Subsequent practice in the ICRC’s updated Commentary on GC IV

Since 2011, the International Committee of the Red Cross (ICRC)  has been engaged in a project to revise and update its Commentaries on the 1949 Geneva Conventions and their 1977 Additional Protocols. Updated Commentaries on the first three Geneva Conventions were released in 2016, 2017 and 2020, and the updated Commentary on the 1949 Geneva Convention (IV) relative to the protection of civilian persons in time of war (GC IV) was released recently in October 2025. One of the key motivations for this mammoth exercise in treaty interpretation is to account for the practice in interpretation and application of the Geneva Conventions and their Additional Protocols that is now available, several decades after their adoption (GC IV Commentary, para. 90). Indeed, one of the characteristics that sets the ICRC Commentaries apart is that they draw on the extensive practice that is documented in the ICRC’s archives (GC IV Commentary, paras 96-97).

As such, the updated Commentaries engage extensively with questions relating to the role of subsequent practice in treaty interpretation. One such question relates to the interpretive significance of practice that is seemingly inconsistent with a treaty provision: in what circumstances and to what extent does such practice influence the interpretation of a treaty provision? In this blog post, we will analyse this question by drawing on two examples from the GC IV Commentary in which seemingly inconsistent practice leads to different interpretive results. Before turning to these examples, we will start by briefly outlining the role of subsequent practice in treaty interpretation.

Subsequent practice in treaty interpretation

The customary (e.g., Case Concerning Sovereignty over Pulau Ligitan and Pulau Sipadan, para. 37) methodology for treaty interpretation reflected in Arts 31-32 of the 1969 Vienna Convention on the Law of Treaties (VCLT) requires: ‘[taking] into account…any subsequent practice in the application of the treaty which establishes the agreement of the parties regarding its interpretation’ (Art. 31(3)(b)). As to what constitutes subsequent practice and how it is to be accounted for in treaty interpretation, a few points – drawn from the recent work of the International Law Commission (ILC) on this topic – bear emphasis.

First, subsequent practice consists of the conduct of one or more parties in applying the treaty after its conclusion: as such, it must indicate the taking of a position regarding the interpretation of the treaty (ILC, conclusions 4 and 6). Assessing whether practice evidences the taking of an interpretive position (and, thus, qualifies as subsequent practice) requires ‘a careful factual and legal analysis’ to determine whether the practice rests on a considered interpretation of the treaty or is motivated by other considerations (ILC, pp. 45-46, paras 7 and 11). The weight of such practice may depend on its clarity and specificity, as well as its repetition (ILC, conclusion 9).

Second, the interpretive relevance of subsequent practice under VCLT Art. 31(3)(b) is conditioned on the practice in question establishing the agreement of all the parties to the treaty (ILC, p. 13, para. 16). (Note, however, that this conclusion is disputed by some scholars – e.g., Hill-Cawthorne). According to the ILC, subsequent practice that does not fulfil this requirement may nonetheless be relevant as a supplementary means of interpretation under VCLT Art. 32 (ILC, p. 33, para. 23).

Third, while subsequent practice certainly plays a significant role in treaty interpretation, it is not conclusive as to the meaning of a treaty provision (ILC, p. 24, para. 4). It is simply one of the elements that must be considered along with others (ILC, conclusion 2(5)) and, as such, its significance may vary in different interpretive contexts (ILC, p. 51, para. 2).

Example 1: grave breaches regime

We will now turn to the first of two examples illustrating the updated GC IV Commentary’s engagement with seemingly inconsistent practice. This example concerns GC IV Art. 146, which is common to all four Geneva Conventions (Art. 49/50/129 in the other three Conventions), and which establishes what is known as the grave breaches regime. Art. 146(1) requires High Contracting Parties to legislate effective penal sanction for persons who commit or order the commission of grave breaches of the Geneva Conventions. Art. 146(2) requires High Contracting Parties to search for persons alleged to have committed or ordered the commission of grave breaches and bring them before its own courts regardless of their nationality.

The plain text of Art. 146 clearly envisages an unconditional obligation to criminalise grave  breaches and to search for alleged perpetrators (through the modes of commission or ordering) of grave breaches and prosecute them (GC IV Commentary, paras 6557 and 6579), including on the basis of universal jurisdiction (GC IV Commentary, paras 6564 and 6583). (The only envisaged exception to the prosecution obligation is for cases where extradition is possible – see, further, GC IV Commentary, para. 6578.) Nothing in the context or drafting history of this provision, nor in the object and purpose of GC IV, undermines this conclusion.

