there is no necessity to act.) Regardless of whether this approach is correct or not, however, for many other people/states the word ‘imminence’ would imply a temporal proximity criterion (‘about to occur’), which creates a significant potential for miscommunication/misunderstanding.

  24. Marty Lederman

    Marko: Just right about how the word “imminence” does almost nothing but confuse and obscure in this context. We shouldn’t forget that the word does not appear in the relevant legal “texts”–Article 51 and the Caroline exchange of letters, in particular. Therefore, it’s a bit strange (I’d say: counterproductive) to expend so much energy explaining/debating what it means–and giving speeches devoted to it, as if we needed to uncover the hidden, meaning of a critical, canonical code. Government officials do themselves no favors by continuing to focus on it. (And that’s not even to account for the confusion engendered by the fact that a similar concept of “imminence” also comes into play w/r/t several other questions governing the legality of a strike, e.g., whether it complies with human rights law, with U.S. constitutional law (when the President acts without legislative approval), statutory questions, possibly regulatory norms, etc. Audiences typically do not know what legal question(s) is (are) being discussed when governments speak of “imminence.”)

    Two other, related points:

    First, if we *were* parsing a legal text, it’s not at all clear that “imminent” would mean “about to occur.” As the U.S. 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.” Impending, menacing, hanging over one’s head: These can describe your solar eclipse and explosion, too.

    Second, in the jus ad bellum context, the use of “imminence” is, of course, a shorthand for an idea contained in an exchange of 1837 diplomatic letters, an idea that does include *some* aspect of temporal proximity, namely, that the need to act in self-defense (not the attack) must be “instant,” leaving “no moment of deliberation.” That can be true even if the attack is sometime later in the future, or its timing is uncertain and unknowable. Might there be time for “deliberation” that might forestall the need to use force? (Webster said yes: There was still a prospect that New York authorities would act against the rebels. Ashburton’s response: Give me a break. We’ve waited long enough, and it’s now manifestly apparent the NY authorities have no desire or design to arrest the rebels. Further deliberation would be pointless.)

  25. Jack McDonald


    WRT to “There is a whole lot of background to the Khan strike, namely the ongoing NIAC to which the UK and ISIL were opposing parties” – as far as I can see, the UK has been at pains to avoid the question of whether or not it was engaged in an armed conflict.

    See, for example:

  26. Kriangsak Kittichaisaree

    Imminence in real life?: