The question of when international humanitarian law (IHL) starts applying is complex enough;Footnote 1 the end of IHL's application perhaps even more so. It is certainly one of those questions to which the relevant treaties provide no clear answer. As we will see, while some provisions of the 1949 Geneva ConventionsFootnote 2 make reference to specific points in time, such as the cessation of active hostilities or the general close of military operations, they do not do so for the purpose of systematic regulation, nor do they indeed define these rather vague terms more precisely. This inherent uncertainty is exacerbated by three further considerations.
First, IHL is not a single, coherent body of law. It had no original designer who thought everything through and tied its loose strands together. Rather, like international law generally, IHL is written on a palimpsest, with layers building upon layers and the new replacing the old, but rarely, if ever, doing so completely. Thus, the Hague law regulating the conduct of hostilities that we still apply today was embedded in the then-customary framework in which “war” was the operative legal concept, rigidly opposed to peace. The various waves of Geneva law then built upon that, with the 1949 Conventions and the 1977 Additional Protocols in particular redefining the thresholds of IHL's applicability. We can then add to this mix the judicial gloss of these thresholds, set out mainly by international criminal courts and tribunals, the developments of customary law that they precipitated, and further developments in State practice in the post-9/11 global arena.
Second, and relatedly, even the factual and objective thresholds of modern IHL are fragmented. One cannot truly speak of the end of application of IHL in general terms, but only of the end of application of international armed conflict (IAC), belligerent occupation and non-international armed conflict (NIAC) respectively. Furthermore, while some IHL rules apply at all times – in other words, even outside armed conflict and occupation (e.g. the obligations to disseminate IHL, mark cultural objects, etc.) – as we will see, the application of others might have started with an armed conflict but need not have ended with the armed conflict (e.g. the obligation to investigate and prosecute grave breaches that occurred in an IAC). And while the development of the substantive customary law of NIACs was frequently based on analogies to IACs, the structural differences between the two types of conflicts may have a bearing on the temporal scope of IHL's application and render such analogies more difficult.
Third, our interpretation of the thresholds of application and IHL's temporal scope will inevitably depend on how, within a particular professional setting – as domestic or international judges, government officials, military legal advisers, humanitarian activists, or academics – we weigh a number of competing, and evolving, policy considerations. In other words, an analysis of the end of IHL's application by any given actor is influenced by whether that actor ultimately wants IHL to continue applying, in light of the consequences of continuation or termination. Thus, for example, in Geneva in 1949, most of the humanitarian community – including the International Committee of the Red Cross (ICRC) – advocated for a broad applicability of IHL, particularly when it came to hitherto almost unregulated NIACs. Most States, on the other hand, wanted to both heighten the threshold for IHL application in case of NIACs and reduce the substantive scope of IHL rules applicable in NIACs, because they sought to preserve their own freedom to suppress rebellion and internal strife as they saw fit.Footnote 3 Today, however, the dovish humanitarians might not want IHL to apply expansively, since they may see it as a departure from concurrently (and if need be extraterritorially) applicable international human rights law. Yet States today might precisely want IHL to apply, since they would see it as empowering rather than constraining them, for instance with regard to targeted killings and preventive detention, allowing them to avoid the more demanding rules of international human rights law.Footnote 4
In yet other situations, that calculus may work out differently. Such is the case, for instance, with regard to the question of whether the occupation of Gaza has ended – a question which turns mainly on whether the continued application of IHL is seen as desirable or not. Similarly, an international criminal tribunal may want to take a very generous approach to IHL's continued application since its own jurisdiction could depend on the existence of armed conflict as a contextual element. The consequences of the end of IHL's applicability, whether they concern the scope of the parties' rights and obligations or individual criminal responsibility, will inevitably be taken into account. This is not to say that even if inevitable, such result-oriented thinking is necessarily fully conscious and deliberate. For what it is worth, this article will strive to weigh the various relevant considerations as transparently as possible.
This article is structured as follows. It will first provide a brief overview of the changes that IHL has undergone through the years – to the extent that they are relevant to understanding the conditions for its end of application – and will set out the general principle governing this process. It will then look at the point at which different IHL rules cease to apply in international armed conflicts, belligerent occupation and non-international armed conflicts. It will subsequently examine the transformative processes that end the application of one IHL sub-regime (IAC or NIAC) but initiate the application of another. It will finish by looking at the (impending) end of the conflict between the United States and Al Qaeda, one of the most current and controversial issues of contemporary IHL.
Brief historical overview and the general principle on end of application
Before we can examine the end of application of modern IHL, we need to take a brief look at the past, as well as the evolution of the thresholds of the beginning of IHL's application, which are dealt with in more detail elsewhere in this issue of the Review.Footnote 5 As noted above, the operative concept in customary international law before the 1949 Geneva reform (and perhaps for some time thereafter)Footnote 6 was “war”. In classical international law, war was defined not merely as a factual situation involving hostilities between two States, but as a legal condition whose initiation and end brought about a host of consequences in the relations between the belligerents among themselves and with third States. War and peace categorically excluded each other, as did the law of war and the law of peace.Footnote 7 War was generally regarded as abrogating all peacetime treaties between the belligerents and triggered the application of the rules on neutrality for non-belligerents. It was also both formal and subjective, requiring not merely the objective existence of hostilities but also the expression of an animus belligerendi. This subjective animus could be proven, or not, by reference to criteria such as the severance of diplomatic relations between belligerents, the existence of a declaration of war or of a notification of the state of war to neutral powers, or the recognition of the state of war by these neutral powers.
