Special agreements in non-international armed conflicts: Lessons from the practice

Abstract

The Geneva Conventions invite the parties to a non-international armed conflict (NIAC) to conclude special agreements (SAs) according to common Article 3(3). However, the lack of definition and insufficient coverage of such agreements by scholars and legal authorities increase the confusion between that mechanism and other agreements occurring in NIACs. This paper offers a contemporary and functional definition of SAs in order to better demonstrate their importance for NIAC regulation based on practice-informed lessons and, incidentally, to advocate for an increased use of those instruments. The data analyzed throughout this article reveal various lessons that will illustrate the potential of SAs. Among the teachings revealed by practice is the expanded material scope of these instruments or the possible ways in which the parties can choose to draft their commitments. In addition, the records of practice reveal how SAs concretely improve the regime of NIACs, and the implementation approaches generally adopted. For a more complete study of the dynamics of SAs, the paper equally draws attention to their formal and personal characteristics. All these features turn SAs into powerful international humanitarian law (IHL) tools that are flexible enough to upgrade the applicable law of NIACs while respecting some necessary boundaries in order to guarantee the validity of the obligations embedded. This is an essential balance for IHL and international legal order consistency.

Introduction

Special or ad hoc ¹ agreements (SAs) rank among the less documented and analyzed international humanitarian law (IHL) resources in jurisdictional and doctrinal releases compared to other instruments governing the regime of non-international armed conflicts (NIACs).² At least four reasons explain this fact. Firstly, the imprecisions of Article 3(3)³ common to the four Geneva Conventions (common Article 3(3), CA 3(3)) on the features of SAs make them hard to identify among the mass of agreements arising during or after a NIAC. CA 3(3) does not give many hints on SAs’ definition when it states that “[t]he Parties to the conflict should further endeavour to bring into force, by means of special agreements, all or part of the other provisions of the present Convention”. The early Commentaries on CA 3 are not of great help either.⁴ The 2020 Commentary on Geneva Convention III (GC III) by the International Committee of the Red Cross (ICRC), for its part, underlines that SAs can restate applicable law (“declaratory”)⁵ or go beyond the Geneva Conventions (“constitutive”)⁶ to cover means of warfare and the conduct of hostilities⁷ or even international human rights obligations when they “help to implement humanitarian law”.⁸

However, these characteristics have little or no justification. For example, the fact that CA 3 stipulates that the agreement should bring about “all or part of the other provisions”⁹ of the Geneva Conventions implies that SAs are not of a “declaratory” nature and do not restate applicable law. In the same vein, it is highly debatable that in a peace or ceasefire agreement, even “relevant provisions … may constitute special agreements under common Article 3”, as claimed by the 2020 Commentary.¹⁰ This claim challenges the ordinary meaning of the word “agreement” by assimilating it to specific provisions. Moreover, the Commentary suggests that a “peace agreement, ceasefire or other accord” can amount to an SA.¹¹ Of course, since the drafters of the Geneva Conventions did not provide for formal characteristics of SAs, nothing precludes that conclusion. Nevertheless, more caution must be taken to ascertain the qualification of an instrument as an ad hoc agreement.

Secondly, the doctrine struggles to capture the essence of SAs because discussions focus on NIAC agreements generally, and rarely on ad hoc agreements as a specific category. Accordingly, legal scholarship often dilutes those particular instruments within ordinary agreements; even direct technical questions (functioning,¹² legal status,¹³ governing law¹⁴ or implementation¹⁵) lack identification methodology or definition for most of them.¹⁶ A third reason – explaining the preceding one – comes from the rare use of SAs in NIACs, as acknowledged by the ICRC,¹⁷ partially because of States’ reluctance to legitimize organized armed groups (OAGs) indirectly through the conclusion of SAs.¹⁸ Indeed, in its 2020 Commentary, the ICRC refers to a dozen instruments that it considers SAs.¹⁹ Even though the ICRC did not claim exhaustivity in its illustrations,²⁰ it was logical to expect more examples from the Commentary on CA 3, especially considering the seventy-one years that the 1949 Geneva Conventions had been in existence at that time.

Moreover, to confirm the scarcity of SAs, the parties to NIACs rarely label their commitments explicitly as a “special agreement” under the meaning of CA 3. Two notable exceptions lie in Agreement No. 1 on the Application of IHL between the Parties to the Conflict in Bosnia and Herzegovina (Agreement No. 1), signed on 22 May 1992,²¹ and the Final Agreement to End the Armed Conflict and Build a Stable and Lasting Peace between the Government of Colombia and the Revolutionary Armed Forces of Colombia – People’s Army (Fuerzas Armadas Revolucionarias de Colombia – Ejército del Pueblo, FARC-EP), of 24 November 2016.²²

This does not mean that an agreement needs the explicit title of “ad hoc” or “special agreement” to reach that status; CA 3(3) does not make it a validity condition. The fact that SAs can bear any label upon which the signatories decide adds to the confusion with other or ordinary NIAC agreements. Still, from what precedes, SAs definitely stand out in a certain way that legal doctrine fails to capture – otherwise, CA 3 would not have provided for such a mechanism. Under the appellation of “NIAC agreements” stand any written (or verbal) understandings occurring during or relating to the armed conflict. Such understandings may not even be between the parties to the conflict, since the latter can conclude agreements with third parties and make those agreements address any matter of concern for the involved actors. In addition, the commitments are not necessarily legally binding, and if they are, only a case-by-case approach can indicate the legal order they belong to.

Hence, without a proper identification method, many NIAC agreements could mistakenly be regarded as SAs. Indeed, the latter have binding force under international law and stand above national legislation as international agreements,²³ but with less implementation burden than treaties in terms of prior domestic reception (dualism). Thus, to determine the importance of these instruments for IHL, a definition is capital. The characteristics of an SA flow from its function: CA 3 shows that an SA’s objective is to enhance the law applicable to the NIAC by including “other” provisions from the concerned Geneva Convention. However, considering the evolution of IHL, that functional approach leads to a broader scope of obligations. This is discussed in the first substantive section of the article below.

Accordingly, the practice resulting from the above definition will help us to decrypt and understand the lessons that such agreements provide for IHL. The main argument is to demonstrate why SAs represent suitable tools to regulate NIACs and deserve more attention from battlefield actors and beyond. The SAs’ legal contribution lies in their enriching nature, their dynamics and their flexibility, which this article will address at three levels. Firstly, in terms of content or substance, they extend the scope of the law of armed conflict (LOAC) as they bring about other regimes of international obligations that share the protective aim of the Geneva Conventions regarding armed conflicts – that is, international human rights and refugee law. This is examined in the second section. To go further, the study of the drafting practices informs on various approaches for streaming IHL obligations into the instrument (third section); moreover, to confirm the functional value of SAs, it is capital to explain how they improve the law applicable to NIACs as required by the provision under analysis (fourth section). Secondly, in terms of implementation, the silence of CA 3 leaves an open choice to the parties (fifth section). Thirdly, the formal and personal features of SAs confirm their openness and accessibility (sixth section).

