Ця стаття також доступна українською мовою тут.

Russia’s war in Ukraine has been marked by what appear to be widespread and systematic violations of the law of armed conflict. Amidst the devastation of the siege of Mariupol, one method of Russian warfighting that has become prominent is the deprivation of food, water, and other essentials.

These tactics implicate international humanitarian law and international criminal law. Starvation of civilians as a method of warfare is prohibited by article 54 of Protocol I Additional to the Geneva Conventions (AP I) and codified as a war crime in article 8(2)(b)(xxv) of the Statute of the International Criminal Court (ICC). Both Russia and Ukraine are party to Protocol I, rendering that treaty straightforwardly applicable in the international armed conflict (IAC) between them. Ukraine has accepted the jurisdiction of the ICC pursuant to article 12(3) of the Court’s Statute, providing the necessary basis for the ICC’s application of article 8(2)(b)(xxv) to anyone who perpetrates the starvation war crime on Ukrainian territory. One or another version of the prohibition is also included in the national war crimes codes of several of the states that have opened investigations into alleged war crimes in Ukraine, including Estonia § 95(1), France art. 461-25, Germany, § 11(1)(5), Liechtenstein § 321e(1)(9), Norway § 106(b), and Sweden § 9(8). Although the criminal codes of Russia and Ukraine lack specific reference to starvation methods, each includes a residual criminal category that could be interpreted to cover such practices (Russia, art. 356(1): “use in a military conflict of means and methods of warfare, banned by an international treaty of the Russian Federation;” Ukraine art. 438(1): “use of methods of the warfare prohibited by international instruments, or any other violations of rules of the warfare recognized by international instruments consented to by binding by the Verkhovna Rada [Parliament] of Ukraine”).

Historically, starvation crimes have often been overlooked. With multiple investigative teams now focused on Ukraine, those involved ought to ensure that these prohibitions are a central part of their work. Four years ago, the U.N. Security Council adopted Resolution 2417, in which it “strongly urge[d]” states to conduct “investigations within their jurisdiction into violations of international humanitarian law related to the use of starvation of civilians as a method of warfare, including the unlawful denial of humanitarian assistance to the civilian population in armed conflict, and, where appropriate, to take action against those responsible in accordance with domestic and international law, with a view to … ensuring accountability” (para. 10).

As investigations develop, it will be crucial that those involved are clear on the law. What is necessary to show that operations in Ukraine constitute the starvation of civilians as a method of warfare? What must be established for the associated war crime to attach? The analysis that follows answers those questions.

Objects Indispensable to Civilian Survival

Starvation methods are prohibited because of their multifaceted, destructive, and torturous impacts on individuals and societies. However, those consequences are gradual and often shaped by a combination of conditions and intervening factors. As such, it can be difficult to trace the specific harms associated with starvation in a particular context to specific military operations. For this reason, the legal keystone of the prohibition of starvation methods is not the infliction of such outcomes, but the heightened protection of “objects indispensable to the survival of the civilian population,” such as food, water, and the systems by which they are produced and supplied (art. 54 AP I). In addition to its place at the heart of IHL’s starvation ban, the deprivation of such objects is the central material element of the associated war crime, which can be established without showing that any specific consequence arose from that deprivation (ICC Elements of Crimes, p.31).

The heightened protection of indispensable objects is best understood in contrast with IHL’s posture on objects generally. Ordinarily, an object with both military and civilian uses—a so-called “dual-use object”—is considered a military objective due to its “effective contribution to military action” (art. 52 AP I). As long as its destruction, capture, or neutralization would return a definite military advantage, such an object is a lawful target as a matter of distinction. Its value to civilians is protected instead through belligerents’ obligations to take all feasible precautions to minimize civilian harm in such an operation (art. 57(2)(a)(ii) AP I) and not to attack if the civilian harm expected would be disproportionate, which is to say excessive in relation to the concrete and direct military advantage anticipated (arts. 51(5)(b), 57(2)(a)(iii) AP I).

Objects indispensable to the survival of the civilian population are different. Even if such objects are used not just by civilians, but also by combatants, two bright-line prohibitions still apply to the attack, destruction, removal, or rendering useless of such objects. As elaborated below, such heightened protections arguably apply also to other forms of deprivation, including encirclement operations that cut civilians off from the external supply of such objects.

