Editor’s Note: The U.S. Department of Defense’s (DoD) Law of War Manual, first published in 2015, sets out the DoD’s authoritative interpretation of international law on matters relating to armed conflict. In recognition of its size, breadth, and the evolving nature of law and conflict, the Manual is also a living document that anticipates needing revisions and updates. In A Living Document, published by Just Security and the Reiss Center on Law and Security, leading scholars and practitioners reflect on how the Manual is keeping pace with its stated purpose and examine specific topics and rules where the Manual may be ripe for clarification or strengthening.
Introduction
Article 4(1) of the Fourth Geneva Convention (hereinafter – GCIV) defines the status of Protected Persons as follows:
Persons protected by the Convention are those who at a given moment and in any manner whatsoever, find themselves, in case of a conflict or occupation, in the hands of a Party to the conflict or Occupying Power of which they are not nationals.
Section 10.3.2.2 of the DoD Manual, which is part of the Manual’s section defining Protected Persons, reads as follows:
To be entitled to the protections provided under the GC for “protected persons,” a person must be located in either (1) occupied territory or (2) the home territory of a party to the conflict. The references in the GC that limit the scope of protections for protected persons to those “in the territory of a Party to the conflict” refer to the home territory of the party to the conflict in whose hands the protected person finds himself or herself.
Section 10.3.2.2 thus seems narrower in scope than the text of Article 4 of GCIV, which on plain reading, does not require that Protected Persons must be either in occupied territory or in the home territory of a Party (hereinafter – the geographic nexus requirement). The status of Protected Persons is important, since the extensive protections afforded according to the Geneva Conventions apply mainly to this category, rather than the more general category of “civilians,” as defined in the Additional Protocols. Acceptance of the narrower version has significant implications for the enforcement of the Geneva Conventions under the grave breaches regime and for the scope of substantive obligations – including rules applicable to invading military forces and tactics such as the forcible transfer of civilians to other countries.
As the key source for the geographic nexus requirement, the Manual relies on an Office of Legal Counsel Opinion by Assistant Attorney General Jack Goldsmith from 2004 (hereinafter – The OLC Opinion). The Opinion concedes that the geographic nexus requirement “does not emerge from Article 4” directly, but rather, can be derived from other provisions of GCIV (p. 6). In particular, the requirement is drawn from the language of the titles of sections of GCIV which deal with Protected Persons: for instance, Part III, Section I of GCIV, addresses “Provisions Common to the Territories of the Parties to the Conflict and to Occupied Territories”; and Section II – which includes inter alia provisions relating to internees –is entitled “Aliens in the Territory of a Party to The Conflict.”
In this article, I first outline the potentially harmful implications of the geographic nexus requirement, and argue that although they can be alleviated somewhat through interpretation, such an interpretation itself would require amendments to the DoD Manual. Then, I argue that the geographical nexus requirement, in fact, is extremely difficult to defend in terms of the lex lata, legal policy, literature or practice.[1]
This piece does not, however, discuss in detail the obligations in fact owed to Protected Persons in situations or territories once the geographic nexus requirement is rejected. It suffices to suggest, for our purposes, that obligations from GCIV may apply in a straightforward, functional manner: protections attach to such individuals commensurate with the level of control a party to the conflict exercises over those persons.
Three Implications and Possible Mitigation
In my view, there are at least three dangerous implications of the geographic nexus requirement.[2] The most immediate one is the potential exclusion from protection under GCIV of persons that find themselves “in the hands” of the rival belligerent during the invasion stage, or otherwise before occupation commences.[3] This problem is accentuated since states almost always deny the existence of occupation, and since there is a wide margin of doubt and disagreement on the point at which occupation commences.[4]
The second implication is a that the geographic nexus requirement might be used to argue that once a Protected Person is removed from the aforementioned territories to a third territory – whether controlled by that Party or otherwise – Protected Person status under GCIV is negated. A belligerent could unilaterally obviate the status of Protected Persons, and absolve its personnel from criminal liability under GCIV for their subsequent mistreatment, simply by crafting ways to transfer them to a third territory it controls or to a third-party non-belligerent state.[5] This problem is aggravated by the DoD Manual’s emphasis that the phrase “in the territory of a Party to the conflict,” as it appears in GCIV, refers “to the home territory of the party to the conflict in whose hands the protected person finds himself or herself.”[6] This would mean, it certainly seems, that if a party to the conflict holds an otherwise Protected Person in sites that it operates in third party states, that person would no longer be protected by the Convention.