While these lessons will highlight the potential of SAs, the paper will not evaluate their effective execution. Each conflict narrative is specific and too variable to capture a global trend, even if low compliance with SAs is an acknowledged weakness.²⁴ The study does not claim exhaustivity either, as the data used for illustration are only a glimpse of a potentially wider practice yet to be uncovered. Nevertheless, in lieu of an analysis of their effectiveness, an account of the specific boundaries applicable to SAs will complete the study. Indeed, as discussed in the seventh section, the validity of SAs depends on imperative and necessary limits to safeguard the essence of IHL and the coherence of the international legal order.

What characterizes a special agreement?

In its Commentary on CA 3 with regard to SAs, the ICRC gives some illustrations and identification clues.²⁵ However, the ICRC does not justify the characteristics that it raises, and their combination does not constitute a comprehensive and consistent method for distinguishing SAs from ordinary agreements between (or with) the parties to a NIAC. The distinction matters for two reasons. The first is treaty clarification: since the features of SAs do not directly appear from a reading of CA 3, careful interpretation will help to identify such instruments and translate the meaning of that provision into recognizable agreements whose drafting composition could inspire actors involved in other existing or future NIACs.

The second reason lies in the legal status and effects of SAs. Not all NIAC agreements are legal instruments anchored on the international plane;²⁶ in contrast, SAs are legally binding because “each Party to the conflict shall be bound to apply, as a minimum”, the rules laid down in CA 3(1) regardless of the fact that SAs “bring into force” further provisions from the Geneva Conventions. Hence, such additional obligations will compel the parties all the more. Moreover, unlike SAs, not every peace or ceasefire agreement is binding under international law,²⁷ nor does it necessarily reflect relevant rules of the Vienna Convention on the Law of Treaties, the quality of “State” notwithstanding.²⁸

Bearing these elements in mind, the letter of CA 3, as well as the objects and purposes of the Geneva Conventions,²⁹ authorizes some deductions to reach a coherent definition and identification method. A functional interpretation of CA 3 discloses the expected content of SAs. Primarily, they are agreements relating to issues governed by the Conventions because the aim of SAs is “to bring into force all or part of the other” Conventions’ provisions. Secondly, the IHL obligations they channel must have an exclusive humanitarian finality, which means that they must not serve to achieve another result or have a conditional construction. This preserves the non-reciprocal nature that these obligations enjoy under customary law and guarantees a necessary functional autonomy. Yet, IHL has evolved beyond the framework of the 1949 Conventions, implying a contemporary reading of CA 3 following the functional approach previously introduced. Thus, SAs can also serve to include other sources of IHL.

Humanitarian matters

CA 3 discloses the function of SAs. Firstly, they are instruments expected to improve the IHL regime of NIACs through “other provisions” of the Geneva Conventions – that is, those obligations which would not apply to that type of armed conflict otherwise and which will thereby enhance the level or the nature of rights and duties between the parties. Secondly, the expression “other provisions” means that SAs must deal with issues handled by the Geneva Conventions – in short, humanitarian concerns. A contextual fact supports this deduction. Since the Geneva Conventions were the only instruments drafted during the 1949 Diplomatic Conference,³⁰ State delegates necessarily envisioned SAs to have a material connection only with the treaties whose provisions would be “copied”.

Accordingly, the function and content of SAs imply that they are NIAC agreements between the parties to the conflict and pertain to one or multiple humanitarian matters. Any question regarding the amelioration of the condition of the wounded, sick or shipwrecked in the field or at sea, the condition of persons detained because of armed conflict, or the protection of civilians in time of armed conflict amounts to a humanitarian concern. These issues must be distinguished from other questions appearing in agreements occurring in NIACs, such as military (e.g., demobilization and reintegration matters), political (e.g., power-sharing, political participation, constitutional organization) or economic issues (e.g., wealth-sharing, farming or rural policies). Of course, many agreements mix those issues with humanitarian concerns, but only agreements where the latter are predominant or exclusive should be considered. Otherwise, the agreement would depart from the legitimate expectation of the Geneva Conventions’ drafters that a “material connection” should exist between SAs and the objects of the Conventions. Thus, CA 3 instils a requirement that sets an additional criterion for SAs.

The purpose of the obligations

IHL obligations must be non-reciprocal and keep their functional autonomy from any other purpose. Non-reciprocity can be inferred from the relevant expressions of CA 3. When the provision invites the parties to “bring into force” the other provisions of the Geneva Conventions, it suggests that NIAC actors have to make them effective or implement them. However, the text of CA 3(1) adds an implicit condition on how to make obligations applicable: “[p]ersons taking no active part in the hostilities, including members of armed forces who have laid down their arms and those placed hors de combat …, shall in all circumstances be treated humanely” (emphasis added). The phrase “in all circumstances” calls for unreserved respect of IHL. In other words, no condition should hinder compliance with the Geneva Conventions, be it by requesting a treatment similar to the right conceded to the other party or by imposing a specific deed in return.³¹

IHL obligations must also be autonomous in that they must not be subject to or pursue a parallel or competing finality or result. Instead, they must aim for their own implementation only. Otherwise, it would create confusion about the motives of their execution by the parties to the NIAC, and there would be no guarantee that those parties would comply in order to fulfil their IHL obligations exclusively. When drafting the obligation, it is important to secure its unconditional nature and functional autonomy.

Indeed, a party could escape the IHL obligation under the pretext that the condition or the result associated with the commitment is not met or achieved. While NIAC actors must respect the LOAC regardless, in practice, framing the obligations of that body of law to reach other goals or submitting them to conditions that entail reciprocity can have the opposite effect. Instead, humanitarian obligations drafted without any link to an expected outcome or a given requirement guarantee an unequivocal humanitarian purpose.

The importance of an IHL-compliant formulation finds justification, again, in the drafting context of CA 3. Since the States expected material connection with the Geneva Conventions and thus IHL, it follows that SAs must contain obligations reflecting that body of law along with the appropriate formulation. Indeed, many NIAC agreements cannot qualify as SAs because of misleading formulations of IHL pledges. This is the case, for instance, with the Agreement on a Temporary Cease-Fire and the Cessation of Other Hostile Acts on the Tajik–Afghan Border and within the Country for the Duration of the Talks.³² In Article 2(b)–(c), the parties defined “cessation of hostilities” as encompassing “[t]he cessation by the Parties of acts of terrorism and sabotage” and “[t]he prevention by the Parties of murders, the taking of hostages, unlawful arrest and detention, and acts of pillage against the civilian population and servicemen in the Republic and other countries”. Although the prohibited behaviours are also forbidden under IHL,³³ the “cessation of hostilities” depends on those conducts, as the provision title indirectly suggests. Therefore, the conflict could erupt again if a NIAC actor infringes on the listed behaviours.

