COMMON ARTICLE 1 OF THE GENEVA CONVENTIONS AND THE METHOD OF TREATY INTERPRETATION

Abstract

In its updated Commentaries on the 1949 Geneva Conventions, the International Committee of the Red Cross (ICRC) embraces the ‘external’ interpretation of Article 1 common to the four Geneva Conventions, according to which States have certain negative (complicity-type) and positive (prevention/response) obligations to ‘ensure respect’ for the Conventions by other actors. This interpretation has been gaining support since the 1960s, though the ICRC's new Commentaries have served as a catalyst for some States recently to express contrary views. This article focuses on two major methodological shortcomings in the existing literature, offering a much firmer foundation for the external obligation under common Article 1. First, it demonstrates the overwhelming support in subsequent practice for external obligations. Previous studies have failed to explain the method by which this practice is taken into account, given the existence of some inconsistent practice. This article addresses this general question of treaty interpretation, critiquing the approach of the International Law Commission that relegates majority practice to supplementary means of interpretation and proposing instead a principled approach that better fits and justifies the judicial practice here. Secondly, the article challenges two common assumptions about the travaux: first, that an original, restrictive meaning was intended, and secondly that the travaux of Additional Protocol I offer no support for external obligations. Given the ubiquity of military assistance and partnering, these findings have far-reaching consequences for the liability of States.

I. INTRODUCTION

Following the adoption in 1949 of the four Geneva Conventions,¹ the International Committee of the Red Cross (ICRC), a key driver behind the Conventions, set to work on four Commentaries to assist with their interpretation.² These Commentaries, known as the Pictet Commentaries after Jean Pictet (the ICRC lawyer that led the project), and the Commentaries on the two 1977 Additional Protocols,³ are frequently relied on as influential, even authoritative, accounts of the drafting histories of the Conventions and Protocols.⁴

The four Commentaries are now the subject of an ongoing updating project by the ICRC Legal Division, with the first three updated Commentaries having been released, respectively, in 2016, 2017 and 2020 (and the fourth due in 2024).⁵ A key rationale for the project was ‘to reflect the experience gained in applying the Conventions and Protocols during the decades since their adoption … [and thereby] ensure that the new editions of the Commentaries reflect contemporary reality and legal interpretation’.⁶ One of the aims, in other words, was to take account of subsequent practice.

An example that was recognized early on in the revision process as being a provision in need of an updated interpretation in light of subsequent practice was Article 1 common to all four Conventions (restated mutatis mutandis in Article 1(1) of the First (API) and Third Additional Protocols to the Conventions (APIII).⁷ According to this provision, ‘[t]he High Contracting Parties undertake to respect and to ensure respect for the present Convention in all circumstances’. As the project lead on the updated Commentaries stated, ‘[t]he application of the obligation to ‘ensure respect’ for humanitarian law has significantly expanded since the 1950s, and current practice in this area indicates more clearly that this obligation extends to both international and non-international armed conflicts’.⁸ Consistent with this view, common Article 1 is given a broader interpretation in the updated Commentaries than in the Pictet Commentaries, which were more equivocal (and certainly less detailed) regarding this provision.⁹ Thus, in the latest Commentaries, whilst the obligation to ‘respect’ is interpreted as reiterating the principle of pacta sunt servanda,¹⁰ the undertaking to ‘ensure respect’ is said to generate obligations for States regarding both persons subject to their authority, such as their populations (the internal dimension),¹¹ and other (State and non-State) actors (the external dimension).¹² The external dimension itself is then said to comprise both negative and positive obligations:

Under the negative obligation, High Contracting Parties may neither encourage, nor aid or assist in violations of the Conventions by Parties to a conflict. Under the positive obligation, they must do everything reasonably in their power to prevent and bring such violations to an end. This external dimension of the obligation to ensure respect for the Conventions goes beyond the principle of pacta sunt servanda.¹³

It must be emphasized that this interpretation by the ICRC is not novel, and the notion that common Article 1 contains ‘external’ obligations for States parties in relation to the conduct of others has been gaining momentum in State and judicial practice, particularly since the 1960s.¹⁴ Nonetheless, this purported external dimension of the ‘ensure respect’ obligation continues to divide opinion. Indeed, whilst scholarly disagreement over what exactly is required by common Article 1 has long existed,¹⁵ there has been a very notable reaction to the ICRC's recent pronouncement.¹⁶ Though many scholars support the notion of an external obligation,¹⁷ there are also some strong dissenters, many of whom argue that the obligation to ‘ensure respect’ refers solely to the internal dimension of ensuring respect by those within the State's jurisdiction and acting on its behalf.¹⁸

A number of States have also recently taken explicit positions on the purported external obligation.¹⁹ Some have done so specifically in response to the ICRC's updated Commentaries.²⁰ Others have expressed their views in the context of the recently concluded work of the International Law Commission (ILC) on the protection of the environment in armed conflict, during which a stark division emerged between States supportive and critical of the purported external obligation under common Article 1.²¹

Academic and governmental disagreement over the content of common Article 1 thus now appears greater than ever. This is particularly concerning in the light of the ever-growing phenomenon of partnered warfare, in which operations by States and non-State actors are inter-connected through cooperation and assistance. The updated ICRC commentaries refer precisely to this situation, where States potentially have a significant influence over the behaviour of their partners, as an example of a situation where the purported external obligation under common Article 1 bites most strongly.²² This context no doubt partly explains the positions recently taken by certain States on this. Coalition forces in Iraq and Afghanistan and the Saudi-led coalition in Yemen immediately come to mind as examples of close partnering in military operations where common Article 1 might especially be relevant. Yet the negative and positive obligations according to the ICRC's interpretation would also be relevant to the vast array of other, less direct forms of assistance commonly given by States to allies, such as intelligence sharing,²³ and provision of weapons and training.²⁴ The ICRC's (and others’) reading of common Article 1 would also extend to State support for non-State armed groups, such as Western support for groups in Syria.²⁵ Moreover, the context-sensitive nature of the due diligence standard with which the ICRC defines the positive obligations thereunder is such that broader attempts to prevent and respond to international humanitarian law (IHL) violations would be required in many other situations.²⁶

A potentially very extensive set of duties would thus follow from an external obligation under common Article 1. States that reject this interpretation argue that it is the secondary rules on State responsibility that determine their ancillary responsibility in relation to the conduct of others.²⁷ As shown below, these secondary rules are much more limited in their scope and content than the purported external obligation under common Article 1 and would fail to capture many of the examples of assistance given above.²⁸ Whether such an external obligation reflects the current lex lata therefore has significant consequences for States’ obligations in armed conflict, especially in light of the increasingly diverse forms of military partnership and assistance.

It is on this question that this article focuses, and it argues that common Article 1 does indeed contain external obligations in relation to the conduct of other actors. The purpose here is not to rehearse the arguments that have already been made in the many contributions on this topic. Rather, the focus is on two fundamental points of method that have been misunderstood or ignored entirely in previous studies, including accounts otherwise supportive of external obligations under common Article 1. The failure to consider these points undermines the reliability of the conclusions those previous studies reach regarding the content of that provision.