This in turn opened the way to situations of widespread and protracted fighting in which, for political reasons, the States concerned refused to recognize the existence of war. A gap opened up between a common-sense, factual understanding of “war” and one derived from the niceties of the law of nations, a gap to be exploited when it served State interests. It not only introduced a large degree of uncertainty with regard to the rights of private citizens, but more importantly created a major obstacle to the application of any humanitarian rules of the law of war. All that a State had to do to avoid the law of war was to deny the existence of war in the technical legal sense, no matter how much blood was being shed in a very real sense.Footnote 8 It was precisely the rigidity of the law of peace/law of war framework and the strict consequences that followed the transition from one to the other that provided the incentive for States to avoid recognizing the existence of war. This led some scholars of the period to argue for the legal recognition of a third, middle category between war and peace – a status mixtus.Footnote 9 Others, in turn, wanted to objectivize war.Footnote 10 However, what the humanitarian lawyers may have wanted and what States thought to be in their interest were not necessarily one and the same. Rather than bringing some resolution, the controversies around the legal nature of war brought even more uncertainty.Footnote 11
Since “war” was a formal business, it also needed to be formally terminated. While hostilities could be interrupted through an agreed-upon truce or ceasefire, or a more comprehensive armistice, the end of war generally required a peace treaty.Footnote 12 It was only upon the conclusion of such a treaty that normal relations between the parties could resume and their earlier peacetime treaties could be revived.
This was still the relevant framework as the Second World War started, but its aftermath saw the addition of more layers to the palimpsest of international law. First, even though the recourse to war as an instrument of national policy was already outlawed under the terms of the 1928 Kellogg–Briand Pact, the adoption of the UN Charter and its general prohibition on the use of force (not just “war”) truly created the jus ad bellum as separate from the jus in bello. Second, the applicability of the 1949 Geneva Conventions was generally detached from the formal concept of war. Article 2 common to the four Geneva Conventions (common Article 2 or CA2) introduced the concept of international armed conflict,Footnote 13 which was designed as an objective, factual replacement for the narrower concept of war while retaining its predecessor's inter-State nature.
The 1949 Geneva Conventions also introduced the first systematic regulation of internal conflicts, through the concept of non-international armed conflict under common Article 3 of the Conventions (CA3). At the time of the Conventions' adoption, the NIAC threshold brought about only the application of CA3 itself and its purely humanitarian provisions protecting persons not taking part in hostilities or rendered hors de combat. That was the sum total of the conventional law of armed conflicts as it applied to NIACs in 1949; for example, it contained no rules on the conduct of hostilities analogous to IACs.
Over time, however, through the adoption of the 1977 Additional Protocols and the evolution of customary law, the law of armed conflict coalesced around the two factual thresholds set out in the 1949 Conventions. Not only is the CA2 threshold also valid for the application of Additional Protocol I (AP I) rules,Footnote 14 but it also became the threshold for application of the customary “Hague law” on the conduct of hostilities. This was also the case with the CA3 NIAC threshold, whereas the gaps in the regulation of NIACs were filled mainly through custom. The 1949 Geneva Conventions' thresholds of application thus became the thresholds for the application of customary IHL more generally. The obsolescence of the traditional concept of “war” became almost complete, with States simply no longer declaring war against one another or formally expressing their view that war exists. Similarly, the distinctions between ceasefires, armistices and peace treaties gradually blurred, with peace treaties in particular becoming increasingly rareFootnote 15 and conflicts frequently ending with less formal instruments such as armistices or unilateral or joint declarations, or simply ending de facto.
But as the law of war was being rewritten into the law of armed conflict, the concept of “war” was still not abandoned formally, even if it was abandoned in fact as a condition for IHL applicability. Looking at the IHL palimpsest, one issue that arises is whether the concept of “war” still has any relevance for our modern debates. Notably, CA2 of the 1949 Geneva Conventions explicitly provides that they “shall apply to all cases of declared war”, thus making it possible for the Conventions (and the relevant custom) to begin applying before a single shot has been fired. But conversely, will their application end only with the end of the war in the formal, technical legal sense, once a state of war has commenced on top of a plain international armed conflict? Just consider a scenario of war with hostilities long having come to an end, yet without a peace treaty or any kind of formal instrument normalizing in full the relations between the parties.
Some authors, most notably Yoram Dinstein, still give significance to war as a legal concept.Footnote 16 Most, however, and here I include myself, would argue that the main point of the 1949 Geneva reform was precisely to do away with the subjectivity and formalism of war, and to make the thresholds of application objective and factual, with this tendency only being strengthened in the intervening years.Footnote 17 Having said that, the possibility of the “old” law still having an influence cannot be categorically excluded, especially because the concept of “war” transgressed the boundaries between jus ad bellum and jus in bello. However, that influence is likely to be minimal. In the case of a “war” in which hostilities have de facto ended with permanence and finality, even in the absence of a formal conclusion of peace, as a matter of jus ad bellum any residual right to resume hostilities would be precluded by the UN Charter prohibition on the use of force,Footnote 18 while in the absence of actual hostilities or prisoners of war or protected persons in the hands of the adversary, IHL would have little or nothing to regulate.Footnote 19
The rules governing the beginning and end of the application of IHL are both customary and conventional in nature. With regard to the conventional rules, of relevance to the temporal application of IHL are also some general issues of the law of treaties regarding entry into force and withdrawal. While the Geneva Conventions have achieved practically universal adoption, the general rules on entry into force are of particular relevance for newly formed States, who can become parties to treaties through accession or (automatic?) succession.Footnote 20 On the other hand, it is perhaps surprising that both the Geneva Conventions and their Additional Protocols contain explicit provisions permitting their denunciation.Footnote 21 No State party has ever denounced one of these treaties, and the inevitable political fallout would render this highly unlikely in the future. The denunciation provisions also contain explicit safeguards that would render any denunciation ineffective with regard to situations arising from an existing armed conflict or with regard to protected persons already in captivity.Footnote 22 The key legal safeguard, however, is the fact that most of the rules in the Conventions and the Protocols have achieved customary status, and would accordingly bind the denouncing State in any event. In the discussion to follow, I will accordingly disregard any potential issues arising from the general law of treaties, with the caveat that conventional rules which do not reflect custom could potentially be affected.Footnote 23
Bearing in mind the evolution of the modern IHL regime, as well as its fragmented nature, we can state at this point the one general principle on the end of application of IHL that will form the basis for further discussion: unless there is a good reason of text, principle or policy that warrants an exception, the application of IHL will cease once the conditions that triggered its application in the first place no longer exist. In other words, if a particular situation can no longer be qualified as an IAC, a NIAC or an occupation, the application of IHL will end.Footnote 24
In the absence of any specific guidance to the contrary, this general principle makes perfect sense in the factual, objective conceptual space of the Geneva Conventions. For IHL to apply, its thresholds of application must continue to be satisfied at any given point in time. In order to elaborate on this general principle, we must of course look at the constitutive elements of each threshold, in the context of those particular scenarios in which these elements might be extinguished. We must then establish whether a departure from the general rule is warranted, and whether some rules continue applying even after an armed conflict has ended. We will then observe certain terminating processes and events, which generally end the application of IHL (but not necessarily all of it),Footnote 25 and certain transformative processes and events, which end the application of one IHL sub-regime (IAC or NIAC) but immediately engage another. In the section below, I will address each threshold in turn.
