Geneva, We Have a Problem: Internationalisation of Armed Conflicts through Indirect Intervention Remains a Dead Letter

Abstract

The article seeks to raise awareness about the non-application of the norms of international humanitarian law (IHL) of international armed conflicts in situations of so-called internationalised armed conflicts – namely, when a non-state armed group (NSAG) that is engaged in an armed conflict against the territorial state enjoys a degree of support from another state. Debates in academic circles and international case law have focused largely on the appropriate test and threshold for establishing the relationship between the NSAG and the supporting state. Practice, however, shows that regardless of the legal test, the foreign state support to the NSAG in a (or an initially) non-international armed conflict is so politically charged that it leads to a complete non-application of the law of international armed conflict by the relevant actors. The article demonstrates its conceptual findings through four case studies: the armed conflicts in Donbas, Nagorno-Karabakh, Democratic Republic of the Congo, and Yemen. Regardless of strong indications of foreign state support to the NSAG in these armed conflicts, no relevant actors applied the IHL norms of international armed conflict. The article provides broader suggestions on the possible avenues for remedying the issue.

1. Introduction

On 24 February 2022, Russia's armed attack on Ukraine shook the world. It is impossible to keep track of all the international legal analyses, texts and scientific articles that have addressed this issue to date. Scholars in the field of international law have produced a sea of texts about the legality of the use of force in this case,¹ the application of international humanitarian law (IHL)² and human rights law (IHRL),³ war crimes⁴ and alleged genocide,⁵ jurisdiction of international courts and international institutional law,⁶ and so on.

However, among those who have contributed to examining the relationship of international law with the armed conflict in Ukraine, there were also those who tried to point out that there are many armed conflicts in the world that failed to provoke the interest either of the experts or of the general public as much as Ukraine has, and that it is necessary to raise the issue of double standards without questioning Russia's violation of international law.⁷ One such example is the conflict in Yemen: it broke out before the armed conflict in Ukraine, it is still ongoing, and it too has disastrous consequences.⁸

There are many similarities as well as many differences between the armed conflicts in Ukraine and those that are taking place in Yemen. One of the more interesting similarities from the point of view of IHL is the issue of foreign participation in these conflicts and how this affects their legal classification. For example, the Rule of Law in Armed Conflict (RULAC)⁹ portal classified the conflict in Yemen as ‘non-international’,¹⁰ and the conflict in Ukraine as ‘mixed’, their position being that there are both non-international and international armed conflicts (IAC)¹¹ in the latter country.¹²

The classification of armed conflicts depends on the actors involved in these conflicts. Thus, the legal classification of armed conflicts in certain situations is determined by the type and degree of participation of foreign states in an already existing non-international armed conflict (NIAC) on the territory of another state. Those addressing IHL have spent much time focused on the question of the type of relationship that must exist between non-state actors and other states in order for a pre-existing NIAC to legally transform into an IAC.¹³ The process of fact-finding and the legal classification of such conflicts is not at all easy, especially while they are still ongoing. In that sense, for example, the question of the relationship between the entities in the east of Ukraine and Russia during the period from 2014 to 2022, or that of the Houthis in Yemen with Iran, has been raised on numerous occasions.

However, even when it is determined, based on the facts, that the relationship between state C and non-state actor B on the territory of state A is such that the requirements for the legal internationalisation of the armed conflict have been met, the question remains as to whether the actors participating in that conflict are implementing the norms of IHL applicable to IAC. In this article we will investigate, precisely, that: whether the actors concerned recognised the existence of an IAC, and committed themselves to respect the IHL of IACs.

Our main argument is that case studies on the legal internationalisation of armed conflicts through indirect intervention show that, although IHL in these situations requires the application of the IAC rules, this did not happen in practice. This conclusion is confirmed by more than 20 cases analysed in our research.¹⁴ Participants in the conflict continued to treat it as a NIAC and to largely apply legal norms applicable under that regime. We identify several different reasons for this, and argue that the most important of these is that the application of such a classification of armed conflicts would mean recognition that a violation of norms of other subsystems of international law occurred, which would threaten the strategic interests of actors and the legitimacy of their military actions as a whole and thus make their execution more difficult. The described lack of application of the law of IAC is perhaps not relevant from a formal standpoint as classification of armed conflicts should be based on objective criteria.¹⁵ However, it is very relevant from the standpoint of the practical significance of IHL and classification of armed conflicts.

The structure of the article follows the main argument. In Section 2 we present the concept of legal internationalisation of armed conflicts and the main theoretical dilemma in IHL related to it, focusing on internationalisation through indirect intervention. In the central section of the article (Section 3) we examine more detailed case studies that confirm the main argument. In doing so, we focus on four case studies (Donbas, Nagorno-Karabakh, Democratic Republic of the Congo, and Yemen), noting that all such studies during this period confirm our initial hypothesis.¹⁶ The article ends with suggesting ways forward to mitigate this phenomenon, and concluding remarks.

We limit ourselves in the article to the situation involving the legal internationalisation of an armed conflict in which there is an indirect intervention by state C against state A, through the action of non-state actor B on the territory of state A. The article does not cover situations involving the direct intervention of state C against non-state actor B who is on the territory of state A without its consent, or the question of indirect occupation through a non-state entity. In addition, the article does not consider situations in which state C is intervening in the already existing NIAC that is taking place on the territory of state A on the side of its armed forces and with its consent (although, at least at first glance, we see no significant reason why our main conclusions would not be applicable in at least some of the scenarios referred to above).¹⁷ Finally, because of length constraints of the article, we only sketch out practical solutions to the problem of the lack of applicability of IHL in the situations of indirect interventions that we have identified. We will address this issue in more detail in our future endeavours.

2. Armed conflicts and their internationalisation in IHL

2.1. Armed conflicts recognised by IHL: IAC and NIAC

IHL recognises only two types of conflict: IAC and NIAC,¹⁸ which do not have legally binding definitions in IHL treaties. The most frequently cited definition of IAC, which derives from the provisions of common Article 2 of the Geneva Conventions, is that of Jean Pictet, from the Commentary on this Article: ‘Any difference arising between two states and leading to the intervention of armed forces is an armed conflict within the meaning of Article 2, even if one of the Parties denies the existence of a state of war’.¹⁹

On the other hand, common Article 3 of the Geneva Conventions, which was adopted after numerous compromises, mentioned armed conflicts not of international character, but their definition was purposefully not offered.²⁰ Later, Article 1(2) of Additional Protocol II provided that it shall not be applied in situations of internal tensions and disturbances without defining them explicitly.²¹ Nevertheless, criteria must be found on the basis of which, on the one hand, a distinction can be made between a NIAC and internal tensions and disturbances and, on the other hand, between a NIAC and IAC. This is essential because important differences still remain in the legal regimes that are applied in these situations regardless of all the changes, the development of customary IHL,²² and human rights and international criminal law. For example, Marco Sassòli mentions that, out of 161 rules of customary IHL mentioned in the ICRC Customary IHL Study, 136 (if not 141) were considered to apply both in IACs and NIACs.²³ At the same time, however, the same author warns that ‘the ILC has advanced a traditional theory of customary law in its attempt to help states identify customary law that would, if applied to IHL, certainly put many of the alleged advances into question’.²⁴ Therefore, there is still a debate over the customary status of certain norms encompassed in the ICRC Study.²⁵ In addition, a relatively low number of IHL treaties are applicable in NIAC. This, inter alia, means that mechanisms provided in the rest of them are not applicable in NIAC even if concrete norms stipulated in them have customary status. Finally, Marko Milanović addresses several aspects of this issue of the difference between the norms applicable in IAC and NIAC, and the process of internationalisation of armed conflicts. He concludes that one of the most evident differences lies in the existence of combatant immunity, POW status, and the ensuing body of norms that offer specific types of protection to this category of individuals.²⁶ Lastly, notwithstanding the developments in international criminal law, there are still differences between the stipulated war crimes in NIAC and IAC. This is confirmed in the statutes of international criminal judicial institutions such as the International Criminal Tribunal for the former Yugoslavia (ICTY) and the International Criminal Court (ICC).

