Home Armed Conflict On Whether IHL Applies to Drone Strikes Outside ‘Areas of Active Hostilities’: A Response to Ryan Goodman

On Whether IHL Applies to Drone Strikes Outside ‘Areas of Active Hostilities’: A Response to Ryan Goodman

Published on October 5, 2017        Author: 

Over on Just Security, Ryan Goodman has an excellent post entitled Why the Laws of War Apply to Drone Strikes Outside “Areas of Active Hostilities” (A Memo to the Human Rights Community). In sum, Ryan argues that human rights activists have been too radical in their critique of US drone strike policy, as reflected in the Presidential Policy Guidance adopted during the Obama administration, and in the context of the Trump administration’s recent proposal to revise this standing policy and relax some of its requirements, especially with regard to the procedure for authorizing lethal strikes. In particular, Ryan argues that human rights activists have been portraying as clearly unlawful decisions which legally fall within the bounds of reasonable disagreement.

In that regard, Ryan argues – persuasively in my view – that the mere fact that a drone strike takes place outside an area of active hostilities under the PPG does not mean that the strike takes place outside armed conflict under IHL. The former, as Ryan correctly notes, is not even a legal term of art. I also agree with Ryan that some US positions that used to be regarded as novel or anomalous have become mainstream with time, in part through the acceptance of these positions by European and other states, by the ICRC and scholars – viz., for instance, the idea of ‘spillover’ NIACs (for more on the operation of this mainstreaming process see here; on spillover NIACs see here).

That said, Ryan in some respects significantly overstates his argument. Yes, states have accepted the idea that they can be engaged in an armed conflict with a terrorist group – but I would say that this really was never in doubt. What was in doubt is whether this NIAC can be global in scope, and this US position has not been mainstreamed – or at least I am unaware of any other state which agrees with it. What do I mean by this?

The US position is that it is engaged in armed conflict with Al-Qaeda and associate forces, full stop; not that there is a series of distinct non-international armed conflicts with Al-Qaeda in Afghanistan or AQI in Iraq or AQAP in Yemen, all of which have their own independent existence after (arguably) having reached the organization and intensity thresholds required by customary IHL, and all of which may independently end. Combined with the completely nebulous concept of associated forces and its further application to entities with a very diffuse organizational structure, this US theory has allowed it to claim that many of its operations were conducted in the course of an armed conflict, when, in fact, under the mainstream view they would not have been. The only body of international law that could have regulated the use of force in these circumstances would be human rights law, in addition to the jus ad bellum.

It is hence this US view of a single, non-territorially localized NIAC with Al-Qaeda which remains problematic. And it is because of that underlying premise that Ryan’s reliance on the acceptance of ‘spillover’ NIACs is also problematic. I personally have no problem in accepting that, for example, an ISIS commander who was directing hostilities against Iraqi and coalition forces in Iraq and who crosses the border into Turkey and continues directing his forces in Iraq from there remains subject to IHL and that the NIAC has spilled-over into Turkey. He would continue being a lawful target under IHL, but Iraq/the coalition would have additional jus ad bellum obligations vis-a-vis Turkey. I also agree with Ryan that it makes no sense to confine this spillover effect to physically neighboring states. That effect would continue if, say, the commander boarded a ship on the high seas or went to Thailand, so long as what he continued doing had a nexus to the ongoing NIAC in Iraq. But this is still a far cry from the US theory that any individual in the world who professed allegiance to Al-Qaeda (or ISIS, or whoever) was a fighter in a global NIAC which has no nexus to specific hostilities in any given place.

So, while I think Ryan is right in saying that we need to be careful in drawing legal conclusions based on the formulations of a US policy guidance, it is not correct to say that all US drone strikes have been conducted in the context of an armed conflict recognized as such under IHL as properly interpreted (to be fair, I’m not sure whether Ryan even thinks that). In other words, if is the set of all historical US drone strikes, it has two sub-sets: DIHL and DIHRLthe latter of which is regulated only by IHRL and has a non-zero content (the jus ad bellum aside, and misguided US positions regarding the non-extraterritoriality of IHRL also aside). Footnote – the best such UK example would be the August 2015 drone strike against ISIS recruiter Reyaad Khan in Syria. Within DIHL there is a further sub-set – let’s call it DIHL+IHRL – in which IHRL not only applies in parallel to IHL (which it does always, in my view), but does so in way that modifies the outcome that would have been correct if IHL had applied alone – for example, by imposing an additional requirement of capture before kill in specific circumstances, as in the Israeli Supreme Court’s Targeted Killings case.

A virtue of Obama’s PPG was precisely that, even though it maintained the rigidity of the US legal position that IHRL did not apply to an extraterritorial taking of life, it substantially reduced the risk of an IHRL right to life violation in both DIHRL and DIHL+IHRL scenarios, by minimizing collateral damage (which IHRL does not categorically prohibit in any event – viz. Finogenov v. Russia) and by moving the use of lethal force against the intended target closer to an option of last resort. (This of course assumes that the PPG was in fact applied as written, which is probably assuming too much). The new Trump policy will at least prima facie escalate this risk, as does the US continued insistence on the global NIAC theory which, again, to my knowledge other states have not endorsed, and which clearly (and I use this word advisedly) has no basis in IHL.

Print Friendly, PDF & Email

38 Responses

  1. Marko, you lost me when you got to Reyaad Khan. Why exactly does IHRL, specifically the ECHR, apply to that strike? And why would it require the UK to capture, rather than neutralise, a member of an opposing organized armed group?

  2. Marty Lederman

    Marko: I agree with much of what you say, but not all. Most importantly, what’s the (legal) source of this idea that one must do a nation-by-nation geographic analysis of the scope of NIACs? Has *any* state argued as much? If there’s an overall conflict between State X and Group Y, what difference does it make *for IHL* (not Charter) purposes in what nation (or area of a nation) the Group Y forces happen to be at any particular moment? Do any states impose such a nation-by-nation determination (again, respecting IHL application, not Charter constraints) w/r/t enemy forces in IACs? If not, why impose it as to enemy forces in NIACs? More importantly, even if one could articulate a reason why such a state-by-state assessment might be appropriate for application of IHL, is there any evidence of any states adopting it–let alone it rising to the level of customary law?