In this configuration, the cessation of hostilities is the broad objective, and IHL undertakings are instrumental. The drafting approach adopted in the above example creates two problems. Firstly, while a breach of IHL by a party normally has no consequences as per the other participant’s obligations, here, violating the listed conducts could provoke a return to the fighting, following the terms of the agreement. In this configuration, non-compliance with IHL entails armed reactions. Secondly, it is not the role of IHL to end (or contribute to ending) hostilities; rather, it is the role of IHL to regulate them, even if, indirectly, IHL contributes to peace.³⁴

The practice of NIAC agreements reveals more cases where the parties have combined IHL undertakings with an expected result or conditioned their performance.³⁵ Nevertheless, when IHL obligations operate to maintain the validity of a “ceasefire”, to be instrumental for the “cessation of hostilities” or to safeguard the agreement itself, it creates an ambiguity that contrasts with the implicit requirement that they must be executed for themselves – that is, without connection to any other requirement. NIAC agreements displaying such drafting approaches should not be regarded as SAs. Thus, to secure non-reciprocity and the autonomy of IHL obligations, and as the second definition criterion, only NIAC agreements whose provisions have mainly humanitarian aims may qualify as ad hoc or special agreements. In conclusion, to simplify the definition, an SA is an agreement addressing humanitarian concerns with a humanitarian purpose.

However, a contemporary reading of CA 3 needs to update this functional view. The material scope of the provision can no longer limit itself to the Geneva Conventions due to the evolution of the IHL regime of NIACs after 1949.³⁶ In coherence with the Conventions’ purpose of protecting the human life and dignity of those not or no longer taking part in hostilities, it is safe to consider other IHL sources as material for SAs because they have the same aim as those four treaties.

Updated material scope of special agreements

There is no need for a detailed examination of the number of provisions of the Geneva Conventions that SAs can bring into force. Instead, it is interesting to note that the Conventions contain model ad hoc agreements for international armed conflicts (IACs) which NIAC actors can copy and adapt to their situation, and that there are other sources of IHL which provide a wide range of possibilities for enhancing the regime of NIACs.

Special agreements embedded in the Geneva Conventions

Some provisions of the Geneva Conventions provide a protective regime or status for areas, zones or facilities to offer shelter and safeguard human dignity. Such “protected zones” were already a concern before the Conventions,³⁷ and the drafters deemed it necessary to secure them because they would contribute to shielding the individuals listed in CA 3(1)³⁸ from the effects of hostilities. Three kinds of protected zones may serve as examples: on the one hand there are “hospital zones and localities”, eventually combined with “safety zones and localities”,³⁹ while on the other hand there is the “neutralized zones” regime.⁴⁰

“Hospital zones and localities” protect not only the wounded and sick from the effects of hostilities⁴¹ but also the “personnel entrusted with the organisation and administration of such zones and localities”.⁴² If referring strictly to the Geneva Conventions, this is an IAC-based regime,⁴³ and there is a Draft Agreement Relating to Hospital Zones and Localities as Annex I of Geneva Convention I (GC I), pursuant to Article 23.⁴⁴ However, the ICRC implemented that regime in relation to a NIAC through the Agreement on the Jaffna Hospital and Safety Zone around It.⁴⁵ The government of Sri Lanka and the Liberation Tigers of Tamil Eelam (LTTE) adopted that SA on 6 November 1990.

“Safety zones and localities” are areas that benefit from the same protection as the previous zones defined above but with a greater focus on vulnerable groups among the civilians, like the infirm, the elderly, children under 15, expectant mothers, and mothers of children under 7.⁴⁶ Furthermore, practice reveals that both protected zones can merge to form “hospital and safety zones and localities”,⁴⁷ whose regime is explained in Geneva Convention IV (GC IV), with a draft agreement as Annex I (consistent with Article 14).⁴⁸ Taken separately, “hospital zones and localities” and “safety zones and localities” are protected zones usually set in safe and peaceful areas far from the battlefield before or after the hostilities, and they may last as long as need be.⁴⁹ Such zones should preferably be created through written agreements despite the silence of Articles 23 and 14 of GC I and GC IV respectively.⁵⁰

As regards “neutralized zones”, they create a sheltered area within the hostilities perimeter and even during the fighting. They have the same protective objective as the previous zones for the wounded, sick and peaceful civilians.⁵¹ However, there is no “neutralized zones” template agreement in the Geneva Conventions, through a framework is laid down in Article 15 (“Neutralized Zones”) of GC IV. As an example of the implementation of that regime, the Yugoslav federal army and Croatia once granted the status of “neutralized zone” to the Vukovar hospital in an agreement witnessed by the ICRC, the Maltese Red Cross and Médécins Sans Frontières on 18 November 1991.⁵² On that occasion, the parties stated that they would “recognise the neutrality of Vukovar hospital during a period covering the evacuation. The hospital will be put under the protection of the ICRC who will advise both parties of the period of neutrality which they require.”⁵³

Furthermore, data records evidence that a combination of the “hospital zones and localities” and “neutralized zones” regimes is possible in NIACs. This occurred in Croatia, with the Agreement Relating to the Establishment of a Protected Zone around the Hospital of Osijek.⁵⁴ The signatories to this agreement regarded “the hospital of Osijek and a zone around it as drawn on the annexed map” as “a protected zone according to the principles of Article 23 of the First Geneva Convention of 1949 and of Articles 14 and 15 of the Fourth Geneva Convention”.⁵⁵ In that sense, they adapted the annexed model agreement of GC I/IV to their needs. This combination shows how the practice of SAs can witness configurations beyond the framework that the Geneva Conventions sought to anticipate with their templates.

