1. Introduction
The expansion of transboundary environmental risks – such as climate change, biodiversity loss, and pollution – has heightened demand for legal frameworks capable of responding to complex, uncertain, and politically contested ecological harm. In this context, the precautionary principle (PP) has emerged as a key tool in international environmental law, shifting the focus from traditional reactive measures to more structured decision-making under uncertainty.Footnote 1 Although the rich body of literature on the PP is not unanimous in how it positions the principle, it is broadly understood to call for action to avert risks of serious or irreversible harm to the environment or human health in the absence of scientific certainty about that harm, and offers the authority to take public policy decisions covering environmental protection in the face of uncertainty.Footnote 2
The PP and its role, overarching tenets, and implications have intrigued academic commentators for years, and most of these debates focus on regional and domestic justice (for example, in the European Union (EU) and the United States (US)),Footnote 3 ecological governance,Footnote 4 and World Trade Organization (WTO) dispute settlement.Footnote 5 As for the application of the PP in international investment disputes, while some scholars have argued for its applicability from perspectives such as cost–benefit analysis,Footnote 6 and Article 25 of the International Law Commission (ILC) Draft Articles on State Responsibility,Footnote 7 existing research remains relatively narrow. Most of it considers the principle from the standpoint of host states, suggesting that precautionary regulations might hinder innovation by increasing costs for firms through required safety and research measures, leading them to invest in countries with less strict safety rules to lower expenses. It also proposes policy suggestions aimed at incentivizing innovation,Footnote 8 or calls for treaty reforms, greater transparency in arbitration mechanisms, and the strengthening of environmental provisions in international law to protect the regulatory autonomy of states based on the PP.Footnote 9
While these contributions are valuable, there are several areas that could benefit from further exploration. Firstly, a systematic case study approach – particularly one that offers a comprehensive analysis of the case-specific use of the PP in international investment arbitration – remains underdeveloped. This has led to a lack of empirical assessment regarding the actual impact of the principle in specific cases, for example, whether it truly influences arbitral outcomes. Secondly, while existing literature often emphasizes the ecological protective function of the principle, there has been limited critical engagement with its operational effect. Few have considered the possibility that the principle may sometimes be invoked as a rationale for reasserting regulatory sovereignty rather than solely serving ecological goals. Such instances should not obscure the principle’s core function as a legitimate expression of sovereign responsibility to protect the environment. Finally, the complex interplay between science, law, and politics tends to be underexplored. Many institutional proposals, such as those advocating treaty reforms, focus primarily on redesigning substantive legal standards. Less attention has been paid to whether procedural approaches can structure decision-making under conditions of scientific uncertainty while mediating between regulatory autonomy and investment protection.
Adopting a functionalist lens, this study systematically examines published investor–state dispute settlement (ISDS) awards that invoke the PP to assess how it is used in adjudication. This review indicates that while the principle is often acknowledged procedurally, its application has yielded divergent arbitral outcomes. A recurring tension is thereby identified: the principle’s formal invocation does not consistently translate into a predictable legal standard for adjudication. This article contends that the inconsistent application of the PP arises not from its inherent vagueness but from a conceptual ambiguity regarding its adjudicative function. The solution lies in reconceiving the PP as a site for managing interactions among politics, science, and law – a discursive space where sovereignty claims, scientific uncertainty, and legal authority intersect. Within this framework, law asserts its normative force through procedural ‘stop rules’, which provide structured means to achieve closure amid competing systemic demands.
The democratic authority to design substantive regulatory frameworks lies primarily with sovereign governments through treaty negotiation and legislative processes. Clarifying treaty language to reflect contemporary understandings of the PP remains the most legitimate and durable route to reform. Yet, treaty reform is inherently slow and politically contested. In the meantime, arbitral tribunals continue to confront complex environmental disputes under older, textually ambiguous investment agreements. The purpose of this article, therefore, is not to bypass states or to advocate new substantive norms. Rather, it seeks to explore how tribunals, operating within the limits of their existing mandates, can employ procedural and interpretive tools, such as calibrated proportionality review and structured allocation of the burden of proof, to adjudicate uncertainty in a disciplined and predictable manner.Footnote 10
The structure of the following chapters is as follows. Section 2 traces the functionalist evolution of the PP, clarifying how the move from rigid rules to goal-oriented governance operates within international environmental law. Through an empirical study, Section 3 examines risk prevention, risk management, and nominalization as three recurrent framings of the PP in ISDS and finds that none of them supplies reviewable standards under scientific uncertainty. Section 4 reframes the PP as a procedural trigger, proposing proportionality review and reallocation of the burden of proof as concrete procedural mechanisms (‘stop rules’). Section 5 concludes by summarizing the central argument of the procedural turn and its broader implications for the theory and practice of international investment law, highlighting how the proposed mechanisms would contribute to preserving legal autonomy and structuring decision-making under uncertainty.