A third, related implication is that negation of Protected Person status might render at least parts of the Grave Breaches regime – which establishes international criminal liability for certain violations –inapplicable to individuals whether in the invasion stage or if removed from the geographical area.[7]
That said, there are potential ways to mitigate at least some of the harmful effects of the geographic nexus requirement, without retiring it altogether. One way is to argue that the law of occupation, or at least parts of it, begin to apply not only when full effective control over territory solidifies, but also functionally during the “invasion stage,” as famously suggested by Jean Pictet in the ICRC Commentaries.[8] Pictet was precisely concerned with possible legal black holes prior to the onset of full effective control, and the potential lack of obligations placed on invading forces despite their exercising some of the powers of belligerent occupation. Adoption of the Pictet approach would have the benefit of also significantly narrowing the effect of the geographic nexus requirement.
Another way to avoid or, more precisely, reduce a legal black hole in the invasion stage is through the Fundamental Guarantees regime established in Additional Protocol I (API), which applies in situations in which any person is “in the power of a Party to the conflict” and does not “benefit from more favorable treatment” under international humanitarian law (IHL) treaty law. The Fundamental Guarantees regime provides for basic protections of individual rights and due process to all such persons, even if not Protected Persons under GCIV. These provisions pertain primarily to the rights of individuals held in custody or situations of detention.
A third way to reduce the effects of the geographic nexus requirement is through the applicability of international human rights law (IHRL) extraterritorially and during armed conflict. Arguably, when individuals fall under the control of parties to the conflict even before occupation commences, or when individuals are under the power of parties to the conflict but in third-state territories, IHRL can fill the gaps and provide for basic protection.
However, the DoD Manual stops short of embracing any of these possibilities as a matter of binding law, and instead opts for a more pragmatic, discretionary approach. First, the manual explicitly rejects the application of the law of occupation to situations of “mere invasion,” while stating that only “as a matter of policy,” the armed forces would seek to apply rules from the law occupation to areas were U.S. forces are passing tor fighting (§11.1.3.1).
Second, the United States is not party to API, and the DoD Manual does not explicitly embrace the Fundamental Guarantees regime as binding. Rather, it supports the application of these guarantees to detained persons as a matter of policy, with some vague allusion to its potential customary international law status. As the Manual states, the “United States has supported and sought to promote as a matter of national policy fundamental guarantees for the treatment of persons detained during international armed conflict that are reflected in Article 75 of AP I” (§19.20.1.3). Elsewhere the Manual appears to cast doubt on the existing customary international law status of Article 75: “Although not a Party to AP I, the United States has stated that the U.S. Government will choose out of a sense of legal obligation to treat the principles set forth in Article 75 as applicable to any individual it detains in an international armed conflict, and expects all other nations to adhere to these principles as well. This statement was intended to contribute to the crystallization of the principles contained in Article 75 as rules of customary international law applicable in international armed conflict.” (§ 8.1.4.2, emphasis added; but see Marco Milanovic’s analysis for an understanding of the underlying U.S. statement involving acceptance of Art. 75 as existing customary law; also, one might claim that this amounts to a binding unilateral declaration). Noticeably, the Manual fails to include previous U.S. sources that more explicitly support recognition of Article 75 as customary international law. For example, the Manual omits Michael J. Matheson’s official remarks in 1987, which included the statement: “We support in particular the fundamental guarantees contained in article 75.” (The Manual relies on Matheson’s similarly worded remarks for many other provisions.) Perhaps more importantly, the Manual also does not reference a Department of Defense “advisory opinion” in 1986 which concluded: “We view the following provisions as already part of customary international law … Fundamental guarantees: Article 75” (memorandum signed by W. Hays Parks, John. H. McNeill, Assistant General Counsel, and others).[9] Third, regarding the application of IHRL extraterritorially and during armed conflict, the DoD Manual generally applies IHL as the lex specialis, and significantly, holds that key IHRL treaties do not apply abroad as a threshold matter (§1.6.3), a long-standing U.S. view that is not specific to the DoD. Even when it does recognize the possible extra-territorial application of an IHRL treaty – such as is the case regarding the Convention Against Torture – this is mainly where States control areas “as a governmental authority,” which would presumably exclude, at least, the invasion stage (§1.6.3.4).