The first elevated prohibition is specific to operations that target indispensable objects for their sustenance value. Pursuant to the terms of article 54(2)-(3)(a) of Protocol I, objects indispensable to survival cannot be targeted for their sustenance value (including their sustenance value to the adversary), unless they provide sustenance exclusively to combatants. In other words, unlike targetable dual-use objects, dual-use sustenance is legally protected from attack. This prohibition holds whether or not the targeting of such objects would leave the civilian population inadequately supplied and regardless of the motive for denying sustenance. It inheres simply in (i) the fact of sustenance denial as a purpose and (ii) the fact that civilians are among those affected.

Second, pursuant to the terms of article 54(3)(b) of Protocol I, objects indispensable to civilian survival can be targeted for reasons other than their sustenance value only if two criteria obtain. First, the objects must provide direct support for military action. This requires a significantly tighter connection than the “effective contribution to military action” test applied to other objects (art. 52(2)). Indeed, “direct” implies a test analogous to that applicable to the targeting of civilians, which is permissible only for such time as they take a “direct” part in hostilities (art. 51(3)). An influential (albeit disputed) analysis by the ICRC understands the latter to attach only when the activity is a single causal step from the infliction of harm on the adversary (pp. 52-54). An oft-cited example in the context of indispensable objects is that of a food-storage barn used for cover by hostile forces (para. 2110). Second, even when the direct support threshold is satisfied, targeting the objects in question still must not be expected to leave the civilian population starving or forced to move. For the prohibition to attach on this basis, the belligerent party need not have created the underlying scarcity or vulnerability to the effects of deprivation. It is necessary only that the operation would be “expected to leave the civilian population with such inadequate food or water as to cause its starvation or force its movement.” (art. 54(3)(b)).

These two alternative prohibitions –  (i) on dual-use sustenance denial and (ii) on actions that would leave civilians starving or forced to move – specify the unique and heightened protection of indispensable objects, even when they provide significant military utility and regardless of whether civilians are the ultimate targets. As rules that reframe the principle of distinction in this context, they attach prior to evaluations of proportionality and precautions (i.e., regardless of military advantage and independent of whether all feasible measures were taken to minimize civilian harm). Moreover, under the first prohibition, operations can be identified as illegal without needing to establish that they caused (or were anticipated to cause) a particular effect within the civilian population. Objects such as food, water, and the systems by which they are produced and supplied are here understood as intrinsically indispensable. The first rule’s purposive focus allows it to capture acts of sustenance denial undertaken before the relevant objects are scarce. Under the second prohibition, the anticipated impact matters, but it is necessary only to show that the operation would be expected to leave the population in starvation conditions (or require its movement), not that the belligerent party caused the conditions of scarcity underpinning that vulnerability. Here, in contrast to the first rule, it matters whether the specific items at stake are themselves indispensable to the survival of civilians in a specific population.

All four of the modalities of deprivation listed in article 54(2) are implicated by Russian actions in Ukraine. Grocery stores, agricultural areas and granaries, and critical infrastructure for the supply of water appear to have been attacked. Reports describe both the destruction and removal of agricultural machinery by Russian troops. And farmland appears to have been rendered useless through Russian forces’ deployment of landmines. There is nothing to suggest that these resources were devoted exclusively to the provision of combatant sustenance. As such, if the actions described were taken with a view to denying sustenance (including to combatants), they violated the terms of article 54(2)-(3)(a) of Protocol I. If the objects were instead targeted because of a direct military contribution other than sustenance (itself a high bar), the lawfulness of those operations would need to be evaluated according to whether they left civilians starving or forced to move, pursuant to the terms of article 54(3)(b). This would require an analysis of the connection between specific acts and the expected food security of the population dependent on the destroyed, removed, or nonviable food or water systems in question.

Encirclement Deprivation – Three Interpretive Frames

Of course, starvation methods are not exhausted by the attack, destruction, removal, or rendering useless of indispensable objects. The most significant method that would appear not to be included in that list is encirclement starvation, whereby civilians are cut off from the external supply of such objects. In Ukraine, this method has been most prominent in the siege of Mariupol, which has been cut off from the beginning of March. On Thursday, Russia claimed to have captured much of the city. However, the Azovstal steel plant remains a holdout and remains besieged. Throughout the encirclement, humanitarian actors have reported insurmountable obstacles to access and there have been repeated allegations that Russian forces have blocked the delivery of aid and shelled promised humanitarian evacuation corridors. In mid-March, the ICRC described “extreme or total shortages of the basic necessities of life.” By late March, U.S. Deputy Secretary of State Wendy Sherman told the U.N. Security Council that people were beginning to die of starvation in the city. In a report published on April 13, the OSCE Moscow Mechanism’s mission of experts referred to reports describing between 150,000 and 300,000 civilians “trapped” in the city with “very little water and food” (p.32).