2.2. Legal internationalisation: The concept of standard internationalisation through indirect intervention

As there is no legally binding definition of an internationalised armed conflict in IHL treaties, we are forced to rely on the doctrine of IHL, having addressed this issue for a long time. For the purposes of this article we accept the definition given by Kubo Mačák, which states that one should distinguish between situations where there is no change in the legal regime that is applied in an armed conflict, even though in the factual sense there has been an intervention by a foreign state in the armed conflict,²⁷ and situations where the intervention of a foreign state in an existing armed conflict leads to a change in its legal classification, and the conflict is thus transformed from a NIAC to an IAC.²⁸

This can happen in various ways but, in the context of this article, we are most interested in the situation that Mačák refers to as ‘standard internationalisation’.²⁹ Here, legal internationalisation occurs because the relationship between state C and non-state entity B on the territory of state A is such that one can no longer speak about the independence of non-state entity B. Instead, non-state entity B becomes so controlled by state C that one can say that the above-mentioned requirements for the existence of IAC – any dispute that leads to the use of armed forces of two or more states – have been met. More precisely, we are interested primarily in situations in which legal internationalisation occurs based on the indirect intervention of state C on the territory of state A through non-state actor B on the territory of state A. This requires the existence of a certain relationship between state C and non-state entity B.

Although the legal internationalisation of armed conflicts also occurred in earlier periods, contemporary doctrinal discussions of this topic usually revolve around the interpretation of two judgments of international courts: the Nicaragua case before the International Court of Justice (ICJ) and the Tadić case before the ICTY.³⁰ In the case of Nicaragua, the ICJ was tasked with reaching a decision on, inter alia, the responsibility of the United States for the IHL violations committed by the contras, a NSAG on the territory of Nicaragua, bearing in mind that it has been proved without a doubt that it was the US that provided them with financial assistance and training. For these purposes, the ICJ developed two tests that would make it possible to answer this question.³¹ The first was the ‘total dependence’ test, according to which the actions of a NSAG are attributable to a foreign state if the former is completely dependent on the latter, and if the state uses said dependence. In that case, all the actions of the NSAG are attributable to the foreign state. If, however, the conditions for the test of complete dependence are not met, then another test developed by the ICJ in this case comes into play: the ‘effective control’ test, which implies that the illegal actions of a NSAG will be attributable to a foreign state if the state, before and during the specific operation, gave instructions to the non-state entity – that is, if it had control at the beginning of the operation and during its execution.³²

On the other hand, at the very beginning of its work, the ICTY was faced with the question of the classification of armed conflicts in Bosnia and Herzegovina; in the case of Tadić, the Court tried to find an answer to this question by applying the jurisprudence of the ICJ from the case of Nicaragua. As several authors have already explained, both the Trial and Appeals Chambers in this case misinterpreted the ICJ judgment in the Nicaragua case, based on different reasons.³³ Without going into details of the complex argumentation, the Appeals Chamber in Tadić concluded that the ICJ had reasoned incorrectly in the Nicaragua case, and that the effective control test was neither adequate for the question of state responsibility nor for the question of the classification of armed conflicts. Instead, the Chamber offered an ‘overall control’ test.³⁴ This test implies that an armed conflict may be legally internationalised if a foreign state is not only equipping and financing the NSAG, but also coordinating and assisting in the general planning and organisation of its activities.³⁵

Later practice of the ICTY has accepted the use of the overall control test³⁶ for the legal internationalisation of conflicts, although it sometimes avoided classifying the armed conflict in question.³⁷ Other courts also have dominantly used the overall control test to classify conflicts.³⁸

It seems that the new commentaries on the Geneva Conventions prepared by the ICRC also accept the use of the overall control test both for the classification of the conflict and for the test of attribution:³⁹

In order to classify a situation under humanitarian law when there is a close relationship, if not a relationship of subordination, between a non-state armed group and a third state, the overall control test is appropriate because the notion of overall control better reflects the real relationship between the armed group and the third state, including for the purpose of attribution.

Sassòli agrees with this opinion, arguing, inter alia, that a different solution could lead to the conflict being classified as an IAC based on the overall control test, without the state involved in the conflict being responsible for the violation of IHL norms.⁴⁰

On the other hand, there are authors who accept the overall control test for the classification of the conflict but point out that it should not be confused with the issue of state responsibility, for several reasons.⁴¹ In a way, the ICJ also left this possibility open when, in the case of Bosnia v Serbia, it asserted that the overall control test could lead to the breakdown of the system of legal responsibility of states, but may be used to classify a conflict.⁴²

Finally, there are authors who believe that there are various problems in the use of the overall control test for the legal internationalisation of armed conflicts. Djemila Carron, for example, first supports those who believe that a distinction should be made between the issue of state responsibility and the legal internationalisation of conflicts. She puts forward that a distinction between tests should be made based on whether third-state intervention occurs from the outset of or during the conflict.⁴³ Carron argues that the first situation requires specific and strict control of state C over non-state entity B (similar to the effective control test), while in the second situation it is necessary that the control be general and strict (but still at a higher level than that of the general control test).⁴⁴ Noam Zamir, on the other hand, first lists numerous problems related to the application of the concept of legal internationalisation of conflicts and the application of the overall control test; he then argues that, although the concept is widely accepted in the practice of international courts, it is still not part of customary international law.⁴⁵ This ultimately leads him to the conclusion that ‘the rules of attribution of the law of state responsibility should be used at least as a benchmark for conflict classification in IHL’, but that it is ‘necessary to take into account the special features, context and purposes of IHL’.⁴⁶

We would like to stress that these contributions focused on the issue of the type of relationship between third states and the NSAG in the other state. Our contribution in this article differs as we are demonstrating that the main problem for the application of norms of IAC is not the type of the above-mentioned relationship or the proper test of control, but the general lack of will of actors to classify the conflict as an IAC, or to apply the norms of IAC in these situations. As is explained in the following section, we are aware that classification of armed conflicts should be carried out based on objective criteria and not the subjective position of interested parties. However, our argument is that parties to the conflict simply ignore this and, because of their strategic political interest, never apply norms of IAC in situations of indirect intervention.

