GC IV Commentary Series – Reprisals in NIACs: What Is the Commentary’s Take?

Editors’ note: The following series considers ICRC’s 2025 updated Commentary on the Fourth Geneva Convention. Rosa-Lena Lauterbach, Liisi Adamson, and Sean Watts’s introduction to the series is available here.

In its last section addressing Common Article 3, the 2025 Commentary of the International Committee of the Red Cross (ICRC) on the Fourth Geneva Convention (GC IV) addresses a matter on which the treaty text is silent: the legality of belligerent reprisals in non-international armed conflicts (NIACs). It provides a two-layered explanation. First, belligerent reprisals in NIACs do not exist. Second, even if they were to exist, belligerent reprisals in NIACs are prohibited, and this prohibition is contained in the article itself.

The Commentary’s take that the concept of belligerent reprisals does not extend to NIACs is grounded in: 1) the lack of State practice confirming an entitlement to resort to reprisals in NIACs; and 2) an explanation as to the low likelihood that States would have left open the possibility of reprisals in NIACs. With this second component, the Commentary implicitly clarifies that there is no space to discuss an imbalanced right to belligerent reprisals. The Commentary elaborates on belligerent reprisals as a tool potentially open to “belligerent Parties” (plural); and speaks of the implications of having international law permit armed groups’ resort to belligerent reprisals. Both elements indicate that the Commentary rejects the in-between understanding whereby belligerent reprisals in NIACs could be lawful but reserved for States.

In fact, the Commentary relies on Professors Milanović and Hadži-Vidanović (p. 16), for whom belligerent equality seems to be essential for rejecting the permissibility of belligerent reprisals in NIACs. This was the view of Cameroon and Nigeria ahead of the Additional Protocol II (AP II) negotiations (as accounted for by Bílková, p. 46); and can be read from certain scholars (Newton, p. 378). By explaining that the chances of having States leave room for armed groups to carry out reprisals are close to zero, the Commentary implicitly concludes that if reprisals were to exist in NIACs, they would need to be available to both States and armed groups.

A Tool “Open to Belligerent Parties”

The Commentary moves on to emphasise that, even in the event reprisals in NIACs are “a tool open to the belligerent Parties,” they would be prohibited in every case where the underlying conduct contravenes common Article 3. This stance, grounded in the absolute nature of common Article 3, from which no exceptions were allowed, was already put forward in the Pictet Commentary, had already met criticism, and is not expanded further here (see, for instance, Bílková, p. 55; De Hemptinne, p. 589; Darcy, p. 148). The Commentary thus argues that acts prohibited by common Article 3 equally cannot be committed by way of reprisals.

That statement invites the reader to examine the rest of the commentary on this article in order to discern the scope of the prohibition of reprisal by considering the interpretation of the acts common Article 3 explicitly prohibits. Which belligerent reprisals does common Article 3 set out to prohibit? And how does this prohibition fare against norms governing reprisals in international armed conflicts (IACs)? These two questions are the topic of this post.

The Prohibition of Belligerent Reprisals: General Aspects

According to the Commentary, common Article 3 implies that belligerent reprisals consisting of violations thereof are prohibited. At first glance, what broader differences can be spotted between this NIAC norm and the IAC prohibitions of reprisals?

On one hand, the Third Geneva Convention (GC III) and GC IV contain a limitation not found in common Article 3, one that relates to their personal scope of application. As the GC IV Commentary reminds, its norm concerning reprisals only prohibits acts against “protected persons,” as defined by the Convention and therefore excludes civilians that are the belligerent’s own nationals. GC III’s provision is equally limited by the Convention’s personal scope, which extends to protecting prisoners of war. By contrast, the scope of common Article 3 is not constrained by nationality or allegiance. It encompasses all persons not taking an active part in hostilities.

On the other hand, one aspect in which common Article 3 would be narrower than the GCs’ prohibitions is in relation to property. Unless the destruction or damage to property is interpreted as contrary to humane treatment or personal integrity, belligerent reprisals taken against the property of persons not taking part in hostilities are not prohibited by common Article 3. By contrast, the protection of protected persons’ property is explicitly enunciated in the IAC framework.

How do the categorical prohibitions of acts themselves compare? Those elements related to the prohibition of inhumane treatment are certainly covered by both IAC and NIAC norms. In this common core, similarities can also be spotted across types of behaviour. The Commentary clarifies that all forms of conduct are covered by the reprisal prohibition in GC IV, including omissions.

The prohibition of murder under common Article 3 requires closer scrutiny, however. Because the Geneva Conventions require that a person be in a belligerent’s power, their reprisal prohibitions would not prohibit acts occurring in the conduct of hostilities. Such a prohibition only came through the body of Additional Protocol I (AP I), and an argument can be made that such a norm, if not for civilian objects, then at least for civilians, is crystallising in customary law. But what about in NIACs to which AP I does not apply?