The contours of the law on encirclement deprivation are more contested than is the law of deprivation listed in article 54(2). Three distinct approaches warrant consideration here: the weaponized civilian suffering approach, the transitive deprivation approach, and the targeted deprivation approach. Even on the narrowest of these interpretation, the siege of Mariupol has likely violated the prohibition.

Approach 1: Weaponized Civilian Suffering

On one end of the interpretive spectrum, the prohibition of encirclement deprivation has sometimes (para. 2089) been (p. 761) understood (p.19) to attach only to operations in which civilians are purposefully starved. On the most restrictive version of this interpretation, a siege would violate the rule only if imposed with the specific purpose of weaponizing the civilian suffering associated with starvation. From this perspective, the absence of encirclement among the modalities listed in the article 54(2) on indispensable objects is thought to preclude the application of that framework to encirclement denial (pp.762-765). Moreover, although the general ban on “starvation of civilians as a method of warfare” in article 54(1) applies to all modalities of starvation, “starvation” is understood in this context to refer to the suffering associated with insufficient food to live. With its use as a “method of warfare” taken to imply its purposive weaponization (p.761), the ban on starvation is therefore thought to attach only to actions that deliberately produce civilian suffering (most obviously as a way to break civilian will). On this view, inflicting starvation conditions on a population of civilians and combatants with the goal of breaking the will of combatants would not be prohibited, because the civilian suffering arising from that action would not be the operation’s purpose (p.19). It may also be noted in defense of this narrow view that the rules on humanitarian access in article 70 of Protocol I preserve significant state discretion by specifying that relief actions in contexts of inadequate supply “shall be undertaken subject to the agreement of the parties concerned” (pp.26-30). On this interpretation, encirclement deprivation would be governed by a standard fundamentally different from that applicable to deprivation via the attack, destruction, removal, or rendering useless of indispensable objects.

Adopting this understanding would mean that encirclement deprivation undertaken with the purpose of starving out combatants would not be prohibited, even if the inevitable consequence of such an operation would be the starvation of the civilian population within which those combatants are embedded. Given the challenges associated with determining whether the encirclement of an area with a mixed population of civilians and combatants is directed toward weaponizing civilian or combatant suffering, it would be difficult to establish violations even when they do occur.

Notwithstanding that difficulty, the siege of Mariupol may satisfy even this high threshold. Most significantly, the aforementioned reports of Russian attacks on civilians seeking to exit via promised humanitarian corridors weigh heavily against interpreting the encirclement deprivation of those that remain as driven by the sole purpose of starving out the combatants within.

However, in the absence of actions of that sort, it would typically be very difficult to establish whether the purpose of cutting off an encircled population is to weaponize civilian suffering or to pursue an alternative military aim, such as combatant starvation. Indeed, on this narrow interpretation, it would be difficult to clearly identify a siege operation along these lines as per se unlawful, even though an otherwise identical operation performed through the attack, destruction, removal, or rendering useless of such objects would plainly violate the terms of article 54(2-3) of Protocol I. The “incidental” civilian starvation would instead be considered in the proportionality analysis (p.19; p.765; and here). The risk, of course, is that the military advantage associated with a successful siege may reduce that “rule’s humanitarian effect, perhaps to the vanishing point” in many contexts. A more protective way of thinking about starvation methods in purposive terms is discussed in Approach 3, below, which includes a discussion of the indiscriminate weaponization of starvation.

Approach 2: Starvation as the Transitive Act of Deprivation

A far better interpretation (which I have developed at greater length here, sections V(D-F)) would narrow that normatively indefensible disparity. As used to describe a method of warfare, the term “starvation” should be understood to refer not to an outcome, but to the transitive act of depriving people of indispensable objects. Understood in this way, “starvation of civilians as a method of warfare” means the deliberate deprivation of objects indispensable to civilian survival as part of a belligerent’s war effort (see, e.g., para 741 of the report of the U.N. Group of Eminent Experts on Yemen). The deprivation of such objects can be purposive in the way suggested by “method of warfare,” without targeting civilians or being undertaken with the goal of weaponizing civilian suffering. Indeed, this is an accurate characterization of the actions prohibited by article 54(2)-(3). On this reading, the role of those latter paragraphs would be to specify the ban through exemplification, while also clarifying how it diverges from the general article 52 framework on object protection. An alternative route to a similar conclusion would be to understand the concept of “rendering useless” more broadly than is suggested by the ICRC Commentary (paras. 2100-2101), such that the obstruction of deliveries would be understood to render those consignments useless to the civilians for whom they were destined.