Nonetheless, when we turn to practice, the picture is rather mixed. Information on national implementing legislation is available for at least 125 out of the 196 High Contracting Parties to GC IV. Clearly, however, a number of States have not complied with the requirements of Art. 146(1), and those that have, have done so in varying ways (GC IV Commentary, para. 6577). Similarly, in relation to Art. 146(2), while there are some examples of prosecution of grave breaches, they remain the exception rather than the norm, especially on the basis of universal jurisdiction (GC IV Commentary, paras 6609 and 6611). This raises questions as to whether inconsistent practice has diluted the mandatory and unconditional nature of Art. 146, perhaps with the result that this provision now authorises, but no longer requires, penalising and prosecuting alleged perpetrators of grave breaches.

The updated ICRC Commentary on GC IV Art. 146 concludes that this is not the case: High Contracting Parties to GC IV remain under an unconditional obligation to provide effective penal sanction for committing or ordering grave breaches and to search for and prosecute (or extradite) alleged perpetrators.

First, supporting practice, although limited, is not absent. Several States have complied with the legislative obligation under Art. 146(1), and there are instances of prosecutions of grave breaches, including on the basis of universal jurisdiction (see examples compiled by the ICRC here (Chapter 44), here (pp. 53-131) and here). In addition to complying with their own obligations under Art. 146, States have also sought to enforce other States’ compliance with Art. 146. For instance, compliance with Art. 146 has recently been specifically sought in two cases at the International Court of Justice (ICJ) (see further GC IV Commentary, paras 6577 and 6610) – Armed Activities on the Territory of the Congo (DRC/Uganda) (para. 258(d)) and Alleged Breaches of Certain International Obligations in Respect of the Occupied Palestinian Territory (Nicaragua/Germany) (para. 67(6))). Similarly, there are well-reported examples of States publicly highlighting failures to investigate and prosecute grave breaches. This supporting practice demonstrates that inconsistent practice does not establish the agreement of all the High Contracting Parties. This means that, at least following the ILC’s understanding, this practice, assuming it qualifies as subsequent practice, does not qualify as subsequent practice under VCLT Art. 31(3)(a), although it may still be relevant as a supplementary means of interpretation under VCLT Art. 32.

Second, it is not clear that inconsistent practice qualifies as subsequent practice in the sense that it does not necessarily reflect the taking of a position on the interpretation of Art. 146. For instance, while grave breaches prosecutions are relatively limited, there is a far broader range of practice concerning the prosecution of war crimes more generally, often in relation to conduct that could have been charged as a grave breach. Indeed, prosecutions of war crimes, including grave breaches, and including on the basis of universal jurisdiction, are actually increasing (GC IV Commentary, paras 6611-6612). While the penalisation or prosecution of grave breaches as war crimes may not reflect perfect compliance with Art. 146 (compare Dörmann and Geiß, pp. 708-710, and Gaeta, pp. 621-624), it certainly does not suggest an interpretation of the obligation as having lost its mandatory and unconditional character. Indeed, it is hard to question the continuing normative significance of the grave breaches regime when it continues to be endorsed by States and international bodies (see further GC IV Commentary, paras 6577 and 6610) – e.g., the ICJ (para. 390); the International Criminal Court (Art. 8(2)(a), the United Nations Security Council (p. 2), the United Nations General Assembly (p. 3) and the ILC (p. 285, para. 5).

Example 2: protecting powers regime

The second example that we will consider is GC IV Arts 9 and 11, which are also common to all four Geneva Conventions (Arts 8/8/8 and 10/10/10 in the other three Conventions). Art. 9 requires that GC IV ‘shall be applied with the co-operation and under the scrutiny of the Protecting Powers whose duty it is to safeguard the interests of the Parties to the conflict’. The 1958 Pictet commentary on GC IV (p. 86) characterised this as a ‘command’ to the Parties to the conflict and to protecting powers, excluding any possibility of ‘protected persons not having the benefit of the services of a Protecting Power or a substitute’. Art. 11 provides for a substitution mechanism in cases where no protecting power is appointed – in particular, Arts 11(2) and (3) oblige Detaining Powers to secure the services of a substitute State or organisation in such a situation. It was understood at the time the Geneva Conventions were adopted that while Arts 9 and 11 do not mandate the appointment of a protecting power or substitute, they clearly presuppose its existence (Pictet commentary, p. 92). This presumption is reinforced by the extensive array of supervisory duties allocated to  protecting powers under various provisions of GC IV (see further GC IV Commentary, paras 1282-1332). In sum, the plain text of Arts 9 and 11 suggests that it is compulsory to rely on protecting powers or substitutes to supervise the implementation of GC IV (GC IV Commentary, paras 1237 and 1465). Nothing in the context or drafting history of these provisions, nor in the object and purpose of GC IV, undermines this conclusion.