    Take the Yamamoto case, for instance. The U.S. kills him in (or over) the Solomon Islands. Did any state–let alone most, or all–argue that IHL (both its restrictive and its permissive aspects) did not apply to that action? I don’t think so.

    OK, same case, but now it’s an al Qaeda commander. The U.S. gets word of his flight plans, and shoots him down over the Solomon Islands. IHL doesn’t apply? Why not? And what state has ever argued this?

    My main point, however, is that although this dispute seems to suck all the air out of the room among LOAC/IHRL lawyers, it is, thankfully, the stuff of law school exams, not actual decisions in the world. The U.S., as well as France, et al., have recently used force exclusively (or close to it, as far as I know) in locations where the enemy is (as we put it in the OLC opinion about Yemen) “conducting terrorist training in an organized manner and has executed and is planning to execute attacks against the United States,” i.e., in safe havens. Has *any* state taken issue with the application of IHL to actions by the US and AQAP in Yemen; to the UBL operation in Abbottabad; to France and ISIL in Syria? I don’t think so. Yet those are the actual sorts of contexts in which the question is raised these days. And in that context, there is not, as far as I know, any real dispute about the application of IHL, which I took to be Ryan’s principal point. (There is, of course, a distinct question, in *any* location, of how IHRL applies in conjunction with or in light of the armed conflict. But that ought to be debated and discussed apart from the question of whether IHL “applies.”)

  3. Fantastic post. What Ryan misses is that critics of the US are not (with the occasional misguided exception) claiming that IHL does not apply to the targeting of an individual outside of a NIAC’s area of active hostilities. They are challenging how the US connects certain individuals located outside of an area of active hostilities to the NIAC itself. Concepts like “associated forces,” “co-belligerency,” and “belonging to” do not apply in NIAC, yet the US consistently relies on them to expand the category of individuals whose targeting is supposedly governed by IHL. To see this, we simply have to turn to Ashley Deeks’ recent post at Lawfare (, in which she says that the US believes it is free to target any member of an OAG that merely provides assistance to an OAG involved in a NIAC with the US. That idea has no basis in IHL, as Annyssa Bellal’s excellent post earlier today on this blog demonstrates.

  4. Marty Lederman

    Ashley doesn’t say that, Kevin (indeed, the Johnson formulation she quotes suggests a very different view). Besides which, that is *not* the issue that Ryan (and those to whom he’s responding) are addressing.

  5. Max Brookman-Byrne

    Well, that was an example of pressing ‘post comment’ a little prematurely… Here it is again, slightly edited.

    Thanks Marco for an interesting response to Ryan Goodman’s initial post. It’s good to see that the distinction between the rules governing the extraterritorial use of force under the PPG and those of LOAC is being firmly restated as the two often appear to be virtually conflated, particularly within the discussion on drones. I agree with you post entirely.

    It’s also pleasing to see more discussion around the discrete nature of the various conflicts in which the US has employed drones vis-a-vis the assertion of a global war on terror. In case it’s of interest, I wrote about this recently in the Netherlands International Law Review, considering in turn the intensity and organisation thresholds of the hostilities in Pakistan, Yemen and Somalia:

  6. Marko Milanovic Marko Milanovic

    Many thanks to everyone for the comments; I’ll try to respond to as many of the points made as possible.

    Aurel, regarding the IHRL applicability question – a drone strike would be covered by the ICCPR, ECHR either because (1) killing someone constitutes authority, power, control over that individual, and thus jurisdiction in the sense of the relevant clauses of these treaties; or (2) because (as is my preferred view) the negative obligation to respect HR is not territorially limited. There is of course the dreary Bankovic decision of the ECtHR under which drone strikes would prima facie not be covered in the absence of territorial control; but that decision is simply wrong and has already been partly overturned.

    As for the substantive question, if IHRL did apply to the Khan operation then IHRL allows the use of lethal force only if non-lethal means are first exhausted, on a reasonable appreciation of the facts (cf. McCann etc). This doesn’t mean that the Khan strike was unlawful/a violation of the right to life, but that his killing has to be justified within that framework, as it can be – if the UK could show that Khan posed a threat to the life of people in the UK, that no capture option was reasonably feasible, and that no other option could present itself (e.g. waiting for him somewhere) then lethal force could be used consistently with IHRL. As indeed happens on a daily basis – the police or even the military in say France have used deadly force against terrorist attackers perfectly lawfully (at least prima facie).


    The legal source for the idea that a NIAC has to be localized the text of Common Article 3, which in fact invented the whole idea of a NIAC as a legal concept, and which speaks of ‘armed conflict not of an international character occurring in the territory of one of the High Contracting Parties.’ There has to be a territory in which ‘protracted armed violence’ is taking place. That territory cannot be the world. There has never been, as far as I known, a claim that such a global NIAC was conceivable before the US Govt espoused that theory. And that theory was a BACKUP theory, which was used after the first theory – that the GWOT was a weird INTERNATIONAL armed conflict between the US and AQ was shot down by the USSC in Hamdan. Every example of a NIAC on record that we have aligns with what I said. So those professing a novel theory have the burden of proving that theory has become legally accepted, not the other way around.

    When it comes to your Yamamoto example I agree that it can be analogized to NIACs – that’s what I said above with regard to the ISIS commander who boards a ship or goes to Thailand. But the US govt went far beyond that – it has for example detained loads of people who clearly had nothing to do with any discrete armed conflict (e.g. the Boumediene litigants who were picked up in Bosnia, or the El-Masri guy who was rendered from Macedonia, etc). And the theory there was precisely that these people were ‘enemy combatants’ in a global war.

