On 23 July 2025, the ICJ delivered its advisory opinion on the Obligations of States in Respect of Climate Change. Although much of the opinion is devoted to the clarification of an extensive set of primary obligations, it also addresses certain key questions in the law of State responsibility stemming from distinctive aspects of climate change, including the diffuse and cumulative character of causes and harms. Moreover, the Court’s determination of some of these questions is of wider application and fundamental importance. In this post, we focus on (I) the Court’s general approach to responsibility; the issues of (II) attribution, (III) causation, and (IV) the erga omnes (partes) character of the primary obligations; and (V) the remedy of guarantees of non-repetition. We conclude by emphasizing the Court’s systematising role in the legal order.
The Court’s General Approach to Responsibility
Question (b) put by the General Assembly provided: ‘What are the legal consequences under [the obligations identified in question (a)] for States where they, by their acts and omissions, have caused significant harm to the climate system and other parts of the environment . . .?’ At the outset, the Court framed Question (b) as only requiring it to set out the general legal framework governing responsibility for the breach of obligations pertaining to the protection of the climate system (for criticism, see Judge Yusuf (paras 5-11)). It did not understand itself to be tasked with establishing ‘the individual responsibility of a State or group of States for the damage caused to the climate system’ (para 406). The Court also explicitly excluded consideration of ‘the rules and principles governing injuries arising out of acts not prohibited by international law’ – also criticized by Judge Yusuf (paras 41-48) and discussed by Judge Nolte (paras 15-17).
As to its overarching approach, the Court found that the rules on State responsibility under customary international law were applicable (paras 405-407). There had been important disagreement on this point amongst participants—disagreement, as the Court put it, on ‘whether the customary rules on State responsibility are excluded by virtue of the application of lex specialis in respect of the legal consequences for breaches of obligations to protect the climate system and other parts of the environment.’ On this question, after accepting the possibility in principle of special rules of responsibility, the Court assessed whether a ‘discernible intention’ of exclusion or inconsistency could be found, with a focus on (a) the loss and damage and compliance mechanisms in the Paris Agreement and (b) the dispute settlement mechanisms in the UNFCCC, incorporated into the Kyoto Protocol and Paris Agreement. Across these instances, the Court found no intention of the parties to exclude or vary the generally applicable rules of responsibility (para 414).
In this respect, we see the Court playing a systemizing role in relation to the law of State responsibility. In fact, that role is two-fold. First, at the beginning of this assessment the Court found that the rules set out in the ILC’s Articles on State Responsibility (ILC Articles) ‘in many respects are reflective of the customary rules on State responsibility’ (para 407, emphasis added). This is not unexpected, and confirms that although, strictly speaking, any assessment of custom must be individually focused on a particular rule, in practice the ILC’s work, as supported and confirmed by the Court, has encoded into the legal system a general framework for responsibility. Second, the Court endorsed a high threshold for specific departures from these general rules, as set out in Article 55 of the ILC Articles.
Attribution (and More)
On attribution, the Court began by distinguishing two senses in which the term ‘attribution’ was used throughout the proceedings: as the ‘process of evaluating the relative contribution of multiple causal factors to a change or event with an assessment of confidence’, and as the ‘operation of attaching a given action or omission to a State under international law’. The former is a concept in the field of climate science, and the latter a concept of State responsibility. The Court notes the latter is one of the elements necessary for responsibility to arise (para 425).
As to legal attribution, the Court’s approach is orthodox, though it does include in paragraph 427 some curious language in relation to fossil fuels that looks more like the articulation of a primary rule than the reiteration of a basic principle of attribution: ‘Failure of a State to take appropriate action to protect the climate system from GHG emissions — including through fossil fuel production, fossil fuel consumption, the granting of fossil fuel exploration licences or the provision of fossil fuel subsidies — may constitute an internationally wrongful act which is attributable to that State.’ (On which, see also the Joint Declaration of Judges Bhandari and Cleveland). That aside, on legal attribution the Court restated the basic principle that States are responsible for the conduct of their organs, including both their acts and omissions. In relation to private conduct, moreover, the Court also reasserted the basic proposition that while private conduct is not attributable to the State, a State may nevertheless be responsible for its failures to exercise due diligence in the regulation of private conduct, including omitting to limit the quantity of emissions caused by private actors under their jurisdiction (para 428).
Beyond this point, this section of the judgment is somewhat underdeveloped, and seemingly addresses matters other than legal attribution under the law of State responsibility. The Court notes the complexity around the existence of a plurality of wrongdoing States (and injured States) but emphasizes that it is ‘scientifically possible to determine each State’s total contribution to global emissions, taking into account both historical and current emissions’ (para 429). It also finds that other courts and tribunals had ‘considered the link between GHG emissions and climate change, the link between climate change and adverse effects suffered by litigants, the link between such harm and the actions or omissions of a particular State, and the attributability of responsibility for such adverse effects.’ Perhaps inevitably, these general statements prefigure future questions.