International armed conflict
As we have seen, the concept of IAC was crafted as an explicit replacement for the concept of war. As with the concept of war, IAC as defined in CA2 is of an interstate nature, a conflict between two sovereigns. In the words of the authoritative Pictet Commentary:
Any difference arising between two States and leading to the intervention of members of the armed forces is an armed conflict within the meaning of [Common] Article 2, even if one of the Parties denies the existence of a state of war. It makes no difference how long the conflict lasts, how much slaughter takes place, or how numerous are the participating forces; it suffices for the armed forces of one Power to have captured adversaries falling within the scope of Article 4 [of the Third Geneva Convention]. Even if there has been no fighting, the fact that persons covered by the Convention are detained is sufficient for its application.Footnote 26
The CA2 threshold is thus remarkably low – all it needs is a difference between two States leading to the intervention of their armed forces.Footnote 27 Whether Pictet was indeed correct in this, or whether a de minimis level of violence needs to occur in order to avoid minor exchanges of firepower between the forces of two States (for instance, a single rifle shot across the border) being classified as IACs, is not the subject of my enquiry at this time. Opinions and practice on this point are conflicted.Footnote 28 Yet, however exactly defined, the IAC threshold is certainly far lower than the NIAC “protracted armed violence” threshold, to which I will come momentarily, since the IAC threshold was not subject to the same sovereignty concerns as the NIAC one. A particular amount of violence may produce an IAC if perpetrated between States, but might not qualify as a NIAC if committed by or against non-State actors.
As we have also seen, the principal distinguishing point between “war” and IAC is the latter's objective and factual nature. The end of IAC should equally be based on purely factual criteria – what matters is that the hostilities between the two parties have ceased. Because the IAC threshold is relatively easy to satisfy, however, and because it would be both impractical and would open the door to abuse to treat every lull in the fighting as an end to an IAC and each resumption of combat as the start of a new one, hostilities must end with a degree of stability and permanence in order for the IAC to be terminated.Footnote 29 Thus, for example, in the Gotovina case, the Trial Chamber stated that:
Once the law of armed conflict has become applicable, one should not lightly conclude that its applicability ceases. Otherwise, the participants in an armed conflict may find themselves in a revolving door between applicability and non-applicability, leading to a considerable degree of legal uncertainty and confusion. The Trial Chamber will therefore consider whether at any point during the Indictment period the international armed conflict had found a sufficiently general, definitive and effective termination so as to end the applicability of the law of armed conflict. It will consider in particular whether there was a general close of military operations and a general conclusion of peace.Footnote 30
This is always a factual assessment, which will vary from case to case, and the exact time at which the IAC ended may be hard to point out. Agreements concluded by the belligerent parties, however called – unilateral statements by either of them, or resolutions of relevant international organizations, such as those adopted by the UN Security CouncilFootnote 31 – may provide evidence that the hostilities have ended with the needed degree of stability and permanence. But it is the fact that the hostilities have ended that ultimately matters, not the precise legal nature of the instrument in question.Footnote 32 Depending on the political and military environment, a ceasefire agreement or an armistice may actually signify the point at which the hostilities have permanently ended, while a formal peace treaty might not be worth the paper it is written on if hostilities continue unabated.
This factual approach is supported by what little we have in the Geneva Conventions regarding the end of their application. Thus, Article 118(1) of GC III provides that “[p]risoners of war shall be released and repatriated without delay after the cessation of active hostilities”. The repatriation obligation would naturally only be acceptable to States if hostilities had ended more or less permanently. Article 6(2) of GC IV, on the other hand, stipulates in the relevant part that “[i]n the territory of Parties to the conflict, the application of the present Convention shall cease on the general close of military operations”. The common provisions on the denunciation of the Geneva Conventions mentioned above refer to yet another point in time when they stipulate that “a denunciation of which notification has been made at a time when the denouncing Power is involved in a conflict shall not take effect until peace has been concluded”.Footnote 33 Here we have three moments in time: the cessation of active hostilities, the general close of military operations, and the conclusion of peace.
Article 5 of GC III and Article 6(4) of GC IV, on the other hand, make it clear that the Convention will continue to apply even after the general close of military operations if protected persons are still in the power of the enemy and they have not been released or repatriated before that time. Article 3(b) of AP I also sets out this principle very clearly:
the application of the Conventions and of this Protocol shall cease, in the territory of Parties to the conflict, on the general close of military operations and, in the case of occupied territories, on the termination of the occupation, except, in either circumstance, for those persons whose final release, repatriation or re-establishment takes place thereafter. These persons shall continue to benefit from the relevant provisions of the Conventions and of this Protocol until their final release, repatriation or re-establishment.
It is unclear whether the drafters of the Conventions were making a firm distinction between the “cessation of active hostilities” standard in Article 118(1) of GC III and the “general close of military operations” in Article 6(2) of GC IV and later in Article 3(b) of AP I – in other words, whether the distinction was deliberate or was the consequence of uncoordinated drafting. What is clear is that the primary motivation behind the GC III formula was to depart from the earlier rule set out in Article 20 of the Hague Regulations, under which the obligation to repatriate prisoners of war started only at the (formal) conclusion of peace. This meant that in several instances in which the conflict had de facto ended but without a formal peace treaty, or with treaty negotiations taking a very long time, vast numbers of prisoners of war continued to be held without any real need to do so.Footnote 34 This of course does not mean that the GC III standard necessarily assures swift repatriation in practice, the lengthy repatriation efforts after the 1980–1988 Iran/Iraq war being a case in point. As for the “general close of military operations” formula in Article 6(2) of GC IV, the Pictet Commentary interprets it as a “final end of all fighting between all those concerned”.Footnote 35 Note that the test is an objective and factual one; as argued above, while an armistice or peace treaty can serve as evidence of the finality of the end of fighting, formal agreements are neither required nor conclusive on the point.Footnote 36 The general close of military operations implies an end of the fighting between all of the belligerents,Footnote 37 even though active hostilities may have ceased between some at a much earlier date (consider, for example, the 1940 surrender of France while the UK continued fighting, or the different times of the surrender of Germany and Japan during the Second World War).