Other sources of IHL

The updated interpretation of the invitation to bring into force “all or part of the provisions of the present Convention” leads to the understanding that SAs can embrace all IHL customary rules and treaties.⁵⁶ Concretely, for the ICRC, there are three supplementary sources that an SA can enforce.⁵⁷ The first are customary norms,⁵⁸ which comprise those exclusively applicable to IACs.⁵⁹ The second are provisions from Additional Protocol I (AP I), as seen in some SAs of the 1990s Yugoslav wars, notably a Memorandum of Understanding (MoU) in 1991,⁶⁰ Agreement No. 1,⁶¹ and the separate but identical undertakings of Bosnia and Herzegovina (BiH), the Croatian entity of Herzeg-Bosna and the Serbian Democratic Party (Srpska Demokratska Stranka, SDS) in the London Conference’s Programme of Action on Humanitarian Issues of 27 August 1992.⁶² Thirdly, the parties can resort to customary principles⁶³ and treaties⁶⁴ of Hague law,⁶⁵ as in the 2016 Humanitarian Appeal for Benghazi,⁶⁶ among other cases.⁶⁷

To summarize, going beyond “the present Convention”, CA 3(3) can be applied by analogy to non-Geneva Conventions sources of IHL. Hence, an SA pertains to any humanitarian concern regulated by the LOAC, assuming it keeps a humanitarian purpose. This means that the key determinants of an SA are the content and purpose of the instrument. Based on that definition, many agreements concluded during or because of a NIAC qualify as SAs, such as in BiH,⁶⁸ Colombia,⁶⁹ Croatia,⁷⁰ Georgia,⁷¹ Indonesia,⁷² Libya,⁷³ the Philippines,⁷⁴ Somalia,⁷⁵ South Sudan,⁷⁶ Sri Lanka,⁷⁷ Sudan⁷⁸ or Syria.⁷⁹ These will serve as examples⁸⁰ to underline the potential of SAs and the lessons they can deliver, among which an extension of the material scope of SAs besides IHL.

Complementary bodies of law

Practice shows that the function of SAs attracts other regimes of obligations that share the IHL objective of protecting human life and dignity from threats that occur during or because of a NIAC. Indeed, some legal regimes can enhance or develop the protection standards of IHL in NIAC – namely, international human rights law (IHRL)⁸¹ and international refugee law, including the treatment of internally displaced persons (IDPs). To simplify, the undertakings originating from these regimes, including IHL, can be regarded as humanitarian obligations with an undeniable legal nature, although they have specific fields of application.

Human rights obligations

In the context of NIAC, human rights obligations could incidentally contribute to implementing IHL if contained in an SA. The combination of IHRL with IHL finds justification on many grounds. Firstly, they both cover vulnerable groups: children,⁸² girls and women,⁸³ and people with disabilities,⁸⁴ to name a few. In the same vein, issues of common interest necessarily create connexions between both regimes,⁸⁵ like the problem of child soldiers.⁸⁶ Many SAs prohibit that practice, sometimes through reference to human rights instruments. In Sudan, an SA concluded by the government and two OAGs condemned the practice, invoking the African Charter on the Rights and Welfare of Children, the Convention on the Rights of the Child (CRC) and the Optional Protocol to the CRC on the Involvement of Children in Armed Conflict.⁸⁷ A joint ceasefire agreement between the Sudanese government and the Liberation and Justice Movement (LJM) contains a similar restriction,⁸⁸ while another ceasefire agreement with the Justice and Equality Movement (JEM) had a more explicit formulation.⁸⁹ Cases of prohibition of the recruitment of child soldiers also extend to the NIAC in South Sudan.⁹⁰

A third reason favouring the complementarity of IHL and IHRL is the protective standard developed by the latter regime regarding torture,⁹¹ enforced disappearances,⁹² detention⁹³ and unfair judicial proceedings.⁹⁴ Some agreements have presented a detailed distinction between IHRL issues and IHL matters, like the Comprehensive Agreement on Respect for Human Rights and International Humanitarian Law (CARHRIHL) between the government of the Philippines and the National Democratic Front of the Philippines (NDFP).⁹⁵ However, from a more general perspective, a few agreements address both regimes of obligations as a combined or indistinctive package. For example, in the Implementing Guidelines on the Humanitarian Rehabilitation and Development Aspects of the Tripoli Agreement on Peace of 2001, the government of the Philippines and the Moro Islamic Liberation Front (MILF) committed themselves to “safeguard[ing] the observance of the international humanitarian laws, respect for internationally recognised human rights and fundamental freedoms for all persons within Mindanao”.⁹⁶ Accordingly, the parties decided to “cooperate in the investigation and prosecution of serious violations of international humanitarian laws and human rights as well as violations of this Agreement”.⁹⁷

Data analysis also reveals particularities in using IHRL in SAs, like acknowledging rebels’ commitment to respecting and protecting human rights. In the Humanitarian Appeal for Benghazi, for example, OAG groups’ leaders in Libya claimed that “[w]ithin the limits of [their] power and influence, [they] commit and call upon all others to … respect IHL and Human Rights”.⁹⁸ The government of the Republic of South Sudan and an OAG went further by considering that the rebels could be on a par with the State in protecting human rights,⁹⁹ whereas the latter is usually the duty bearer.¹⁰⁰ Despite the singularity of that last example, practice shows IHRL’s contribution to enriching the scope of SAs, as confirmed by some scholars.¹⁰¹

International refugee law

Although IHL does not seem to address the condition and treatment of refugees and IDPs directly,¹⁰² several sources prove its interest in the fate of those categories of individuals.¹⁰³ In its Commentary on CA 3(2), the ICRC recognizes that the “protection” mentioned in that provision must (also) cohere with refugee law.¹⁰⁴ Equally, Article 17 of Additional Protocol II (AP II) defines the restrictive conditions under which a population displacement can occur. Finally, the ICRC Customary Law Study also covers those questions under Rules 129 and 131–133.

Considering all these connections, it is unsurprising that SAs serve to protect refugees and IDPs as well, sometimes to the point that those agreements seem to cover the material scope of international refugee law in some respect. The parties to a NIAC can address the situation of refugees and IDPs as a whole, notably in the wake of the conflict. This was the case in the NIAC within BiH, with the London Conference Programme of Action on Humanitarian Issues stipulating that “refugees and displaced persons should be allowed to return voluntarily and safely to their places of origin”.¹⁰⁵ Likewise, consider the Protocol on the Improvement of the Humanitarian Situation in Darfur between the government of Sudan, the Sudan Liberation Movement/Army (SLM/A) and the JEM,¹⁰⁶ among other examples.¹⁰⁷

In a few situations, an SA can address the fate of IDPs within specific provisions. For instance, in Agreement No. 1, Article 2(3)(4)¹⁰⁸ is very close to the text of Article 17 of AP II. SAs involving the government of Sudan further illustrate this trend, whether with the LJM¹⁰⁹ or the JEM.¹¹⁰ Of course, nothing precludes an SA from dealing entirely with refugees and IDPs, and the Quadripartite Agreement on Voluntary Return of Refugees and Displaced Persons signed by Georgia and the Abkhaz side on 4 April 1994 is an excellent example in that respect.¹¹¹

The additional obligation regimes illustrated so far increase the importance of SAs for IHL. Indeed, such agreements enhance the applicable law of NIACs and embrace other dimensions of human dignity in armed conflicts. Accordingly, they surpass the expectations of CA 3 as initially designed by the drafters of the Geneva Conventions. This first lesson justifies the consideration of SAs as humanitarian agreements based on their content, though this general label does not mean that they all mobilize humanitarian obligations similarly.