    As for Osama bin Laden, it’s at the very least not obvious to me that IHL applied to that operation, and if it did on the basis of what theory. There’s the US global war view on the one hand; another would be that he (arguably) took an active part in commanding AQ efforts in Afghanistan, thus creating a nexus to that armed conflict. But even if IHL did apply, this is paradigmatic example of the +IHRL scenario I pointed to above, in which he could have been captured rather than killed quite feasibly. See this old chestnut for more:

  7. Marko Milanovic Marko Milanovic

    Btw, as for the ICRC position on this matter – see, e.g., this report, p. 10:

    “the ICRC does not share the view that a conflict of global
    dimensions is or has been taking place. … the ICRC
    has taken a case by case approach to legally analyzing and classifying the various situations of violence that have occurred in the fight against terrorism.”

  8. Marko, thanks for the reply. On applicability, I assume that you would also regard the CoA decision in Al-Saadoon to be wrong on the question of control.

    Without entering into an extended discussion on control or bringing in lex specialis, just one follow-up on the substantive issue. Accepting, for the sake of argument, your suggestion that IHRL requires the exhaustion of non-lethal means before a lethal strike against Khan is permissible, I wonder whether the use of extra-territorial lethal force could ever be permissible in a situation, like that of Khan, where the State seeks to avert an extra-territorial threat against targets located in its own territory. Would the geographical distance not always suggest that the threat is too remote and therefore not sufficiently immediate? Meaning that the anticipatory use of force, as in Khan, could not be justified under IHRL? Meaning that a State would in effect have to wait until the threat becomes more immediate, for example by entering its execution phase?

  9. Marko Milanovic Marko Milanovic

    Thanks Aurel. No, I don’t exactly think that the CoA was wrong in Al-Saadoon – as I explain here both the High Court and the CoA were right in a sense. The HC was right that on any principled conception of jurisdiction as authority and control over individual, jurisdiction has to include killing. The CoA was right that the ECtHR didn’t want to go so far in Al-Skeini; as I explained before they tried to artificially prevent the personal model of jurisdiction from collapsing in order to preserve the result of Bankovic.

    As for the substantive issue: I see no reason using force an individual in defence of another group of people should depend on the location of those threatened, e.g. whether Khan posed threat to the people of UK or the people of Syria or the people of Zanzibar. What matters is whether the threat was real/imminent. My understanding of the facts is that Khan was actively planning numerous attacks on UK soil, some of which had already been attempted. He wasn’t simply recruiting others to join ISIS. If the UK had concrete information that he was for example preparing to travel to an area in which he could feasibly have been captured, then the UK would have had to wait. If he was located in a territory whose government could have done something to stop him, the UK would have had to wait. But from what we know his killing could plausibly be justified under the IHRL framework – depending on a more detailed examination of the facts which we do not fully have in our possession. It is precisely the fact that he was located in Syria that reduced the non-lethal options that the UK might have had available.

  10. Kevin Jon Heller

    Wrong again, Marty. Quoting Ashley’s post: “Approach 3 – Assistance by Group Y to Group X in Group X’s NIAC against State A is enough to render Group Y a functional co-belligerent, even if Group Y does not directly engage in hostilities against State A…. This approach may more accurately describe the U.S. approach in practice. Not only has the government cited the Goldsmith/Bradley article in its briefs, but one commentator also suggests that the Executive “may accept some level of support to al Qaeda . . . as sufficient to render a group a co-belligerent.”

  11. Kevin Jon Heller

    That comment, of course, follows Ashley’s statement that “At first glance, the U.S. government appears to employ this [ie, Johnson’s] approach.

  12. I’m not so sure on the substantive issue, Marko. Active planning to me does not seem to meet the IHRL imminence standard as the threats did, say, in McCann, Finogenov or Tagayeva.

    If Khan was planning to personally carry out an attack on UK soil, it is difficult to see how the absolute necessity standard could be met before he actually began to travel to the UK. Using lethal force against him before that time would in effect amount to a pre-emptive killing.

    The situation is different if he had access to a weapon that he could deploy in Syria to achieve effects in the UK (eg cyber). The imminence requirement may be met, as killing him may be absolutely necessary to prevent him from deploying that weapon.

    But this must be distinguished again from a situation where Khan is controlling accomplices in the UK, as in this case the threat emanating from Khan is indirect. Again, I’m not confident that the absolute necessity requirement would be met.

    Of course, this all presumes that the ECHR applies in the first instance (which I dispute in the actual Khan case). Also, I accept that hypotheticals only get us so far. Whether or not the standards are met will depend on the facts of each individual case. Nonetheless, I think that these hypotheticals illustrate that location does play a role, since it either requires the person posing the threat to travel and potentially enter the territory of the target State, in which case the target State will have a broader range of non-lethal options available to deal with the individual, or it will strech the causal chain, in which case imminence and necessity become far more contentious.

    The bottom line is this: if the ECHR applies, the use of lethal force against overseas targets presenting a threat to a State inside its own territory may be difficult, or impossible (depending on the facts), to justify under IHRL without recourse to in bello and ad bellum.

  13. Jens Iverson

    I’d be very curious to see a further amplification of this…

    “When it comes to your Yamamoto example I agree that it can be analogized to NIACs – that’s what I said above with regard to the ISIS commander who boards a ship or goes to Thailand. But the US govt went far beyond that – it has for example detained loads of people who clearly had nothing to do with any discrete armed conflict (e.g. the Boumediene litigants who were picked up in Bosnia, or the El-Masri guy who was rendered from Macedonia, etc). And the theory there was precisely that these people were ‘enemy combatants’ in a global war.”

    It seems to me your underlying logic is relying at least as much contingent on your analysis of the nature of the organization of armed group (esp. re membership) as it is on the nature of the territorially-bounded armed conflict. If the individuals in the Balkans were undoubtedly members of an organized group and that group was in a NIAC with the US outside of the Balkans, would IHL be applicable to the conduct between the US and the individuals in the Balkans? How do you distinguish this with your ISIS commander in Thailand example?

  14. Marty Lederman

    Thanks very much, Marko. I’m still confused, however. I’m intrigued by your invocation of CA3, which I don’t recall as a move others have made in this debate (although I’m sure I’m overlooking something). In any event, CA3 does not say anything about scope of the conflict, or “protracted armed violence,” let alone that “a NIAC has to be localized.” It merely provides that CA3 itself applies to NIACs that occur “in the territory of one of the High Contracting Parties.” The NIACs in question do, of course, “occur” in the territories of several such Parties, such as Afghanistan, Syria and Yemen, and thus CA3 is applicable (that’s the holding of Hamdan)–which is, of course, a very good and humanitarian conclusion, one that the Bush Administration fought tooth and nail.