Prohibition of Murder Under Common Article 3 and the Conduct of Hostilities

The Commentary engages with the scope of the prohibition of murder under common Article 3 (paras. 614-17). The text provides an account of the two legal positions as to whether common Article 3 governs the conduct of hostilities, and concludes by identifying the ICRC view. One view maintains a strict differentiation between “Geneva law” and “Hague law” and sets out that common Article 3 would only apply to acts that presume that a person is in (enemy) power. This position is also supported by the ICRC. For belligerent reprisals, this would entail that, save for discrepancies in the scope noted above, the common Article 3 prohibition would match the prohibitions enshrined in the Geneva Conventions.

A second view, one the International Criminal Tribunal for the former Yugoslavia (ICTY) has endorsed, reads conduct of hostilities rules into common Article 3, and maintains that a violation of a conduct of hostilities rule constitutes, in addition, a violation of common Article 3. What are the implications of this position for the scope of the reprisal prohibition under common Article 3?

First, this expanded understanding of common Article 3 equalises the scope of prohibited belligerent reprisals in relation to the protection of civilians, as well as persons covered by special protection. Belligerent reprisals directed, for instance, against medical personnel, are prohibited under both treaty and customary law of IACs. Further, the treaty (if not also customary) law of IACs equally prohibits belligerent reprisals whereby civilians are made direct objects of attack. While not containing explicit prohibitions of belligerent reprisals, the law of NIACs equally envisages special protection of medical personnel and embodies the cardinal principle of distinction governing conduct of hostilities. Targeting medical personnel, or civilians, as an act that does not comply with international humanitarian law (IHL), amounts to a violation of common Article 3 as well and therefore comes within the scope of this norm’s reprisal prohibition.

The analysis is more complex in cases beyond those where these persons are directly targeted. Consider an instance in which a civilian object is targeted, and a civilian dies as a result of that attack. Or suppose that the object of attack is a military objective (therefore, now complying with the principle of distinction), but the attack causes disproportionate loss of civilian life. Does the civilian death qualify as murder?

At least in principle, when there is a causal link between an unlawful attack and the death, this death could qualify as murder under common Article 3. The analysis would turn on the question of intent. For some bodies, including the International Criminal Court (ICC), murder would only exist if the perpetrator was “at least aware that, in the ordinary course of events, the occurrence of such crimes was a virtually certain consequence of the implementation of the common plan.” Recklessness, or dolus eventualis, would not be covered. (Bemba, PTC, 2009, para. 369 as cited by Knuckey, p. 459). For others (the ICTY and inquiry mechanisms), this intent can be indirect. What must be established is that “an agent who, without being certain of a particular result, accepts the possibility of it happening” (Knuckey, p. 458-59). This latter view would bring the death occurring in our scenarios within the common Article 3 prohibition of murder.

Depending on how intent is understood, the NIAC prohibition of reprisals could be seen as broader than its IAC parallel in this aspect. The NIAC prohibition of violations of common Article 3 (and corresponding reprisals) would cover cases where death is “a probable consequence” (Knuckey, p. 459). On the other hand, IAC reprisals prohibit those attacks where a specific object, or person, is made “object of;” they speak of attacks “against” civilians, or civilian objects. The civilian loss cannot be “merely” incidental, it ought to be “deliberate and desired” (Sari, p. 230). O’Keefe’s analysis confirms the striking difference with IACs. Commenting on the prohibition of reprisals against cultural property, he underlines, “as long as reprisals against surrounding property remained lawful, cultural property remained at risk” (O’Keefe, p. 135).

Further, NIAC belligerent reprisals would be prohibited in other cases where objects enjoy special protection and, in IACs, explicit protection against being made object of reprisals. The argument would thus equally stand if “civilian object” in the first example were replaced by “cultural property” or “object indispensable for the survival of the civilian population,” for instance. Finally, using a prohibited weapon to target a military objective in circumstances where there is virtual certainty (for the ICC) or the probability (for the ICTY) of death of civilians not taking part in hostilities, would qualify as a prohibited belligerent reprisal in NIACs. By contrast, as the Convention on Cluster Munitions does not contain an explicit prohibition on their use in reprisals, nor is there evidence of such a customary law norm, in principle, such a reprisal would not raise concerns in IACs.

Customary IHL Norms and Reprisals

While this outcome would not bring about a complete prohibition of belligerent reprisals in NIACs by way of common Article 3, similar concerns would arise. Professor Sivakumaran, for instance, reminds us of: the difficulties surrounding the negotiation of the prohibition of reprisals against civilians and civilian objects in AP I; reservations submitted by certain States; and the disputed status of an equivalent customary law norm. It is thus highly questionable whether even this prohibition, of direct attacks (let alone other aspects mapped above), can be extended to a traditionally less regulated context of NIACs with such ease, purely based on an already questioned interpretation of common Article 3, and absent any explicit provision regulating any belligerent reprisals in NIACs. What is dubbed as the “minimum yardstick” thus comes to outlaw, at least in some aspects, a larger scope of belligerent reprisals than all IAC reprisal provisions combined.