This conceptual approach is consonant with Karl-Heinz Ladeur’s account of stop rules, which emphasizes that legal systems must establish cognitive boundaries in decision-making to prevent scientific disputes from producing indefinite delay.Footnote 11 Building on this insight, the article reframes the ‘solution’ as a procedural spectrum: a family of adjudicatory governance devices ranging from calibrated rational-connection test and targeted burden-shifting to ongoing data-collection duties, reporting regimes, and periodic reassessment, rather than as a pair of isolated doctrinal techniques.
2. From Rules to Goals: The Materialization of Risk Governance
This section traces the PP’s shift from rule-based regulation to goal-oriented governance and identifies the tensions this shift creates in investor–state arbitration. Environmental principles are supposed to provide guidelines for government action and judicial review.Footnote 12 The PP has quickly become a central concept in international environmental law over the past three decades. As Ellen Hey observes, this principle ‘emerged not as a static rule but as a dynamic process of legal evolution responding to the epistemic challenges of ecological uncertainty’.Footnote 13 Some scholars contend that its emergence signals a paradigm shift in international environmental law, transitioning from a predominantly anthropocentric and economically focused framework to one that places ecological concerns at the forefront.Footnote 14 This shift can be attributed to two primary factors. Firstly, the implementation of the PP has transformed environmental law, moving it from providing mere regulatory oversight to serving as a proactive form of risk prevention.Footnote 15 Secondly, its rise is part of a broader trend in modern legal systems towards deformalization, or materialization, where law becomes more goal-oriented and flexible. Yet, as the following part will show, this same flexibility also exposes the PP to structural tensions when transplanted into the fragmented field of investment arbitration.
The emergence of the PP in German law and subsequent international codification reflect a functionalist commitment to anticipatory regulation. Designed to ‘anticipate and avoid environmental damage before it occurs’, the principle first emerged in Germany during the 1970s.Footnote 16 As it became embedded in German law, it evolved into a cornerstone of the country’s environmental legal framework.Footnote 17 Beginning in the 1980s, references to precaution and to the PP or precautionary approach started to find their way into international instruments, including a significant number of treaties.Footnote 18 The principle was codified in clear form for the first time at the international level in Principle 15 of the non-binding Rio Declaration on Environment and Development, which states:
In order to protect the environment, the precautionary approach shall be widely applied by States according to their capabilities. Where there are threats of serious or irreversible damage, lack of full scientific certainty shall not be used as a reason for postponing cost-effective measures to prevent environmental degradation.Footnote 19
Climate change has many remaining scientific uncertainties in terms of its timeline and the manifestation of specific climate impacts.Footnote 20 Moreover, unlike localized types of environmental harm, climate effects are diffuse, global, and protracted, with the result that governance tools originally designed for project-level impacts cannot be directly transposed to systemic risks of this kind.Footnote 21 The PP asserts that where there is a threat of serious or irreversible damage, the lack of full scientific certainty should not be used as a reason to delay cost-effective measures to prevent environmental harm. Despite its widespread acceptance, there is ongoing debate about the status and application of the PP in international law.Footnote 22 Critics argue that it is often too vague to guide action effectively, leading to paralysis in decision-making, or that it is anti-scientific, promoting irrational fear.Footnote 23 Current different understandings of the PP, informal scientific levels of certainty or uncertainty, and legal standards of proof may introduce subjectivity into the judgements. These criticisms stem from the general and malleable nature of the principle, which can lead to inconsistent application, overreaction to minor risks, or resistance to technological innovation based on unproven concerns.Footnote 24