    The question on the table, however, is not the application of CA3, but the application of the customary LOAC. And just as LOAC applies to the parties’ use of force in an IAC wherever it may occur (e.g., Yamamoto), I don’t see any reason why it should not likewise apply to the parties’ use of force against one another’s forces wherever they happen to find them in a NIAC. Surely states always have considered themselves bound by the laws of war in such cases, including before there was a legal idea of an “armed conflict” as such–e.g., the British against the rebels in the Caroline; the U.S. against the Viet Cong in Cambodia; etc. And, more to the point, I am not aware of any State that has *denied* that IHL would apply in such a case, not that many have had any occasion to opine on the matter. (As you know, I think this is, in practice, a salutary and humanitarian thing, because of the *restrictive* aspects of IHL–there’s a reason the Bush Administration was so resolute in trying to deny application of the LOAC–but that’s a discussion for another day.)

    Now, I surely agree that the Bush rhetoric of a “global NIAC” or GWOT was ill-advised, counterproductive, and deeply misleading, in large measure because, as Obama tried to stress, it doesn’t describe anything the U.S. or others are doing or are likely to do. Other constraints — the Charter; IHRL; the everyday realities of diplomacy and international relations — conspire to ensure that the U.S. will rarely if ever engage in status-based targeting (or even “self-defense” targeting of a specific threat) over 99% of the Earth. The real, practical question that confronts us, that is to say, is the use of force in the safe havens (Syria, Yemen, the FATA) from which the enemy plans and operationalizes attacks. And there, if I’m not mistaken, you agree that IHL applies (and you have not indicated that any State has said otherwise).

    Even as to the rare, hypothetical Yamamoto-type case, however (“the ISIS commander who boards a ship or goes to Thailand”), you seem to agree that IHL would apply! In which case I don’t think we differ (except, perhaps, that I would consign such aberrational cases to law school exams). You then write, however, that “the US govt went far beyond that” in detaining the Boumediene litigants who were picked up in Bosnia and the “El-Masri guy who was rendered from Macedonia.” Those were aberrational cases early on, which the Obama Administration did not repeat or endorse because it concluded, properly, that ordinary law enforcement tools can and should be used in such cases, regardless of what IHL would allow. I suspect that even the Bush Administration, however, would never have targeted such persons with lethal force in those places, which almost surely would have violated the Charter. More to the point, as a matter of *IHL* (not the Charter or IHRL), why are those cases “far beyond” our hypothetical AQ commander in Thailand, the targeting of whom, we agree, would be governed by IHL (but not to the *exclusion* of the Charter and IHRL)? Aren’t they, in fact, less extreme cases, because they don’t involve targeting?

    As you know, and as others can probably tell from this exchange, I think the real question, if and when such cases outside safe havens ever again arise, is not whether IHL applies–of course it does, thankfully–but whether IHRL should be understood to impose additional constraints in the rare, perhaps hypothetical, instances in which the host state consents to the other state’s use of lethal force in circumstances where the host state could ameliorate the threat through ordinary nonmilitary means. I think it’s absolutely clear that the host state itself would violate IHRL by giving consent to the use of lethal force in such a case, even if the acting state itself probably would not (the hypo being that it has no other means of dealing with the threat).

    Again, however, this is truly a hypo that rarely if ever occurs in the world. The recurring *practical* issue is whether it is permissible to use force against an NSA with which one one is otherwise (indisputably) in an armed conflict, when that NSA takes refuge in an ungovernable space in a new state, from which it then engages in its belligerent activities. I don’t think we — or any state, for that matter — disagrees about the answer to that question. (The Khan case, of course, is more difficult precisely because the UK probably was *not* in an armed conflict with ISIL when it decided to target him.)

  15. Kevin Jon Heller

    Marty continues to attack a strawman. Few if any US critics disagree with his claim that there is no reason IHL “should not likewise apply to the parties’ use of force against one another’s forces wherever they happen to find them in a NIAC.” Marko doesn’t. I don’t. I’m quite certain Gabor Rona doesn’t. The question that Marko’s post addresses and that I addressed in my comment — and the question that Marty seems to believe is irrelevant — is who qualifies as a member of “another’s forces” in “a NIAC.” Or, to be legally accurate, the question is who qualifies as a member of an organized armed group engaged in a NIAC with the US such that it is lawful to target him anywhere at any time. (And as Marko rightly points out, hostilities have to be localized somewhere for a NIAC to exist, even though IHL is aspatial with regard to targeting.) If an individual is not a member of “another’s forces” in “a NIAC,” his targeting is governed by IHRL, not by IHL. And as many scholars have pointed out, including Marko in this post and Annyssa Bellal in yesterday’s post, the US’s view of who qualifies as “another’s forces” in “a NIAC” is profoundly at odds with IHL.

  16. Marty Lederman

    Kevin, Marko, Ryan, Monica, et al., surely can speak for themselves, but I’ll simply say that I do not read their debate to be about “who is a member?” in the slightest — it’s about the geographic parameters, if any, of IHL in a NIAC.

  17. Marty, what makes you think that the UK may not have been in an armed conflict with ISIS when it targeted Khan?

  18. Marko Milanovic Marko Milanovic


    The question of membership, as you put it, is in fact related to the issue of geographical scope, because it is a criterion in asking whether a specific individual can be targeted or detained under IHL. It is in other words impossible to answer the question whether IHL applies to the killing of individual X without also asking the question whether X is a member of an armed group/has a continuous combat function within that group/is a civilian directly participating in hostilities within the specific armed conflict at issue. The two questions are analytically separate, but they come together whenever we apply the law to a given operation.

    To further bring together yours and Jens’ comments, let’s develop a hypo. Imagine my ISIS commander in Syria gets a hold of a ship, equips it with missiles and crew and they start sailing towards the United States from a Syrian port, with the intent of striking at targets in the US. The commander and the crew clearly satisfy whatever membership/CCF/participation option one prefers, as above, so we can focus on the geographical scope of IHL issue.