Approaches for streaming IHL obligations into special agreements

SAs help regulate NIACs by using the Geneva Conventions as their first resource, which the parties may refer to differently. The parties can expressly list the relevant provisions that they seek to apply in their SA, or only make a general reference to the treaties. The Agreement on the Release and Transfer of Prisoners during the Yugoslav wars is a good example of the former practice:¹¹² this SA cites Articles 50, 51, 130 and 147 of GC I, II, III and IV respectively to define the conditions for releasing prisoners in Article 3 of the agreement.¹¹³ As regards the approach of general reference to the Geneva Conventions, the aforementioned MoU of 1991 provides a clear illustration.¹¹⁴ The signatories agreed that “[a]ll wounded and sick on land shall be treated in accordance with the provisions of [GC I]”, that “[a]ll wounded, sick and shipwrecked at sea shall be treated in accordance with the provisions of [GC II]”, and that “[c]aptured combatants shall enjoy the treatment provided for by [GC III]”, pursuant to Articles 1–3 of the MoU respectively.¹¹⁵ Of course, the practice also applies to other IHL treaties like AP I.¹¹⁶

Furthermore, from an extensive interpretation of CA 3(3), it is not compulsory for the parties to expressly link the obligation to an IHL treaty; silent or unreferenced matches with treaties and customary rules are commonplace. However, even within this perspective, there are variations. To begin with, the SA can effectively match an initial treaty obligation. In the SA between the Republic of the Philippines and the MILF introduced previously,¹¹⁷ the parties adopted a definition of “mine”¹¹⁸ which espouses the formulation used in Protocol II to the Convention on Certain Conventional Weapons (CCW), the Protocol on Prohibitions or Restrictions on the Use of Mines, Booby-Traps and Other Devices.¹¹⁹ Equally, the parties conferred the same meanings to “explosive remnants of war” and “abandoned explosive ordnance” (AXO) as are used in Protocol V to the CCW.¹²⁰

Alternatively, the SA can indirectly match IHL treaty obligations when their substance is the same despite a different formulation. For example, an equivalent rephrasing of CA 3(2) and Article 18 of AP II in the agreement could consist of committing “to respect the rules of operation of the [ICRC] and the United Nations High Commissioner for Refugees and the organization of the delivery and distribution of humanitarian aid, [and] to facilitate the delivery of such assistance”.¹²¹ Likewise, as the Geneva Conventions’ common Article 1 commands parties “to respect and to ensure respect for the present Convention in all circumstances”, the actors of a NIAC could choose to reformulate that obligation by agreeing “to undertake all necessary measures to prevent any violations of international humanitarian law”.¹²²

Finally, an updated reading of CA 3(3) suggests that SAs’ provisions can also validly reflect customary rules, even if some have a treaty nature in parallel, especially through AP II. An agreement on cessation of hostilities between the government of South Sudan and the Sudan People’s Liberation Movement/Army (SPLM/A)¹²³ gives some examples of this practice. In Article 3, paragraphs 1 to 4 match various customary rules. For example, the second paragraph forbids “acts of rape, sexual abuse and torture”, which is a combination of the rules on the prohibition of rape and other forms of sexual violence,¹²⁴ but also a reference to the ban on torture and cruel, inhuman or degrading treatment.¹²⁵

It is not rare to mention customary principles regarding the conduct of hostilities, such as distinction, precaution and proportionality. In Article 3(1) of the same South Sudan SA, the signatories agreed to prohibit “attacks on the civilian population”, reflecting the principle of distinction.¹²⁶ To refer to the principles of proportionality and precautions in attack, the participants of the 2016 Humanitarian Appeal for Benghazi pledged, “[w]ithin the limits of [their] power and influence, [to] commit and call upon all others to … avoid collateral damage induced by the war and fighting”.¹²⁷

The lessons from these different modalities of incorporating humanitarian obligations demonstrate the freedom of the parties to draft their agreement. These teachings will ease our understanding of the functional value of SAs regarding regime enhancement.

How do special agreements improve the law of NIACs?

SAs compensate for the imbalance of IHL regulation for NIACs¹²⁸ compared to IACs. CA 3(3) does not explicitly impose that all of the provisions of an SA go beyond the applicable law in order for the instrument to become “constitutive”¹²⁹ of IHL; thus, a few obligations or even a single obligation may suffice to that effect. There are two approaches to regime enhancement disclosed by practice, each based on a specific interpretation of the CA 3 expression “the present Convention”. Firstly, the parties may insert obligations from treaty or customary law that do not apply to the NIAC (yet). The most straightforward interpretation of the CA 3 wording “the present Convention” favours that meaning. Nevertheless, since treaty and customary IHL have evolved after the 1949 Geneva Conventions, the expression could be equivalent to “any non-Geneva Conventions source of IHL”. This is an external upgrade because the obligations brought into force originate outside of CA 3.

However, SAs show that the parties can equally reinforce the existing IHL regime for NIACs. In that case, the improvement is internal. In this interpretation, “the present Convention” obliges parties to consider “any IHL obligation more favourable or protective than CA 3”. For a better understanding of both interpretations, the practice offers different examples.

The external regime upgrade with new obligations generally describes two situations. The first case is where the parties bring the IAC treaty or customary law into force. The SA achieves its function if the signatories invoke those sources or their relevant obligations – if, for example, the SA indicates that the signatories will address “‘graves breaches’ as defined in the 1949 Geneva Conventions”,¹³⁰ it becomes clear that they welcome a concept usually invoked in IAC instruments.¹³¹ It indicates that the signatories have decided to embrace the four treaties, some of them, or other conventions that apply to IACs exclusively.¹³²

A second option of external regime improvement is when the agreement bears obligations from conventions fitted for NIACs but which the State has not (yet) ratified, like AP II or a treaty of Hague law that is also applicable to NIACs. In the earlier example of an SA about the ban on landmines in the Philippines,¹³³ the parties addressed mines and UXO. However, the State did not ratify Protocol V of the CCW, an international instrument on UXO and AXO that also applies to NIACs.¹³⁴