Be that is it may, what has to be emphasized is that even if we assume that these “policy mitigations,” so to speak, are applied in good faith by the United States, nothing prevents other actors from embracing the DoD Manual’s restrictive interpretation of Protected Persons status as a matter of law, without adopting any of these policy-based limitations. This, of course, adds to the harmful implications of this approach: international law is not “owned” by any single State, and permissive interpretations by one can equally serve others. It also undermines efforts to hold other actors accountable, the ability for the United States to have a legal basis to criticize other military forces, and the Geneva Conventions system for repressing grave breaches.
With this in mind, I now turn to discuss the geographic nexus requirement in terms of the lex lata.
Language and Context
In my view, the text of Article 4(1) of GCIV, read in good faith in light of its context and object and purpose, leads to the conclusion that GCIV does not include the geographic nexus requirement. Article 4(1) uses broad language, referring to Protected Persons as those who “at a given moment and in any manner whatsoever” find themselves “in the hands” of a belligerent (emphasis added). The ordinary meaning of such language clearly conveys that it encompasses any and all situations in which a person is “in the hands” of a belligerent, even situations that have not been explicitly mentioned in the Convention.[10] Notably, this language seems to clearly apply to persons in the hands of belligerents during the invasion stage. (The grave breaches provisions would then apply to those people in those situations.)
Article 4(1)’s context also supports this conclusion. First, Article 4(2) includes express and detailed limitations on the definition of Protected Persons, by excluding several categories of persons from its ambit.[11] However, these do not refer to any geographic limitations, be it not-yet-occupied territory during the invasion stage, or third-party territory in cases of removal. It would have been expected that had the drafters intended to exclude additional cases from the definition, this would have been explicitly stated. Granted, due to the reality of most wars, the common situations would indeed be ones in which Protected Persons would find themselves in territories that fulfil the geographic nexus requirement; but from the mere fact that these are the “main classes of protected person”[12] it cannot be inferred that other cases are per se excluded. Those classes of cases include civilians caught up by invading forces in World War II, which informed the drafting of GCIV, and in today’s conflicts, alleged war crimes committed in Russia’s invasion phase in Georgia, for which the ICC has issued arrest warrants. It is easy to contemplate other historical and recent examples where people find themselves under belligerent control before the onset of “full” occupation. For example, Israel’s detention of Palestinians in Gaza, if we assume that Gaza was not occupied prior to Israel’s ongoing ground operation there, and that these detentions took place before effective control over the territory has been solidified (if, of course, we consider Gaza occupied, the question does not arise).