2.3. Problems in applying the concept of standard internationalisation through indirect intervention in practice: Strategic interests of actors

Zamir, referred to above, is one of the authors who also pointed out that there are certain doctrinal and practical problems related to the application of the concept of standard internationalisation of armed conflicts in the case of indirect intervention.⁴⁷ Among them, he points to the fact, already referred to, that the concept of internationalisation is not defined by treaties in the area of IHL; that states are generally reluctant to recognise members of a NSAG as POWs; that states’ application of norms of IHL, valid in the case of an IAC based on the fact that a non-state entity is acting on behalf of another state, is controversial, and so on.⁴⁸

Along similar lines, in a recent text about the conflicts in the east of Ukraine, Sassòli draws attention to the fact that Russia has been denying its participation from the very beginning of the conflict, while Ukraine has always claimed that the actors in the east of Ukraine were nothing but Russian agents.⁴⁹ That, however, does not mean that Ukraine has applied the norms of IAC. In this respect Sassòli warns that the test of internationalisation developed by the Appeals Chamber in the Tadić case creates enormous practical and theoretical problems if one tries to apply it in situations such as that in Ukraine, and that ‘unrealistic rules do not protect anyone; they undermine the credibility and, therefore, the protective force of the entire IHL regime with the fighting parties’.⁵⁰

Prior to Zamir and Sassòli, other authors also kept pointing to problems related to the application of the concept of standard internationalisation. For example, Andrew Carswell first gave the following warning regarding the process of classification of armed conflicts: ‘Given the inherently political nature of such a determination, it would be naïve to presume that states will undertake it with complete objectivity’.⁵¹ More specifically, with regard to proxy wars, Carswell had the following argument (and it is worth quoting at length):⁵²

[G]iven the fact that states will rarely admit their responsibility for a third-party armed group, the exercise of classifying such a conflict can take place in a political minefield. Even if the requisite level of control is objectively established, the non-state actor will most likely have an incentive to deny that it is being controlled by the state in question, since an acknowledgement of that relationship could engage both the political and international legal responsibility of its closest allies. As such, the non-state actor would have a disincentive to publicly apply the more fulsome body of IHL related to international armed conflict, even if it is objectively applicable.

Although other authors have also noted incidentally the political sensitivity of the classification of armed conflicts, especially the issue of standard internationalisation, it was probably Carswell who went the furthest and was the most precise in this sense. Still, neither he nor other authors considered this question as a central technical problem when it comes to the classification of internationalised armed conflicts, but rather as incidental. The case studies discussed below show that political interests of the actors cannot be treated as incidental or technical because it is obvious that they completely prevent the proper application of IHL in cases of standard internationalisation of armed conflicts through indirect intervention. Consequently, solutions such as Carswell's – about ‘“fortifying” the formally applicable law through military doctrine’ – are proving to be insufficient.⁵³

Despite the rule of the strict separation between jus in bello and jus ad bellum as one of the foundations of the application of modern IHL,⁵⁴ the question of the application of IHL norms in connection with standard internationalisation cannot be viewed separately from the overall context of military operations that are carried out in such a case and the right to use force – jus ad bellum. Namely, the recognition of actors that they are in fact providing financial assistance and participating in training or arming – not to speak of complete dependence or effective control of the NSAG on the territory of another state – would mean that a state is admitting to a violation of international law, which states almost never do. In other words, states would be admitting that they were violating the principle of the prohibition of intervention in the internal affairs of other states, as well as, in most situations, the already mentioned norms of jus ad bellum. Additionally, not even NSAGs are willing to admit that they belong to another state – that they completely depend on it or are under its control – because this would threaten the interests of the actors on which they are completely dependent. In fact, some of the rules that would be triggered in the application of IHL norms of IAC arguably would benefit the NSAG. This would be the case, for instance, for combatant immunity and the consequent shielding from prosecution for mere participation in hostilities, should it be the case that the NSAG members fulfil the criteria outlined in Additional Protocol I or Geneva Convention III.⁵⁵ On the other hand, the evident drawback would be that complying with the law of IAC by the NSAG would trigger the question of its capacity in fact to do so. Nonetheless, admitting the existence of a relationship of control would ultimately displace the obligation of compliance from the NSAG towards the intervening state as the party to that IAC. The choice of the NSAG then to still deny the links with the intervening state testifies to the fact that other interests prevail over those inherent in the more ‘beneficial’ international legal framework. It is therefore unrealistic to expect – as will be shown by the case study analysis presented in the next section – either an intervening state or a non-state actor under its influence to admit at any time that the internationalisation of an armed conflict through indirect intervention occurred. In addition, not even states on the territories of which indirect interventions take place are ready to accept the application of IAC norms, even though they usually claim that non-state groups are just proxies of another state. One of the main reasons is their unwillingness to accept POW status for members of a NSAG in their territory.⁵⁶ It is reasonable, therefore, to conclude that the three main actors in the situation of standard internationalisation through indirect intervention – the intervening state, the NSAG, and the territorial state – have their own strategic interests in not applying the norms of IAC in these situations. On the one hand, the intervening state and the NSAG do not even accept the existence of the intervention and classification of the armed conflict as IAC. On the other hand, even if the territorial state accepts the internationalisation of the armed conflict, it refuses to apply the whole range of norms applicable in this type of conflict. As will be illustrated in case studies featuring in this article, the above-mentioned strategic interests are of fundamental importance for these actors.

The insistence on the strategic interests of actors in IHL, however, is not entirely new, and it is likely that those who work with IHL will respond with several counter-arguments to the argument that these strategic interests influence the application of IHL norms. It is possible, first of all, to claim that conflict classification is undertaken based on objective facts, that it is not based on subjective statements and positions of the actors in the armed conflict, and that – from the IHL point of view – it is not relevant that the actors are not admitting to the existence of standard internationalisation through indirect intervention. However, what is relevant is whether the actors admit to applying relevant norms of IHL. The problem with application stems from the fact that those who directly participate in the conflict are the only ones who can ultimately decide on the application of IHL norms in the domain of classification of armed conflicts. In addition, this classification is especially important during the conflicts themselves. It is not disputable that those who refuse to acknowledge the existence of IAC, despite the fact that the objective criteria for doing so have been met, are in violation of IHL. This, however, is no consolation if it turns out that all, or almost all, actors do not apply rules of IAC during an armed conflict as this would undermine the accomplishment of the main objectives of IHL. An additional problem lies in the fact that, for the reasons we have just stated, actors will do almost anything in their power to hide the facts – that is, evidence of their own participation in cases of indirect intervention or dependence on that state – which makes the task more difficult, even for those who classify the conflict in accordance with the principle of good faith.⁵⁷

The second counter-argument could be that even if actors do not classify the conflict properly, this does not constitute a special exception in the field of IHL. Namely, one could offer an argument that there are numerous examples of violations of IHL norms as ‘IHL represents the extreme end of the spectrum of international law where the most fundamental interests and even the very existence of the state may be at stake’.⁵⁸ Nevertheless, the problem in situations of standard internationalisation through indirect intervention is the fact that actors never apply norms of IAC in such situations, which is confirmed by our analysis of relevant case studies. Therefore, this situation is different from those in which we have examples of both violations and compliances.⁵⁹ It must be stated therefore that, in relation to standard internationalisation through indirect intervention, the problem is that it is almost impossible to find an example of the application of proper IHL norms in this respect, as is shown in the rest of this text.