As to the issue of multiple contributing States, the Court found that the general rules of responsibility can address the problem. Citing its Reparations Judgment in Armed Activities, the Court held: ‘in certain situations in which multiple causes attributable to two or more actors have resulted in injury . . . responsibility for part of such injury should [be] allocated among [the] actors’ (para 430). By way of aside, it may be noted that this was in contrast to the other possibility mentioned in Armed Activities, that ‘in certain situations in which multiple causes attributable to two or more actors have resulted in injury, a single actor may be required to make full reparation for the damage suffered’ (para 98). The allocation of responsibility amongst multiple actors remains under-developed in the practice of international law, although scholarship has addressed these issues in some detail. At least at the level of principle, and even if much is still to be determined, the Court’s approach is significant.
Consequentially, the Court’s approach entailed the conclusion that: ‘in the climate change context, the Court considers that each injured State may separately invoke the responsibility of every State which has committed an internationally wrongful act resulting in damage to the climate system and other parts of the environment.’ (para 432). All ‘[f]actual questions arising in the context of attribution’ – presumably here meaning scientific attribution on the Court’s own definition – and ‘apportionment of responsibility’ are pushed for resolution on a case-by-case basis.
Causation
The AO is also important in relation to the question of causation, a central issue in contention among participants. As a matter of first principles, the Court holds that: (i) ‘causation of damage is not a general requirement for the determination of responsibility’ and (ii) that causation is relevant to reparation, insofar as it is necessary to establish a causal link between the State’s conduct and damage to the injured party. Whether causation is necessary thus turns on the relevant primary rule and the remedy sought. The Court then makes four remarks on this issue.
First, the Court rejects the view that it is impossible to establish causation due to the diffuse nature of climate change causation, and also rejects the idea that causation might be presumed in this context (para 435). That damage results from multiple causes does not exempt a State from responsibility for its contribution to the damage; but, in all cases, causation needs to be proven.
Second, while acknowledging that causal standards may vary depending on the relevant primary rule, the Court held that the standard it has developed in its case law is ‘capable of being applied’ in the context of climate change obligations. The standard requires the existence of ‘a sufficiently direct and certain causal nexus between the wrongful act . . . and the injury suffered by the Applicant’; a standard which, according to the Court, is ‘flexible enough’ to address the challenges arising in the context of climate change (para 436).
Third, the Court explained that in the context of climate change, causation requires a two-step process:
First, whether a given climatic event or trend can be attributed to anthropogenic climate change; and second, to what extent damage caused by climate change can be attributed to a particular State or group of States. While the second element must be established in concreto in respect of specific claims brought by States in respect of damage, in many cases the first element may be addressed by recourse to science. (para 437)
Fourth, in line with comments made in the context of attribution, the Court notes the difficulties in establishing causation in respect of environmental damage: such damage may result from multiple causes, or the scientific knowledge on the causal link may be uncertain. However, the Court takes the view that while the causal link between State conduct and harm arising from climate change is ‘more tenuous than in the case of local sources of pollution’, establishing causation in this context is not impossible. The assessment must be made in concreto in light of the elements outlined by the Court (paras 437-438). The Court thus seems to allow that a ‘more tenuous’ connection may be sufficient to meet the standard of a ‘direct and certain’ link.
The Court’s observations on causation are laconic and, aside from clarifying in general terms the applicable legal standard, do not add much in substance. Two issues would have benefitted from further consideration. First, the Court does not address the principles of factual and legal causation applicable in cases of multiple causes, which can involve particularly difficult assessments in connection with omissions. On this point, the Court simply asserts that climate science is capable of addressing these situations. As Judge Nolte points out in his separate opinion, climate science can assist (to an extent) with assessments of fact, but only partially with resolving the legal question whether a certain harm was caused by a specific State’s wrongful act (para 20). As Judge Nolte’s puts it: ‘Natural science cannot help in determining which GHG emissions since the beginning of industrialization have resulted from internationally wrongful acts, and natural science can help only to a limited extent in determining whether a given wrongful act translates to a particular quantity of GHG emissions. In particular, the question whether a State has undertaken its best efforts to reduce its GHG emissions is not one that natural science can answer; nor can natural science determine whether a particular level of GHG emissions amounts to a violation of the obligation to undertake best efforts’.