Importantly, the “in the hands of” requirement itself does not seem to be geographically bound in a stringent way as the OLC Opinion implies. To see this, consider two examples: in cases of occupation, Article 49 of GCIV prohibits – except in very narrow circumstances – the removal of Protected Persons from the occupied territory to third party states or territories. However, in the extremely rare case when it does allow such evacuations – as is the case in situations enumerated in Article 49(3) – it continues to treat persons removed as enjoying that status,[13] and provides for the occupant’s continuing obligations toward them, although they are no longer in the occupied territory, and perhaps not even under its direct control.[14]
Additionally, Article 45 of GCIV provides strict conditions under which Protected Persons who are found in the territory of the opposing belligerent can be transferred to another Power. However, GCIV clearly continues to treat such transferred persons as Protected Persons, both in the text of Article 45,[15] and in the requirement that the receiving Power possesses the “willingness and ability … to apply the present Convention” to these persons.[16] Crucially, the 1958 Commentary is clear that the receiving Power must apply to transferred persons the protections afforded to protected persons, including those that are found under the same Sections of GCIV which the Opinion claims are applicable only where the geographic nexus exists.[17]
The above implies that, in fact, the geographic nexus is not as determinative as the OLC Opinion states, and that persons remain protected by GCIV even in the rare case in which the law allows their evacuation to third states – and this remains true notwithstanding the titles of the Sections of GCIV which the Opinion invokes. It is almost redundant to say that this would also be the case when belligerents remove people from territories unlawfully.
Object and Purpose and Legal Policy
In my view, the geographic nexus requirement is also contrary to the object and purpose of GCIV, since it arbitrarily deprives individuals of Protected Person status – although they might be in the hands of the enemy for all practical purposes –without a clear rationale justifiable under the principles of IHL. In this context, the International Court of Justice held, in the Wall Advisory Opinion, that since the object and purpose of GCIV is to protect civilians in war, it should not be read in a restrictive manner absent a clear contrary intention.[18] This reasoning clearly pushes for the more expansive reading of the “in the hands of” requirement, which applies both to the invasion stage and to cases in which persons are held in third party territories.
Another legal policy reason militates against the geographic nexus requirement. There is an imbalance between the United States’ relatively wide interpretation of the geographic application of the targeting rules of IHL,as extending beyond “hot battlefields,”[19] at least in the context of non-international armed conflicts, and the narrow interpretation of the status of Protected Persons as requiring a geographic nexus. This type of inconsistency should be avoided, since it can be understood as an attempt to utilize IHL in order to expand state power, while shirking its restraining potential.
Scholars and Practice
The OLC Opinion cites some scholarly authorities in support of the geographic nexus requirement, but in my view these are at best inconclusive.[20] The ICRC’s 1958 Commentary on GCIV refers to the geographic limitations suggested in the Opinion as comprising the “main classes” of protected persons but does not per se exclude other classes.[21] Articles by Baxter and by Yingling and Ginnane only mention this issue tangentially without substantive analysis.[22] Parkerson, in fact, highlights the tension between Article 4(1) – which defines Protected Persons broadly as all persons in the hands of the enemy – and the language of the Section titles relied upon by the Opinion.[23] Be that as it may, all of these sources are decades old and weight should be given to the purposive interpretation of GCIV advanced by the ICJ in the Wall Opinion.
Furthermore, leading contemporary analyses of IHL do not embrace the geographic nexus requirement. Sometimes, they only assert the common cases of Protected Persons as being either in the territory of the party or in occupied territory.[24] Marco Sassòli, for his part, proceeded from the assumption that anyone “in the hands” of the enemy – who is not otherwise excluded from that status under Article 4 GCIV – is a Protected Person, and applied this reasoning to persons held at Guantanamo.[25] In Sassòli’s more recent treatise, the geographic nexus is not envisioned.[26] In a 2021 article, Van Schaack explicitly objects to the geographic nexus, mainly on account of its incompatibility with the object and purpose of the Geneva Conventions and IHL.[27]
Unlike the DoD Manual, recent military manuals by peer states have generally not made an explicit point about the geographic limitation of Protected Person status, nor do they address in detail the definition of this status.[28] This could be because the category of civilian established in API – as well as its Fundamental Guarantees regime – has in many ways superseded in importance the category of Protected Persons in these jurisdictions. Nonetheless, the UK 2004 Manual repeats Article 4(1)’s wide formulation,[29] adding that “[i]n the hands of” means “simply that the person concerned is in territory which is under the control of the state in question” – wording which could apply also to territories of third states under the control of a Party to the conflict but are not occupied strictly speaking, such as Guantanamo. The UK Manual also refers to persons falling under the geographic nexus as mere “common groups” within the definition, suggesting that other categories are not excluded.[30] Australia’s 2006 Manual generally follows suit.[31] Canada’s 2001 Manual also repeats Article 4(1)’s broad language with no geographic specification.[32] Germany’s 2013 Manual does not deal with the general definition of Protected Persons at all,[33] as is the case in New Zealand’s 2017 Manual.[34] In sum, military manuals are at best inconclusive regarding this question, if not opposed to the geographic nexus requirement altogether.