3. Case studies

This section seeks to showcase that the concept of standard internationalisation through indirect intervention has never been applied in practice. An examination of over 20 armed conflicts in which there were indications of a certain level of indirect intervention by foreign states by way of support for the NSAG, and where the question of internationalisation was posed by scholars,⁶⁰ demonstrates that in none of these situations of armed conflict was the concept of an internationalised armed conflict applied by all the parties. There are at least two reasons to believe that the inapplicability of the concept of internationalisation was neither a consequence of disagreement between the parties on the appropriate test to be used to determine control, nor the appreciation of whether in fact the test, whichever it was, had not been met in the case in question. The first reason is the absence of evidence of any such assessment by the relevant actors at the time of the conflict. The second can be inferred from the very fact that the behaviour of the parties always corresponded with the model presented above, regardless of whether the assessment by scholars or international tribunals spoke in favour of or against any threshold of control being met. In other words, both in the case studies where it was, and was not, found that a third state intervened indirectly by exercising control over a NSAG engaged in an armed conflict against the territorial state, the NSAG and the third state were denying the existence of such control. Further, the territorial state was either also denying the existence of an IAC, or was accusing the outside state of exercising control over the NSAG but did not (according to the available evidence) accordingly apply the norms of IHL applicable in IAC.⁶¹ In the case of Yemen, for instance (as will be shown in the section below), reports and international scholarship suggest that this relationship did not amount to one of overall control. Nonetheless, regardless of this factual pattern, the stances of the relevant actors involved in the conflict were the same as in cases where control was (at least post facto) established (such as in the cases of Donbas, Nagorno-Karabakh, and the Democratic Republic of the Congo). This occurrence leads us to the conclusion that the reason behind this inapplicability of the concept of internationalisation generally does not lie in the disagreement between the parties as to the appropriate test to be used to determine control nor the factual assessment of whether the appropriate legal test has been met, but in the legal and political implications that such findings have for the interested sides. As explained above, the quintessential implications would arguably include the violation of jus ad bellum by the third state, and the probable loss of legitimacy and claim for its cause by the NSAG. For the territorial state, the concept would possibly trigger the necessity to extend immunity from prosecution for participation in hostilities to its own nationals fighting against it.

The limits inherent in producing an article of this length do not allow us to engage in a more comprehensive and in-depth analysis of all the situations referred to. Thus, we examine four case studies to illustrate the theoretical remarks made in the earlier sections: (i) Donbas, (ii) Nagorno-Karabakh, (iii) Democratic Republic of the Congo, and (iv) Yemen.⁶² The four case studies reflect the patterns of behaviour which were also spotted in other analysed situations – all leading to the non-application of the full scope of IHL norms applicable in IAC.⁶³ However, the final choice of these four scenarios was made based on the volume of information being available in the public domain. Such access to reports allowed us to look at both the factual circumstances and the pronouncements of the participants in more detail. In fact, part of the problem of classification of internationalised armed conflicts lies in the difficulties in accessing information that would allow a more detailed assessment of all the situations, and monitoring the evolution throughout the different stages of the armed conflict. Having evidence and data on the possible relationships of control deliberately hidden and destroyed by the parties to the conflict hinders researchers and scholars in engaging in a bona fides, neutral inquiry into the conflict classification as per the criteria derived from international jurisprudence. The fact that violations of international law in these four armed conflicts were examined in court proceedings also contributed to the accessibility of information on the factual circumstances of the conflicts, as well as insight into the views of parties on conflict classification. The four studies represent situations in which concerns were raised as to both the involvement of an external state in the conflict and whether the nature of support of such state altered the classification of that conflict. In selecting the case studies, we also aimed to depict different circumstances where the issue of classification might arise, with the fact that internationalisation was not applied across the spectrum, thus confirming the hypothesis of this article of internationalisation being just ‘a dead letter’. Although the article presents the four case studies in more detail and chooses the cases for the reasons listed above, the information available from other examples encountered in international practice does not undermine the overall thesis of the article.

3.1. Donbas (2014–2022)

The issues surrounding classification against the background of strong indications of indirect support by an intervening state arose in the context of the armed conflicts involving Ukraine and the People's Republics of Donetsk (DPR) and Luhansk (LPR), and occurring in the territory of Ukraine between 2014⁶⁴ and 2022.

A large authoritative stream of international scholarship found that the support provided to the DPR and LPR by Russia was not of such a degree so as to render it a relationship of control and thus give the conflict an international nature. For instance, the RULAC portal⁶⁵ and Human Rights Watch⁶⁶ found evidence of equipping and training the armed groups in their assessments; however, unlike Amnesty International,⁶⁷ having examined the information that follows in this text, they concluded that the type of assistance had not sufficed to reach the threshold of control. The ICRC reached a similar classification of the conflict in 2014.⁶⁸

Further, various authors also point to the shifting dynamics in these relations. At the beginning, contrary to the vast convergence in the political sphere, military support was confined to the provision of weapons and training, but was rated as periodic and dependent on satisfaction with the behaviour of the local authorities.⁶⁹ Towards the end of 2016, reports found that Russia came to exert influence through persons in charge of the republics’ trade and finance, the local military forces as well as Russian advisers pertaining to Federal Security Service (FSB) agents and deployed on the ground (kurators).⁷⁰ Additionally, the scope of Russia's support provided to the republics widened with the internal divisions around the creation of a Novorussiya,⁷¹ as well as willingness to engage in fully fledged hostilities⁷² having been settled. This assistance ranged from an increase in the budget allocated to the salaries and social benefits of local military and civil authorities,⁷³ military supplies encompassing ammunition, weaponry and fuel,⁷⁴ as well as strengthened economic ties,⁷⁵ increasing the dependence of the DPR budget on Russia to as much as 90 per cent.⁷⁶ According to analysts, such relations led to all key strategic military decisions being imported from Russia, with no autonomy on the part of the republics to challenge them.⁷⁷ In her assessment, Natia Kalandarishvili-Mueller underlines that the initial autonomy the rebels held in their relations with Russia disappeared in around 2019⁷⁸ when internal divisions of the republics came to an end.⁷⁹ Training and equipping the republics, coupled with the role played by Russia in the planning of the groups’ activities by way of influencing strategic decisions, could well be found to reach the threshold of overall control. Even more so, the adoption of military decisions by Russia on behalf of the republics could arguably prove the existence of the more stringent effective control exercised over the republics, which requires the third state to issue precise instructions on committing specific acts.

Regardless of the arguably fluctuating relationship of support between the de facto republics and Russia, this changing factual pattern was not met with any amendments in the positions of the relevant actors as to the classification of the conflict. From the early stages of the conflict in 2014,⁸⁰ Ukraine classified it as an international conflict, deeming the People's Republics as mere proxies of the Russian Federation, and the situation in Donbas as one of temporary occupation exercised by Russia. Ukraine declared its loss of control over the two oblasts in April 2014.⁸¹ Such pronouncements were later accompanied by legislative changes, defining the territory under occupation as including that controlled by Russia's occupying administration, and referring to the self-proclaimed republics.⁸² Interestingly, however, as Sassòli underlines, this denomination of the People's Republics did not prompt Ukraine to apply the IHL of IAC.⁸³ Ukraine's behaviour well illustrates the scholars’ findings that, even in cases of denouncing its domestic rebels as proxies of foreign powers, the territorial states do not invoke the (full) application of IHL of IAC towards them.⁸⁴ The rule featuring most prominently in distinguishing between the legal framework of the law of IAC and NIAC is that of combatant immunity in the former. Conversely, the existence of overall control of Russia over the DPR and the LPR would prompt Ukraine to be required at least to examine whether members of such armed units comply with the conditions set out in AP I and GC III, and consequently classify for combatant immunity and merit POW status. Such status would have attached the immunity from prosecution for mere participation in hostilities.⁸⁵ However, reports do not suggest that there had ever been such an appraisal by the Ukrainian side. In fact, when it comes to members of armed groups affiliated with Russia in the new phase of the conflict, with the ties between the entities being even more evident, Ukraine subjected such persons to criminal charges pursuant to domestic law for crimes belonging to the realm of national security.⁸⁶