The ICRC Commentary to AP I tried to draw more of a distinction between the “general close of military operations” and the “cessation of active hostilities” formulas. Thus, it held that military operations can be taken more broadly than actual combat as including “the movements, manoeuvres and actions of any sort, carried out by the armed forces with a view to combat”, and that “[t]he general close of military operations may occur after the ‘cessation of active hostilities' referred to in Article 118 of the Third Convention: although a ceasefire, even a tacit ceasefire, may be sufficient for that Convention, military operations can often continue after such a ceasefire, even without confrontations”.Footnote 38 The distinction is also supported by academic commentary.Footnote 39 However, depending on the circumstances on the ground, the distinction may actually be hard to draw – in other words, the cessation of active hostilities and the general close of (combat-oriented) military operations may occur at the same time, or very close to one another.
In sum, we can conclude that an IAC would end with a general close of military operations, with no real likelihood of a resumption in hostilities. The end of the IAC will also end the application of those rules of IHL regulating the conduct of hostilities. Further, it will end any IHL-granted authority to detain combatants or civilians preventively purely on grounds of security.Footnote 40 However, as we have also seen, some obligations under IHL will survive the end of the armed conflict, and indeed may be triggered by the conflict's (imminent) end, as with the obligation to repatriate prisoners of war. Persons who remain in the power of the enemy will continue enjoying the protections of IHL until their repatriation or release, including inter alia the right of access by the ICRC, even if IHL no longer authorizes their continued detention. They will also continue benefiting from fundamental guarantees in Article 75(6) of AP I, under which “[p]ersons who are arrested, detained or interned for reasons related to the armed conflict shall enjoy the protection provided by this Article until their final release, repatriation or re-establishment, even after the end of the armed conflict” (emphasis added). These protections would normally be complemented by human rights law, and to the extent that IHL allows any departures from human rights, for instance by virtue of the lex specialis principle, such departures would no longer be permitted with the end of the conflict.Footnote 41 Similarly, Article 33(1) of AP I provides that “[a]s soon as circumstances permit, and at the latest from the end of active hostilities, each Party to the conflict shall search for the persons who have been reported missing by an adverse Party” (emphasis added). This and other obligations with regard to missing and dead persons, such as facilitating access to gravesites, will continue applying after the end of the conflict, as would the obligations to investigate and prosecute grave breaches of the Conventions and AP I.
There are thus a number of exceptions to the general principle on the end of application. The exceptions are not themselves temporally limited; further passage of time after the end of the conflict cannot by itself terminate the extant obligations. For example, if a State detains a prisoner of war for decades after the conflict, he or she will still be protected by IHL. The obligation will only terminate if its functional predicate is discharged, for instance when the prisoner is repatriated.
As with IHL generally, this interaction between the general principle and the exceptions thereto strikes a balance between military necessity and considerations of humanity. Clarity and predictability require that the end of a conflict should not be presumed lightly. Rules on the conduct of hostilities stop applying only once the hostilities have definitively ended, while the cessation of hostilities will initiate the obligation to repatriate prisoners of war and release any civilian internees, because this is when the obligation can be realistically complied with and the need for such measures ceases. The protective rules on treatment in detention will continue applying while the detention lasts, which is of the greatest importance in cases of delayed repatriation. The need to protect persons deprived of liberty for reasons related to an armed conflict does not end with the conflict itself, nor would any serious violence against these persons be any less of a war crime. Similarly, the obligation to repress grave breaches can at times be implemented even more effectively in the post-conflict period, as investigations and any prosecutions can take place unhindered by active hostilities. I will now briefly look at the end of belligerent occupation as a subspecies of or a threshold complementary to IAC.
Belligerent occupation
The end of occupation is again a complex topic, examined not long ago, for instance, by an expert meeting on occupation convened by the ICRC.Footnote 42 I will address it only very briefly. As with IACs and NIACs, we can start off with the general principle that IHL will normally cease to apply once its threshold of application – here, belligerent occupation – is no longer met. If we define occupation as effective control by a State of the territory of another State without the latter's consent,Footnote 43 it follows that there are two basic modalities through which an occupation can end: loss of control by the occupant, or the occupant being granted valid consent by the displaced sovereign.
This basic position was not spelled out in the 1907 Hague Regulations, but it again follows from the very definition of the concept of occupation in Article 42: “territory is considered occupied when it is actually placed under the authority of the hostile army. The occupation extends only to the territory where such authority has been established and can be exercised” (emphasis added). The inability to exercise authority would consequently terminate the occupation; the difficult question is what factors are to be taken into account in establishing whether the occupant lost control over a territory or a substantial part thereof.
Article 6(3) of GC IV provided that in occupied territory,
the application of the present Convention shall cease one year after the general close of military operations; however, the Occupying Power shall be bound, for the duration of the occupation, to the extent that such Power exercises the functions of government in such territory, by the provisions of the following Articles of the present Convention: 1 to 12, 27, 29 to 34, 47, 49, 51, 52, 53, 59, 61 to 77, 143.