    A US drone hovers above the ISIS ship while it is in port. Would the strike there be covered by IHL/fall within the scope of the pre-existing NIAC? Clearly. What then if the ship sails off, and the strike takes place within Syrian territorial waters? The answer should again be yes. The ship is now on the high seas – should the answer change? I think not. The ship is now in the middle of the Atlantic – should the answer change? Again I think not. The ship is now in US territorial waters… again, you see my point. A different question is whether at any given point IHRL+ kicks in. But accepting that IHL can move with the fighters – just as if it would cover all hostilities between the US and North Korea, no matter their exact location – doesn’t mean the US global NIAC theory is correct. There still has to be a nexus to a specific, localized armed conflict.

    And btw, if we look at Al-Qaeda Central in particular, its involvement in most actual, specific armed conflict has ended, has it not? Where, exactly, at this moment in time, are hostilities/’protracted armed violence’ taking place between the US and AQ? It seems to me that the answer is nowhere.

    Anyway, for a fuller examination of this I highly recommend this paper by Lubell and Derejko – I agree with almost every word thereof:


    1) On what precise theory do you reject the extraterritorial application of IHRL to kinetic operations? There has to be a normative argument in that regard, and I at least am yet to see one even remotely persuasive. Do you have one?

    2) Re Khan, again on precisely what theory do you argue that the UK and ISIS were in armed conflict when it targeted Khan? Where did the ‘protracted armed violence’ take place between the UK and ISIS in mid-2015?

  19. Marty Lederman

    Marko: Now I am fairly confident that we are in virtually complete and stubborn agreement! I concur in virtually everything you write in that comment — including, of course, that the question of membership is very significant in the *application* of LOAC to a particular use of force. (But it is not, I think, directly germane to the geographic scope question we’ve been debating and that Ryan was addressing.) Moreover, I agree that Noam’s article with Derejko is, not surprisingly, fantastic. I’m fairly certain that, like you, I agree with almost every word of it. ( I remain agnostic on whether “The application of IHL is dependent upon the existence of an armed conflict,” but that’s a topic for another day.)

    But here’s the thing that might surprise you: I think the U.S. agrees with all of your post and most of Noam’s article, too–or it did until January, anyway. 😉

  20. Marko, let me reply briefly.

    1) I do not reject the extra-territorial applicability of IHRL to kinetic operations as such. Rather, I dispute that the ECHR applied to the strike on Khan. The UK did not control the territory in which the strike took place. It did not conduct the attack with the consent, acquiescence or at the invitation of the Syrian authorities. Nor did it do so on the basis of an assumption of authority based on a Security Council resolution, as the ECtHR found the Netherlands did in Jaloud, given that the strike was undertaken in the exercise of the right of self-defence. That leaves the question whether the UK exercised effective authority and control over Khan, within the meaning of the ECtHR’s jurisprudence. In my reading of that case-law, what is required under this standard is effective, factual control. My argument is that the conduct of a lethal strike that does not, and cannot, establish an agent-principal relationship between the target and the attacking force does not constitute control in the sense that the term is understood in international law. I am aware of Isaak and Andreou, but these cases can be distinguished without too much difficulty.

    2) Regarding the existence of a NIAC between the UK and ISIS at the time of the Khan strike (21 Aug 2015), it is not so much a theory that I am relying on, but facts. I assume that no one is doubting that ISIS constituted (or had at its disposal) an organised armed group. The question therefore centres on intensity. Operation Shader began in September 2014. The number of RAF strikes in its first few months were relatively low, but they quickly picked up in terms of pace and scale. By August 2015, strikes were carried out on an almost daily basis. The data is available here: Note that these numbers only deal with attacks, but do not account for other operations, such as ISTAR, which of course also form part of the conduct of hostilities against ISIS. I’m content to conclude that the criteria for the existence of a NIAC are satisfied based on these figures. However, Operation Shader did not take place in isolation, but at the invitation of the Iraqi Government, which undoubtedly was engaged in a NIAC with ISIS (an AP II NIAC no less) and as part of a broader coalition effort. That coalition effort included the US, which carried out a greater number of strikes than the UK during this period. Applying Tristan Ferraro’s ‘support-based’ approach, I am even more content to conclude that the UK was party to a NIAC against ISIS at the relevant time. The fact that the strike against Khan occured in Syria has no impact on the existence the NIAC between the UK and ISIS. Some might think that it had an effect on the legal status of Khan. I would disagree on the basis that the strike against Khan took place in the context of the pre-existing NIAC and formed an integral part of it. This strike therefore does not engage the question of spillover at all, as the NIAC involving ISIS was translational (covering both Iraq and Syria) from the beginning. The fact that the UK for some time chose to engage ISIS only in the territory of Iraq does nothing to alter this.

  21. Ryan Goodman

    Despite all the discussion here about which side of this debate bears the burden of proving State practice and States’ views of their legal obligations, I thought to underscore one point. Given how much esteem I have for Marko, it is encouraging to see how much we agree on (though his initial post does not fully recognize the scale of agreement).

    Marko’s framework clearly supports the application of the laws of war to the US decision to strike members of AQAP in places like Yemen. Make no mistake. That is a game changer in this debate, and legal experts will align on one side or the other on that question.

    Here’s an easy way to show our agreement: I can’t find anything meaningful to distinguish Marko’s hypothetical “ISIS commander” from Anwar al-Aulaqi. Under existing international law, Marko accepts that if his ISIS commander goes to a faraway third country [in his example, Thailand] and continues to direct attacks carried out in the NIAC [in his example, forces in Iraq] that individual, in Marko’s words, “remains subject to IHL and … would continue being a lawful target under IHL” even though he is a continent away from Iraq.

    Here’s what the declassified Justice Department’s Office of Legal Counsel opinion says about al-Aulaqi:

    “al-Aulaqi [think: ISIS commander] has been involved, through his operational and leadership roles within AQAP, in an abortive attack within the United States [think: Iraq] and continues to plot attacks intended to kill Americans from his base of operations in Yemen [think: Thailand].”