On the other side, Russia denied any such agency,⁸⁷ let alone direct military involvement,⁸⁸ in the Donbas events, and repeated its role as the guarantor of the Minsk agreement and not a party to the armed conflict.⁸⁹ Quickly abandoning its initial support for the Novorossiya project,⁹⁰ Russian officials stated that they considered the Donetsk and Luhansk oblasts to be part of Ukraine,⁹¹ and placed hope that future talks would lead to their preservation within that country.⁹² Nearing the end of this ‘non-international phase’ and days before directly intervening in the armed conflict (thus, certainly elevating it to an (additional) international conflict), Russia recognised the People's Republics as independent states.⁹³

Finally, as certain sources claim, the DPR and LPR themselves largely have maintained publicly that Russia's influence on their decision making was minimal,⁹⁴ and rather was confined to a potentially significant intervention in terms of financial and military aid in the event of a greater humanitarian catastrophe.⁹⁵ Following the initially Russian-endorsed uprising,⁹⁶ the results of referenda held in the two regions led to the republics’ proclamation of independence from Ukraine.⁹⁷ However, in the light of the personal ties between the republics’ leadership and Russia, the DPR higher officials initially sought integration with Russia,⁹⁸ rendering their attitude towards the relationship with Russia ambiguous.

Such a denial of a relationship of control on both sides is not surprising. On the side of the NSAG, admitting to such a relationship could invoke state responsibility of its decisive ally. Similarly, such findings could greatly diminish their claims to the establishment of an independent country, given that it would create a perception of the NSAG being a mere puppet of another state. On the side of the intervening state, the political interest it holds in the situation that prompted it to opt for indirect intervention in the first place would render the entire effort futile, were it then to be led by objective criteria as to the conflict classification. Direct intervention, one might argue, would have then been a much simpler endeavour. Consequently, Russia's behaviour goes along the lines of prevailing state practice, which shows that in no instances did the ‘intervening’ state classify such armed conflicts as of an international nature.

Having multiple human rights bodies tackle the situation in Ukraine, while jumping over or avoiding a conclusive classification of the conflict,⁹⁹ prompts us to ask whether classification efforts are indeed idle. The Donbas conflict illustrates the difficulties of conflict classification in cases involving indirect intervention. In addition to the problems occurring as the conflict unfolds, equally concerning is the fact that the availability of information had not progressed for eight years after the beginning of the conflict, and that debates among both academics and international institutions still hold up.¹⁰⁰ However, if classification may take years to figure out, how do we expect the soldier to know which body of law to implement at the instant of the conflict?¹⁰¹ Even with this post facto perspective, data (and the lack thereof) shows us that we are unable retrospectively to monitor the entire evolution of armed conflicts and the extent of such relationships of support. In the context of Ukraine, the decision of the European Court of Human Rights (ECtHR) in the case of Ukraine and The Netherlands v Russia sheds light on the facts of the case and comes to the conclusion that Russia exercised jurisdiction over the territory of Eastern Ukraine from 2014, based on the military, economic and political support provided to the DPR and the LPR.¹⁰² However, even the posterior clarification of the extent of the support relationship between the ‘de factos’ and Russia and the moment in question does not remedy the fact that the applicable norms of IHL were not applied throughout the conflict, because this relationship was consistently denied by the relevant participants.

3.2. Nagorno-Karabakh (1991–2020)

The territory of Nagorno-Karabakh was placed initially as an autonomous region within the Soviet republic of Azerbaijan. As a result of it being mainly comprised ethnically of Armenians, it requested its integration with the Soviet republic of Armenia; this was denied in 1989. With the dissolution of the USSR, Nagorno-Karabakh remained under the formal authority of Azerbaijan. However, backed by Armenia, the region restated its secessionist claims and declared an independent Republic of Nagorno-Karabakh (NKR) in 1991. An attempt to end the hostilities was made in 1994, but to little avail, as skirmishes and periodic resurgences of fighting continued until 2020. The armed conflict broke out again in September 2020, with open hostilities lasting for two months.¹⁰³

The following elements were found to corroborate the existence of an exercise of control by Armenia over Nagorno-Karabakh forces:¹⁰⁴

• • overlaps in the personal exercise of authority in the official apparatus of both the state of Armenia and the separatist forces;

• • an agreement between the two entities providing for the possibility of exercising military service in either of them;

• • the high level of integration between the two militaries;¹⁰⁵

• • the decisive role of military support provided by Armenia for the exercise of territorial control¹⁰⁶ and for the exercise of border control;¹⁰⁷ as well as

• • essential financial assistance.¹⁰⁸

RULAC underlined Armenia's role in equipping, financing, training and providing operational support, as well as coordinating and helping the general planning of military activities.¹⁰⁹ Hence, according to RULAC's assessment, this factual pattern would satisfy the overall control test for the conflict's internationalisation. In addition, the ECtHR noted Armenia's political support, the presence of its law enforcement agents, as well as the issuing of passports to inhabitants of Nagorno-Karabakh, to depict the multi-layered relationship between the two entities.¹¹⁰

Unsurprisingly, the stance that actors took up aligned with their political and strategic interests in the conflict. The view of Azerbaijan was that Nagorno-Karabakh was under the occupation of Armenia.¹¹¹ Upon its third-party intervention in the Chiragov case before the ECtHR, it stated that the NKR was subordinated to Armenia and could survive only thanks to its extensive political, economic and military support.¹¹² Azerbaijan firmly accused Nagorno-Karabakh of being Armenia's puppet and brought this alleged occupation of its territory to the attention of the international community. Along those lines, it recognised only Armenia as its interlocutor in conversations addressing the territory's status.¹¹³

On the other hand, Armenia insisted on the independence of the NKR, stating that the relationship between the two states was nothing more than mere cooperation, and that it sees its role as guarantor of the security of the region.¹¹⁴ Similarly, during the Chiragov proceedings, although Armenia admitted to the existence of the 1994 Military Agreement, it stated that the percentage of Armenians actually performing military service in the NKR was insignificant.¹¹⁵ Beyond assisting with specific infrastructure-related projects, including rebuilding schools and hospitals,¹¹⁶ and providing humanitarian assistance, Armenia has adamantly denied providing any support to the NKR, especially with regard to its military efforts.¹¹⁷ This denial of ties with the separatist forces is also reflected in the disagreement between Azerbaijan and Armenia on the meaning of the provision of the Ceasefire Agreement that stipulates the withdrawal of Armenian forces.¹¹⁸ Armenia insisted on the relationship with the NKR authorities as being one of de facto alliance. It underlined Armenia's position as a security guarantor for Nagorno-Karabakh and characterised Azerbaijan's actions as aggression on Nagorno-Karabakh.¹¹⁹ Such a position precludes Armenia from accepting any position as an occupying force through a proxy given that such an exercise of authority would trigger its responsibility for ensuring public life and order in that territory. Further, its resolute position on the independent nature of the NKR provides insight into the (un)likelihood of Armenia accepting responsibility for any persons detained by the NKR, as well as, consequently, recognition of being the detaining power.