The first, and most significant, concerns the role of subsequent practice in interpreting common Article 1. Given the lack of clarity regarding the content of common Article 1 that is left by the text, context and object and purpose, previous studies rely heavily (often decisively) on subsequent State practice.²⁹ Whilst State practice does overwhelmingly support an external element to common Article 1, as explored below, there is also some clearly inconsistent State practice that challenges that interpretation. On the orthodox account of subsequent practice in treaty interpretation, such inconsistency prevents the majority, supportive practice from being taken into account under Article 31(3)(b) of the Vienna Convention on the Law of Treaties (VCLT), given the lack of ‘agreement’ amongst the parties to the Conventions (as required by Article 31(3)(b)).³⁰ Many previous studies relying on subsequent practice to found an external obligation under common Article 1 do not consider this apparently fatal methodological flaw.³¹ The few that do, follow the approach of the ILC in its draft conclusions on subsequent agreements and subsequent practice,³² arguing that supportive State practice is still admissible under Article 32 of the VCLT notwithstanding the existence of some dissent, but with the necessary consequence that it carries less weight than practice establishing the agreement of all the parties and admissible under Article 31(3)(b).³³ This is also said to be the ICRC's general method in its updated Commentaries.³⁴ Yet the lesser interpretive weight to be given to such practice according to that method sits uneasily with the apparently decisive weight given by those previous studies to practice supportive of the external obligation.

This article engages squarely with this general question of treaty interpretation. It argues that the approach adopted by the ILC with respect to subsequent practice that includes some dissent is conceptually flawed and misreads previous judicial practice on which it purports to be based. The consequence, it is argued, is that such practice can still be admissible as an authentic means of interpretation under Article 31(3)(b) of the VCLT, instead of relegating it to Article 32 of the VCLT. In this respect, the findings in the article are significant for the role of subsequent practice in treaty interpretation generally, with a novel and principled account offered of the admissibility and weight of subsequent practice where some dissent is present. Indeed, this methodological issue regarding subsequent practice arises frequently across different regimes of international law.³⁵ Regarding common Article 1, the argument made in this article offers, for the first time, a principled basis for taking account of subsequent practice that supports the external obligation in a way that other studies have failed to do, thereby placing that obligation on a much firmer footing.

The second key contribution of this article relates to the travaux of common Article 1 and Article 1(1) of AP I. Whilst it is often said that the travaux show little or no support for the external obligation, they are re-examined here with clear support unearthed for the external obligation, particularly during the drafting of AP I. Though necessarily carrying less interpretive weight than the authentic means under Article 31 of the VCLT, these findings regarding the travaux offer an important rebuttal to those who make claims for an original restrictive meaning of ‘ensure respect’ under common Article 1 and Article 1(1) of AP I.

In elaborating upon these under-explored issues, the article has two goals: first, to consider the precise meaning of the obligations in common Article 1; secondly, and more broadly, to use common Article 1 as a case study on the relevance of consent and dissent in treaty interpretation. The argument proceeds as follows. Section II considers some preliminary matters for the interpretation of common Article 1 (and Article 1(1) of AP I) (the ordinary meaning, context and object and purpose) before moving on in Section III to the analysis of subsequent practice. Section IV then considers the travaux of the two provisions, as supplementary means for their interpretation. For completeness, Section V then considers the other key question regarding the scope of common Article 1, namely its application in non-international armed conflicts. Section VI concludes the analysis with an alternative way of thinking about the precise elements of the external obligation.

II. ORDINARY MEANING, CONTEXT, AND OBJECT AND PURPOSE

As often noted in the literature, the fundamental importance of subsequent practice, and potentially the travaux (as supplementary means), for interpreting common Article 1 arises from the uncertainties flowing from the application of the other authentic means of interpretation (codified in Article 31 of the VCLT).³⁶ The purpose of this section is to confirm this uncertainty, whilst clarifying certain misunderstandings that pervade the existing literature.

It must immediately be acknowledged that little can be gleaned regarding the scope of the undertaking to ‘ensure respect’ for the Conventions and Protocols from any ‘ordinary’ meaning alone.³⁷ Both the restrictive interpretation (that views this duty as being entirely internal, but which goes beyond pacta sunt servanda and the duty to ‘respect’, ie to ensure respect by the State's own armed forces and population) and the expansive interpretation (that views this duty as also having an external dimension) can reasonably be read into the term ‘ensure respect’.³⁸ Similar language in human rights treaties³⁹ has been interpreted as requiring States to take positive steps to protect individuals from rights violations by others (including other States).⁴⁰ Yet that can hardly form the basis for suggesting that the same interpretation constitutes the ordinary (or special⁴¹) meaning to be given to the term in common Article 1.⁴² Most fundamentally, positive obligations under human rights treaties frequently extend only to victims who are within the jurisdiction of the State, as opposed to creating general obligations which can be enforced against other contracting parties.⁴³

In truth, Article 31(1) of the VCLT speaks not of a stand-alone ‘ordinary meaning’, but rather of the ‘ordinary meaning to be given’ to the provisions of the treaty ‘in their context and in light of its object and purpose’.⁴⁴ Regarding the context, some have argued that the Conventions and AP I explicitly set out elsewhere when States have obligations in relation to the conduct of others, apparently suggesting a more restrictive interpretation of common Article 1 and Article 1(1) of AP I.⁴⁵ In truth, those other provisions are much more specific regarding their scope and content, and they in no way undermine the claim that common Article 1 contains broader, less prescriptive obligations on all States. Focarelli, for example, refers to Article 7 of AP I, which states that the depositary ‘shall convene a meeting of the High Contracting Parties, at the request of one or more of the said Parties and after the approval of the majority of the said Parties, to consider general problems concerning the Application of the Conventions and of the Protocol’.⁴⁶ Yet this differs from Article 1(1) of AP I, not only in its focus on multilateral meetings of States parties, but also in its limitation to ‘general’ problems concerning the application of the Conventions and Protocol, which, as the drafting history confirms, refers to issues of application other than specific conflicts or violations.⁴⁷

Article 89 of AP I (providing for individual or joint action in response to serious violations⁴⁸) is similarly looked to when arguing a contrario that Article 1(1) of AP I could not have been intended to contain broad external obligations.⁴⁹ Yet that provision was intended to confirm that States could not act inconsistently with the United Nations (UN) Charter (eg by reprisal) when responding to violations of the Protocol.⁵⁰ As discussed below, the travaux of AP I more generally confirm that many of these specific enforcement provisions were seen as particular mechanisms for ‘ensuring respect’ for IHL.⁵¹ This understanding of common Article 1 (and Article 1(1) of AP I) is supported by their being the first provisions of the Conventions and Protocol, suggestive of their having a broad and general character.⁵²

With respect to the object and purpose of the Conventions and Protocol, it is often said that they reflect a careful balance between humanitarian considerations and considerations of military necessity.⁵³ There is some truth in this. It is the fact of the underlying armed conflict which helps to explain the presence of considerations of military necessity in IHL and its more permissive approach to targeting and detention than under human rights law, for example.⁵⁴ Yet such general claims regarding the underlying values of IHL are not necessarily normatively helpful as factors influencing interpretation.⁵⁵ To use Gardiner's terminology, this balancing that is said to underpin IHL might be seen more as its ‘spirit’ than as its legally relevant object and purpose.⁵⁶

The object and purpose should rather be sought from the text itself, including, in particular, the preamble.⁵⁷ Though the Geneva Conventions do not contain any substantive preambles, both the idea and possible content of a preamble to each Convention were discussed at great length at the 1949 diplomatic conference, and the various proposals all referred in an unqualified way to the humanitarian goals of the Conventions.⁵⁸ In the end, no substantive preambles were adopted due to disagreement over the degree to which they should include specific prohibitions, as well as whether they should include a reference to a divine source for the rules.⁵⁹ But there appeared to be no disagreement on the solely humanitarian aims of the Conventions.⁶⁰ Indeed, their full titles are suggestive of this.⁶¹ The substantive provisions of the Conventions (and AP I) also clearly focus on protecting individual war victims,⁶² along with mechanisms for monitoring and enforcement.⁶³ Similarly, it has persuasively been argued that the object and purpose of AP I are entirely humanitarian, notwithstanding that it contains so-called ‘Hague rules’ on the conduct of hostilities that are seen as being tempered by considerations of military necessity.⁶⁴ This is supported by the preamble to AP I, which did make its way into the final treaty.⁶⁵ This reading of the Conventions’ and Protocol's object and purpose cannot alone indicate how to interpret common Article 1. Yet an interpretation that recognizes the existence of external obligations for States to ensure respect by others for the Conventions and Protocol certainly advances their humanitarian aims by offering a potentially far-reaching means of enforcement.⁶⁶