Following the approach of internal enhancement, the SA can go beyond the applicable law if it offers more protective standards than the NIAC regime without necessarily copying the rules applicable to IACs. It could consist of a tighter interpretation of an existing concept to strengthen the obligation. For example, “incidental loss” (“collateral damage”) in IHL is acceptable only when the attack is not excessive in relation to the concrete and direct military advantage anticipated.¹³⁵ In contrast, if the SA states that the parties “commit themselves to … avoid[ing] acts that would cause collateral damage to civilians”,¹³⁶ it leaves no exceptions for incidental losses. This formulation creates a higher standard than the classical LOAC. Another way to strengthen applicable law is by providing more elaborated definitions of some notions than treaty law. Consider, for example, the definition of “explosive ordnances” in the 2010 SA between the Philippines and the MILF, which is more extensive than its description in Protocol V of the CCW.¹³⁷

A last trend appears halfway between the two approaches detailed above – namely, the creation of new (legal) concepts by the SA. A good example of this is the Philippines–MILF agreement mentioned above; this SA defines “victim-activated mines”, “humanitarian mine action” and “joint mine/UXO clearance”,¹³⁸ but classical IHL does not acknowledge these notions.¹³⁹ As a final remark, the above improvement scenarios can of course be combined in the same SA. For example, in their Recommitment to Humanitarian Matters, the government of South Sudan and the SPLM/A (in opposition) pledged that they would “[t]ake all possible measures to respect human rights and protect the civilian population from indiscriminate attacks, rape or any other form of abuse”.¹⁴⁰ Regarding regime improvements, firstly, the parties to this SA refer to the legal corpus of human rights, which does not belong to the regime of NIACs. Secondly, the signatories commit indistinctly to human rights, while OAGs are not the traditional addressees of that body of law. As for classical IHL obligations, the above quotation reflects the principles of precaution and distinction, and compliance with the spirit of CA 3(1) generally.

While the divide between “external” and “internal” helps to structure the above analysis, what really matters in practice is the enhancement of the NIAC regime as illustrated. These scenarios shed light on the role of SAs as vectors of (new) humanitarian obligations. This background understanding allows us to examine SAs’ functioning more deeply, starting with an explanation of the approaches that the signatories can choose to bring the obligations of their SA into force.

Implementation patterns

CA 3 does not specify how the parties to an SA must perform their undertakings. Practice shows different types of implementation, the most simple form remaining direct enactment by the signatories of the SA. However, indirect or remote execution can also occur when the parties create and empower an entity or mechanism to achieve some humanitarian obligations on their behalf. In either case, nothing precludes the invitation of a third party to assist them in the same task by granting rights or imposing duties on the new actor. The ICRC is an interesting example in this regard.

Direct implementation

A plain reading of the CA 3(3) wording “[t]he parties to the conflict” makes it logical to consider those actors as the principal addressees of the invitation to form an SA. In that sense, they can draft their agreement to perform humanitarian obligations themselves, with no or very few intermediaries. Practice discloses two combinatory configurations in this respect. The first and most simple is when the State and an OAG commit themselves to achieving certain deeds. For example, they can ensure the “[p]rotection of humanitarian workers”,¹⁴¹ guarantee humanitarian access¹⁴² or pledge that “[t]he wounded and the sick shall be collected and cared for”.¹⁴³ Simultaneously, they can also abstain from adopting certain behaviours, like acts of violence against civilians,¹⁴⁴ including IDPs and refugees,¹⁴⁵ or civilian objects – that is, their belongings or goods, civilian facilities and the resources necessary for the survival of those categories of protected persons.¹⁴⁶

A second and more complex composition is where many actors come into play. For example, in the context of the South Sudan NIAC, three groups of actors formed the signatories of the 2017 Agreement on Cessation of Hostilities, Protection of Civilians and Humanitarian Access. First, there was the Transitional Government of National Unity (TGoNU), then many armed opposition groups like the SPLM/A (in opposition), the National Salvation Front, the National Democratic Movement and the Federal Democratic Party/South Sudan Armed Forces, among others.¹⁴⁷ Finally, political parties participated – notably, as listed in the agreement, the “SPLM Leaders – Former Detainees (SPLM-FDs), Civil Society, Umbrella Party and Faith Based leaders in the High Level Revitalization Forum”. Regardless of their differences, the parties exchanged humanitarian obligations jointly (“Each Party shall”,¹⁴⁸ “the Parties shall”¹⁴⁹), with no intermediary; however, records show that this is not always the case.

Entities or mechanisms set up by the agreement

As an alternative to a direct exchange of rights and obligations in their SA, the parties to the conflict can resort to an intermediary to achieve (some of) their obligations on their behalf. Following this approach, the signatories generally create a joint body or mechanism to handle their commitments. In this framework, the object of the agreement is exclusively the setting up of such an entity and the assignment of its humanitarian functions by the parties to the NIAC. It is essential to clarify that agreements dealing with the internal rules and the functioning of such entities still qualify as SAs because they indirectly help the parties to bring other humanitarian obligations into force. Indeed, the vagueness of CA 3 leaves space for creativeness and adaptation as long as the ad hoc agreement keeps its functional purpose. This is why indirect enforcement can take different forms and apply to various humanitarian issues.

Concretely, forming a specific body or mechanism can help to prepare, secure and coordinate humanitarian relief, as was the case with the Indonesian government and the Free Aceh Movement in 2000. They created a Joint Committee on Security Modalities to enact their SA, the Joint Understanding on Humanitarian Pause for Aceh.¹⁵⁰ From a different perspective, the protagonists of the 1991 armed conflict in Croatia established a Joint Commission to Trace Missing Persons and Mortal Remains (Joint Commission) along with its Rules of Procedure and Plan of Operation.¹⁵¹ In that double-sided SA, the parties assigned their obligations regarding the missing, the dead, and mortal remains to a body created for that mission, the Joint Commission. While the object of the instrument fulfils IHL requirements,¹⁵² its sixteen rules explain the regime of the Joint Commission, from membership conditions to the charging of expenses. The Annex to the Rules of Procedure details an explicit Plan of Operation to effectively address situations of deceased and missing persons by indicating which information to record and how to archive files. It is a set of technical guidelines completing the main agreement, with the same compelling value as the latter.