Conclusion
In sum, in my view there is no convincing basis, whether in law or policy, to limit Protected Person status only to situations where the individual is in occupied territory or in the home territory of the opposing party. Such a limitation might result in a legal black hole either in the invasion stage, or if a person is transferred to a third state, whether lawfully or unlawfully.
To reduce the scale of this problem in the DoD Manual, two main approaches are available. The first is to accept, as a matter of law, that rules stemming from the law of occupation apply during the invasion stage; or that, again as a matter of law, the Fundamental Guarantees regime applies to all persons; or that IHRL applies extraterritorially to all persons under the power of a belligerent. But this is a second-best solution by far. It would leave unrepaired the influence of the Manual on the scope the grave breaches regime and the protections guaranteed civilians in situations outside the geographic nexus.
The other is to amend Section 10.3.2.2 in a manner that reflects that there are exceptional situations in which a person may be “in the hands” of the enemy even when located outside territories falling under the geographic nexus requirement, whether in the invasion stage, or in cases where a party to the conflict transfers or evacuates a protected person to third party states, whether lawfully (in rare cases of permitted evacuations) or unlawfully.
Creating legal black holes – alleviated by flexible policy standards – might be tempting for states that seek to maximize their latitude in the field. However, it should always be remembered that at the end of the day, nothing prevents other states from capitalizing on the same interpretations. Accordingly, for both moral and instrumental reasons, the integrity of the laws that provide basic protection for individuals in wars should be upheld.
[1] To complete the picture, it should be added that the Manual’s interpretation of “find themselves” in the hands of the party to the conflict, in §10.3.2.1, is also overly narrow, by requiring an element of “happenstance or coincidence.” This adds to the overall exceedingly narrow interpretation of the status of Protected Persons in the Manual. The UK Supreme Court rejected this “happenstance” theory, which had also been articulated by Jack Goldsmith in a separate OLC opinion.
[2] For additional detail see Beth Van Schaack, Animating the U.S. War Crimes Act, 97 Int’l L. Stud. 1542, 1568 –1569 (2021).
[3] Van Schaack, at 1565, 1568.
[4] On problems relating to the beginning of occupation see Eliav Lieblich and Eyal Benvenisti, Occupation in International Law 59–74 (2022); Eliav Lieblich, Germany Goes 19th Century with New Statement on Law of Occupation — On Status of US Forces in Syria, Just Security (Feb. 19, 2020).
[5] On this practice see generally Gary D. Solis, Extreme Measures: The U.S. Removal of ‘Protected Persons’ From Occupied Iraq and Afghanistan, 10 G’town J. Int’l Aff. 121 (2009).
[6] DoD Manual, §10.3.2.2.
[7] GCIV, art .147. See Van Schaack, at 1565 (2021) (analyzing specifically how this problem would affect the ability to prosecute under the U.S. War Crimes Act).
[8] ICRC, Commentary on IV Geneva Convention 60 (Jean S. Pictet ed. 1958) [hereinafter: ICRC Commentary 1958].
[9] I thank Ryan Goodman for pointing out these previous sources on the status of Article 75.
[10] As the 1958 Commentary on GCIV states at p. 47: “The words ‘at a given moment and in any manner whatsoever’, were intended to ensure that all situations and cases were covered.”