Finally, as regards the other authentic means of interpretation, Article 31(3) of the VCLT refers, in addition to subsequent practice (discussed below), to subsequent agreements (Article 31(3)(a)) and other relevant rules of international law (Article 31(3)(c)). In the case of the Geneva Conventions and AP I, no subsequent agreements exist,⁶⁷ and no extraneous rules can be drawn on to provide a useful indicator of the intention of the States parties (as the discussion above regarding similar concepts found in certain human rights treaties demonstrates). Some commentators have drawn on Article 31(3)(c) of the VCLT to suggest that, since it is rare for international law to place obligations on States in relation to the conduct of others, common Article 1 should be read restrictively.⁶⁸ However, this is not an accurate application of Article 31(3)(c).⁶⁹ Rather, the principle requires a consideration of specific extraneous rules that are relevant to the interpretation of the treaty terms at issue.⁷⁰ It does not enable such general claims (much less general claims concerning the absence of rules) to influence interpretation, and in this case such an approach is all the more untenable given that similar obligations do exist elsewhere in international law.⁷¹

The above analysis confirms that the text, context and object and purpose leave the content of the ‘ensure respect’ obligation unclear, and this remains unresolved by recourse to extraneous agreements or rules. Therefore, subsequent State practice takes on a crucial role when interpreting common Article 1 and Article 1(1) of AP I/AP III. This is recognized by others who argue in favour of the external obligation and who rely heavily (sometimes decisively) upon subsequent practice.⁷² It is, therefore, essential that any analysis of subsequent practice is based on a sound, principled methodology. Previous studies have failed to do this.⁷³ The following section explores these issues in detail.

III. SUBSEQUENT PRACTICE

A. State Practice Regarding Common Article 1

It is well established that subsequent State practice can shed light both on the original common intentions of the parties in adopting a particular treaty term and on any later intention, which may have developed over time.⁷⁴ With this in mind, this section will demonstrate that the notion of external obligations under common Article 1 finds overwhelming support in State practice, which has been growing particularly since the 1960s.⁷⁵ Previous studies have drawn on examples of State practice, which show support amongst States for a reading of common Article 1 that requires the international community to ensure respect for IHL by others, although some of the examples referenced are not clearly linked to common Article 1.⁷⁶ Without repeating relevant examples given in that literature, the purpose of this section is to evidence this support in subsequent practice more fully. It will demonstrate the true extent of support for at least a positive external obligation that requires States to respond to IHL violations. The focus here is on recent practice, and it is especially instructive that practice continues to support this external obligation, given the growing attention placed on common Article 1 in light of the ICRC's updated commentaries and the ILC's work on the protection of the environment in armed conflict.

Clear support for the external element under common Article 1 can be found in many UN General Assembly (UNGA) resolutions. For example, in UNGA Resolution 77/126, adopted in December 2022, States asserted the following:

Recalls, in this regard, the statement of 15 July 1999 and the declarations adopted on 5 December 2001 and on 17 December 2014 by the Conference of High Contracting Parties to the Fourth Geneva Convention on measures to enforce the Convention in the Occupied Palestinian Territory, including East Jerusalem, welcomes in this regard initiatives by States parties, both individually and collectively, in accordance with article 1 of the Convention, aimed at ensuring respect for the Convention and accountability, and calls upon all High Contracting Parties to the Convention to continue, individually and collectively, to exert all efforts to ensure respect for its provisions by Israel, the occupying Power, in the Occupied Palestinian Territory, including East Jerusalem, and other Arab territories occupied by Israel since 1967.⁷⁷

This same position has been taken in many other recent UNGA resolutions relating to Israel and Palestine.⁷⁸ These resolutions have a long pedigree, with similar resolutions referencing common Article 1 having been adopted by the UNGA for decades.⁷⁹ Focarelli argues that ‘it is possible to assign to these resolutions a mere recommendatory meaning’, given the lack of prescriptiveness in the phrase ‘in accordance with article 1’.⁸⁰ Such a reading, however, is difficult to reconcile with the clear, prescriptive language of common Article 1 (‘undertake’). Moreover, some of these UNGA resolutions explicitly refer to an external obligation on States to ensure respect by others under common Article 1. For example, in UNGA Resolution ES-10/4, adopted in 1997, the UNGA:

Reiterate[d] its recommendation to the High Contracting Parties to the Geneva Convention relative to the Protection of Civilian Persons in Time of War, to take measures on a national or regional level, in fulfilment of their obligations under article 1 of the Convention, to ensure respect by Israel, the occupying Power, of the Convention, as well as its recommendation to Member States to actively discourage activities that directly contribute to any construction or development of Israeli settlements in the Occupied Palestinian Territory, including Jerusalem, as these activities contravene international law.⁸¹

Other UNGA resolutions refer positively to the Wall advisory opinion of the International Court of Justice (ICJ), in which the same external, obligatory reading was given to common Article 1.⁸²

Through their explicit endorsement by large majorities in these UNGA resolutions, States have consistently reaffirmed that common Article 1 contains external obligations requiring all States to respond to and attempt to halt IHL violations committed by others. It is well established that States’ voting patterns in the UNGA can constitute subsequent practice for the purposes of treaty interpretation.⁸³ This is accepted in scholarship⁸⁴ and in the practice of international tribunals.⁸⁵ The limits on what is to be admitted as subsequent practice come from the other elements of Article 31(3)(b) of the VCLT, ie that practice must be ‘in the application of the treaty’ and must establish ‘the agreement of the parties regarding its interpretation’.⁸⁶ The former is clearly met, given the explicit reference to common Article 1 in the above resolutions and the fact that the very act of condemning IHL violations or collectively demanding non-assistance in such resolutions might partly fulfil the obligation to ensure respect.

As regards the requirement of ‘agreement of the parties’, not all States parties need actively participate in the particular practice but may instead be presumed to acquiesce through silence ‘when the circumstances call for some reaction’.⁸⁷ Such circumstances include voting in particular fora where non-objecting States are aware of the position of others, particularly where that position has a bearing on their own rights and obligations.⁸⁸ Abstentions to UNGA resolutions recognizing external obligations for all States under common Article 1 can, therefore, reasonably be counted as practice supporting those external obligations where abstaining States have not otherwise dissented.⁸⁹ Whether the few negative votes, together with other dissenting practice, nonetheless prevents an interpretive ‘agreement’ from arising is explored in detail below.

States have continued to express support for external obligations under common Article 1 in other fora also, including the UN Human Rights Council (HRC). In Resolution 52/35, adopted in April 2023, the HRC reminded ‘all States of their legal obligations as mentioned in the advisory opinion of the ICJ of 9 July 2004 … including … to ensure compliance by Israel with international humanitarian law as embodied in the Fourth Geneva Convention’.⁹⁰ Member States of the European Union (EU) similarly endorsed the external reading of common Article 1 in the ‘User's Guide’ to its Common Position on arms exports.⁹¹ The limited membership of the HRC and EU means that these particular outputs do not themselves represent sufficient practice for interpreting common Article 1 (and one arguably could not expect non-Member States to respond should they object).⁹² However, such examples from the HRC and EU offer further support for the external obligation alongside the UNGA resolutions.

