Beyond Headcounts

Editor’s Note: This post is part of a symposium relating to the ICRC’s customary international humanitarian law study, featured across Articles of War and Völkerrechtsblog. The introductory post is available here. The symposium highlights presentations delivered at the young researchers’ workshop, Customary IHL: Revisiting the ICRC’s Study at 20, hosted by the Institute for International Peace and Security Law (University of Cologne) and the Institute for International Law of Peace and Armed Conflict (Ruhr University Bochum) on September 18-19, 2025.

The ICRC’s 2005 Customary International Humanitarian Law study (“the Study”) adopted a deliberately universalist method to identify customary IHL. It downplayed the “specially affected States” doctrine on the premise that in IHL, “all States have a legal interest” in compliance and therefore the practice of all States should count equally. In effect, the Study treated state practice and opinio juris as a global headcount and on that basis declared a wide range of humanitarian rules to be customary. Critics, above all the U.S., argued this misdiagnoses custom by treating as equivalent the practice of States with little combat experience and that of States whose forces routinely operate under fire, even though the latter generate more, and more operationally probative, evidence. The risk of an arithmetic approach is blatant: it can produce formal majorities while obscuring whether those most exposed to a rule actually accept and follow it.

On the doctrine’s outer edge, some readings treat the role of specially affected States as close to a veto – if “important actors do not accept the practice, it cannot mature into a rule of general customary law” (p. 26). The U.S. did not frame it that starkly: in its response to the Study, it faulted the Study for “fail[ing] to pay due regard to the practice of specially affected States,” invoking the ICJ’s North Sea Continental Shelf formula that such practice must be examined and carry particular weight (paras. 73f.).

This post sketches a middle path. It retains broad inclusion but proposes calibrated weighting: “specially affected” here refers to states with sustained, transparent operational exposure to the specific rule’s domain (not merely great-power status), whose practice/opinio juris – when shared by a plurality – warrants greater evidentiary weight in diagnosing custom. Two operator-heavy clusters, reprisals against civilians and environmental protection, show how calibrated weighting yields a more accurate and legitimate account of customary IHL than headcounts alone.

Reprisals against Civilians: Treaty Consensus, Customary Contestation

The ICRC’s 2005 Study presents a prohibition on reprisals against civilians as customary. Building on the treaty ban in the 1977 Additional Protocol I (Art. 51(6)), the ICTY’s Kupreškić judgment posits those provisions had “subsequently been transformed into general rules of international law” binding even on non-parties, relying on the Martens Clause and the ‘demands of humanity’ notwithstanding ‘scant or inconsistent’ usus (paras. 527–531).

Since 2005, practice and opinio juris among high-operational-capacity states have not converged; if anything, they do not match the universalist account advanced by the Study and Kupreškić. The U.S., never a party to AP I, has long maintained that it retains a right of belligerent reprisal in extreme cases; in 1987 its State Department Legal Adviser warned that renouncing even the option “removes a significant deterrent,” and the position is reflected in the 2015/16 DoD Law of War Manual’s (updated in 2023) refusal to treat AP I’s reprisal prohibitions as customary. The UK ratified AP I but entered a reservation preserving a last-resort reprisal option; UK practice accepts a policy constraint but does not concede an absolute customary ban in all circumstances. By contrast, Germany’s Joint Service Regulation (ZDv) unambiguously prohibits reprisals against civilians, objects indispensable to their survival, and the natural environment, and Canada’s manual reflects an analogous prohibition. Several other states that once hedged – e.g., Italy, Egypt, France – now affirm the ban in national doctrine. The result is a sharp split: a very large treaty-backed majority professes that reprisals on civilians are unlawful, while a small but militarily consequential minority (led by the U.S. and UK) refuses to concede a universal customary rule.

A headcount method would downplay that minority and declare the custom settled. A calibrated approach asks instead who is specially exposed, whether they have been transparent about their legal view, and whether there is a plurality of such states. On reprisals, the principal operators – especially the U.S. and UK – are the ones articulating dissent, and they have done so consistently and publicly. They are not entirely alone: Israel and Türkiye are not parties to AP I, and neither state’s publicly available materials clearly adopt an AP I-equivalent blanket customary ban. Weighting that practice for capacity and transparency does not hand anyone a veto, but it does counsel a more candid description: a strong, near-universal trend toward prohibition, supported by broad treaty acceptance and the moral logic emphasized in Kupreškić, yet a contested customary status vis-à-vis a small, transparent plurality of specially affected States. That conclusion neither licenses reprisals nor ignores reality: in practice, even dissenters have refrained from declared reprisals against civilians for decades. Accounting for their stated positions, rather than pretending they do not exist, improves descriptive fidelity and – by identifying precisely where the friction lies – may ultimately strengthen the norm’s compliance pull by focusing engagement on the hold-outs.

A word on specially affected States and persistent objector status. The SAS point goes to whether a general customary rule has crystallized at all: under North Sea Continental Shelf, practice must be “extensive and virtually uniform,” including that of “States whose interests are specially affected.” If a transparent plurality of such operators maintains contrary opinio juris, the threshold for general custom is not met – no veto, but no crystallization either. The persistent-objector rule is different and applies once a general rule forms: a state that has objected from the outset and continuously is not bound by that rule. On reprisals, the U.S. would likely qualify as a persistent objector as well (from the late 1970s through its 1987 Legal Adviser statement, 2006 comments, and 2016/2023 DoD Manual). The UK, by contrast, is treaty-bound by AP I and therefore not a classic persistent objector; rather, considering its ratification statement, whatever one makes of its legal effect, it has been custom-skeptical while accepting the treaty ban. Either way, their weighted dissent matters at the formation stage because the North Sea formula requires convergence by the specially affected for a rule to be described as general custom.