This approach is more coherent with the broader legal context. It makes sense of the notion (expressed in the ICRC Commentary) that the framework banning the attack, destruction, removal, and rendering useless of indispensable objects in article 54(2) “develops the principle formulated in paragraph 1” and “describes the most usual ways in which this [general principle] may be applied” (para. 2098). That framing that makes little sense if the prohibition in paragraph 2 is understood to far exceed a narrow understanding of the ban in paragraph 1 or to exhaust all the ways in which the principle may be applied. It similarly accounts for the codification of “starvation of civilians as a method of warfare” in the first paragraph of article 54, which is entitled “protection of objects indispensable to the survival of the civilian population.” This approach is also consistent with the avoidance of any normative distinction between modalities of the deprivation of indispensable objects in article 54. In particular, it coheres with the focus on the deprivation of indispensable objects as the core material element of the starvation war crime (p.31), which is explicit in including willful encirclement denial as a form of that deprivation (ICC Statute, art. 8(2)(b)(xxv)). And it is similarly compatible with the U.N. Security Council’s landmark Resolution 2417, which “strongly condemns …. depriving civilians of objects indispensable to their survival, including willfully impeding relief supply and access” (para. 6). Willfulness in this context is understood ordinarily to describe not purposive action, but rather “the attitude of an agent who, without being certain of a particular result, accepts the possibility of it happening.” (ICRC para. 3474; see also IT-01-42-A paras. 270, 277; IT-98-29-A para. 140; IT-98-30/1-A para. 261).

Assuming that encirclement deprivation should indeed be understood to be regulated by the same framework as other modalities of deprivation, the first legal question in such a context would be whether the supply of indispensable objects is being deliberately blocked. If yes, as appears to have been the case in the siege of Mariupol, the next question is whether the purpose is to deny sustenance (even to combatants) in a context in which that denial will also impact civilians. If yes, the operation would be prohibited. In the case of a siege of a military base or holdout without a civilian presence, on the other hand, the prohibition would not apply. It appears unlikely that the latter is an appropriate characterization of the encirclement of Ukrainian forces within the Azovstal steel plant, with thousands of civilians reported to have been sheltering there earlier in the war and as many as one thousand reportedly still there earlier this week.

Assuming access to be blocked for reasons other than sustenance denial, the alternative question (suggested by article 54(3)(b)) would be whether those other reasons include the prevention of direct support to military action. This test is unlikely to be satisfied in the context of food consignments, at least as long as they can be subject to comprehensive and effective search and examination. However, it could in principle be met in the context of either (i) agricultural equipment that could equally be used for military operations or (ii) objects, such as fuel and electricity, that are necessary to the preservation or supply of food, water, and heating. Even in that context, the denial of access would be prohibited if conditions at the time were such that the civilian population would be left starving or forced to move. Notably in this respect, having determined (controversially) that Israel is no longer the occupying power in Gaza, the Israeli High Court of Justice evaluated the delivery of fuel and electricity into Gaza with reference to the customary analogues of articles 54 and 70 of Protocol I, citing approvingly the interpretation that “a party to a conflict may not refuse to allow the passage of foodstuffs and basic humanitarian equipment necessary for the survival of the civilian population” (HCJ 9132/07, para. 14). Similarly, on the issue of encirclement deprivation, the ICRC Commentary on Protocol II reasons that refusing humanitarian relief in a context in which the survival of the population is threatened “would be equivalent to a violation of the rule prohibiting the use of starvation as a method of combat” (para. 4885). On this interpretation, such operations would be prohibited regardless of whether civilians were the targets of deprivation.

Approach 3: Targeted Deprivation

One might accept the characterization of “starvation of civilians as a method of warfare” as describing the deprivation of indispensable objects, rather than the weaponization of an outcome, and yet insist that the central place of “civilians” in the concept means that the general prohibition should be understood to attach only to deprivation targeted at civilians. This interpretation would render the general ban (and thus the ban on encirclement deprivation) narrower than the broad prohibition on the attack, destruction, removal, or rendering useless of indispensable objects. However, as I have elaborated elsewhere (part VI) it would still avoid the difficulties created by the weaponization of civilian suffering approach. Three features of civilian status and protection are important to understanding why. First, civilians who remain in a besieged area retain full civilian status and protection under the law, regardless of whether they refused to exit or were denied safe exit. Second, a population that is composed predominantly of civilians is a civilian population and has legal protection as such in its aggregate form. Third, the indiscriminate targeting of civilians and combatants may be understood to include within it the targeting of civilians. Consider these points in turn.