Article 6(3) of GC IV focused on those situations in which the occupation outlives the IAC that created it, with the application of most of the Convention ending one year after the factual end of the conflict, and only a number of core humanitarian provisions applying thereafter, and even those only to the extent that the Occupying Power exercises the functions of government in the occupied territory. The provision was essentially tailor-made for the transformative Allied occupations of Germany and Japan after the Second World War, which were extensive and prolonged, followed the unconditional surrender of the Axis powers (or debellatio), and ended gradually through the creation of new institutions of self-government.Footnote 44
Article 3(b) of AP I dropped the one-year limit from GC IV, providing that it would cease to apply “in the case of occupied territories, on the termination of the occupation”, whereas, as we have seen, above persons still in captivity would continue to be protected by the Protocol.Footnote 45 Whether customary IHL would follow Article 3(b) of AP I and displace Article 6(3) of GC IV even for States not parties to AP I is a difficult question which I will not address here – suffice it to say that while GC IV's applicability qua treaty remains governed by Article 6(3), it is perfectly possible for treaties to be supplanted by supervening custom, if the existence of the customary rule is indeed established.Footnote 46
Occupation can end through loss of control in a variety of scenarios: unilateral withdrawal; defeat of the occupying forces by the displaced sovereign or other outside intervention; or loss of control due to an insurgency in the occupied territory. The enquiry into loss of control should always be objective, factual and contextual, taking into account all of the circumstances on a case-by-case basis, and subject to two basic principles. First, as with IACs and NIACs, the end of occupation should not be presumed lightly. In particular, once the occupation is established, the maintenance of occupation might not require overt and frequent displays of military strength by the occupant, especially if the occupation faces little or no resistance and there are no competing authorities on the ground. Second, not every temporary lapse in control would terminate the occupation.Footnote 47 No matter how powerful the occupant, it may be impossible for it to control every single bit of the occupied territory all the time, especially in the case of an insurgency (one need only consider post-2003 Iraq, or the Nazi occupation of Yugoslavia in the face of partisan resistance). So long as the lapse in control is only temporary or very localized, and so long as the occupant has the full capacity to re-establish its control, the occupation should be considered to be uninterrupted.Footnote 48 Clearly, opinions will differ on the facts of specific cases as to whether an occupation has ended through loss of control, as with the Israeli withdrawal from Gaza, but the basic principle is relatively uncontested.
End of occupation through loss of control has parallels with the extraterritorial applicability of human rights law in the occupied territory – one issue raised before British courts and the European Court of Human Rights in the Al-Skeini litigation was whether the UK possessed effective control for the purposes of European Convention on Human Rights (ECHR) Article 1 jurisdiction in Basra, due to the level of sustained insurgency there, and despite the fact that the UK was formally the occupant in Southern Iraq. The House of Lords held that the UK did, in fact, lose effective control for the purposes of ECHR Article 1, even if it formally remained the Occupying Power for the purposes of IHL,Footnote 49 while the Grand Chamber of the European Court of Human Rights avoided the issue altogether.Footnote 50 The issue of whether the effective control thresholds for occupation and for the application of human rights treaties differ or not remains unresolved.Footnote 51
With regard to occupation ending by the occupant obtaining consent, it is generally possible for the displaced sovereign to agree to the presence of the (former) occupier, whether by way of a peace treaty or some other kind of formal or informal agreement.Footnote 52 Note in this regard that Article 7 of GC IV prohibits special agreements between States Parties when these agreements adversely affect the situation of protected persons. However, such agreements are only prohibited in situations in which GC IV itself applies, and Article 7 of GC IV does not preclude the general provision of genuine consent by a State to the presence of foreign forces, assuming that a government exists which can validly provide that consent.Footnote 53 That expression of consent would be subject to the general principle articulated in Article 52 of the Vienna Convention on the Law of Treaties, which provides that a treaty is void if it is procured by coercion of a State by threat or use of force contrary to the UN Charter.Footnote 54 It is also possible for the government of the displaced sovereign (whose territory may or may not have been occupied in its totality) to change, often at the instigation or at least with the input of the occupant, and then provide consent. Article 47 of GC IV makes it clear that any changes in the (local) government of the occupied territory, or any purported annexation of that territory by the occupant, cannot alter the applicability of the law of occupation.Footnote 55 But what of those situations in which the government of the displaced sovereign is wholly destroyed or changed, as was the case in Iraq post-2003? In this respect, we can observe at work similar considerations as with the transformative process of the “internalization” of an IAC into a NIAC, which I will deal with in detail below. What entity has the sufficient capacity and legitimacy to provide meaningful consent on behalf of a State may be a difficult and controversial issue.
Non-international armed conflict
This brings us to the end of application of IHL in NIACs. We have seen that the end of IHL application in IACs and occupation is complex enough. This complexity is exacerbated with regard to NIACs by the structural differences between IACs and NIACs and the almost complete lack of any textual guidance. CA3 says nothing about its end of application, while Article 2(2) of AP II appears to endorse the general rule that the application of IHL will end with the conflict which initiated it when it provides for the humanitarian exception from the general rule regarding persons who remain in captivity:
At the end of the armed conflict, all the persons who have been deprived of their liberty or whose liberty has been restricted for reasons related to such conflict, as well as those deprived of their liberty or whose liberty is restricted after the conflict for the same reasons, shall enjoy the protection of Articles 5 and 6 until the end of such deprivation or restriction of liberty.
In short, drawing analogies with IACs will be difficult.
As it stands today, NIAC is a plural legal concept, defined differently under different treaty regimes. The basic (non-)definition of NIAC, which encompasses all others, is that found in CA3. Its terms were famously elaborated by the International Criminal Tribunal for the former Yugoslavia (ICTY) Appeals Chamber in the Tadić Interlocutory Decision on Jurisdiction,Footnote 56 which has been widely accepted as reflecting custom. Under Tadić, CA3 requires “protracted armed violence” – a threshold of intensity and possibly duration, rising above mere riots or disturbances – between a State and an armed non-State actor or between two such non-State actors, which are sufficiently organized to conduct hostilities.Footnote 57 On the other hand, the heightened threshold of Article 1(1) of AP II, which is only applicable to conflicts involving a State and a non-State actor, but not two such non-State actors, requires the non-State actor to have an organizational structure with a responsible command, to control a part of the State's territory, and to have the ability to conduct sustained and concerted military operations and to implement AP II. I will leave aside the question of to what extent exactly the AP II threshold is really higher than the customary CA3 one when applied to particular facts.