Second, the Opinion seems to overlook the distinction between the two roles causal analysis might play in the law of State responsibility. Causation may first be relevant to establishing the existence of a wrongful act in cases where the primary rule requires the occurrence of a particular harm or consequence. Once a wrongful act has been established, causation may also be relevant in the determination of the extent of a State’s compensatory obligation in light of its individual contribution to the harm. These two steps of causal analysis are well-known in domestic legal systems, acknowledged in the relevant literature in international law (eg, Besson, 311; Nedeski and Nollkaemper), and were (possibly) distinguished by the Court in Bosnian Genocide (para 432). The standards applicable in each step may differ, with the first step often involving a lower standard like foreseeability (for wider discussion, see Lanovoy). This clarification would have been significant in this context, especially in connection with the prevention duties recognised by the Court in relation to Question (a). Only the first step is necessary to establish the breach of these duties, which is sufficient to trigger obligations of cessation and guarantees of non-repetition by the responsible States.
Erga Omnes Obligations
The Court also found that States’ obligations ‘pertaining to the protection of the climate system and other parts of the environment from anthropogenic GHG emissions, in particular the obligation to prevent significant transboundary harm under customary international law’ are obligations owed orga omnes. Similarly, treaty obligations under the UNFCCC and Paris Agreement are erga omnes partes, on the basis that they ‘protect the essential interest of all States in the safeguarding of the climate system, which benefits the international community as a whole’ (para 440). The Court connects the erga omnes character of these obligations to the nature of the environment as a ‘global common good’ (see also Judge Tladi’s declaration at para 34). Notable here is the sweeping character of these findings, relating to whole bodies of law.
The important consequence, which is explicitly flagged by the Court, is that all States or all State Parties have an interest in compliance such as to enable them to invoke the responsibility of other States for failing to fulfil these obligations. Of course, such measures of invocation might be diplomatic, but the judgment also opens up the possibility of further contentious litigation before the Court itself. Certainly, this aspect has been captured in some popular reporting, with the BBC headline reading: ‘Top UN court says countries can sue each other over climate change.’ In addition to general instruments such as the Pact of Bogotá, one possible route arises in relation to States which have recognized the jurisdiction of the Court as compulsory under the optional clause. In addition to the requirement of reciprocity, of particular importance will be the existence of reservations ratione materiae, whether in relation to climate change or the environment more generally (see e.g. Romania and Slovakia). As the Court itself notes, any such claim by a non-injured State cannot be for reparation for itself (para 443). Perhaps as to be expected, the Opinion makes no mention of the possibility of countermeasures, by either injured or non-injured States.
The Court thus for the first time explicitly acknowledged the legal standing of all States in relation to obligations erga omnes (in the sense of Article 48(1)(b) ASR). Until now, the Court’s decisions on collective obligations had been confined to obligations erga omnes partes arising under treaties (see Urs). Acknowledging the legal standing of all States in relation to obligations erga omnes under customary law was, to be sure, the logical next step (or even one implicit in the existing approach). But until now, the Court had not clarified this point. Moreover, whereas the Court had until now refrained from citing Article 48 of the ILC’s Articles in this context, at paragraph 442 the Court quotes this provision. It seems clear that this Opinion cements the customary status of the jus standi of States other than an injured State in relation to erga omnes obligations.
Guarantees of Non-Repetition
One final, albeit somewhat narrow, point arises. In its discussion of legal consequences, the Court held: ‘Additionally, in appropriate circumstances, a responsible State could be required to offer appropriate assurances and guarantees of non-repetition (see ILC Articles on State Responsibility, Article 30 (b)).’ Here, in relation to assurances and guarantees of non-repetition, we see (again) a mutually reinforcing relationship between the Court and the ILC. In LaGrand, Germany (para 6.71) had demanded ‘formal assurances that the United States will bring its practice in conformity with the requirements of international law, without laying out in detail whether these modifications are to be brought about by formal changes in its domestic law or simply by changing the practical application of its respective legislation.’ In the dispositif in LaGrand, the Court took note of the United States’ assurances to ensure implementation of measures adopted in performance of its obligations and found this commitment sufficient to meet Germany’s request (for discussion, see e.g. Stoica; Tams). Shortly hereafter, citing LaGrand, the ILC included assurances and guarantees of non-repetition in Article 30(b) ASR, which has now been confirmed (implicitly as a customary rule) by the Court in the AO. Much as in relation to necessity as a circumstance precluding wrongfulness, the Court and ILC act here as partners in the articulation and development of general rules of international law.
Conclusion
Even leaving aside the Court’s extensive clarification of primary obligations, there are many other points of interest in the opinion, including its recognition that primary rules may give individuals the right invoke responsibility (para 111, again with reference to the ILC Articles [33(2)]) and the Court’s reiteration, following Armed Activities, of the possibility of the award of compensation in the form of a global sum in the face of ‘uncertainty with respect to the exact extent of the damage caused’ (para 454). More widely, in keeping with its approach to questions of sources, in relation to responsibility the Court placed great emphasis on generally applicable rules. Moreover, as set out above, it continues to draw on and consolidate the ILC’s work, acting to build and systematise fundamental aspects of the international legal system. Of course, it remains to be seen how these general frameworks will be applied (and adapted) as climate change is litigated in other fora.