    I suppose Kevin or others may raise the question of AQAP’s relationship to Al-Qaida central as an “associated force.” But is that relevant here? From the same OLC opinion: “a decision-maker could reasonably conclude that this leader of AQAP forces is PART OF al-Qaida Forces.”

    Somewhat more difficult questions can arise, but I believe Marko and I (and Marty) may agree on these too. For example: (1) what if the ISIS Commander directed his attack against the US embassy in Ankara instead of back at Iraq? (2) what if the ISIS Commander directed his attack against the US embassy in Bangkok instead of back at Iraq? (3) what if the ISIS Commander directed his attack against a US military vessel on the high seas? Based on Marko’s analysis, it sure seems hard to say the ISIS commander isn’t also covered (and that includes protected) by the laws of war across these cases as well.

    [Note: I wrote the above before Marko and Marty’s most recent entries in this thread. In their exchange in this thread, Marty has drawn out the “virtually complete” agreement between Marko and Marty. Plus, as Marty says, what might come as a surprise (but it shouldn’t) that Marko’s analysis fits well with the U.S. position.]

  22. Noam Lubell

    Jumping in here a little late, but all my friends seem to be having an IHL party without me, so I’ll just add a few quick thoughts. I fully agree with Ryan, Marko and Marty as to the difficulties in trying to geographically restrict the application of IHL. The closing sentence of my article with Nathan Derejko was “IHL is not in and of itself pre-determined as applying to a limited geographical scope, and its applicability is designed to follow the prevailing hostilities wherever they may spread, rather than vice versa”. Those that wish to find more solid ground to critique some of the operations under discussion, would be better served by focusing on the following aspects:

    1. The ius ad bellum rules and in what circumstances they might restrict uses of force on the territory of another state.

    2. The interplay with human rights law once we are talking about operations that occur further away from an active zone of hostilities. There’s a case to be made that a) IHRL applies to all targeting operations – it does not depend on control of territory or physical control of the person, but rather on direct control over a person’s right to life. Otherwise a State could send an agent into another State on an assassination mission, and so long as the executions take place through the scope of a sniper rifle, IHRL would never be applicable. This would be an absurd result, and there is plenty of IHRL case law to support extraterritorial applicability in lethal use of force even without holding on to the person physically. b) the interplay between IHL and IHRL is context dependent, and a context in which we are looking at an operation against an individual far from any active hostilities may well require that IHRL plays a greater role and that lethal force might be judged against IHRL/Law enforcement standards rather than IHL status-based targeting.

    3. The ‘associated forces’ or who is a party to the conflict debate. This one often gets confused in a mixture of internal US authorisation to use military force debates, as opposed to ‘pure’ IHL questions. While the above ad bellum and IHRL points are contentious, there are at least clear arguments to be had and case law to assist us. But this third point is the messiest of them all and we’ll probably be arguing (and figuring it out) for a while.

    I’ve written a little about all of this on but there’s plenty more to argue over!

  23. Kevin Jon Heller

    Marty says: “I remain agnostic on whether ‘The application of IHL is dependent upon the existence of an armed conflict’, but that’s a topic for another day.”

    Now that is a post I look forward to!

  24. Marko Milanovic Marko Milanovic

    This really turned out to be quite the IHL party. Let me first (again) express my agreement with Noam’s points above. And also agree with Mary and Ryan that we are in agreement on the narrow question of geographical scope. We may also be in agreement on some of the questions of the jus ad bellum and IHRL that Noam also points to (even if the USG is not).

    But I have to stress that we are NOT in agreement on other fundamental questions, most importantly the definition of the armed conflict itself, against which any nexus requirement will be measured. So unlike Ryan I don’t think there’s much of a game-changer here. First, regarding the global NIAC, I reiterate my question – where, exactly, are the hostilities between the US and AQ taking place today, and do they reach the intensity threshold? As far as I understanding it AQ is no longer significantly involved in Afghanistan, so again, where is that NIAC taking place? ‘All over the world’ is not, in my view, a legally plausible answer.

    Second, regarding Al-Awlaki, here were run into the ‘part of’ the armed opposition group problem, because the diffuse, network-like organization like AQ, which does not operate under a single chain of command and control, is being treated like a unified entity. For me it is at the very least non-obvious that AQAP is a part of AQ, or that Al Shabaab is a part of AQ, etc., etc. I am therefore not sure, on the facts as are available, that Al-Awlaki is really analogous to my ISIS commander hypo. But even if HE WAS, I am sure that there have been loads of people detained or targeted by the US in the GWOT who are not so analogous (as e.g. the Boumediene people).

  25. Marko Milanovic Marko Milanovic


    Re 1 – you did not actually answer my question. I asked you for a normative argument as to why the ECHR should not apply to kinetic operations without territorial control, and you responded by citing ECtHR case law. I already conceded that under that case law, per Bankovic as modified by Al-Skeini, a drone strike would likely not be covered. But my whole point is that this case law is incoherent/internally inconsistent and draws entirely arbitrary distinctions. So what you’d have to do is have an argument which would demonstrate that the distinctions drawn are NOT arbitrary, and with respect, I don’t think you’ve done that. There is no non-arbitrary way to say that jurisdiction is authority, power or control over a person, and that killing that person does not constitute authority, power, control. (This was the central argument of my book on the subject, and again I have not seen a response to that argument that would be normatively persuasive).

    Re 2 – your argument, as I take it, is that Khan was a fighter in the UK/ISIS NIAC in Iraq, in essence analogous to my ISIS commander hypo? That’s a possibility, I’ll admit that, but it’s a factually contingent one. What is your evidence that he in any way participated in the Iraq NIAC/that what he did had a nexus to that conflict? If say he was running his terrorist planning activities from Germany or even in the UK rather than Syria(which would have different ad bellum implications), would in your view he still have been a fighter in the Iraq NIAC? Is any none-lone wolf ISIS member in say Europe in your view a fighter in that Iraq NIAC?

    (FWIW, your position on this point is shared by the UK Government – see, e.g., . Again, I don’t think it’s implausible, but it is factually contingent in such a way that I’m not sure that the answer you give is the right one).