Thus, Armenia denied involvement in the domestic affairs of the other state. For Armenia to claim otherwise would lead it to admit violating the principle of non-intervention. Similarly, in line with its strategic and political claims to independence and preserving the support of Armenia, Nagorno-Karabakh denied having any ties with the latter. Nagorno-Karabakh Armenians see themselves as a separate, sovereign nation, which was economically and culturally oppressed while living alongside the Azeris throughout the existence of the USSR.¹²⁰ It also identifies itself as independent of Armenia and considers it necessary for NKR representatives to engage in direct negotiations with Azerbaijan.¹²¹

Nonetheless, even with the rigorous scrutiny of the ECtHR in Chiragov on the ties between Armenia and the NKR, many pages were written of dissenting opinions, and scholarly critiques¹²² were written over the ‘watering down’ of the criteria previously applied by international courts to assess the existence of a control relationship between a state and a non-state entity.¹²³ For instance, Judge Pinto de Albuquerque engaged in an extensive rebuttal of the existence of a relationship of control across various spheres, including by underlining the independent external representation of the NKR, the voluntary rather than automatic application of Armenian legislation in the NKR and a discrepancy in the court system of the two entities, the exceptional and non-regular possibility of having NKR citizens issued with Armenian passports, as well as the flawed argument relating to financial support provided to the NKR taken in the light of modern trends in international financial cooperation. Thereupon, Judge Pinto de Albuquerque pointed to the findings of the 2005 fact-finding missions of the Organization for Security and Co-operation in Europe, corroborating the finding that available evidence did not substantiate the conclusion that Armenia was directly involved in the administration of the NKR.¹²⁴

This discrepancy is obvious not only at the time of the conflict itself, but also, as Chiragov demonstrates, after having the hostilities moved to the courtroom. Although amusing for subsequent legal analysis, the Chiragov proceedings similarly show us that the disagreement over the legal concept of an internationalised armed conflict is twofold. At the first tier, discords arise in determining the adequate test for scrutinising the relationship between two actors. At the second tier, scholars and practitioners diverge in seeking the manner by which to apply the agreed test to the facts on the ground. Using arguably the same corpus of information available in the public domain, multiple reports mentioned above have argued for different conclusions as to whether support amounted to control. Finally, the Nagorno-Karabakh case demonstrates that it was in the interests of none of the actors to claim, and apply, conflict internationalisation. The alignment between political interests and the legal classification by the parties should thus be noted.

3.3. Democratic Republic of the Congo (Zaire, 1996–97, 1998–2003)

The armed conflicts in the Democratic Republic of the Congo (DRC) in each of its multiple phases provides a front-row seat in observing the lack of application of the concept of an internationalised armed conflict throughout the multiple armed conflicts fought and the changes in their parties. In both instances of the armed conflicts in 1996 and 1998, in addition to their direct military engagement, Uganda and Rwanda held a supporting role for certain armed factions acting in opposition to the governing regime of the DRC.¹²⁵

The question of classifying the DRC clashes only came to the fore in July 1996.¹²⁶ The DRC tried to raise international attention regarding the direct incursions of Uganda, Rwanda and Angola, and their military activities, on its territory in the summer of 1996, but to no avail. Further, from the outset, the DRC continuously upheld that it was party to an IAC against Rwanda and Uganda, which exercised control over the rebels of the Alliance of Democratic Forces for the Liberation of Congo-Zaire (Alliance des Forces Démocratiques pour la Libération du Congo – AFDL).¹²⁷ The support of both Rwanda and Uganda for the insurgent AFDL was documented in numerous sources¹²⁸ and provided the basis for reasoning on the exercise of overall control,¹²⁹ whereby both countries not only supplied the rebels with weapons and ammunition, but also provided them with training and offered strategic support in planning their operations.¹³⁰

What is original about Rwanda's involvement in the insurgence in the DRC is the explicit recognition of such an indirect intervention by that country's highest officials. Rwanda's behaviour went in contrast to almost all other instances of foreign indirect intervention, where the intervening country would be adamant in denying its ties with the rebel forces operating on the ground. To the contrary, Rwanda's President, Paul Kagame, underlined that overthrowing the DRC President Mobutu Sese Seko through the military and logistical support of insurgents (the AFDL) was a strategic objective of the country.¹³¹ However, his open pronouncement was not followed by any determinations in terms of repercussions on applicable IHL norms.

Additionally, there was little appetite on the part of the wider community outside the region to engage in assessing the international elements of the conflict in Zaire. It is interesting, however, to note that, in making its assessment, the UN Commission for Human Rights Joint Mission felt that it had become appropriate to admit to an existence of a relationship of control between the AFDL on one side and Rwanda and Uganda on the other, only at the point when the factual circumstances of the conflict had become more obvious, and officials of the two countries had publicly acknowledged their role in the conflict.¹³² The armed conflict in the DRC again points to the inappropriateness of the instantaneous application of the concept of an internationalised armed conflict, seeing that clarification of the facts of the case and the relationships of control only arises well after the end of hostilities.¹³³ Further, while authors point to the convoluted nature of the relations among actors and the lack of available data in the Donbas case, the extract from the Mapping Report might rather point to the additional political implications in the event that the UN Commission for Human Rights Joint Mission had classified the conflict as international from the outset.¹³⁴ Although it is known that the classification of an armed conflict is made by the parties themselves, scholars repeat that such an endeavour should be grounded on an appraisal of the facts in good faith. For the Mission to then give such weight to the public statements of the highest leaders upon engaging in conflict classification might risk having political criteria loom over the objective criteria in determining the nature of the armed conflict and acquiesce to the classification being a political instead of a factual exercise.

The Congolese situation a couple of years later, in 1998, showed the flip side of the coin with regard to the behaviour of the territorial and the intervening states. This time, it was the supporting states, Rwanda and Uganda, that were largely turning a blind eye to the relationship of control between them and the rebel forces of the Movement for the Liberation of the Congo (Mouvement de Libération du Congo – MLC) and the Rally for Congolese Democracy (Rassemblement Congolais pour la Démocratie – RCD) during the period from 1998 to 2003.¹³⁵ Contemporary reports indicate that such a relationship of control arising from the degree of support and coordination did in fact exist between the RCD and Rwanda, if not also in the case of the MLC and Uganda.¹³⁶ Such a conclusion is made based on evidence which suggests that the RCD actions were coordinated by Rwanda (thus satisfying the limb of the third state playing a role in organising, coordinating or planning military activities); as well as the role of Rwanda's support in establishing the RCD (hence meeting the second limb of the overall control test on equipping, financing, training or providing operational support for the group).¹³⁷

The RCD splintered soon after the beginning of the conflict: one group being based initially in Kisangani and supported by Uganda (RCD-ML), and the other being based in Goma and enjoying Rwanda's assistance (RCD-Goma).¹³⁸ Rwanda, for its part, initially denied any involvement and denounced the armed conflict as a purely domestic matter for the DRC. The RCD also underlined its battle as a ‘struggle of the Congolese people’¹³⁹ and autonomous from Rwanda.¹⁴⁰ Uganda similarly rejected any allegations of playing a role in the rebellion for the first month of the armed conflict,¹⁴¹ although it was evidenced that the country provided assistance to both the RCD-ML branch and the MLC through recruitment, training and armament.¹⁴²

