The NIAC Threshold

At least three things trouble me about Adil Haque’s recent post over at Just Security about how to determine when armed violence crosses the threshold from ordinary criminality or the like to non-international armed conflict (NIAC), such that the law of armed conflict applies. As Adil rightly notes, much rides on the question. On one hand, recognition of a NIAC imposes on all parties to the conflict an obligation to comply with, at a minimum, the humanitarian provisions of Common Article 3 to the Geneva Conventions (prohibiting torture, cruelty, and much else). On the other hand, under the law of armed conflict (LOAC), a state party can use force anytime and against any member of an opposing force. In armed conflict, and in no other circumstance, killing is lawful as a first resort. For this reason, among others, Adil’s suggestion that we should lower the threshold for recognizing the existence of a NIAC, i.e. apply the law of armed conflict even for nominal levels of violence involving non-state actors, merits careful attention. So here are some initial concerns…

First, I am not convinced by Adil’s description of the existing definition of NIAC as requiring a “bilateral trigger,” meaning, I take it, that both (or all) parties to the conflict must be actively engaged in the use of armed force, and that the LOAC of NIAC is not triggered until forces under attack “respond by military means.” (Adil relies on the brief discussion of the definition of NIAC in the new GCI commentary (here, para. 423 et seq), as well as a 2008 ICRC opinion paper, and the relevant rulings of the ICTY/ICTR. I accordingly review the same sources.) It is certainly true that existing definitions of NIAC require that there be at least one party on each side of a NIAC organized enough to be identifiable as a specific “party,” a party at least capable of wielding armed force itself and functioning through some sort of unified command. Existing NIAC definitions also require that armed violence between the parties reach a certain degree of prolonged “intensity.” But the requirements that there be at least two organized parties, and that there be intense and sustained violence, are both conceptually distinct from a requirement that says, as Adil suggests, both parties must be engaged in armed violence. Imagine, for example, a state launching a military attack against an internal opposition group – a group though boasting a command structure and plenty of arms, finds itself so overmatched and otherwise incapacitated by the state’s sustained bombardment over a period of weeks (bombardment using heavy equipment that destroys entire tracts of property and drives thousands of civilians from their homes), the non-state group is incapable of deploying its arms to any effect (or indeed at all). There is no question both parties have the requisite degree of organization. More, although the violence in this scenario is carried out entirely by one side, I find it hard to imagine the Tadic court would dispute characterizing these events as a NIAC. It is true existing definitions of NIACs are summarized as requiring intense violence and at least two organized parties, but the tests as applied invariably look to an array of factors and circumstances to help assess both intensity and organization, and expressly resist claims that the presence or absence of any one criterion is dispositive.

Second, now consider the scenario that troubles Adil, in which ISIL “fighters pour over the Syria-Iraq border, killing Iraqi civilians, capturing Iraqi territory and taking over Iraqi government institutions. Iraqi forces flee, offering no resistance.” As I have described the existing threshold, such an incursion might well amount to a NIAC, depending on (among other things), how many are killed, how much territory captured, etc. But let’s assume I’m wrong in my description above. What follows? Adil is concerned about the inability to charge ISIL forces with war crimes; and it is of course true, no war, no war crimes. But how much this matters depends I think on a number of factors other than NIAC classification per se, first and foremost on the availability of other channels for criminal prosecution if and when accountability becomes possible. As it stands, several states might plausibly accuse ISIL forces of violating a range of domestic criminal laws depending on which particular heinous act is at issue; UNHCR has suggested ISIL may have committed crimes against humanity; and universal jurisdiction prosecutions are also at least conceivable. But the ISIL v. Iraq conflict (assuming the parties in the scenario alone) is at best a NIAC, not an IAC; while the obligation to prosecute-or-extradite is clear in the latter circumstance, it (among other things) is far from settled in the former. All this to say – there are downsides to expanding the definition of what counts as a NIAC (more on this below). It is at least worth asking in detail how much of an accountability upside is there really in adding more substantive offenses to the list of manifest wrongs ISIL has already committed.

Third, Adil argues that instead of requiring a prolonged and intense level of violence before recognizing a NIAC, we should accept the same “nominal intensity threshold” for NIACs as is currently required for conflicts between states (international armed conflicts or IACs). Here again, Adil’s primary concern seems to be (in addition to preserving the possibility of subsequent war crimes prosecution) to expand the circumstances in which Common Article 3 protections attach. The ICRC has over the years expressed similar sensible concerns. Yet as even the ICRC has noted in the post-9/11 age, because LOAC “rules governing the use of force and detention for security reasons are less restrictive than the rules applicable outside of armed conflicts governed by other bodies of law,” it is “inappropriate and unnecessary to apply” LOAC to circumstances not amounting to armed conflict. To reduce the argument to its core, this is because – contrary to what I take Adil and maybe a few others to suggest – it is not possible as a matter of law to reconcile the basic human rights law prohibition on killing with the basic LOAC acceptance of the power to kill as a first resort. It may well be that other human rights law obligations are complimentarily defined by LOAC, such as the prohibition against arbitrary detention. It may well be that a given military in a given conflict concludes that it is inconsistent with military necessity to kill when capture is possible. But this judgment is not compelled by the law of armed conflict. It is only compelled by the law of human rights. The lower the threshold for recognizing a NIAC, the fewer the circumstances in which that compulsion applies.