In addition to confirming their interpretation of common Article 1 through resolutions and other outputs of different institutions, a number of States also continue to do so in intergovernmental debates. As already noted, the recently concluded work of the ILC on the protection of the environment in armed conflict, for example, was a catalyst for some States to express their views concerning the content of common Article 1.⁹³ For example, Switzerland and Palestine expressed support for the view that it contains external obligations for all States.⁹⁴ Similarly, during a January 2022 UN Security Council (UNSC) debate concerning the protection of civilians in urban conflict, Sweden, Denmark, Finland, Iceland and Norway jointly expressed their position as follows:

We share the concerns of the Secretary-General and call upon all parties to armed conflict to prevent civilian harm resulting from the use in populated areas of explosive weapons, particularly those with wide-area effects … That increasingly pressing problem deserves the full attention of the Security Council and adequate monitoring of the implementation of the relevant Security Council resolutions. The Council and the international community as a whole have a shared responsibility to fully uphold and respect international law, including international humanitarian law and international humanitarian principles. We recall our joint obligation to respect and ensure respect for international humanitarian law, as enshrined in article 1 common to the Geneva Conventions.⁹⁵

Here, UNSC monitoring of armed conflicts is clearly seen in the context of the obligation of all States under common Article 1, which is thus read as including an external element. During a May 2021 UNGA debate on the situation in the Middle East, Azerbaijan, on behalf of the Coordinating Bureau of the Non-Aligned Movement, similarly stated:

The Movement calls for action to ensure accountability for all of Israel's violations against the Palestinian people. Israel's continued non-compliance with the law warrants collective action in line with various obligations under international law, including article 1 of the Fourth Geneva Convention, on respecting and ensuring respect for the Convention in all circumstances.⁹⁶

In a June 2020 UNSC debate on children in armed conflict, Peru also argued that common Article 1 obliges all States and the UNSC to adopt measures aimed at ‘revers[ing] the trend’ of IHL violations involving children.⁹⁷ Belgium too expressed strong support for there being an external obligation on all States to ensure respect for IHL in an August 2019 UNSC debate on IHL.⁹⁸ These are just recent examples in a long history of debates during which States express support for external obligations under common Article 1.⁹⁹

Finally, States have also expressed support for the external obligation outside particular fora or debates. For example, Armenia appeared to do so when it submitted reports from the human rights ombudsman and Ministry of Foreign Affairs of the so-called Republic Artsakh to the UN Secretary-General (requesting that they be forwarded to Member States) in relation to its 2020 conflict with Azerbaijan. Both reports referred to external obligations of the international community under common Article 1 to ensure respect for IHL by Azerbaijan.¹⁰⁰

In contrast to this widespread, repeated support over time for the existence of an external obligation under common Article 1, negative practice that unequivocally rejects this has been relatively rare and is, partly, a recent phenomenon. Again in the context of the ILC's work on the protection of the environment in armed conflict, Canada, Israel, the United States (US) and United Kingdom (UK) argued against there being an external element to common Article 1.¹⁰¹ Canada¹⁰² and the US¹⁰³ have expressed this same view elsewhere in recent years, with Israel apparently having taken this position for some time now.¹⁰⁴ Yet there is some inconsistency, with Canada,¹⁰⁵ the US,¹⁰⁶ Israel¹⁰⁷ and the UK¹⁰⁸ at other times all expressing support for the external obligation under common Article 1.

Australia's practice is more equivocal, but it might still be considered a dissenting State. It appeared to reject the notion of an external obligation under common Article 1 when responding to an ICRC questionnaire in 1973 and has abstained from voting on many of the UNGA resolutions cited above.¹⁰⁹ More recently, Australia seems to have rejected the purported positive obligations under the external element of common Article 1.¹¹⁰ It has, however, been inconsistent, voting in favour of some earlier resolutions invoking common Article 1 in the Israel–Palestine conflict.¹¹¹ Even so, it is clear that only a small minority of States explicitly dissent from the proposition that common Article 1 contains external obligations applicable to all States in relation to the conduct of others. This interpretation of common Article 1 thus has widespread support.

Finally, it has been argued that the failure of States to allege that other States which do not respond to breaches of substantive IHL rules are themselves in breach of common Article 1 is an indication that it does not, in fact, include external obligations.¹¹² However, States frequently do condemn alleged violations of IHL in specific conflicts and call on the parties to comply with IHL.¹¹³ Indeed, as the many examples above confirm, common Article 1 is specifically invoked by States when calling on others to respond to IHL violations.

More generally, there are various political or strategic reasons why States who are silent in the face of substantive IHL violations are not censured for their own apparent breach of common Article 1. As noted above, silence can constitute acquiescence (and thus subsequent practice) ‘when the circumstances call for some reaction’.¹¹⁴ Drawing an analogy with custom, such circumstances might include cases where practice (or the lack of practice condemning substantive IHL violations) ‘affects interests and rights of an inactive state’.¹¹⁵ This is unlikely to be the case here. Whilst it is often said that common Article 1 (and other IHL rules) create obligations owed erga omnes partes, this merely creates a legal right to invoke the wrongdoing State's responsibility;¹¹⁶ it does not mean that all States have an actual interest in the way other States respond to substantive IHL violations. Thus, one cannot reasonably assign normative weight to the failure of a State to condemn others for allegedly violating common Article 1.

B. Consent and Dissent in Treaty Interpretation

There is therefore widespread and long-standing practice at least in support of a positive, external obligation under common Article 1 requiring States to respond to and seek the cessation of violations of the Conventions.¹¹⁷ But how far might the existence of the contrary, dissenting practice noted above affect the extent to which this supportive practice can influence the interpretation of common Article 1?

The orthodox reading of Article 31(3)(b) of the VCLT is that practice is inadmissible thereunder if there is explicit dissent, given its requirement that subsequent practice ‘establish[] the agreement of the parties’.¹¹⁸ As Georg Nolte, the ILC's special rapporteur on subsequent agreements and subsequent practice, put it in his Hague Lectures, subsequent practice under Article 31(3)(b) embodies ‘the will of all parties to a treaty. Thus, a practice by only one party, or even the practice of almost all parties, is not subsequent practice under Article 31, paragraph 3.’¹¹⁹ Indeed, at the time of drafting what became Article 31(3)(b), the ILC viewed it as referring to practice that reflects the understanding of ‘the parties as a whole’, justifying its inclusion amongst the authentic means of interpretation on the basis that ‘it constitutes objective evidence of the understanding of the parties as to the meaning of the treaty’.¹²⁰ On this basis, the widespread practice supporting the external element of common Article 1 would not be admissible under Article 31(3)(b) due to the inconsistent dissenting practice.¹²¹

Notwithstanding the considerable weight placed on subsequent practice in other studies supporting the external element to common Article 1, most fail to acknowledge this potentially fatal methodological flaw.¹²² The few that do acknowledge this problem adopt the approach of the ILC in its draft conclusions on subsequent agreements and subsequent practice.¹²³ Here, the ILC argued that subsequent practice that does not establish the agreement of all of the parties to the treaty (eg where some dissent is present) can still be admissible as a supplementary means of interpretation under Article 32 of the VCLT.¹²⁴ Two key consequences follow from this. First, as the ILC argues, if practice is admissible under Article 32 of the VCLT, it would be accorded less interpretive weight as a supplementary means of interpretation than it would as an authentic means under Article 31(3)(b).¹²⁵ Secondly, in contrast to Article 31(3)(b), practice under Article 32 could only be drawn on to confirm an interpretation reached by the other authentic means, or where those other means yield an unclear or manifestly absurd result.¹²⁶

The ILC's bifurcation of subsequent practice into Articles 31(3)(b) and 32 finds some support amongst scholars.¹²⁷ Indeed, during the drafting of what eventually became Articles 31 and 32 of the VCLT, the ILC alluded to the possibility that practice establishing the agreement of only some parties concerning the interpretation of a treaty provision might be considered a supplementary means under draft Article 70 (which would become Article 32).¹²⁸ By reinvoking this argument in its recent work on subsequent practice, the ILC sought to explain the tendency of many international courts and tribunals to refer to State practice as an interpretive aid even where that practice does not represent the agreement of all States parties.¹²⁹ Sometimes, one can infer that this is the sense in which a tribunal relies upon subsequent practice, for example, when it is invoked to confirm an interpretation reached by other means.¹³⁰ In many cases, however, including most of those on which the ILC relies, international courts and tribunals appear to rely on subsequent practice that is explicitly not accepted by all parties pursuant to Article 31(3)(b) of the VCLT.