[11] For a detailed analysis see Marco Sassòli, International Humanitarian Law 289–292 (2019).
[12] ICRC, Commentary on IV Geneva Convention 46 (Jean S. Pictet ed., 1958).
[13] Art. 49(3) states “The Occupying Power undertaking such transfers or evacuations shall ensure, to the greatest practicable extent, that proper accommodation is provided to receive the protected persons” [emphasis added].
[14] Commentary, 281-282.
[15] For example, “Nevertheless, if that Power falls to carry out the provisions of the present Convention in any important respect, the Power by which the protected persons were transferred shall, upon being so notified by the Protecting Power, take effective measures to correct the situation or shall request the return of the protected
persons” (emphasis added).
[16] GCIV, art. 45(2).
[17] For instance, by referring to the See, e.g., GCIV, arts. 27, 28, 30–34, as well as the provisions relating to internment. See 1958 Commentary, at 269 (“if the persons transferred are interned, which will most often be the case, a great many of the provisions in the chapter on internment will also have to be borne in mind, in particular those dealing with civil capacity, maintenance, food, clothing hygiene and medical attention, religious and intellectual activities, correspondence and relief. In the treatment accorded to protected persons, the application of all these provisions is essential.”).
[18] Legal Consequences of the Construction of a Wall in the Occupied Palestinian Territory, Advisory Opinion, 2004 I.C.J. Rep. 136 ¶95 (Jul. 9); see also Van Schaack, at 1567-1568 (highlighting that a key objective of the Geneva Conventions is that nobody under enemy hands would be outside the law).
[19] DoD Manual, n. 58.
[20] Opinion, n. 8 (citing 1958 Commentary, at 45–46; Richard R. Baxter, So-Called “Unprivileged Belligerency”: Spies, Guerrillas, and Saboteurs, 28 Brit. Y.B. Int’l L. 323, 328 (1951); Raymund T. Yingling & Robert W. Ginnane, The Geneva Conventions of 1949, 46 Am. J. Int’l L. 393, 411 (1952); John Embry Parkerson, Jr., United States Compliance with Humanitarian Law Respecting Civilians During Operation Just Cause, 133 Mil. L. Rev. 31, 74 (1991). For a similar view see Van Schaack, at 1570.
[21] 1958 Commentary, at 45-46.
[22] Baxter, at 328, Yingling & Ginanne at 411.
[23] Parkerson, at 74 (“Section I consists of ‘Provisions Common to the Territories of the Parties to the Conflict and to Occupied Territories.’ The choice of this title is unfortunate because it signals the ambiguities in the protections that are provided for those ‘protected persons’ who are defined in Article 4.”).
[24] See e.g., Elizabeth Salmón, Who Is a Protected Civilian? In The 1949 Genevan Conventions: A Commentary 1135, 1142 (Andrew Clapham et al eds., 2015).
[25] Marco Sassòli, The Status of Persons Held in Guantanamo Under International Humanitarian Law, 2 J. Int’l Crim. Just. 96, 100–101, 103–105 (2004).
[26] Sassòli, 289–292.
[27] Van Schaack, at 1564–1571.
[28] See Danish Ministry of Def., Military Manual on International Law Relevant to Danish Armed Forces in International Operations 191 (2016).
[29] U.K. Ministry of Def., The Joint Service Manual of the Law of Armed Conflict §9.17(2004).
[30] Id. §9.17.2.
[31] ADDP 06.4, Law of Armed Conflict (Australia) §9.55 (2006).
[32] Joint Doctrine Manual, Law of Armed Conflict at the Operational and Tactical Levels (Canada) §1103 (2001).
[33] Law of Armed Conflict – Manual – Joint Service Regulation (ZDv) 15/2 (Germany) (2013).
[34] New Zealand, DM 69 (2 ed.) Manual of Armed Forces Law, Volume 4 Law of Armed Conflict (2017).