The ICTY endorsed the view that the prohibition of murder covers conduct of hostilities cases, as well as the stance that belligerent reprisals against civilians are prohibited in NIACs. The ICRC Commentary does refer to the relevant ICTY jurisprudence. However, this reference is limited to the discussion of the change in ICTY’s case law as to whether belligerent reprisals against civilians were prohibited under customary IHL of NIACs. What the Commentary does not say, however, is that under its own understanding, the reprisals that were the topic of discussion in the cases before the ICTY would not come within the scope of prohibition under common Article 3 (nor the customary law norm, whose scope hinges upon the content of common Article 3 and Article 4 of AP II).

The Commentary stops short of entertaining this conversation. Having argued that common Article 3 does not address deaths in conduct of hostilities, it emphasises that this scope limitation does not mean that the conduct would be lawful; and it may well constitute a different violation (namely, of conduct of hostilities rules). This signposting, also present in sections addressing other norms (see, for instance, para. 3373, concerning art. 53 of GC IV), in the Commentary is a useful reminder that a norm “does not stand in isolation from other rules of international humanitarian law.” The same could be said for belligerent reprisals in NIACs. The absence of a norm addressing NIAC reprisals, or at least NIAC reprisals in conduct of hostilities, does not leave belligerents with a carte blanche.

First, a customary IHL norm operates in parallel to the purported common Article 3 prohibition. Professor Sivakumaran asserts that several prohibitions could potentially already form part of customary law: of reprisals against medical personnel; cultural property; and captured fighters (Sivakumaran, p. 452).

Second, and also arguably under customary IHL, any reprisal in NIAC would need to conform to the general conditions that reprisals must meet, conditions largely considered to reflect customary law (see rule 145); being taken with the purpose of inducing the adversary’s compliance; being carried out as a measure of last resort; proportionate to the violation it aims to stop; decided at the highest level of government; and terminated as soon as compliance by the adversary is achieved. While certain scholars warn about certain issues in how these conditions would be transposed to the context of NIACs (see De Hemptinne, p. 590-91), challenges in their application merely mean that the rules, or their interpretation, might need to be adapted to the NIAC context, not that the conditions are non-existent altogether.

Third, while the absence of a prohibition in international law could be interpreted as the freedom to act, an argument can be made that this freedom is confined by IHL’s own guardrails. The Martens Clause could be the first such brake, and bar belligerents from acting in a way that is contrary to dictates of public conscience. The Martens Clause allowed the ICTY to loosen the requirements to prove that a customary IHL norm prohibiting reprisals against civilians had formed, allowing for opinio juris, coupled with dictates of humanity or public conscience, to trump a potential lack of State practice (see Kupreškić, 2000 judgment, paras. 527-28, 534, 536). Thus, it is not unimaginable that, even without “pushing” for a customary law rule, the Martens Clause would still bar States from exercising their Lotus-style “freedom of action” in the absence of a specific prohibitive rule.

Furthermore, norms of other branches of international law could bar such behaviour. Darcy, in that vein, argues that the international human rights law non-derogable right to life would bar belligerent reprisals against civilians (Darcy, p. 180).

Conclusion

If belligerent reprisals are prohibited by common Article 3, how does the content of this norm shape the scope of the prohibition? More specifically, how would the scope of prohibited reprisals vary according to whether common Article 3’s prohibition of murder extends to conduct of hostilities, or is confined to situations where a person is in a belligerent’s power? By setting out that the prohibition covers only the latter, does that imply that belligerent reprisals are “a tool open to belligerent parties” during conduct of hostilities?

The post has argued that this is not the case. Even beyond the reminder that any such act, if not already in violation of another norm, must comply with the general requirements governing reprisals, the post sought to explore how any claim of having located a legal vacuum would fare against the general nature of IHL, a branch in which the general freedom to act in the absence of a prohibition is not unbounded. Reinforcing this “safety net” allows us to reach the same protective outcome, without stretching the interpretation of a norm beyond its breaking point.

***

Mina Radončić is a PhD candidate at the University of Geneva and a Teaching Assistant at the Geneva Academy of International Humanitarian Law and Human Rights.

The views expressed are those of the author, and do not necessarily reflect the official position of the United States Military Academy, Department of the Army, or Department of Defense. 

Articles of War is a forum for professionals to share opinions and cultivate ideas. Articles of War does not screen articles to fit a particular editorial agenda, nor endorse or advocate material that is published. Authorship does not indicate affiliation with Articles of War, the Lieber Institute, or the United States Military Academy West Point.

Photo credit: U.S. Army, Army Sgt. Charlyn Abuyo