In Loizidou v Turkey, for example, the question for the European Court of Human Rights (ECtHR) was whether Articles 25 and 46 of the European Convention on Human Rights, providing for acceptance of the jurisdiction of the Commission and Court, permitted territorial restrictions of the type found in Turkey's declarations. The Court referred to the practice of the vast majority of Convention parties of not including territorial restrictions in their Article 25 and 46 declarations as ‘confirm[ing]’¹³¹ the interpretation that was ‘strongly support[ed]’¹³² by other factors, such that territorial restrictions were not permitted. Though this might suggest that the Court was referring to such practice as supplementary means under Article 32 of the VCLT, its stated methodology indicated that it took account of subsequent practice only in the sense of Article 31(3)(b).¹³³ This was so, even though the practice clearly did not reflect an interpretation shared by all parties. Importantly, the respondent State, to whom the Court's interpretation was opposable, explicitly rejected it, and had done so consistently in its own practice (in the form of the territorial restrictions contained in its declarations). The Court did not think this prevented it from relying on the (lack of) practice of the other Contracting Parties, apparently on the basis of Article 31(3)(b). For the Court, Turkey's contrary practice, together with the unclear practice of two other States, ‘do not disturb the evidence of a practice denoting practically universal agreement amongst Contracting Parties that Articles 25 and 46 … do not permit territorial or substantive restrictions’.¹³⁴

In other cases, the ECtHR has also drawn on general, though not universally supported, practice of Contracting States as evidence of a ‘European consensus’,¹³⁵ apparently considering it admissible under Article 31 of the VCLT. Thus, in Magyar Helsinki Bizottság v Hungary, a case concerning the right to information, the Court stated that ‘[t]he consensus emerging from specialised international instruments and from the practice of Contracting States may constitute a relevant consideration for the Court when it interprets the provisions of the Convention in specific cases’, considering this as practice under Article 31 of the VCLT, and looking separately at ‘supplementary’ means in the following paragraph.¹³⁶ The Court concluded that, ‘since the Convention was adopted the domestic laws of the overwhelming majority of Council of Europe member States … along with the relevant international instruments, have indeed evolved to the point that there exists a broad consensus, in Europe (and beyond) on the need to recognise an individual right of access to State-held information’.¹³⁷ Importantly, this interpretation was rejected by Hungary (the respondent) and, as intervenor, the UK. The ECtHR thus takes a broader view of subsequent practice under Article 31(3)(b) than is permitted by the orthodox account.¹³⁸ This seems to be a consistent approach of the Strasbourg Court.¹³⁹

The Inter-American Court of Human Rights (IACtHR) has also explicitly relied on widely shared practice amongst States parties pursuant to Article 31(3)(b), even in the face of some minority dissent. In Artavia Murillo v Costa Rica, the Court referred to the ‘generalized practice’ of ‘most’ States parties permitting in vitro fertilization (IVF) as a basis for concluding that Article 4(1) of the American Convention on Human Rights on the right to life should not be interpreted as prohibiting IVF, referring explicitly to Article 31(3)(b) of the VCLT:

The Court considers that, even though there are few specific legal regulations on IVF, most of the States of the region allow IVF to be practiced within their territory. This means that, in the context of the practice of most States Parties to the Convention, it has been interpreted that the Convention allows IVF to be performed. The Court considers that this practice by the States is related to the way in which they interpret the scope of Article 4 of the Convention, because none of the said States has considered that the protection of the embryo should be so great that it does not permit assisted reproduction techniques and, in particular, IVF.¹⁴⁰

Again, this interpretation was opposable to the respondent State, Costa Rica, whose own practice in prohibiting IVF on the basis of its alleged violation of the right to life was inconsistent with this majority position.¹⁴¹

The International Criminal Tribunal for the former Yugoslavia (ICTY) also at times took a much broader view of admissible subsequent practice under Article 31(3)(b) than the orthodox approach would permit. In its 1999 Tadić judgment, for example, the Appeals Chamber, when interpreting the requirement in Article 4A(2) of the Third Geneva Convention (GC III) that paramilitary groups must ‘belong[] to a party to the conflict’ to benefit from prisoner of war status, placed decisive weight on its reading of State practice, which it considered to set a ‘control’ test for this purpose.¹⁴² It did not, however, examine the extent to which State practice converged here and referred only to a single domestic court judgment as evidence.¹⁴³ Nonetheless, in other cases the Appeals Chamber has been stricter in its application of Article 31(3)(b).¹⁴⁴

This practice of the ECtHR, IACtHR and ICTY might be viewed as representing a sui generis approach to treaty interpretation in the context of international tribunals performing a broader rule of law function, in contrast to traditional inter-State dispute settlement tribunals.¹⁴⁵ However, the more liberal approach to Article 31(3)(b) is also to be found in some ICJ jurisprudence. Thus in the Namibia advisory opinion, when interpreting Article 27 of the UN Charter, the Court placed decisive (indeed, exclusive) weight on the consistent practice of the (limited) UNSC membership whereby abstention by a permanent member was not treated as barring the adoption of a resolution.¹⁴⁶ Although it referred to this practice as being ‘generally accepted by Members of the United Nations’, the Court cited no evidence indicating the acceptance of all members, as one would have expected if the Court was adhering to the orthodox reading of Article 31(3)(b).¹⁴⁷ Similarly, in the Wall advisory opinion, the Court recognized a change in the interpretation to be given to Article 12(1) of the Charter arising from subsequent practice, such that the UNGA and UNSC could deal with matters simultaneously.¹⁴⁸ The Court relied, again decisively, on UNGA resolutions that had received a number of negative votes.¹⁴⁹ The decisive weight given to subsequent practice in these cases (and the automatic recourse thereto) suggest that the Court drew on it pursuant to Article 31(3)(b), rather than Article 32.

Various explanations have been given for these ICJ cases. Arato views them (and the similar approach of the ECtHR) as examples of international courts interpreting their own constituent instruments, though the different approach of the World Trade Organization Appellate Body militates against generalizing this argument.¹⁵⁰ Whilst this might explain the stricter approach of the ICJ in Whaling in the Antarctic,¹⁵¹ there is nothing in the Court's (or ECtHR's) reasoning that implies such a distinction, and it is not clearly grounded in principle. It has also been argued that the distinctive feature of these cases is that they invoke the practice of organs of international organizations (not State practice as such), either as an element in interpreting those organizations’ constituent instruments,¹⁵² or as ‘established practice’ of the organization influencing its internal rules.¹⁵³ This may explain the similar approach of the Court in the Certain Expenses advisory opinion, in which it explicitly noted its methodology of relying on the consistent practice of UN organs,¹⁵⁴ and went on to rely on UNGA resolutions with significant negative votes.¹⁵⁵ Yet in the Namibia and Wall advisory opinions the Court's focus seemed clearly to be on the practice of States (as opposed to the practice of organs), in the form of their voting on and reaction to resolutions.¹⁵⁶