  26. Marko, to continue our little fringe event to the main IHL party.

    On 1), I guess I need to work harder to convince you. This is for another occasion, but in the meantime two quick points. First, I am less concerned about the arbitrary or non-arbitrary nature of the distinctions than you are. Second, that said, I do think that a normatively defensible distinction can be drawn here by distinguishing power from control. Lethal targeting is undoubtedly an exercise of coercive power, but control requires a bilateral relationship and a feedback loop, which does not exist when you are dropping a Hellfire on someone.

    On 2), of course Khan’s status is factually contingent, but so is the battlefield status of every person. More importantly, I am not arguing that Khan was a lawful target because he had a nexus to the Iraq NIAC, but because he was (if we go by the info in the public domain) a member of ISIS, an OAG against which the UK was at the relevant time engaged in a NIAC. That OAG operated across both Iraq and Syria. Under LOAC, its members constituted legitimate targets for UK forces irrespective of whether they were present in Iraq or Syria. Again, not because of any spillover theory, but because the NIAC was not confined to a single country in the first place. My point is that it makes little sense to talk about an ‘Iraq NIAC’ between ISIS and the UK and a separate ‘Syria NIAC’ between the same two parties. It was a single NIAC between ISIS and the UK that took place across Iraq and Syria. (Think about it: if the conflict had not been transnational in character, Iraq could not have invited the UK to come to its assistance in the exercise of its right of collective self-defence, which requires the existence of a right of individual self-defence, which in turn requires an armed attack that is international (rather than purely domestic) in nature. Before I am being accused of conflating in bello and ad bellum, my point is simply that the ad bellum position taken by the States involved underlines the transnational character of ISIS and thus the transnational nature of the conflict waged against it. Or, to put it differently: the Iraq NIAC and Syria NIAC are one and the same.)

    To answer your question: if Khan had been active in Germany or the UK, rather than in Syria, would he still have been a fighter (which I understand to mean member of OAG) in this NIAC? Absolutely.

  27. Ryan Goodman


    Question 1. When in your view did there stop being an armed conflict with AQ in Afghanistan?

    (Even if such an armed conflicted ended at some point, you’ve conceded the game up until that point.)

    Question 2. Do you think there’s not an armed conflict in Libya, in Syria, and in Somalia today such that IHL does not apply to US operations in those areas?

  28. Ryan Goodman

    * I meant in Libya, in Yemen, and in Somalia

  29. Marko Milanovic Marko Milanovic

    Hi Ryan,

    Now we’re getting into very fact-specific questions, on which my own knowledge cannot compare with those who have access to intelligence of have worked in-theater. Re Q1, I’m not sure what you mean exactly by ‘conceding the game’ – even before the (arguable) end of the NIAC with AQ in Afghanistan, not every individual targeted or detained by the US in the GWOT was a member/fighter of that organized armed group – again AQAP was/is (to my mind at least) a different group than is AQ Central, as was AQ in Iraq, as is /was definitely Al Shabaab. And again I would just refer to the Boumediene claimants and other people in a similar situation.

    As for whether the conflict with AQ actually ended in Afghanistan my understanding is that since 2010 US officials said that AQ had no more than 50-100 fighters remaining in Afghanistan. That may be subject to dispute, and there has been at least one operation in which there were more AQ people – and again this is a factually contingent question, which I have no way of answering without a proper evidence base to which I have no access. (See in that regard here:

    As for Q2 sure those conflicts exist, but again as distinct conflicts between different armed groups/governments. The OAG AQ which fought or fights a NIAC in Afghanistan is not a party to any of these conflicts. They have their own parties, their own independent existence, a beginning, an evolution and eventually an end, which is not related to a supposed global NIAC against AQ.

    So, again, I repeat my question – where today is the United States engaged in intense hostilities against Al Qaeda, an organized armed group in the IHL sense? Still in Afghanistan, maybe? Anywhere else?

  30. Marty Lederman

    Alas, this “party” appears to be ending, but on a hopeful note: If I’m not mistaken, there is broad agreement among Marko, Ryan, Noam and me–and the United States!–on the “where does LOAC apply?” question that prompted this rich discussion in the first instance. (And, I’d add, there’s still no evidence of any states that *disagree.*) We also broadly agree that this does not mean that status-based targeting can or will take place over the vast majority of the planet, because other sources of law (e.g., the Charter; IHRL; the “soft law” of conditional consent and diplomatic relations) will preclude it, and will counsel if not demand the use of ordinary law enforcement methods, which are in fact the coin of the realm in virtually all locations outside those “safe havens” in ungovernable spaces that have been directly at issue in this and similar discussions (e.g., Yemen, Syria, the FATA).

    Marko, however, does raise at least two other very important, fact-intensive questions. First, is the United States correct that AQAP is part of AQ (i.e., subject to its direction and control), or at the very least an “associated force” acting in coordination with AQ in the conflict against the U.S. and its allies? Alas, I’m not at liberty to provide any details on this question–note that the first 11 pages of the OLC memo are redacted, for reasons that are probably obvious–nor am I privy to the intel that would bear on this question over the past seven years. The most I can say is that, at least as of 2010, rest assured it is a question that the President and his advisers took very seriously. That does not, of course, mean that you should take their word for it; I do think it’s telling, however, that, as far as I know, there does not appear to be much of a debate about that question among experts who follow such things closely.

    Second, and of much greater current importance, Marko wonders whether the conflict with AQ continues if, as the U.S. says so often, we have largely degraded its capabilities to fight and engage in terrorist activity. That’s an excellent question. Alas, I don’t know any more about the facts bearing on the answer than Marko does. But even if Marko is right in his suggestion that there is no longer any “intensity” of hostilities coming from AQ’s direction, that would only raise an important legal question that I believe remains unsettled: What are the legal rules that govern the U.S.’s continuing use of force against AQ?