In practice, however, there are only five known instances in practice in which protecting powers been appointed in accordance with Art. 9 (GC IV Commentary, para. 1337). Similarly, substitutes have not been formally appointed when they should have been in accordance with a plain text reading of Art. 11 (GC IV Commentary, para. 1466). A number of reasons may explain these failures (Kolb, pp. 557-558) – for instance, the difficulty of identifying a neutral intermediary acceptable to both parties; reluctance of protecting powers to take on the onerous and expensive responsibilities entailed in the function; or, concerns that appointment of a protecting power may imply recognition of the existence of the adversary or of the applicability of the Geneva Conventions.

As a result, the updated ICRC Commentary recognises that the subsequent practice of States interprets the supervisory role of protecting powers and their substitutes under Arts 9 and 11 as optional (GC IV Commentary, paras 1340 and 1466). Beyond the extent of the inconsistent practice, three factors support this conclusion.

First, the failures of the protecting powers regime were expressly recognised by High Contracting Parties to the Geneva Conventions during the negotiation and drafting of the 1977 Additional Protocols. Indeed, Art. 5 of Additional Protocol I (AP I) was intended to address the shortcomings of the protecting powers system under the Geneva Conventions (GC IV Commentary, para. 1338; AP I Commentary, paras 177-179); in particular, by bridging the gap between the appointment and existence of protecting powers by mandating their appointment and clarifying the procedure for doing so. To be clear, AP I Art. 5 reinforces the compulsory nature of the protecting powers system and was intended to remedy the difficulties raised by the system as set out in the Geneva Conventions (an endeavour in which it seems to have failed). Nonetheless, the negotiations relating to AP I Art. 5 provide important confirmation that many of the High Contracting Parties of GC IV saw failures to comply with Arts 9 and 11 as reflecting problems in the protecting powers system rather than breaches of these obligations.

Second, while the obligations set out in Arts 9 and 11 may have become optional, this does not mean that the protecting powers regime itself has become obsolete and fallen into desuetude (GC IV Commentary, paras 1265 and 1466). This has two implications. One is that it remains possible for Parties to conflicts to appoint protecting powers or substitutes, and there have, in fact, been attempts to appoint protecting powers under the Geneva Conventions in recent armed conflicts. The other is that the optional nature of the reliance on protecting powers or substitutes does not detract from the value of their function – the effective supervision of the Conventions. While High Contracting Parties to GC IV now arguably have more discretion as to how the objective of appointing protecting powers or substitutes is to be achieved, they should still ensure this objective is achieved, perhaps by inviting an impartial humanitarian organization, such as the ICRC, to perform these functions (GC IV Commentary, paras 1340-1341 and 1470).

Third, in practice, the humanitarian consequences of the absence of protecting powers and substitutes have been mitigated to a great extent through the activities of the ICRC. Acting pursuit to its right of humanitarian initiative under GC IV Art. 10 (also common to all four Geneva Conventions – Art. 9/9/9 in the other three Conventions), the ICRC performs many – and in some situations most – of the tasks normally entrusted to Protecting Powers.

Reconciling the different significance of inconsistent practice in the interpretation of Art. 146 and Arts 9 and 11

The different interpretive roles of inconsistent practice in the cases of the grave breaches regime and the protecting powers regime may seem contradictory. In both cases, the unconditional and mandatory text of the obligations is undermined by a large body of inconsistent practice. Yet, in one case the updated ICRC commentary considers that this practice has not diluted the mandatory nature of the obligation, while in the other it reaches the opposite conclusion.

However, once we look past the mere fact of inconsistency of practice, it becomes apparent that any similarity is merely superficial. In the case of Art. 146 and the grave breaches regime, the interpretive significance of inconsistent practice is undermined by the nature of the inconsistent practice (it is not clear that it reflects the taking of a position on the interpretation of Art. 146) and by a countervailing body of supportive practice including challenges to inconsistent practice. In the case of Arts 9 and 11 and the protecting powers regime, there are no similar countervailing factors. Instead, there is acknowledgment of widespread non-compliance with the ordinary meaning of the text. Furthermore, the lack of compliance and dilution of the obligation has not frustrated the underlying purpose of the regime (effective supervision of implementation of the Geneva Conventions), in particular through the fulfilment of the functions of a protecting power by other means, for example by accepting the services of an impartial humanitarian organization, such as the ICRC.

As such, the different conclusions in these two examples do not point to a methodological inconsistency. Instead, they emphasise the importance of placing subsequent practice in its proper context as part of a ‘single combined operation’ (ILC, p. 51, para. 2) of interpretation, and demonstrate that subsequent practice may play different roles in different interpretive contexts. Indeed, the different outcomes in these two cases provides a valuable reminder that although subsequent practice plays a critical role in treaty interpretation, it is not determinative of the meaning of treaty provisions. Its interpretive significance must be assessed with care and nuance, and in conjunction with other interpretive elements.

Note: This post forms part of a joint symposium with the International Committee of the Red Cross (ICRC) and the editors of Just Security, sharing expert contributions on selected topics addressed in the updated ICRC Commentary on the Fourth Geneva Convention.