As with IACs, the termination of NIACs is a matter of factual enquiry, but the intensity and organization thresholds in CA3 and AP II make that enquiry even less straightforward than in IACs. As we have seen, for IACs to end, the hostilities themselves need to end with a certain degree of permanence or finality. One option would be to treat NIACs in exactly the same way – so long as some hostilities continue, so would a NIAC.Footnote 58 This seems to be the implication of the insistence of the ICTY in Tadić that IHL applies “in the case of internal conflicts, [until] a peaceful settlement is achieved”.Footnote 59 This makes perfect sense from the standpoint of an international criminal tribunal, which wants to stabilize its jurisdiction and bring to account as many perpetrators of war crimes as possible. Thus, for instance, with respect to the NIAC between Serbia and the Kosovo Liberation Army in 1998, the first ICTY Trial Chamber judgment in the Haradinaj case found that “since according to the Tadić test an internal armed conflict continues until a peaceful settlement is achieved, and since there is no evidence of such a settlement during the indictment period, there is no need for the Trial Chamber to explore the oscillating intensity of the armed conflict in the remainder of the indictment period”.Footnote 60 The ICTY's task was accordingly much easier.
Another option, however, and to me more logical from a purely IHL standpoint, would be to take into account the heightened NIAC intensity threshold when compared to IACs, since this is where the analogy with IACs may be at a breaking point. Any resumption of hostilities between States would in any event reconstitute an IAC, and it therefore makes sense to wait for the complete end of all hostilities for the IAC to be terminated. In NIACs, by contrast, it could be enough for the hostilities to fall below the threshold of “protracted armed violence” with a certain degree of permanence and stability so as to enable us to establish that the hostilities have, in fact, ended. As with IACs, once the threshold is met, there should be a presumption that it continues to be met absent strong evidence to the contrary – as a matter of policy, a NIAC which peters in and out of existence on a daily basis would be undesirable. But unlike in IACs, in NIACs the hostilities would not need to end altogether. What would matter is whether the intensity of the hostilities or the organization of the non-State actor factually eroded to such an extent that the threshold is no longer met.Footnote 61 For example, looking at the post-2003 conflict in Iraq, which involved the new Iraqi government and its foreign allies on the one side and several organized armed groups on the other, it could be argued that this NIAC (or set of NIACs) ended at some point in late 2009, as the capacity of insurgent groups was degraded and the level of armed violence decreased (even if the violence never ended completely). In August 2010, the US military withdrew its combat troops from Iraq. But after a period of relative calm, violence rapidly escalated in 2012 and 2013, as armed groups such as the Islamic State of Iraq and the Levant (ISIL) regrouped and reorganized, and a new NIAC was initiated.
As with IACs, NIACs can end through an agreement between the parties, a stalemate, or one of the parties' defeat and surrender, but again the only legally relevant question would be whether the threshold continues to be satisfied. Similarly, AP II NIACs could end or transform into simple CA3 NIACs if the non-State actor fighting the State becomes so structurally compromised that it no longer has control over territory or the ability to conduct sustained and concerted military operations or implement AP II. The disadvantage of this approach is the greater likelihood of multiple transitions between peace and a NIAC.
For an example of a NIAC ending through the complete defeat of an adversary, we need only look at the Sri Lanka conflict. An example of a NIAC petering out would be the post-2007 surge in Iraq. Again, the enquiry is purely factual. In some cases it will be relatively easy to determine the exact moment when the conflict ended, especially when there is a peace agreement (as long as the agreement is, in fact, observed). In others, it might be exceedingly difficult to pinpoint the exact time when the hostilities fall below the “protracted armed violence” threshold, with Iraq again being a case in point.
Note again how diverging policy considerations can influence the contextual determination of whether the NIAC has ended, depending on the identity and the goals of the actor making the determination. For example, the end of the conflict would signal the end of any authority to detain individuals preventively and the resumption of the normal human rights regime (to the extent that IHL was actually capable of displacing it). If the actor making the determination cares about the possible arbitrary exercise of State power, it might be inclined to see the end of the NIAC more quickly. If, on the other hand, it cares about the arbitrary exercise of power by the non-State actor, which is bound by IHL but probably not by human rights law, then that calculus may turn out differently.
Transformation of conflicts: internationalization (NIAC to IAC) and internalization (IAC to NIAC)
Until now we have looked at terminative events or processes, which end the application of IHL. But we also need to examine transformative processes, which end the application of one IHL sub-regime but start another.Footnote 62
Let us first look at internationalization, one of the most controversial topics of modern IHL. In my view, the concept of internationalization is only legally useful if it is defined as the transformation of a prima facie NIAC into an IAC, thereby applying to this conflict the more comprehensive IAC legal regime.Footnote 63 The most important of these legal consequences is the grant, in principle, of combatant immunity and potential prisoner of war status to combatants on both sides of the conflict. As for the mechanism of internationalization, we have seen that under CA2, IACs are defined as differences leading to the use of armed force between two States. Accordingly, there are two basic ways of internationalizing a NIAC. First, a prima facie NIAC can be subsumed under the existing CA2 definition. In other words, what at first glance looks like a conflict between a State and a non-State actor is, on a deeper look, actually a conflict between two States. This can either happen because a third State exercises control over a non-State actor at some point during a NIAC, with the most controversial issue here being what test or standard of control is to be applied, or because a non-State actor involved in a NIAC emerges as a State during the conflict – a rare occurrence, but one that we can perhaps observe with regard to some of the conflicts in the former Yugoslavia.
Secondly, a NIAC can be internationalized through the redefinition of IAC in terms of its party structure. In these cases a treaty or customary rule changes the CA2 definition so as to potentially include some non-State parties. Internationalization under this heading would clearly require proof of a specific rule to that effect. One such rule can be found in Article 1(4) of AP I (self-determination conflicts against colonial, occupying or racist regimes), subject to a declaration under Article 96(3) of AP I, while another possible candidate is the customary doctrine of the recognition of belligerency, which has fallen into disuse but perhaps not desuetude.Footnote 64
The second transformative process is internalization, or de-internationalization, and it again flows from the inter-State CA2 definition.Footnote 65 It is easy to say that IACs are fought between States, and statehood may even be uncontested in a given case, but who gets to represent the State may turn out to be a very difficult issue. Not only is this question important for the initial qualification of a conflict, but it may also prove to be crucial for its requalification or transition from one type to another. Consider all those conflicts during which some kind of regime change takes place, whether in Afghanistan, Libya, Iraq or the Côte d'Ivoire. In all of these conflicts there was some form of foreign intervention coupled with a reversal of roles between a government and a rebel group, with a new government extending its invitation to the intervening State or States to assist it in fighting the former government.