Two points are clear from this survey of case law. First, it is wrong to suggest, as have some,¹⁵⁷ that subsequent practice is inadmissible as a means of interpreting a treaty provision where there is a small amount of inconsistent practice. Secondly, and contrary to the understanding of the ILC and others,¹⁵⁸ the practice of a majority of States parties (even in the face of some explicit dissent) has often been treated by international tribunals as an authentic means of interpretation under Article 31(3)(b), rather than a supplementary means under Article 32 of the VCLT. That the ILC's approach cannot explain this practice is especially clear in those cases where the respondent was one of the dissenting States, since the interpretation reached is opposable to them. If, as the ILC (and others subscribing to the orthodox account) argue, this cannot happen under Article 31(3)(b), it is difficult to see how it can happen under Article 32.¹⁵⁹

It is submitted that this judicial practice should be accepted on its own terms and that it should be recognized that subsequent practice may be admissible as an authentic means for interpreting widely ratified multilateral treaties, even where there is a small amount of dissenting practice. Several reasons justify this. First, whilst it is sensible that universal (or widely supported) practice carries more weight than practice that is supported only by a small minority of States parties, this does not depend on a binary approach of classifying practice under Articles 31(3)(b) or 32 of the VCLT. Indeed, treating only practice from which there is no dissent as authentic, and relegating all other practice to Article 32 (thus according it less interpretive weight), gives a false impression of the weight of subsequent practice under Article 31(3)(b). The ILC itself recognizes that the weight accorded to Article 31(3)(b) practice depends on a number of factors,¹⁶⁰ even referring at one point in its commentary to the extent of consensus amongst the States parties as one such factor.¹⁶¹ Moreover, subsequent practice in the Article 31(3)(b) sense is not necessarily decisive; rather, its influence depends on its interaction with the other means of interpretation under Article 31 of the VCLT.¹⁶²

Secondly, the categorical distinction drawn by the ILC does not create a workable framework. According to the ILC, categorizing practice that does not establish the agreement of all parties to the treaty under Article 32 means that less weight must be accorded to it than is given to practice which does establish such agreement. Yet it is not known how to quantify this lesser weight.¹⁶³ This approach thus introduces additional questions, without providing any tools for discovering the answers. Indeed, when considering the factors that might influence the weight to be given to different types of subsequent practice, the ILC draft conclusions appear to draw no particular distinction between practice under Articles 31(3)(b) and 32.¹⁶⁴

The third reason in favour of a broader reading of Article 31(3)(b) relates to the role of consent in treaty interpretation. The broader reading proposed here might be objected to on the basis that the strict, orthodox account protects party intentions and the consensual nature of treaty obligations.¹⁶⁵ Whilst it is true that determining party intentions is the goal of treaty interpretation, it is an objective or presumed, common intention that is to be determined through the application of the Vienna rules.¹⁶⁶ Indeed, the move away from the pre-Vienna interpretive approach that relied heavily on subjective State will (eg the travaux) towards objective rules for determining this presumed or objectivized collective intent was key to the evolution of the ILC's project on the law of treaties and the final VCLT.¹⁶⁷ No enduring, dispositive function therefore resides in the subjective will of individual parties.¹⁶⁸ As such, to argue that the consistent practice of a large majority of parties might not provide an authentic means of interpreting widely ratified multilateral treaties due to the dissent of a small minority of States appears to endow the subjective will of those few States with considerable weight and is arguably at odds with the general approach reflected in the VCLT rules.¹⁶⁹ The requirement of an ‘agreement’ in Article 31(3)(b) should thus be read more broadly than under the orthodox account. Indeed, there does not appear to be any principled basis for determining a particular degree of ‘agreement’ that is necessary for admissibility under Article 31(3)(b); as long as there is some agreement between at least two of the States parties, it is submitted that the influence (or weight) of that practice should be determined by the extent of the agreement and its interaction with the other means of interpretation.¹⁷⁰

The problems with the ILC's approach become particularly clear when applied to the practice set out above regarding the external obligation under common Article 1. Those that apply the ILC's approach categorize the majority practice supporting the external element under Article 32, claiming to accord it less interpretive weight than practice under Article 31(3)(b).¹⁷¹ Yet no further indication of the relative weight to be given to this practice is offered and it is simply argued that ‘[o]n balance, the available practice is sufficient’ to prove the existence of the external element.¹⁷² Ultimately, subsequent practice is given decisive weight, which sits uneasily with the requirement that less weight be attached to it than if it were admissible under Article 31(3)(b). This is a necessary consequence of the ILC's binary approach, which requires lesser weight to be attached to Article 32 practice without explaining what that actually means.

Moreover, if dissent does indeed render practice inadmissible under Article 31(3)(b), it is not clear why it should be admissible under Article 32. One cannot reconcile the supportive and dissenting practice regarding external obligations under common Article 1; there is no compromise solution which allows for lesser weight to be given to the supportive practice under Article 32. The risk is that Article 32 essentially becomes a way of circumventing the stringent requirements of Article 31(3)(b), whilst claiming to show fidelity to State consent, with practice ostensibly admitted under the former having precisely the same interpretive effect as it would under the latter.

The approach proposed here avoids these methodological problems by recognizing that the requirements of Article 31(3)(b) are less stringent than the orthodox account suggests. Thus, it has been seen that practice in support of external obligations under common Article 1 is extensive and long-established, with dissenting practice constituting a small minority. It therefore seems reasonable automatically to consider such practice alongside the other authentic means of interpretation, rather than referring to it only if the conditions of Article 32 are met.¹⁷³ Moreover, treating such practice as admissible under Article 31(3)(b), as opposed to under Article 32, means the starting point need no longer be that such practice somehow carries less weight. Instead, this approach allows for a more nuanced consideration of its weight, based on the usual factors, including the level of support or dissent and the interaction with the other authentic means of interpretation.

Given the widespread nature of the supporting practice and the relatively recent (and inconsistent) nature of the dissenting practice, there does appear to be broad agreement in favour of external obligations under common Article 1. As regards the interaction of this practice with the other means of interpretation, the weight assigned to subsequent practice may be influenced by its consistency with those other means of interpretation, meaning that practice which is inconsistent with the object and purpose of a treaty, for example, might be expected to require the support of all parties to the treaty if it is to influence interpretation.¹⁷⁴ As shown above, an external obligation under common Article 1 that is supported by subsequent practice is fully consistent with the object and purpose of the Conventions (as well as with the other authentic means of interpretation).¹⁷⁵ In the light of this, the most persuasive interpretation of common Article 1 is one that recognizes that it contains external obligations relating to the conduct of others.

This argument means that it is not strictly necessary to consider the drafting history of common Article 1 or Article 1(1) of AP I. However, the following section will investigate the travaux given that some claim that there was an original, restrictive meaning to common Article 1. Such an original meaning might be another factor influencing the weight of non-universal subsequent practice (in contrast to universal practice).

IV.

TRAVAUX PRÉPARATOIRES

Previous analyses of common Article 1 tend to conclude that the travaux either do not reveal any intention either way regarding an external element to that provision,¹⁷⁶ or that they demonstrate that such an obligation was explicitly rejected by the negotiating States (with the words ‘ensure respect’ referring to obligations vis-à-vis a State's own population).¹⁷⁷ As regards the travaux of Article 1(1) of AP I, this question is often ignored altogether.

