GC IV Commentary Series – Lost at Sea: Naval NIACs

by | Nov 18, 2026

NIACs

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

The 2025 Commentary on the Fourth Geneva Convention states that non-international armed conflicts (NIACs) may “take place, and have taken place, in part or in entirety, at sea or on other bodies of water” and that “[w]ithin the framework of the Geneva Conventions, Article 3 applies to such conflicts.” As phrased, the statement suggests that NIACs can be entirely sea-borne and that Article 3 applies at sea without restrictions. These claims are difficult to reconcile with the historical record and with the terms of Common Article 3 (CA 3).

The Historical Record

As the updated Commentary to the Second Geneva Convention observes, “[n]ot many non-international armed conflicts with a naval dimension have taken place since the adoption of the Geneva Conventions.” This Commentary offers the Sri Lankan civil war as the most prominent example, though the only authority it cites in support is an arbitral award concerned with combat operations on land (paras. 61-64).

Beyond Sri Lanka, the examples are few indeed. The Newport Manual mentions the Spanish civil war, which of course pre-dates the Geneva Conventions of 1949, Israel’s blockade of Gaza, and the Houthi attacks on shipping in the Red Sea. Others have cited the blockades of Abkhazia by Georgia, of East Timor by Indonesia, and of Yemen by the Saudi-led coalition. The classification of some of these conflicts as non-international is open to question. In any event, none of these conflicts were entirely sea-borne, since in each case the action at sea formed part of a conflict with a base on land. In fact, no entirely sea-borne NIAC appears to have ever taken place.

Perhaps the closest thing to an entirely sea-borne NIAC is the campaign of strikes the United States has conducted since September 2025 in the Caribbean and the eastern Pacific against vessels it says are operated by drug traffickers. The United States has portrayed this campaign as part of a NIAC between itself and two drug cartels, the Tren de Aragua and Cartel de Los Soles. Putting aside the question of whether the conditions for the existence of a NIAC are actually satisfied in this case, the United States itself has not claimed that the conflict is entirely sea-borne, because its position rests on the argument that the two cartels have caused the deaths of tens of thousands of Americans within the United States. The Commentary’s claim that NIACs have taken place entirely at sea is therefore not borne out even in this instance.

The Meaning Of “Territory”

The suggestion that CA 3 applies at sea runs into another difficulty. CA 3 applies to an armed conflict not of an international character “occurring in the territory of one of the High Contracting Parties.” Whether “territory” extends to the sea is not obvious. The answer turns on the rules of treaty interpretation in Articles 31 to 33 of the Vienna Convention on the Law of Treaties, which the Commentary adopts as its guiding methodology (para. 98). The general rule of interpretation under Article 31(1) requires a treaty to “be interpreted in good faith in accordance with the ordinary meaning to be given to the terms of the treaty in their context and in the light of its object and purpose.”

In its ordinary meaning, the word “territory” refers to tracts of land. The Commentary gives the term a wider sense for the purposes of the law of armed conflict, suggesting that “a State’s territory includes not only its land surface but also rivers and landlocked lakes, the territorial sea, and the national airspace above this territory” (n. 191). The term is sometimes used in this wider sense in international law, so that territory is treated as coextensive with State sovereignty. For instance, the phrase “territorial integrity” in Article 2(4) of the UN Charter is commonly understood in this way. In Corfu Channel, the International Court of Justice (ICJ) described the United Kingdom’s principal justification for sweeping mines in Albanian territorial waters as a form of intervention to secure evidence “in the territory of another State” (p. 34). In Paramilitary Activities, the Court held that attacks on Nicaragua’s ports were “infringements of the territorial sovereignty of Nicaragua” and that unauthorized overflights violated the same principle (paras. 212, 251).

This wide reading is not universal, however. The sovereignty that a State exercises over its territorial sea and airspace derives from its dominion over land, since, in the ICJ’s words, “the land dominates the sea” (para. 96). The “defined territory” that Article 1 of the Montevideo Convention requires a State to possess thus means a defined area of land. No State exists at sea or in the air. Nor can “defined territory” within the meaning of the Montevideo Convention be expected to include territorial waters, since then landlocked countries such as Paraguay, a party to the Convention, would not be covered by its Article 1. The Chicago Convention defines a State’s territory in Article 2 as “the land areas and territorial waters adjacent thereto,” which leaves out the airspace over which Article 1 of the Convention recognizes its sovereignty. The Convention on the Territorial Sea and the Contiguous Zone of 1958 does not include the sea within the concept of territory when it distinguishes between infringements that take place within a State’s “territory or territorial sea” in Article 24(1).