The DRC, on the other hand, held a rather ambiguous position with regard to the domestic rebellion. While, at the political level, refusing to engage in direct negotiations with the insurgents, and denominating both the RCD and MLC as the two states’ proxies, on the ground the DRC still applied the law of NIAC towards the two non-state armed groups. It was a year into the conflict when the DRC authorities referred to some of its provinces as occupied by Rwanda and Uganda in their statements towards international organs.¹⁴³ The fact that towards the end of the conflict, DRC President Kabila decided to grant amnesties to members of the armed groups for acts such as carrying weapons against one's own nation precludes the possibility that the DRC applied the IHL of IAC with regard to the conflict with the rebel movements. Indeed, if the DRC were to have done so, and if the NSAG members were to have fulfilled the criteria for combatant and POW status set out in AP I and GC III, it would precisely be the participation in hostilities that would be considered permissible, and no amnesty would be necessary for such activities in the armed conflict.¹⁴⁴

In the ICJ proceedings instituted by the DRC, Uganda admitted to providing assistance to the RCD, which consisted of political advice initially, and limited military support and assistance in administrative and governance matters in the later stages of the conflict; but it sought to justify this support under the cloak of self-defence.¹⁴⁵ The MLC and RCD pronouncements quoted by Uganda were perfectly in tune with their statement of enjoying support, but one that was limited and in line with their self-defence requirement.¹⁴⁶ Interestingly, the DRC and Uganda also disagreed on the requisite degree of control for finding a relationship of control over the NSAG which conducted the armed attack so as to find the state responsible for it. Therein, the DRC took a stricter stance and set out four conditions for such a finding to be made, which included ‘substantial and active involvement in these forces’ activities’ by the controlling state. Disagreement was voiced in the background of the counter-discussion of whether, and the extent to which, the DRC was also found to be supporting armed groups operating against Uganda.¹⁴⁷

Both the belatedness and the framing of the admitted support provided to the rebels by the two intervening states showcase that both states were wary of the implications of this conduct on state responsibility for the use of force and unlawful interference in the internal matters of another state. As Longman comments, ‘Rwandan leaders apparently felt that they could not admit an extraterritorial intervention so clearly in violation of international law until they had prepared the international community to accept it’.¹⁴⁸

Finally, from the outset, the classification of the armed conflict as an international conflict by external actors was embedded in statements of UN officials, calling states to refrain from interference in the domestic affairs of other countries.¹⁴⁹ In fact, international opinion oscillated between findings of a relationship of control of Rwanda and Uganda over the rebels, and thus an occurrence of one IAC involving the DRC on the other side; and a parallel occurrence of an IAC and a NIAC.¹⁵⁰

With the evolving situation in the DRC, similar divisions among both scholars and practitioners were later present before international courts aiming to establish either individual criminal¹⁵¹ or state responsibility,¹⁵² particularly with regard to the events having taken place in Ituri, a northeast region of the DRC. The factual circumstances surrounding the events in Ituri were particularly difficult to decipher because of the parallel exercise of territorial control by the Union of Congolese Patriots (Union des Patriotes Congolais – UPC), which did receive external support from Rwanda, and on-the-ground direct occupation exercised by Uganda.¹⁵³ A corollary point showcased by the armed conflicts in the DRC and their various stages is the changing dynamics of armed conflicts both in terms of parties to the conflict and the relations among different actors. This convoluted and rapidly evolving picture of armed conflict further brings attention to the need to have clear-cut instructions at the time of the conflict as to which body of law applies in which circumstances.

3.4. Yemen (2004–ongoing)

The conflict in Yemen between the government and the Houthis was classified as non-international in scholarship.¹⁵⁴ Among other support relationships surrounding that conflict, the one subject to scrutiny was the nature of support provided to the Houthis by Iran. Analyses suggest that Iran's channelling of support to the Houthis was one of the three domains comprising the country's broader warfare strategy. Namely, reports elaborate on Iran providing the Houthis with weapons (such as missiles, unmanned aerial vehicles, rockets and air defence systems) and technology, training in military tactics (including the modification and further improvement of weapons systems),¹⁵⁵ as well as missile assembly and weapons use.¹⁵⁶

However, the support provided to the Houthis by Iran was not considered to surpass the threshold of control (regardless of the test) that would internationalise the conflict.¹⁵⁷ For instance, reports suggest that despite this military and financial support,¹⁵⁸ Iran had no leverage in the Houthis’ decision making,¹⁵⁹ and that its overall influence over the group was somewhat marginal.¹⁶⁰ Likewise, analyses show that the political and military support received by the Houthis from Tehran was not decisive in its successful takeover of Sanaa in September 2014, explaining the much greater leverage obtained through the group's alliance with the former President Saleh.¹⁶¹ Reports analysing the nature of this relationship also included a comparative analysis between Iran's behaviour towards the Houthis and Hezbollah, and concluded by stating that, while there was evidence of arms transfers¹⁶² as well as the provision of military advice, the relationship with the former was notably less close.¹⁶³ Although an overt strengthening of ties between Iran and the Houthis was witnessed after taking Sanaa, this cooperation was reflected in the arrangement of regular frequent flights between Sanaa and Tehran,¹⁶⁴ and humanitarian aid being sent to the region.¹⁶⁵ Numerous scholarly analyses indeed elaborate on the Houthis not being a mere Iranian proxy,¹⁶⁶ and the level of support not reaching that of overall control exercised over the group.¹⁶⁷

Thus, while the information at hand would indicate Iran's relationship with the group in terms of equipping, training, financing and providing operational support, the second limb of the overall control test¹⁶⁸ – the state's role in military activities – would seem to be missing. This conclusion can be deduced at least from Iran's lack of influence on Houthi decision making, including the group's disregard of advice given on the military operation in Sanaa.

The relevant actors involved were not as consolidated as the international reports on the subject of classification. Yemen ascribed Houthis’ actions to Iran.¹⁶⁹ In its pronouncements, Yemeni officials elaborated on the political backing of the group,¹⁷⁰ weapons and money flow carried out via sea, visits abroad, and through transfers. Additional sources suggested that military training was provided to Houthis in Iran and Lebanon.¹⁷¹ Apart from the United States, which justified its involvement in the conflict on the basis of Houthis being mere proxies of Iran,¹⁷² the Gulf Cooperation Council allies had described Iran's role in Houthi activities as significant.¹⁷³ Additionally, the United Kingdom also classified the conflict as of an international nature.¹⁷⁴ However, such pronouncements do not provide further insights into the reason behind this consideration, and do not suggest that this classification was made based on a relationship of control between Iran and the Houthis.¹⁷⁵ Similarly, claims made by Yemen and its allies were not followed by possible changes in the legal framework that would be applicable to captured Houthi members. While the law of IAC would not preclude, and rather would demand, prosecution for the commission of international crimes, the fact that the group's members were put on trial for such acts as attacks against military officers, planting of improvised explosive devices, and the launching of missile and drone strikes targeting military camps testifies to the lack of combatant immunity, and suggests rather that the domestic counter-terrorism framework was applied.¹⁷⁶ Were the Houthis deemed to be acting under Iran's control, their entitlement to POW status, pursuant to either AP I or GC III, would at least need to be assessed.