The draft of common Article 1 received very little attention at the 1949 diplomatic conference, and the few interventions that were made do not establish a clear consensus for or against an external obligation. Italy's delegate to the conference argued that ‘the terms “undertake to ensure respect” should be more clearly defined. According to the manner in which they were construed, they were either redundant, or introduced a new concept into international law.’¹⁷⁸ Moreover, whereas the delegates for Norway, the US, France and Monaco (the only other States to comment) said that they ‘considered that the object of this Article was to ensure respect of the Conventions by the population as a whole’,¹⁷⁹ the ICRC delegate noted in reply that ‘the Contracting Parties should not confine themselves to applying the Conventions themselves, but should do all in their power to see that the basic humanitarian principles of the Conventions were universally applied’.¹⁸⁰ Though there is some disagreement as to what was meant here by ‘universal application’, it seems that it was used in contradistinction (and in addition) to application within a specific State and thus referred to promoting compliance by others.¹⁸¹

As is often the case, the travaux do not resolve the conundrum of how to interpret common Article 1.¹⁸² There is value, nonetheless, in noting this, as it demonstrates, contrary to the claims of some, that there was no ‘original [restrictive] meaning’ of common Article 1 agreed by the parties.¹⁸³

The original ICRC commentaries on the Geneva Conventions, which might be seen as an elaboration of the official travaux,¹⁸⁴ offer some further support for the external element. Thus, the Commentaries on the First, Second and Fourth Conventions state that it follows from the undertaking to ‘ensure respect’, ‘that in the event of a Power failing to fulfil its obligations, the other Contracting Parties (neutral, allied or enemy) may, and should, endeavour to bring it back to an attitude of respect for the Convention’.¹⁸⁵ Whilst this has been read as recommendatory, rather than binding,¹⁸⁶ the Commentary on GC III omits the word ‘may’ altogether: ‘each of the other Contracting Parties (neutral, allied or enemy) should endeavour to bring it back to an attitude of respect for the Convention’.¹⁸⁷ Even though these references reflect the ICRC's endorsement of an external element, rather than that of States, they do indicate the crystallization of the idea at the time.¹⁸⁸

In the years between the adoption of common Article 1 and its restatement in Article 1(1) of AP I, two developments confirmed the growing acceptance of there being an external dimension to the undertaking to ‘ensure respect’. First, common Article 1 was no doubt what States had in mind when they adopted the preambular paragraph to Resolution XXIII at the 1968 Tehran Conference on Human Rights, which stated that ‘States parties to the Red Cross Geneva Conventions sometimes fail to appreciate their responsibility to take steps to ensure the respect of these humanitarian rules in all circumstances by other States, even if they are not themselves directly involved in an armed conflict.’¹⁸⁹ Whilst this language has been said by some to be equivocal (‘responsibility’ as opposed to ‘duty’),¹⁹⁰ its adoption prior to the reiteration of the obligation in Article 1(1) of AP I is instructive.

Secondly, even more fundamental for understanding what was in the minds of the delegates when adopting Article 1(1) of AP I are the responses of States to an ICRC questionnaire in 1973 concerning means of encouraging compliance with the Conventions, issued following a recommendation by the first session of the 1971 Conference of Government Experts, and which informed the 1974–1977 diplomatic conference.¹⁹¹ Many of the responses from States made clear that not only was the external element of common Article 1 taken for granted, but it was also considered to be a legal obligation binding all contracting parties; this was the case even for States that more recently have rejected this interpretation.¹⁹²

The travaux of AP I itself have either been overlooked in previous scholarship on the subject or noted only in passing, given the very limited substantive discussion of Article 1(1) at the 1974–1977 diplomatic conference.¹⁹³ However, a number of delegates at other points in the negotiations gave clear indications of the measures that common Article 1 (and Article 1(1) of AP I) required. In particular, when discussing drafts of what became the ‘external’ enforcement provisions in AP I,¹⁹⁴ some delegates made it clear that they considered these to be methods for ‘ensuring respect’ for the Conventions and Protocol.¹⁹⁵ Indeed, the delegate of Pakistan explicitly stated that common Article 1 imposed on States parties an obligation to ensure respect for the Conventions by others: ‘Article 1 common to the four Geneva Conventions of 1949 … implied that, if a Party failed to carry out its obligations, the other Contracting Parties were bound to endeavour to bring it back to an attitude of respect for its engagements.’¹⁹⁶ Elsewhere, the Pakistan delegate referred approvingly to the ICRC's commentary on common Article 1 (though without specifying which commentary) when explaining Pakistan's own detailed proposal for a permanent commission,¹⁹⁷ which would investigate alleged violations by States and ‘endeavour to bring back to an attitude of respect for and obedience to the provisions of the Conventions and this Protocol’.¹⁹⁸ No other delegate appeared to reject this view. The idea that ‘ensure respect’ contains an external dimension thus appears to have been assumed by many States during the drafting of these new enforcement provisions of AP I.

This reading of the travaux of AP I is supported in the ICRC Commentary on the Additional Protocols, which quotes the more resolute language from the GC III Commentary.¹⁹⁹ Some still maintain that the use of the word ‘should’ (as opposed to ‘shall’) in that Commentary means that the ‘ensure respect’ element of common Article 1 and Article 1(1) of AP I is recommendatory, in contrast to the obligatory ‘respect’ element.²⁰⁰ However, not only is this inconsistent with the use of the word ‘undertake’ in relation to both elements, but the ICRC Commentary on the Additional Protocols explicitly referred to the ‘ensure respect’ element of Article 1(1) of AP I as a ‘duty’ for contracting States.²⁰¹

These findings relate to the earlier discussion of subsequent practice in two ways. First, the travaux of common Article 1 show that there was never an original, restrictive meaning to the provision.²⁰² Subsequent practice under Article 31(3)(b) of the VCLT can certainly lead to a change in the meaning of a treaty term from that originally intended.²⁰³ However, it might be argued that where subsequent practice is accompanied by some dissenting practice which reflects the original meaning and has been consistent over time, then that subsequent (majority) practice should have less weight.²⁰⁴ In the case of common Article 1 there is no such original meaning,²⁰⁵ and thus the value of the majority practice is undiminished.

Secondly, it is clear that the idea of an external obligation, at least in the sense of an obligation to respond to violations, became increasingly accepted by many States in the period leading up to the adoption of Article 1(1) of AP I, including States that have more recently rejected that interpretation. The travaux may thus be taken as confirming the interpretation of Article 1(1) of AP I to which the other, authentic means point.²⁰⁶

V. APPLICATION IN RELATION TO NON-INTERNATIONAL ARMED CONFLICTS

Before turning to the precise content of the external element of common Article 1, attention must be paid to the other key controversy concerning its interpretation, namely its application in relation to non-international armed conflicts.²⁰⁷ This is important given the contemporary prevalence of such conflicts.²⁰⁸

The text of common Article 1 certainly appears comprehensive, calling on States to ‘respect and ensure respect for the present Convention in all circumstances’ (emphasis added), suggesting no limit to the substantive provisions or situations to which the obligation attaches. Moreover, its place as the first provision, followed by common Articles 2 (international armed conflicts) and 3 (non-international armed conflicts) supports a broad reading of its scope of application. In addition, extending common Article 1 (including its external dimension) to non-international armed conflicts is clearly consistent with the humanitarian purpose of the Conventions, offering a means of enforcement in relation to what are today the prevalent types of armed conflict.²⁰⁹ Indeed, that purpose would be frustrated if the obligation to ensure respect does not apply in the majority of armed conflicts.

However, whilst there is ample subsequent practice in support of an external obligation under common Article 1, most of this concerns international armed conflicts (particularly, though not exclusively, the Israeli occupation of Palestinian territory). Nonetheless, there is some State practice supporting the applicability of the external obligation under common Article 1 to all armed conflicts.²¹⁰ Practice rejecting this is also very limited. It was in the context of the non-international armed conflict in Yemen that both the Canadian government and the Federal Court of Canada appeared to express the view that common Article 1 did not apply to such conflicts.²¹¹ Yet the Court's conclusion here was unreasoned, and both the government and Court (citing prior Canadian case law) appear to have recognized that there might be circumstances in which common Article 1 imposed obligations on States concerning the conduct of others in non-international conflicts.²¹²

The supporting practice here is certainly more limited and less well established than in the case of international armed conflicts. Whilst explicit negative practice is minimal, it is unclear whether the silence of many States amounts to acquiescence in the supporting practice (and thus constitutes supporting practice itself).²¹³ The degree of ‘agreement’ (in the Article 31(3)(b) sense) reflected in this practice thus appears to be considerably less than in the general practice supporting the external element. Importantly, however, the application of common Article 1 in relation to internal armed conflicts clearly furthers the object and purpose of the Conventions, and this might help to compensate for the more limited supporting practice.