Environmental Protection: Baselines Accepted, Absolute Bans Contested

The ICRC’s universalist approach is on full display in Rule 45, which folds AP I’s environmental provisions into custom: Article 35(3) bars means or methods of warfare expected to cause “widespread, long-term and severe” damage; Article 55 requires care for the environment and forbids environmental reprisals. The Study treated these as binding beyond AP I, citing state manuals, official statements (including from nuclear powers), and the ICJ’s Nuclear Weapons advisory opinion, which underscored that environmental considerations form part of IHL assessments. The thrust is a baseline duty on all states to avoid severe ecological harm in armed conflict.

State practice and positions among major operators are more qualified. The U.S. never accepted Article 35(3) as custom, criticizing in 2006 the Study’s elevation of “gap-filling” treaty rules to custom and insisting that environmental harm be assessed through necessity, distinction, and proportionality rather than a free-standing absolute ban. It has consistently resisted any view that nuclear use would be per se unlawful on environmental grounds, reiterating since 1977 that AP I was not intended to govern nuclear weapons. The UK and France, both permanent members of the UN Security Council and AP I parties, ratified with understandings that effectively preserve nuclear options in “extreme circumstances,” signaling that they did not regard AP I’s environmental (and related) limits as unqualified customary constraints in every scenario.

Outside the P5, several states push for universalization. New Zealand has long argued (including before the ICJ) that AP I’s environmental rules reflect binding custom for all states, and states such as Sweden, Switzerland, and Pacific Island countries have pressed in UN and ICRC fora for recognition of a no-environmental-destruction norm. Russia and China, both AP I parties, have voiced support for strengthening protection without clearly endorsing Article 35(3) as already customary. The UN International Law Commission’s 2022 Draft Principle 15 (“attacks against the natural environment by way of reprisals are prohibited”) captured that ambivalence: the commentary notes uncertainty whether it codifies existing custom or advances the law, and the Commission framed it as promoting development. Germany appreciated the inclusion, aligning with its doctrinal view; the U.S. urged deletion of any categorical rule, maintaining there is no current practice supporting a blanket ban on environmental reprisals.

Read through a calibrated lens, the pattern is consistent with a split outcome: broad acceptance of core IHL-based environmental constraints (no targeting the environment as such absent military objective, no disproportionate incidental damage), coupled with continuing contestation by key, specially affected States over an absolute AP I-style prohibition on “widespread, long-term, severe” harm or on environmental reprisals in all circumstances – especially in nuclear contexts. This does not deny protection; it marks the outer edge of consensus and shows where engagement is still needed.

Applying calibrated weighting yields a more textured account of custom. Start with capacity/exposure: foreground the practice of states with the means to inflict large-scale environmental harm – nuclear powers and advanced militaries (U.S, Russia, China, the UK, France, India, etc.). Their record and doctrine do not show a settled practice of forgoing severe environmental effects when core military aims are at stake. Many have been transparent about their legal views: the U.S. rejects an absolute “no widespread, long-term and severe damage” rule; the UK and France signalled on ratification that they did not consider themselves wholly foreclosed from environmentally damaging measures in extreme circumstances. On plurality, this is not a lone outlier: several states with significant force-projection capacity share similar hesitations about elevating AP I’s environmental provisions to universal custom.

The calibrated takeaway is two-part. First, core protections grounded in general IHL – no deliberate destruction of the environment as an end in itself; no disproportionate incidental damage – are widely accepted and function as customary baselines (even the U.S. frames environmental effects through necessity, distinction and proportionality). Second, the absolute AP I-style prohibition on widespread, long-term, and severe harm, and a categorical ban on environmental reprisals, remain contested among those most ‘in a position to be bound’ – i.e., states most exposed to and capable of generating W-L-S effects. Rather than declaring Rule 45 ‘custom, full stop,’ a candid diagnosis is that we are looking at an emerging general custom accompanied by a strong (but rebuttable) operational presumption against operations expected to cause W-L-S environmental damage and against environmental reprisals. That presumption is binding for AP I parties as treaty law and is operationally applied, and asserted as law, by many other States in their manuals and statements. It is not opposable to any State that qualifies as a persistent objector, and its status as general custom remains contested by a small, transparent plurality of specially affected States. The ‘moral momentum’ behind the presumption is evidentiary rather than constitutive: it refers to the humanitarian reasons states themselves invoke when articulating opinio juris, which help explain the direction of travel without doing the normative work on their own. In short, calibrated weighting affirms a customary core of environmental protection, while the outer limits, including an AP I-style absolute ban in extreme scenarios (e.g., nuclear use), remain unsettled pending convergence between the broad majority and the dissenting, specially affected states.

Methodological Takeaway

Calibrated weighting is not a veto for any one state. By conditioning extra interpretive weight on operational exposure, transparency, and plurality, only genuine, relevant, collectively significant dissent affects the identification of custom. The question is no longer merely how many states support a rule, but which states, how they actually behave, and why they take their positions. Read that way, calibrated weighting improves descriptive accuracy and the perceived legitimacy of IHL: it avoids proclaiming ‘universal’ rules that influential operators openly resist – proclamations that can erode, rather than bolster, compliance.

Acknowledging the perspectives of specially affected States is not deference; it is a method for turning friction into convergence. As scholars have suggested, a modernized, criteria-based version of the doctrine can promote optimal development and protect the legitimacy of custom by forcing high-operational-capacity states either to justify their practices or to adjust them to shared humanitarian values. An ICRC Study 2.0 could integrate this refinement, preserving the Study’s universalist ethos while recognizing that not all states shape or experience the law equally. Custom in IHL should be neither arithmetic nor great-power dictate; calibrated weighting offers the credible middle course.

The “Bofaxe” series appears as part of a collaboration between the IFHV and Völkerrechtsblog.