What is at stake here is a process of transformation from an IAC into a NIAC. Looking at the competing policy considerations, we can see what is not enough for such internalization to occur. That the incumbent government of a country is defeated cannot by itself transform the conflict, nor can the establishment of a proxy government by the victors, as this would allow them to effectively strip by force the protections granted in IACs to the remaining combatants of the defeated State, depriving them of combatant immunity and POW status. Similarly, that a rebel group is recognized as the new legitimate government of the country cannot of itself transform the character of the conflict, as this would again allow the intervening States to unilaterally do what they will by installing their own proxies as the new State government.
In my view, both considerations of policy and recent practice support a rule consisting of the following three elements: a conflict would transform from an IAC into a NIAC only when (1) the old regime has lost control over most of the country, and the likelihood of it regaining such control in the short to medium term is small or none (negative element); (2) the new regime has established control over most of the country, and is legitimized in an inclusive process that makes it broadly representative of the people (positive element); and (3) the new regime achieves broad, although not necessarily universal, international recognition (external element). None of these elements is enough by itself, but jointly they take into account both questions of legitimacy and factual developments on the ground while providing safeguards against abuse by an intervening power. With regard to both the positive and the negative elements, the degree of control would be looked at holistically, taking into account not just troops on the ground but also direction over the State's institutions more generally, its economic assets, the media, or the nature of the internal legitimizing process (for instance, reasonably free and fair elections, or some other representative process such as the convening of the Loya Jirga, the grand tribal assembly, in Afghanistan). The external element would depend not just on bilateral State action but also on the decisions and conduct of the relevant international and regional organizations.
The exact mix of the three elements will inevitably vary from case to case. Internal and external legitimization will be especially important when the new government is able to wield control in the country only because it is being propped up by its foreign sponsors. The purpose of the formula is to find a solution which will, on the one hand, prevent foreign interveners from using puppet governments, established by force alone, to claim that the IACs/belligerent occupations have transformed into NIACs or mostly unregulated pacific (or non-belligerent) occupations, yet will, on the other hand, in some cases allow a transition to occur, but will do so in line with the rights of all peoples to self-determination, and on the basis of safeguards which will hinder a transition on the basis of pure force (especially when that use of force is ad bellum unlawful, as was the case in Iraq).
Obviously, it may be difficult to pinpoint the exact moment of internalization in any given case, and thankfully in most cases it may be unnecessary to do so, but it is necessary for us to be aware of the relevant elements and their interplay. In doing so, we must also be aware that the internalization of a conflict has as its consequence a possible reduction of various protections under IHL. While the legal regimes applicable to IACs and NIACs have largely been brought together through the development of custom, significant differences and uncertainties still remain in areas such as grounds for detention, treatment and procedural safeguards in detention, combatant immunity, and status-based targeting. Thus, though transformative processes do not end the application of IHL altogether, their practical consequences should not be underestimated. These practical consequences are probably the most significant with regard to detention which begins before, but continues after, the moment of transformation from one type of conflict to another. The key issue here is the existence or modification of any authority to detain; while such authority may exist directly under IHL in IACs, it likely requires the intervention of domestic lawFootnote 66 in NIACs.Footnote 67
End of the conflict with Al-Qaeda?
This brings us to one of the most vexing questions of contemporary international relations and IHL – that of the existence of an armed conflict between the United States and Al Qaeda, and its seemingly impending end. In order to analyse this question, we must first understand some relatively recent shifts in the underlying legal and political dynamics.
As originally designed, and through most of its history, IHL was seen by States as a system of limitations on their sovereignty and freedom of action, particularly so when it comes to the law of NIAC.Footnote 68 In other words, the baseline for international regulation from the classical period onwards was essentially the Lotus presumption: States were at liberty to do anything that international law did not expressly prohibit, including the unrestrained freedom to wage war, both against each other and internally. The law of war evolved precisely to impose such restraints, first and foremost in the inter-State context. States resisted the international regulation of internal conflict because of the fear – founded or not – that it would impose limits on how they could deal with rebels, and would confer on these rebels some rights in international law. It is this sovereignty-induced concern of States that explains the IAC/NIAC dichotomy in modern law and the higher thresholds of applicability for NIAC rules, in terms of intensity and organization of the non-State parties. Hence, it was rarely if ever in the interests of a State embroiled in a NIAC to recognize the existence of such a conflict – such a State simply did not want IHL to apply.Footnote 69 This may explain why, for instance, the United Kingdom consistently denied the existence of a NIAC in Northern Ireland, claiming that the euphemistically termed “Troubles” were an internal affair of criminal law enforcement despite the fact that the CA3 threshold was arguably reached,Footnote 70 and why it made a declaration regarding Article 1(4) of AP I that explicitly excluded acts of terrorism, whether concerted or in isolation, from the scope of armed conflict.Footnote 71
Now, however, human rights have gradually replaced, or are in the process of replacing, the idea of unrestrained freedom of action as the baseline for regulation, as much culturally as formally.Footnote 72 Instead of IHL being the only set of limitations on States, a more rigid, demanding and legalistic set of limitations has emerged, particularly in the internal context. States, or at least some States, have accordingly stopped seeing IHL as a constraining body of rules whose application they want to avoid in their engagements with non-State actors.Footnote 73 Rather, they have increasingly started seeing IHL as an authorizing body of rules liberating them or derogating from human rights or other constraints, often on the dubious basis of the lex specialis principle.Footnote 74 For example, while IHL targeting or detention rules evolved as limitations (for instance, while people will inevitably be killed in wartime, deliberately targeting civilians is prohibited), they are now seen as permissive rules authorizing departures from human rights (for instance, while an enemy fighter may be killed even if he or she does not pose an imminent threat, preventive detention for reasons of security is authorized even if human rights law generally prohibits preventive detention).Footnote 75
To see how these dynamics have evolved, we need only look at US policy post-9/11.Footnote 76 The US government from the outset decided to cast the Al Qaeda threat and the US response thereto as a “war” for both domestic and international purposes, in order to get the detention and targeting authority that it thought it needed and avoid having its hands tied by any applicable rules of domestic constitutional law as well as international human rights law. The moniker “global war on terror” denoted a supposed single conflict under IHL between the US on one side and Al Qaeda and its affiliates on the other. Initially, the US characterized this conflict as an IAC, albeit a strange sort of IAC which transcended the Geneva Conventions CA2 definition since one of its parties was not a State. In the US government's view, the CA2 definition did not create “field pre-emption”, Footnote 77 or in other words was not all-encompassing; put in more traditional international legalese, the notion of IAC was wider under customary law. This was a completely ahistorical argument; the notion of IAC was invented in the Geneva Conventions and replaced the equally inter-State notion of war, while there was no evidence that it was redefined either through treaty or through custom in this particular fashion.Footnote 78 The administration also considered that the conflict could not be a NIAC as it transcended the borders of a single State.Footnote 79 In Hamdan, the US Supreme Court rejected the government's arguments and found that the conflict with Al Qaeda could not be an IAC since it was not inter-State, and in an ambiguous holding apparently qualified it as a NIAC.Footnote 80 The position of both the Bush and Obama administrations post-Hamdan has hence been that the conflict with Al Qaeda is some sort of single, global NIAC, which is territorially unlimited in scope.