The Meaning of “Territory” in CA 3

Whether “territory” carries a narrow or a wide meaning in CA 3 is therefore not immediately clear. The context points to a narrow sense. For example, Article 14 of the Fourth Convention allows the parties to establish hospital and safety zones and localities “in their own territory and, if the need arises, in occupied areas.” The territory meant here is an area of land. Hospital and safety zones would be impractical at best at sea and impossible to establish in the air. There is no precedent for them in either domain, as distinct from “neutralized zones” that have been established at sea.

The rules on belligerent occupation refer repeatedly to “occupied territory.” Occupation is a concept of land warfare. The Hague Regulations of 1907, which deal with belligerent occupation extensively, are entitled the “Regulations concerning the Laws and Customs of War on Land.” Nothing in the Fourth Convention itself suggests that occupied territory means anything other than land territory. The Commentary itself derives its concept of occupation from Article 42 of those Regulations and measures its extent by the occupant’s capacity to station troops and project its authority over territory (paras. 368, 385–87).

Turning to object and purpose, the Commentary declares that CA 3 “has its own object and purpose,” which it suggests is to “protect persons not or no longer participating in hostilities in situations of non-international armed conflict” (para. 116). This departs from the mainstream view that object and purpose belong to the treaty as such, rather than to its individual provisions, and that they can be determined only by an examination of the treaty as a whole (ILC Guide to Practice, guideline 3.1.5, para 11; Chicken Cuts, para. 238; also Gardiner, p. 133; Klabbers, p. 152–53; Dörr, p. 546).

Whereas the Commentary describes CA 3 as a “Convention in miniature,” it is not in fact a treaty in its own right, but simply one provision that the four Conventions share (para. 427). Accordingly, it may be questioned whether Article 3 can have its own object and purpose at all. In any case, the object and purpose of CA 3 as defined by the Commentary cannot settle how far “territory” reaches. It would be circular to prefer a wide reading of “territory” for the purposes of CA 3 over the narrower sense it carries in other provisions of the Fourth Convention on the ground that this better protects persons not or no longer participating in hostilities “in situations of non-international armed conflict,” since what is at issue is precisely where such situations exist in the first place.

Subsequent practice also does not support a broad meaning. Practice counts as subsequent practice under Article 31(3)(b) of the Vienna Convention only where it “establishes the agreement of the parties regarding its interpretation,” meaning the parties as a whole (art. 27, para. 15). Given the universal character of the Geneva Conventions, any subsequent practice must be similarly universal (p. 133). There is no evidence of consistent practice on whether “territory” in CA 3 extends to the sea, let alone any agreement of all the parties. Military manuals diverge on where a NIAC takes place and no manual seems to distinguish between land, internal waters, and the territorial sea when addressing CA 3.

The preparatory work, which is a supplementary means of interpretation under Article 32 of the Vienna Convention, does not decide the point. A draft considered before the Diplomatic Conference of 1949 referred to conflicts “which may occur in the territory of one or more of the High Contracting Parties.” The words “or more” did not survive. The record gives no reason for their removal and the Commentary accordingly draws “no conclusion” from their absence (para. 543). This aspect of the drafting history has no bearing on the meaning of “territory” and nothing else in the debates on CA 3 sheds any light on the matter.

However, there is one relevant pointer in the drafting history of what is now Article 39 of the Second Geneva Convention. The provision applies to situations where medical aircraft alight involuntarily “on land or water in enemy or enemy-occupied territory,” which implies that enemy or enemy-occupied territory includes bodies of water. The Drafting Committee emphasized that Article 39 does not make reference to “territorial waters” because the Committee did not wish “to venture onto such uncertain ground” and “did not regard as part of its business the definition of the rules applying to territorial waters” (p. 141, 205). This suggests that the drafters of Article 39 understood water to mean internal waters, but not necessarily to include territorial waters. This understanding does not, of course, govern the interpretation of CA 3, which is a different provision. It does suggest, however, that it would be wrong to simply assume without any evidence that the drafters understood territory to carry a broad meaning under CA 3 that includes the territorial sea.