    Imagine that the following is true: AQ was once engaged in “intense” and persistent uses of force against the U.S. and its allies, more than sufficient to create an armed conflict. Now, however, it no longer has the capability of acting as frequently, or as effectively/destructively, as it once did–but it has not abandoned its designs or objectives (there hasn’t been a surrender or armistice, etc.), and it continues to do all it can to plan and carry out attacks. Moreover, left to its own devices–if the U.S. and allies were to cease their efforts against the group–AQ would gradually (perhaps rapidly?) rebuild and revive its belligerent capabilities. Must the U.S. now abandon all status- and infrastructure-based targeting, just because the Tadic test (or whatever the proper test for AC might be) would no longer be satisfied as we sit here today? And in answering that question, shouldn’t we bear in mind that there is a great benefit in insisting that the U.S. (and all other actors) must continue to abide by the much larger corpus of *restrictive* rules of the LOAC, in targeting and detention matters alike? (That is to say: If the standard view is correct that the LOAC do not apply outside armed conflict (a view I continue to resist, or at least question), then perhaps folks should be careful what they wish for when they ponder whether to insist that the armed conflict with AQ is over–the upshot might be much less humanitarian protection rather than more, although thankfully the U.S. armed forces are committed to compliance with the LOAC as a matter of policy in all their operations.)

    I genuinely do not know quite what the customary law is on this “when does LOAC expire?” question, let alone what it *should* be. (Deborah Pearlstein, among others, has begun to examine it in greater detail.) I do agree with Marko, however, that it’s a question of deep importance and will only become more pressing in the years to come.

  31. Marko Milanovic Marko Milanovic

    Hi Marty,

    Thanks for this wrap-up. Regarding the legal requirements for ending a NIAC – I wrote a paper in the IRRC on this:

    I argue that ending a NIAC depends on the same criteria as starting it – intensity and organization. If the OAG is degraded to such an extent that it has lost the minimum required for engaging in hostilities, the conflict is over. There is no conflict without an organized adversary. Regarding intensity, there are two possible options: (1) either a NIAC ends if there is a general close of military operations, with no real likelihood of a resumption in hostilities (which is the IAC standard); or (2) the NIAC ends if the violence does not abate completely, but falls below the ‘protracted armed violence’ threshold required to initiate the NIAC, essentially by petering out.

    Whatever option one argues for, I would just caution that our primary point of reference should be the paradigmatic, ‘ordinary’ NIACs (e.g. Sri Lanka/Tamil Tigers), rather than the weird/anomalous example of the US and AQ, however properly qualified. That is, we should not reason about the rule by starting from the hard or exceptional cases.

  32. Gabor Rona

    I tried to stay away, since most of what needs be said has been said. But I’m falling off the abstention wagon simply to remind readers where we began. In February of this year, the US Director of National Intelligence issued a report asserting that in all of 2016, the US undertook 53 strikes against terrorist targets outside areas of active hostilities, resulting in 431-441 combatant deaths, and 1 – yes that’s right, 1 – “non-combatant” death. In a Just Security post of February 7, I took issue with several assertions and legal assumptions inherent in the DNI report. Ryan Goodman responded then, and again this past week. While I stand by my assertions of February 7, I want to point out that no answer to the question of whether IHL CAN apply outside areas of active hostilities settles the matter of whether or not IHL DOES apply to any, let alone all, of the 53 admitted strikes. As others have noted, in addition to the question of whether, as a matter of fact, there continues to be an armed conflict, there are questions about whether the particular targeted individual is targetable under IHL on theories of “membership” in an armed group, “associated forces,” “direct participation in hostilities,” or, as some, including the ICRC would have it, “continuous combat function (CCF).” Even accepting the controversial CCF notion, it requires magical thinking to conclude that all but one person killed by the US “outside of areas of active hostilities” was a civilian.

  33. Gabor Rona

    Oops. I meant “only one,” not “all but one.”

  34. Kriangsak Kittichaisaree

    One practical consideration seems to be missing in this lengthy debate.

    If an ISIL member escaped to Thailand, the US etc. would request the Thai authorities to apprehend him/her (and if the person fought back to resist his/her arrest, the Thai authorities could resort to the use of force according to standard rules of engagement). In other words, drone strikes against ‘terrorist targets outside areas of active hostilities’ can happen in practice only where the said areas are located in failed States and/or States unwilling to take necessary action against the terrorists.

    Any drone strike against an ISIL terrorist in Thailand without its prior consent, for example, would be considered a violation of territorial sovereignty, Art 2(4) of the UN Charter etc.

  35. Marko Milanovic Marko Milanovic

    Kriangsak – I completely agree, that’s why I repeatedly referred to additional constraints imposed by the jus ad bellum, as did Noam for example.

  36. Marty Lederman

    Thanks, Kriangsak; but that practical consideration isn’t missing at all–it’s precisely the point I (and Noam, and others) have been stressing, and it’s the basis for Obama and Brennan insisting that there is not, in fact any “global” war–over the vast majority of the planet, it’d be preposterous and unlawful (and damaging) to use military force where ordinary criminal-law means will do.

  37. JS

    Fascinating. So there’s a degree of consensus on attributing certain drone strikes outside areas of hostilities to extant armed conflicts elsewhere if there is still a sufficient nexus. Geography is relevant to nexus but not necessarily dispositive.

    But is there any consensus regarding a strike against an organised armed group that cannot be attributed to an extant armed conflict anywhere, which is legitimate as a matter of jus ad bellum and which would otherwise conform to jus in bello rules were they formally applicable? Might IHRL be modified by the jus ad bellum in the absence of formal modification by the jus in bello?

    Alternatively, and in response to Marko’s invitation for a normative justification for avoiding IHRL extraterritorial jurisdiction for such uses of force, could we not point to the absence of the intervening state’s “assumption of public powers”? That is the route by which state agent control and authority is usually made out for ECHR cases in order to trigger jurisdiction. Having justified the necessity and proportionality of the strike under the jus ad bellum you would have had to address that very same issue, namely your inability to deal with the threat by alternative more peaceful means (i.e. control of the criminal justice system).

  38. […] of the theories they have been advancing in the pursuit of the global conflict against terror. Just a few weeks ago we had just such an “IHL party” on the blog, provoked by a post of Ryan Goodman on Just […]