The complexity of SAs’ configurations can go even further, however. On 22 June 2001, the government of the Philippines and the MILF concluded a peace agreement providing for a mechanism to monitor their undertakings, the International Monitoring Team (IMT), but they also conferred IHL missions to a sub-organ, the Civilian Protection Component (CPC), attached to that monitoring entity. The latter was the object of two SAs, the Agreement on the Civilian Protection Component of the International Monitoring Team ¹⁵³ and the Terms of Reference of the Civilian Protection Component of the International Monitoring Team.¹⁵⁴

In Article 6 of the Terms of Reference, the parties entrusted the CPC with ensuring “that [they] respect the sanctity of places of worship namely mosques, churches and religious places and social institutions including schools, madaris, hospitals and all places of civilian nature”.¹⁵⁵ In addition, based on the same provision, the CPC was to assess “the needs of the IDPs and the delivery of relief and rehabilitation support efforts in conflict affected areas in Mindanao”, pursuant to Article 6(2)(c), and “acts of violence against civilians in conflict affected areas”, according to Article 6(2)(e). Following the model of the CPC, the parties added another body within the IMT, the Humanitarian, Rehabilitation and Development Component, tasked with the self-explanatory mission to “[m]onitor the observance of international humanitarian law and respect human rights”.¹⁵⁶

Because all the missions attributed to those mechanisms relate to treaty or customary IHL obligations, the agreements fulfil the objective of CA 3. Using IHL compliance mechanisms is consistent with the Geneva Conventions’ rationale.¹⁵⁷ Indirect and direct implementation SAs differ from one another even if they have the same purpose. The object of the former is to create and empower executive bodies, while in the latter, the parties commit themselves and can, in addition, set up an implementation entity for some of their obligations or a monitoring body for the whole agreement.¹⁵⁸ In both cases, there is always the possibility of calling upon a third party.

Acknowledgement or participation of a humanitarian actor: The example of the ICRC

An ad hoc agreement can support or extend the mandate of a humanitarian actor like the ICRC.¹⁵⁹ In many situations, the parties to a NIAC have assigned rights and obligations to the ICRC. However, the demonstration will be limited to SAs. Returning to the example of the SA on the Joint Commission and its Rules of Procedure and Plan of Operation,¹⁶⁰ it appears that the signatories gave essential roles to the ICRC: notably, it was the permanent adviser of the Joint Commission (Rule 1), which it chaired (Rule 17), and it was allowed to schedule extraordinary meetings (Rule 6), permit exceptional disclosure of confidential deliberations (Rule 10), and determine the admission of any external members to the meetings (Rule 10).¹⁶¹

The Agreement Relating to the Establishment of a Protected Zone around the Hospital of Osijek is evidence of an identical practice. The signatories entrusted the ICRC with (almost) the entire functioning of the zone.¹⁶² Likewise, an SA in the form of a joint communiqué between the government of Colombia and the FARC-EP called upon the participation of the ICRC “to locate, identify and respectably deliver the remains of persons deemed as missing within the context [of] and due to the armed conflict”.¹⁶³

That said, the warring parties can also pledge their active cooperation with and protection of the ICRC’s mandate in their SA in order to guarantee unhindered achievement of the operations of that humanitarian actor.¹⁶⁴ To give an example, this was the position of Croatia, Serbia, Slovenia, Montenegro and Macedonia when they jointly agreed to “unconditional support for the action of the ICRC in favour of the victims” in the Hague Statement on Respect of Humanitarian Principles.¹⁶⁵

In conclusion, ad hoc agreements show great potential for flexibility, adaptation and even the creation of humanitarian obligations either by the parties, through built-in mechanisms or via (humanitarian) third parties. There are also noteworthy features that practice shows in terms of formal requirements and the signatories of SAs.

Format and participants

The text of CA 3 does not provide information about an SA’s form or external features. Hence, the identification and validity of SAs are disconnected from their format, their title or the number of instruments composing them. As regards the participants, CA 3 only stipulates that the parties to the conflict are invited to conclude SAs. This requirement is confirmed in practice, even if it constitutes only a minimal condition.

Multi-form agreements

As noted earlier, the key determinants of an SA are its content and purpose.¹⁶⁶ The Geneva Conventions do not impose the need to have written agreements¹⁶⁷ or a particular written format;¹⁶⁸ however, paper-based agreements are usually preferred for apparent clarity, transparency and evidence reasons in case of claims, especially when the content is dense.¹⁶⁹ Moreover, the Geneva Conventions expected some SAs to assume a written form – notably, agreements on “safety zones or hospital zones”.¹⁷⁰ Indeed, Article 23 of GC I indirectly orients the parties to a specific format for agreements on establishing hospital zones and localities, as the Convention contains a draft agreement for that purpose.

As far as written agreements are concerned, practice in NIACs reveals that, formally, an agreement can qualify as an SA regardless of the number of instruments it combines. Hence, it could be a single piece¹⁷¹ or a main agreement completed by protocols.¹⁷² Practice also highlights that SAs do not exclusively pertain to agreements concluded during armed conflict – pauses, ceasefires or post-conflict written undertakings can amount to ad hoc agreements content-wise.¹⁷³ This fact challenges some positions claiming that only agreements concluded during an armed conflict can be regarded as SAs, because they pertain to “active hostilities”.¹⁷⁴ Even if this argument makes sense, it describes an incomplete scope of IHL obligations, as agreements aiming to end, resolve or address the causes of the conflict can also channel humanitarian commitments covering post-hostilities questions like de-mining,¹⁷⁵ exchange or release of persons detained because of the armed conflict,¹⁷⁶ the dead, the missing and mortal remains.¹⁷⁷ More generally, this also means that the denomination of the instrument as a “humanitarian agreement”, “ceasefire agreement” or “peace agreement” does not matter. Thus, regarding format, structure or composition, SAs can take the shape of any instrument. The distinction with other NIAC agreements lies in the humanitarian content and purpose.

Of course, the signatories can explicitly use the title “special agreement”, as observed on rare occasions.¹⁷⁸ However, that subjective perception does not prevent a close perusal of the agreement’s content. For example, among the six articles of Agreement No. 1, only Article 2, “Special Agreement”, seems to reflect the invitation of CA 3. However, that drafting choice may mislead us into thinking that the other provisions of the SA (Article 3, “Red Cross Emblem”, Article 4, “Dissemination”, and Article 5, “Implementation”) do not contribute to increasing the applicable law. In fact, the status of ad hoc agreement can extend to the whole instrument and not only Article 2. The identification of an SA must rely on an objective appraisal.

Open participation

Generally, the executive branch concludes the SA with an OAG, but more complex compositions may occur, as discussed earlier.¹⁷⁹ Ratione personae, CA 3 establishes the minimum requirement for the existence of an SA: to be a party to the NIAC. The double political and military leadership of some OAGs does not affect that requirement, however. A government (or an OAG) can validly conclude an SA with the political branch of a rebel movement,¹⁸⁰ its military side,¹⁸¹ or both subdivisions of the OAG¹⁸² because each wing is presumed to engage the whole entity when signing the agreement (though this does not mean that dissidence cannot occur). This unified subjecthood prevents legal insecurity in the event of a breach of the SA: if only the branch that made the deal were bound by the agreement, it would lessen the responsibility of the whole entity towards IHL. Regardless, in practice, there is no specific divide as to the nature of the undertakings pledged by the respective organs of an OAG: the military wing does not exclusively conclude agreements on military questions and vice versa.¹⁸³

Two observations arise from this minimal personal scope. Firstly, CA 3 excludes agreements concluded only between one actor in the armed conflict and a third party. There must be a connection with the opposing actor on the battlefield, even if the SA takes the form of a separate but parallel agreement with the same third party.¹⁸⁴ Secondly, nothing precludes the involvement of other entities (e.g., humanitarian actors, States, international organizations), either as active participants¹⁸⁵ or simply as witnesses or moral guarantors,¹⁸⁶ and without limit of number. Therefore, practice proves SAs to be an open tool for both parties and third parties, especially considering the teachings on indirect enforcement explained above.¹⁸⁷ However, this account of SAs’ functional, formal and personal characteristics commends a balanced appraisal. Indeed, the agreements must respect some essential conditions.