Opinions of course differ on whether the idea of such a global NIAC is legally tenable. In my view, even under a flexible interpretation of the IHL framework, which would allow for various kinds of cross-border NIACs, the idea of a global NIAC is an oxymoron. Any NIAC requires the existence of protracted armed violence which by definition has to take place somewhere, i.e. has to be localized at least to the territory of one State. That violence can spill over to the territory of another State (which need not necessarily be adjacent to the primary State), but there has to be a nexus to the protracted violence in the primary State for IHL to apply to that violence.Footnote 81 Thus, while one can safely speak of a NIAC (or NIACs) between the United States and the Taliban and other armed groups in Afghanistan, and while that conflict can spill over into, say, Pakistan or any other country – and arguably still be regulated by IHL – the existing legal framework does not seem to allow for a construction as amorphous as a planet-wide NIAC,Footnote 82 particularly one in which a loose terrorist network such as Al Qaeda is treated as a single organizational entity and belligerent party.Footnote 83 All of the difficulties in squaring the US conflict with Al Qaeda with the NIAC legal regime stem precisely from the fact that this regime was not designed to regulate anything like it.Footnote 84
In short, Al Qaeda as a non-State actor was and perhaps still is a party to a NIAC in Afghanistan. Its offshoots and affiliates have also been involved in other NIACs, as in Iraq or Yemen. But the attacks by Al Qaeda elsewhere have (thankfully) been so sporadic, and its control over its allied non-State actors so loose, that it would be exceedingly difficult to say that the intensity and organization criteria for a NIAC were satisfied for the purpose of establishing a single armed conflict which is global in scope.Footnote 85 One simply cannot aggregate all terrorist acts motivated by Islamic fundamentalism coupled with professed allegiance to Al Qaeda all across the world in order to satisfy the twofold intensity and organization test. Nor does it seem justifiable to depart from the Tadić criteria in the case of Al Qaeda but continue applying them rigorously to “normal”, common NIACs.
Again, opinions on this issue will differ. But even if the conflict with Al Qaeda could legally be qualified as a single NIAC, albeit a very unorthodox one, that conflict may well be approaching its end. The degradation that the US military operations have inflicted on the “core” Al Qaeda organization further threatens to push it below the NIAC organizational threshold, even if the threshold is applied more flexibly. Nor can this be prevented by using the concept of “co-belligerency”, which was imported from the law of IAC without much consideration as to whether the analogy can actually be drawn.
US policy-makers are of course well aware that the construction of a NIAC with Al Qaeda, which is sustained internally because it has broad bipartisan political appeal, is legally compromised by the continued degradation of core Al Qaeda and the impending US drawdown from Afghanistan.Footnote 86 But this does not end the US desire to keep using targeted lethal force, for instance through drones, when it considers such force to be necessary – nor, even more importantly, does this make the government's life any easier with respect to those individuals that it still holds in preventive security detention, in Guantanamo or elsewhere, despite efforts to draw such detention to a close.
Its immense political importance notwithstanding, the US “war on terror” or its rebranded NIAC with Al Qaeda and associated forces is so idiosyncratic and mixed up with parallel questions of US domestic law and policy that we should exercise extreme caution in drawing from it any lessons regarding the end of application of IHL.Footnote 87 As explained above, the better view is that IHL did not even apply in the first place to many situations within this supposedly single conflict. A significant number of individuals were either targeted or detained outside the framework of any discrete, localized, legally cognizable NIAC. The individuals who were detained in the context of an actual NIAC, e.g. in Afghanistan, and remain in captivity, would benefit from the protection of IHL. The end of the conflict would spell the end of any preventive detention authority (and such authority probably does not even exist in NIACs for either party),Footnote 88 while the protective rules of IHL regarding treatment and procedural safeguards would continue applying so long as the person remains detained.
Finally, it is important to note that the end of IHL's application does not necessarily leave a regulatory void. This is not the place for any extensive argument about how human rights law would apply to all these situations, especially those unregulated by IHL, but for our present purposes I would simply caution against a dogmatic juxtaposition between “war” and “law enforcement”. Not every targeted use of lethal force, nor for that matter every preventive detention, would necessarily be unlawful under the relevant human rights treaties, even in the absence of an armed conflict.Footnote 89
Conclusion
We have seen how the approach to the end of IHL's application needs to be objective and factual, as with its beginning – but that is easier said than done. The basic principle is that the end of armed conflict also ends the applicability of those parts of IHL that regulate the conduct of hostilities, while substantial parts of IHL actually survive the armed conflict. And while the end of a conflict should not be presumed lightly, with hostilities having to end with a degree of permanence and stability, any analysis of the ostensibly factual question of whether an IAC, NIAC or occupation has ended will at least partially be driven by external policy factors and the consequences of any finding regarding termination. Those factors are also significantly influenced by how IHL interacts with other branches of international law, such as international criminal law and human rights law. Accordingly, while the basic rules may be simple to state, they can be very complex to apply in practice, and that, at least, is inevitable.