On the other hand, in spite of the sympathies expressed towards Iran,¹⁷⁷ the Houthis denied even those reports that suggested financial or material support by Iran over the entity, let alone control.¹⁷⁸ In a similar vein, Iran was seen to show support for the Houthis as the legitimate authority in Yemen,¹⁷⁹ and openly provided political support for the Houthis.¹⁸⁰ However, Iranian officials denied the allegations that the country was exercising control over local officials,¹⁸¹ as well as that it was providing them with training, money,¹⁸² or weapons.¹⁸³ It also denied Yemen's allegations with regard to weapons found on a ship seized by Yemeni authorities,¹⁸⁴ as well as dhows carrying heavy weapons, interdicted in February 2021.¹⁸⁵ Similarly, it denied having produced rifles and RPG-7 launchers found in another cargo, which the Panel of Experts on Yemen found to resemble the technical characteristics of weapons produced in Iran; and stated that the mere resemblance in appearance could not be conclusive as to the weapons’ origin.¹⁸⁶

4. Ways forward: Fostering humanitarian protection without altering the conflict classification

The previous sections of this article have showcased why classifying armed conflicts in the modern era has become increasingly difficult, and why there is a heavy burden of proof when it comes to establishing that a certain armed conflict amounts to an internationalised conflict. If, among others, the objective of raising the classification of the conflict to the international level is one of having a broader range of humanitarian norms that apply in IAC with NIAC, the fallacies of the ‘internationalisation’ method lead to the question of whether one can resort to other avenues for increasing compliance with IHL. Several mechanisms come to mind, and can be discerned according to the actor they target: the non-state actor, the supporting state, or the territorial state.¹⁸⁷

When it comes to the NSAG and the territorial state, one mechanism that was regularly resorted to by parties to the armed conflict was that of special agreements, envisaged explicitly in Common Article 3 of the Geneva Conventions, and the provisions of which also might include more protective norms stemming from the IHL of IAC.¹⁸⁸

Another option that could be explored is to abandon the application of the concept of internationalisation in cases of indirect intervention. Instead, focus could be placed on resorting to IHL provisions applicable to actors not parties to the NIAC. Common Article 1 of the Geneva Conventions is understood to presume a due diligence obligation of all contracting states to take all feasible measures to ensure compliance with IHL of parties to an armed conflict.¹⁸⁹ In cases of external support, the scope of feasible measures would be broader in comparison with other, ‘neutral’ states. On the other hand, the benefit of insisting on this more covert mechanism of supporting states influencing parties’ behaviour is the absence of political implications for the latter. Thus, for the intervening state, the narrative would be changed from that of control and participation in the conflict to that of positive influence and improving IHL compliance.

Additionally, a possible avenue might be precisely to not focus on clarifying the law but on strengthening the pathways to clarify the facts.¹⁹⁰ A caveat with this proposal lies in the lack of political will for such a step, which is reflected in the very rudimentary norms enshrining the possibility of entrusting fact-finding missions and enquiry bodies with unravelling and publishing the factual state of affairs, conditioning every step on the consent of the parties involved. Nonetheless, removing this politically charged issue from the parties onto a neutral body by perhaps broadening the powers of bodies such as the International Humanitarian Fact-Finding Commission (IHFFC), and allowing it to be more agile and immediate in reacting, would facilitate reducing the influence of the parties’ reasoning embodied in political and strategic considerations as well as having a comprehensive and objective understanding of the reality on the ground. In turn, certainty of the facts may, if through nothing more than through international pressure, reduce the actors’ margin of manoeuvre in backing up their claim that they are not involved in an indirectly internationalised armed conflict.

Alternatively, and considering the insistence of scholars on differentiating between questions of classification and responsibility, one might consider whether indeed issues relating to external participation in the conflict are better off being tackled before the courts and discussed in terms of state responsibility. Additionally, perhaps one should explore the reinforcing of accountability mechanisms and introducing analogous modalities on NSAG collective responsibility;¹⁹¹ it would be possible to separate questions of classification and responsibility completely, whereby a conflict could remain non-international, but the supporting state would be held accountable for its violation of jus ad bellum. In the long term, politicisation and disregard of the applicable body of IHL by always falling back on collective accountability of all actors could be prevented.

Finally, one should not forget that in times of armed conflict the applicability of IHL does not exclude that of other branches of international law. Where IHL on its own does not provide an adequate response to matters on the ground, IHRL can complement it and serve as a mechanism to advance a more protective interpretation of the legal framework. In the first line, IHRL obligations are most obvious for the territorial state. However, an important stream of human rights jurisprudence has demonstrated that control over an actor may also trigger the extraterritorial application of human rights treaties,¹⁹² and thus be employed to also seek the accountability of the third state. Finally, it is more established in the international arena that, at least when speaking of the NSAG exercising de facto authority over an area, such an entity also holds human rights obligations pursuant to customary international law.¹⁹³ Hence, perhaps even without insisting on applying the IHL of IAC, an increase in invoking IHRL and its enforcement mechanisms might foster reaching a more protective legal framework.

5. Conclusion

The internationalisation of armed conflicts through indirect intervention is a long-term tendency of contemporary armed conflicts, and it is here to stay. This means that the issue of internationalisation is not just theoretically interesting, but also in practice is very relevant for all IHL scholars and practitioners. We have addressed the issue of legal internationalisation, with our focus being on one subtype: internationalisation of armed conflicts through indirect intervention. We have established that too much ink has been spilled on the issue of how internationalisation occurs (the exact relationship between an intervening country and the NSAG on the territory of the state: tests of control) and almost none on the critical issue of the virtual non-application of this concept in practice.

We have also argued that this lack of implementation is not just a technical issue, but is a substantial issue closely connected to the strategic interests of both states and NSAGs. Namely, we have demonstrated that the implementation of the concept of internationalisation through indirect intervention would mean that intervening states would recognise their violation of other norms of international law and would jeopardise their strategic interests. Related to this, we have demonstrated that NSAGs, considering their close relationship with the intervening state, will also have a strong inclination not to confirm dependence on that state. Finally, even territorial states refuse to apply norms of IAC in the situation of indirect internationalisation as that could mean that members of the NSAG would become POWs if relevant criteria are fulfilled. All case studies confirm these conclusions, and we have focused on four of them in our detailed analysis (Donbas, Nagorno-Karabakh, the DRC, and Yemen).

At the same time, we have tried to respond to some of the possible critiques of our approach – from the argument that classification of armed conflicts is an objective process and that the position of involved states is not important, through to the argument that this situation is just one of those in which IHL is violated more often than not. Our central argument in this respect has been that, in short, the law does not work. This is not just an ordinary situation as we were practically unable to find the application of IAC rules in these situations of internationalisation through indirect intervention.

With these conclusions in mind and the fact that there are very slim chances that the behaviour of states and NSAGs in this respect will change, it is our firm belief that future efforts in this particular field of IHL should be aimed at finding proper practical solutions for overcoming shortcomings that we identified. We have sketched out some of the possible ways forward that could potentially mitigate the issues surrounding classification of conflicts involving indirect intervention.