The necessary boundaries of special agreements

While an SA aims to enhance the regime of NIACs, its validity depends on some restrictions from IHL and the international legal order for the instrument to keep its value. That said, CA 3(4)¹⁸⁸ introduces a limitation that deserves clarification to avoid confusion with a denial of OAGs’ international legal personality.

The limited regime upgrade

From a general perspective, NIAC SAs are subject to the same limitations as in IACs, as expressed in Article 6/6/6/7 of the Geneva Conventions.¹⁸⁹ More specifically, SAs’ function of improving the IHL regime of NIACs is subject to four conditions. Firstly, they do not override the standards applicable; they only exist in complement or development thereof. Further, “even if the Parties have agreed to a more limited number of additional provisions, they nevertheless remain bound by all applicable humanitarian law norms”.¹⁹⁰ Also, since SAs incorporate further obligations, it is not (always) necessary for the parties to recall those which already compel them. Accordingly, the absence of fundamental rules does not mean that the agreement lowers the ordinary IHL regime in force or takes over. What the SA does not mention still applies as the minimal regime.

Secondly, the humanitarian undertakings and the agreement, in general, should not derogate the standards (i.e., relevant treaties and customs) by weakening the regime in force, as underlined by the ICRC.¹⁹¹ In other words, the commitments should not infringe or lower the minimal protection guaranteed to some categories of persons or their belongings in NIACs. For example, the agreement should not set abusive or unlawful exclusions from the basic NIAC protection regime. Likewise, when the applicable law unequivocally commands an indiscriminate treatment of an issue at stake, the humanitarian obligation in the agreement should not create abusively restrictive and burdensome procedures or requirements. Another possible downgrading of the IHL regime could include imposing deadlines or conditional requirements for performing an obligation falling under NIAC applicable law while that regime does not require so.¹⁹²

In parallel, the non-derogatory character must apply to the Geneva and Hague law.¹⁹³ Even if there are no explicit provisions or principles in the law of warfare to prevent an adverse impact of SAs on the conduct of hostilities, those agreements should not be the way to derogate to the Hague law. It would be odd to draw a permissive pretext from the silence on possible derogations in Hague law, whereas the prohibition on derogations holds for the Geneva law. Even though there is no positive law to prevent such derogations for the conduct of hostilities, it is not easy to establish what would justify a different fate for SAs under the Hague law. Therefore, we consider that an analogous restriction must apply, and it is even more justified as the travaux préparatoires of the Geneva Conventions did not specify which branch of IHL SAs would apply.¹⁹⁴

Finally, ad hoc agreements cannot derogate from or oppose jus cogens.¹⁹⁵ IHL is a branch of international law, and it seems coherent that agreements promoting IHL and the relevant obligations therein obey the same limitations that the international legal order imposes upon its norms.

The denied upgrade of status for organized armed groups

A superficial reading of CA 3(4)¹⁹⁶ seems to prevent OAGs from reaching an international legal status,¹⁹⁷ including in the advent of an SA. The signatories of some SAs have even recalled that provision in their agreements.¹⁹⁸ The ICRC Commentary on CA 3 can help clarify paragraph 4’s objective to shed light on its meaning. The aim was to avoid potential claims of rebels to the status of “belligerents”, especially if the agreements allowed the enforcement of all the Geneva Conventions.¹⁹⁹ The treaty drafters did not want OAGs to enjoy the benefits of the status that such a regime would entail – notably, the status of prisoner of war.²⁰⁰

The ICRC also explains that CA 3(4) is not intended to resolve the broader question of rebels’ (international) legal personality.²⁰¹ In other words, preventing the attribution of a status of belligerent to outlaws does not necessarily equate to a bar from any international legal standing.²⁰² OAGs can enjoy legal subjectivity on the international plane without reaching the level of belligerent; CA 3(4) only precludes them from being on a par with States, notably thanks to SAs. A lower standing remains possible, as supported by many scholars.²⁰³ Therefore, the CA 3(4) only clarifies that IHL compliance is separate from any claim to a (new) legal status by a party due to the conclusion of an SA or the application of the whole of CA 3.²⁰⁴ Bearing that clarification in mind, an SA can validly produce its effects.

Conclusion

To address the lack of analysis of CA 3 agreements in legal theory, this paper has tried, under a functional approach, to capture their features. This was necessary for the ultimate aim of explaining their dynamics for framing humanitarian obligations and engaging the parties to NIACs. What type of commitments can SAs cover? From which sources? Under which forms? How do SAs improve applicable law, and how are they implemented? The answers to these questions represent the lessons that this paper has tried to highlight. In addition, the fact that SAs can embody any NIAC agreement (e.g., “ceasefire”, “peace agreement”, “pause”) makes them formally attractive to every party on the battlefield, since their (distinctive) essence lies in their humanitarian content and purpose. Finally, while CA 3 calls upon the participation of NIAC actors, SAs remain open to third parties.

Therefore, ad hoc agreements are essential instruments for fostering IHL compliance and streaming humanitarian obligations in various ways beyond the frameworks of the Geneva Conventions and even IHL. Indeed, SAs have a valuable impact on beneficiaries in the humanitarian field, notably children,²⁰⁵ persons deprived of their liberty,²⁰⁶ refugees and IDPs,²⁰⁷ and civilians in general,²⁰⁸ but also (families of) the dead and the missing.²⁰⁹ Third parties, especially humanitarian actors, can also rely on ad hoc agreements. Furthermore, some agreements contribute to the performance of UN peacekeeping²¹⁰ or humanitarian missions.²¹¹

Of course, the functionality of SAs is guaranteed only under certain conditions regarding the betterment of applicable law and the status of the parties, mainly OAGs. That said, the above teachings from the practice could enlighten the parties to ongoing NIACs and any actor involved in conflict areas such as humanitarian professionals or negotiators. The examples provided could inspire the drafting of more SAs, given their flexibility, and generate (more) IHL “good practices” with more protective standards.