Everyone who has not joined the armed forces of a party to an international armed conflict, an armed group belonging to such a party, or a levée en masse is a civilian in that armed conflict (art. 50(1) AP I). Civilians are protected as such, including by the rules on starvation and indispensable objects, “unless and for such time as they participate directly in hostilities” (art. 51(3) AP I). Pursuant to this standard, the prior or current availability of egress through humanitarian corridors (as have been provided successfully at certain moments during the Ukraine conflict) is irrelevant to the status of those who remain. Whatever one makes of debates regarding the status of voluntary human shields (e.g. HCJ 769/02 para. 36, ICRC pp. 56-57), it is implausible to argue that such a characterization could extend to those who decide not to abandon their homes (whether during a siege or otherwise). The error in asserting an obligation to leave is only amplified when purportedly humanitarian corridors would require passage into Russia or Belarus, potentially implicating the crime against humanity of forced deportation (art. 7(1)(d) ICC Statute).

Plainly, the status of those forced to stay by the tactics of the besieging party is also unaffected by the fact of remaining. In scenarios in which the besieged force were to prevent civilian exit, it may violate both its obligation to take “necessary precautions to protect the civilian population [and] individual civilians … under [its] control against the dangers resulting from military operations” (art. 58(c) AP I) and potentially its duty not to use civilians’ presence to “render certain points or areas immune from military operations” (art. 51(7) AP I and art. 8(2)(b)(xxiii), ICC Statute). However, those violations would neither change the status of those civilians nor release the besieging party from any of its obligations to them (art. 51(8) AP I).

It is now well-established that the presence of combatants or persons participating directly in hostilities affects neither the civilian character of the population as a whole nor its protection as such under IHL (art. 50(3) AP I). The International Criminal Tribunal for the former Yugoslavia found that the encircled population during the siege of Sarajevo retained its civilian status, notwithstanding the presence of ensconced combatants, because it was “predominantly civilian” (ICTY IT-95-5/18-T para. 4610 n. 15510). The same would be true of any predominantly civilian population in Ukraine.

Aggregate population status aside, perhaps the most foundational guiding principle in contemporary IHL is that parties to the conflict must “at all times distinguish between the civilian population and combatants” and “direct their operations only against military objectives” (art. 48 AP I). The reference to “operations” rather than “attacks” indicates the direct applicability of this principle to encirclement, even in the absence of acts of violence. Additionally, and of normatively analogical significance, the elaboration of that rule as applied to attacks further provides that such actions must be targeted at discrete and specific military objectives and that they may not be launched indiscriminately at civilian areas, even if there are military objectives within those areas (art. 51(4)(a-b), (5)(a), Protocol I). Although lacking a codified war crime along these lines, the ICC has determined that an attack directed indiscriminately at “civilians and fighters alike” can qualify legally as an attack on civilians (ICC-01/04-02/06-2359 paras. 921-923; ICC-01/04-02/06-2666-Red paras. 418, 424, 491; ICC-01/04-01/07-3436-tENG paras. 801-802). This line of reasoning may also suggest that indiscriminate weaponization of starvation could also qualify as the intentional use of starvation against civilians under a modified version, or well-reasoned application, of Approach 1.

Taking these points together, even if the siege of Mariupol were shown to be driven by the ultimate objective of starving out combatants, the pursuit of that ultimate objective required the adoption of a predicate purpose—namely, depriving the population as a whole of food, water, and other essentials. Insofar as that population remained civilian in its aggregate character, the ultimate objective of starving out combatants was necessarily pursued via the targeted deprivation of a civilian population. Combatant starvation (and capitulation) in such a scenario is pursued via the targeted starvation of civilians as a method of warfare. Approaching the issue from a slightly different angle, an encirclement operation along these lines can also be understood to be an indiscriminate deprivation of food and water to all within the encircled area. This may appropriately be understood as deprivation targeted at both combatants and civilians, and thus qualify as the targeted starvation of civilians as a method of warfare.

War Crimes

Beyond the question of whether Russian methods have violated the IHL prohibition of starvation of civilians as a method of warfare, a second question is whether any of the relevant conduct could qualify as criminal starvation. The appropriate starting point is the ICC Elements of Crimes document. In addition to requiring material and mental elements relating to the association with an international armed conflict, it provides that the crime attaches whenever a perpetrator both “deprived civilians of objects indispensable to their survival” and “intended to starve civilians as a method of warfare” (p.31). Crucially, it is not necessary to establish that the impugned conduct caused any particular form of civilian harm, suffering, or death. Rather, the focus is on the deprivation of indispensable objects and the intent to starve civilians.