The travaux shed little light on this point, as the possibility of common Article 1's application in non-international armed conflicts did not seem to occur to the delegates at the 1949 diplomatic conference. Indeed, the travaux indicate that the reference to ‘in all circumstances’ in common Article 1 was a vestige of the 1929 Conventions, intended to reinforce the omission of the clausula si omnes,²¹⁴ and to emphasize that the Conventions applied in both peace and war.²¹⁵ Nonetheless, it does appear that common Article 1 was inspired in part by the ICRC's desire leading up to the 1949 conference to see States encourage their populations to respect IHL, thereby promoting its application in civil wars.²¹⁶ This suggests that the application of common Article 1 in non-international armed conflicts might have some basis in its drafting history.²¹⁷ The absence of an identical provision in Additional Protocol II (AP II) (in contrast to AP I) does not, as some have suggested,²¹⁸ contradict this, given that the AP II draft was significantly reduced in scope at the eleventh hour of the 1974–1977 diplomatic conference in order to avoid the negotiations collapsing.²¹⁹ As such, one cannot place much weight on the absence of particular provisions from AP II.²²⁰

To conclude, it is not immediately clear from State practice that common Article 1 extends to non-international armed conflicts. However, the text and context of the provision, together with the object and purpose of the Conventions, all point in favour of it doing so, with some practice supporting this. An arguable case can thus be made for this interpretation. This is also supported by the ICJ and others.²²¹ To be clear, the applicability of common Article 1 to non-international armed conflicts does not affect the scope of application of substantive provisions of the Conventions.²²² It simply means that States must respect the provisions of the Conventions themselves and ensure that they are respected by others in accordance with their defined scope of application (which, in the context of non-international armed conflicts, is limited to common Article 3).²²³ What is more, and as explained below, the due diligence standard by which the positive obligations under common Article 1 must be defined, means that those obligations would often easily be discharged in non-international armed conflicts for third States that have no involvement therein.²²⁴

VI. CONCLUDING REMARKS

This article has addressed some significant limitations with the existing literature on common Article 1 and the purported external obligation of States parties to ensure respect for the Geneva Conventions and AP I/III by others. The most significant concerned the decisive weight placed on subsequent practice by those studies supportive of the external obligation, notwithstanding the dissent of some States. These previous studies have failed to justify the relevance placed on such practice, given that orthodox accounts of subsequent practice require that it establishes the agreement of all States parties. This is partly due to the inadequacy of attempts to rationalize this issue more generally in international law, as exemplified by the ILC's bifurcation of practice into Articles 31(3)(b) and 32 of the VCLT. The major part of this article has engaged squarely with this general question of treaty interpretation, offering a principled account of the interpretive relevance of subsequent practice that includes some dissent, according to which such practice can still be admissible under Article 31(3)(b), with its weight then determined by the usual factors, including its consistency and interaction with the other means of interpretation. Importantly, it has been shown that this approach better fits and justifies existing jurisprudence than the ILC's approach, allowing for a more nuanced analysis of interpretive weight. This novel approach is important for treaty interpretation generally, beyond common Article 1.

The second key contribution involved reassessing the travaux of common Article 1 and Article 1(1) of AP I. Previous studies suggest that there was an original, restrictive meaning to common Article 1, with the drafting history of AP I rarely considered in detail. Section IV has challenged these accounts, demonstrating that no such original restrictive meaning existed and, importantly, that there was clear support for the external element during the drafting of AP I. This was shown to be important, particularly in relation to the interaction between such findings from the travaux and the other means of interpretation.

This article has therefore placed the purported external obligation under common Article 1 on a much firmer footing. There remains, however, the question of what exactly is required of States under the external obligation.²²⁵ The ICRC takes a taxonomic approach, arguing that the ‘ensure respect’ element of common Article 1 contains both negative (no encouragement/aid/assistance) and positive (prevention and response) obligations.²²⁶ Given that the external obligation is grounded heavily in subsequent practice, its precise content should ideally be manifest in that practice.²²⁷ Most of the recent practice explored supports the claim that common Article 1 requires all States to take action in the face of ongoing violations of the Conventions with a view to cessation, but the specific context of much of that practice means that the other purported elements are often not mentioned.

Moreover, whilst this obligation to respond to violations was, as noted above, supported in the travaux of AP I, the purported negative obligations, and the obligation of prevention, were apparently not considered. These additional negative and positive obligations certainly seem consistent with the object and purpose of the Conventions and Protocol.²²⁸ Indeed, the obligation not to aid or assist in violations seems to follow a fortiori from the obligation to respond to and bring to an end violations.²²⁹ And whilst much of the practice referred to concerns responses to specific violations, some does provide support for a broader set of obligations under common Article 1.²³⁰

As an alternative to seeking support in practice for each purported element of the external obligation under common Article 1, its content could instead be framed by generalizing from the specific instances of practice identified, in a manner similar to the inductive method typically applied when identifying rules of custom.²³¹ On this basis, it could be argued that practice offers general support for the existence of ‘external’ obligations of States under common Article 1 (and Article 1(1) of AP I/III) to ensure respect for IHL by other actors. Rather than attempting to delimit the specific elements of this obligation, one might instead accept that its content remains open and that the obligation to ‘ensure respect’ might be engaged by a number of different actions (or omissions), both forward-looking (preventive) and backward-looking (responsive), depending on the particular context.²³² Moreover, as noted, many of the policy and pragmatic concerns to which the positive (prevention and response) elements of the obligation give rise could be mitigated by their context-sensitive nature as obligations of conduct (whose content is determined by a due diligence standard).²³³ How onerous such obligations will be therefore depends on a number of factors. Where a State is a co-party to a conflict with another State or non-State actor, for example, their positive obligations will justifiably require more of them in relation to the conduct of their co-parties than would be required of a third State adopting a posture of neutrality.²³⁴

These negative and positive obligations are familiar to international lawyers, and yet the interpretation endorsed in this article goes beyond other primary and secondary rules that arise for States in relation to the conduct of others. The idea of negative obligations prohibiting aid or assistance mirrors similar obligations found, for example, in the Chemical Weapons Convention,²³⁵ and in the secondary rules on State Responsibility.²³⁶ Yet common Article 1 is broader than both in various ways. For example, it extends to all obligations under the Conventions (in contrast to the narrower aid/assistance prohibitions under specific primary rules), and it places obligations on States in relation to the conduct of other States and non-State actors (in contrast to Article 16 of the ILC Articles, which is generally seen as applying only in the inter-State context).²³⁷ The latter is especially (though not only) important in non-international armed conflicts, where an arguable case can be made that common Article 1 applies.

Positive obligations of prevention and response are similarly found in other rules, including human rights treaties and under the special regime of responsibility for serious breaches of peremptory norms set out in Article 41 of the ILC Articles on State Responsibility.²³⁸ Again, however, the positive obligations that might be required in particular circumstances under common Article 1 go beyond these existing rules. Common Article 1 is not limited by jurisdiction clauses as are human rights treaties,²³⁹ and it applies in relation to any rule under the Conventions and not only those considered to contain the most important obligations.

That common Article 1 places obligations on States in relation to the conduct of others thus represents a significant addition to existing rules of international law. As explained at the outset, States increasingly assist partners in the planning and execution of military operations in a myriad of ways. It is essential, now more than ever, that they consider these potential avenues of liability when making decisions on military cooperation.