Overall, the context, object and purpose, subsequent practice, and drafting history of CA 3 do not establish that the word “territory” featured in that provision should mean anything other than its ordinary meaning, which is a tract of land. Even if a broad reading is adopted (despite what the rules of treaty interpretation suggest), the term cannot be stretched to include the high seas. Whereas the sovereignty of coastal States extends to their territorial sea (art. 2(1)), the law of the sea provides that “[n]o State may validly purport to subject any part of the high seas to its sovereignty” (art. 89). Waters that no State may subject to its sovereignty simply cannot form part of “the territory of one of the High Contracting Parties” for the purposes of CA 3.

The object and purpose of the Convention cannot be invoked to reach a different outcome. Object and purpose serve to “shed light on the terms actually used in their context, rather than introduce an alternative option for finding the meaning.” They are not a vehicle for rewriting a treaty. As the ICJ noted in Interpretation of Peace Treaties, “[i]t is the duty of the Court to interpret the Treaties, not to revise them” (p. 229). Reading “territory” to include the high seas would blatantly cross that line and nothing in the practice of the High Contracting Parties supports such a move.

Accordingly, applying the rules of treaty interpretation, it is open to doubt whether CA 3 extends to the territorial sea, while there is no doubt whatsoever that a “case of armed conflict not of an international character occurring in the territory of one of the High Contracting Parties” does not cover the high seas. The Commentary’s unqualified statement that NIACs may take place at sea and that CA 3 applies to them therefore cannot be reconciled with its interpretative methodology.

Can a NIAC Exist at Sea at All?

Writers who accept that a NIAC can exist at sea typically take one of three routes. The first runs through custom, suggesting that “State practice seems to provide sufficient evidence that there is no rule of customary international law prohibiting the parties to a non-international armed conflict from engaging in hostilities against each other in high seas areas.” Others have argued that customary international law “does not prohibit the participants in a NIAC from engaging in hostilities against each other on the high seas if they take due regard for the rights of other States,” and still others have defended blockade in a NIAC on the footing that “a lack of precedents does not normally create a prohibition in international law.” Each of these authors reasons from the absence of a prohibition, rather than arguing that the requisite level of State practice and opinio juris exists to establish the existence of customary rules. This reasoning may explain why it is permissible for States to engage in hostilities on the high seas against organized armed groups, but it fails to provide a legal basis for any restrictions governing the conduct of such hostilities.

The second route omits location as a definitional element of a NIAC. Supporters of this route have taken the view that “there is no geographic limitation that applies to the initiation of a NIAC.” However, the definition of a NIAC these authors rely on is that offered by the Appeals Chamber in Tadić, which speaks of “protracted armed violence between governmental authorities and organized armed groups or between such groups within a State” (para. 70, emphasis added). The Tadić definition clearly does not extend to the high seas.

The third route toward accepting that a NIAC can exist at sea is through the idea of spillover. On this view, a NIAC located in the territory of a party may extend to the sea if the hostilities spill over from the land into that domain. The Newport Manual takes this position by recognizing “the potential for NIAC to spill over into the conflict State’s territorial sea, and then into adjacent EEZ and high seas areas.” This stance is echoed in the view that a conflict “that is fought on territorial waters but which extends to the high seas” remains covered by CA 3.

This approach can be reconciled with the terms of CA 3 without too much difficulty if the requirement that a NIAC must occur in the territory of one of the High Contracting Parties is understood to apply to the commencement of the conflict, without it being read as imposing a geographical restriction on the continued application of CA 3. In other words, for CA 3 to be engaged, a NIAC must occur in a party’s territory, but its requirements continue to apply should hostilities spill over into other territories or domains. In fact, this reading finds some textual support in the equally authentic French version of the Convention, which refers to “conflit armé ne présentant pas un caractère international et surgissant sur le territoire de l’une des Hautes Parties contractantes” (emphasis added), that is, to armed conflicts “arising” or “emerging” in the territory of a party.

Conclusion

The Commentary’s unqualified statement that NIACs may take place at sea and that CA 3 applies in such situations is unconvincing. CA 3 does not apply to entirely sea-borne hostilities. Rather, CA 3 extends to the sea only if a pre-existing NIAC on land spills over into the maritime domain. In practice, this has implications for the classification of certain acts of violence at sea, such as the Houthi attacks on shipping in the Red Sea. Based on the spillover view, those attacks would form part of or qualify as a NIAC only if they were the naval extension of protracted armed violence taking place on land between the Houthis on the one side and the relevant flag States on the other side.

***

Dr Aurel Sari is a Professor of Public International Law at the University of Exeter.

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

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

 

 

 

 

 

 

 

 

 

Photo credit: NAVCENT Public Affairs