For the reasons already articulated, the material element is very likely satisfied in various contexts in Ukraine, as Russian troops appear to have engaged in multiple forms of the deprivation of indispensable objects. The critical question is whether the individuals involved in those actions “intended to starve civilians as a method of warfare.” This is, of course, a question that requires individualized analysis. Nonetheless, it is important to be clear about the necessary legal threshold. Interpretive questions parallel to those applicable to the underlying IHL prohibition arise here, too.

On the most restrictive interpretation, the use of “intent” (perhaps together with “method of warfare”) might be understood to imply that this element would be satisfied only when the perpetrator acts with the purpose of weaponizing the civilian harm or death associated with starvation. This approach would apply the narrow IHL understanding discussed above not only to encirclement deprivation but also to the attack, destruction, removal, or rendering useless of indispensable objects in the war crimes context. Here, too, it is not a compelling approach. Certainly, the reference to “intent” in article 8(2)(b)(xxv) does not require narrowing the scope of the war crime in this way. On the contrary, a careful reading of the term suggests a more comprehensive criminal ban.

First, if “starvation of civilians as a method of warfare” is understood to refer to the transitive act of deprivation, then even an understanding of “intent” that is limited to direct (or purposive) intent would require establishing only that the deprivation was purposive, not that this entailed the purpose of weaponizing civilian suffering. Moreover, as discussed above, even if it were thought necessary to show that the deprivation was targeted at civilians, precisely that could be shown whenever the impugned acts were either targeted at a population that is civilian in aggregate or targeted indiscriminately at civilians and combatants. Indeed, the ICC has indicated that attacks targeted indiscriminately at combatants and civilians can satisfy the specific criminal threshold for the war crime of “intentionally directing attacks against the civilian population or against individual civilians” (ICC-01/04-02/06-2359 paras. 921-923). Other tribunals have reached similar conclusions (Extraordinary Chambers in the Courts of Cambodia, para 310; International Criminal Tribunal for the former Yugoslavia, para. 260).

Second, even if “intentionally using starvation of civilians as a method of warfare” were thought to imply the intentional infliction of a particular form of suffering on civilians (i.e., to implicate starvation as an outcome, rather than as a transitive act), it does not follow that that outcome must be the purpose of the impugned act for the crime to attach. On the contrary, as it relates to consequences, “intent” is defined in the ICC Statute to include not only purposive intent, but also awareness that the proscribed consequence will occur in the ordinary course of events (art. 30(2)(b), ICC Statute). This form of oblique intent has been understood to include a perpetrator acting with virtual certainty that the consequence in question will occur, whether or not she seeks that result (ICC-01/04-01/06-3121-Red, paras. 447-450). Although that definition is technically applicable only to the default mens rea standards, the oblique form of intent is far from anomalous in criminal law generally, and the ICC has recognized that crime-specific uses of “intentionally” should be read in light of that general meaning unless there is a good reason not to do so (ICC-01/05-01/13-2275-Red, para. 677). Given that article 54(3)(b) of Protocol I is explicit in prohibiting the deprivation of indispensable objects when that may be “expected” to result in the starvation of civilians, this approach also coheres with the underlying regime.

In short, where the material elements are established, criminal liability ought to attach whenever the acts of deprivation were purposive and applied indiscriminately or to a population that is civilian in its aggregate character. Even if a narrower interpretive approach were adopted such that intent vis-à-vis civilian starvation as an outcome were required (despite its absence from the material elements), this could be satisfied through showing that outcome to have been either the purpose or the inevitable consequence of the deprivation actions.


Civilian starvation is among the most devastating features of contemporary armed conflict, with torturous implications for individuals and communities. It arises not from intractable scarcity, but from particular choices about how to conduct war. The legal framework necessary to hold perpetrators to account is now available in many of the jurisdictions examining the situation in Ukraine. The interpretive questions that such jurisdictions will face are real, but they are not insurmountable. The immediate imperative is to collect and preserve evidence to provide the foundation for accountability, however those questions are answered.

Photo credit: Russian soldiers walk along a street in Mariupol on April 12, 2022, as Russian troops intensify a campaign to take the strategic port city. (Alexander NEMENOV/AFP via Getty Images)