Despite these criticisms, the PP seeks to achieve its functionalist goal by promoting the ‘goal-oriented’ nature of legal rules.Footnote 25 It is often discussed along two influential orientations: risk prevention and risk management.Footnote 26 While differing in emphasis, both share a fundamental characteristic: they conceptualize the PP as a normative principle with adjudicative functionality, tasked with guiding decision-making in the face of uncertainty and when clear standards are absent. In this sense, both paths view the PP as a tool for making substantive judgements where the legal framework lacks explicit guidelines, thus making it a goal-oriented legal norm. At its core, the PP serves not only to regulate and prevent risks but also to structure legal decision-making when scientific evidence is incomplete: rather than deferring to science for proof of harm, it recognizes that the identification of the threshold at which action becomes necessary is a matter of policy, informed by science.Footnote 27
This adjudicative conception encounters structural limits in the decentralized architecture of ISDS and at the interface of science, politics, and law. Unlike domestic systems with unified regulators and enforcement architectures, investment arbitration does not operate under a single codified substantive code but within a decentralized network of treaties the provisions of which are not uniform.Footnote 28 Arbitration decisions rely heavily on the interplay of political, technical, and legal discourses specific to each case. In particular, in disputes involving environmental and public health issues, scientific uncertainty is no longer just a challenge for fact-finding; it becomes the point of intersection between political discretion and legal adjudication.Footnote 29 In these cases, arbitral tribunals face a structural tension: on the one hand, scientific knowledge is fluid and open to interpretation; on the other hand, state decisions increasingly reflect a growing bias towards sovereignty. Meanwhile, the law itself must maintain normative legitimacy and procedural autonomy.
Given these dynamics, the intricate relationship between scientific uncertainty, political discretion, and legal reasoning makes it exceedingly difficult to formulate a single substantive rule for the PP that could apply uniformly across all cases. The core of the challenge is systemic: each dispute requires a delicate, case-specific balancing of economic, social, and ecological considerations, making it nearly impossible for a one-size-fits-all substantive obligation to provide a just and predictable outcome. This complexity suggests that attempting to resolve such disputes through predetermined substantive duties alone may be a Sisyphean task.
This, in turn, raises a crucial question: if substantive approaches consistently falter in the face of complexity, could a procedural lens offer a more productive pathway for operationalizing the PP in investment arbitration? To explore this hypothesis, it is essential firstly to ground the analysis in arbitral reality. The following sections therefore undertake a systematic review of the case law, examining how the principle is actually treated in practice to diagnose the precise nature of the adjudicative challenge and to assess whether a procedural turn is warranted.
3. The Precautionary Principle in ISDS: Three Recurrent Treatments in Arbitral Practice
Building on this conceptual backdrop, this section examines how tribunals have handled the PP in practice and whether those treatments yield reviewable standards under uncertainty. Depending on the statute or regulation at issue, adjudicators may lack guideposts for how to navigate scientific uncertainty.Footnote 30 Jaye Ellis, from a functionalist perspective, sharply identifies a legal paradox within the PP: the legal system must draw on scientific expertise to address environmental complexity; yet, it must not defer to science to such an extent that it loses its capacity to reach autonomous legal judgments – a risk of colonization by science and politics.Footnote 31 Arbitral practice, as discussed below, reflects this tension: the PP is variously pulled towards expert dominance, bent to policy convenience, or left only nominally acknowledged.
Inspired by this insight, this study draws on a comprehensive review of investment awards conducted via the Jus Mundi and Investment Arbitration Reporter databases to map these adjudicatory patterns. This systematic search identified a broad corpus of disputes in which the precautionary principle is expressly invoked (Table 1). Crucially, this wider set of cases – including recent awards such as Gabriel Resources v. Romania Footnote 32 and Eco Oro v. Colombia Footnote 33 – reinforces the tripartite typology proposed here: the additional awards consistently align with the recurring trajectories of (i) strong-form risk prevention, (ii) managerial risk balancing, or (iii) nominal acknowledgement.
Overview of International Investment Disputes Involving the PP
Table 1 Long description
The table has 17 case rows and 3 columns. The columns are labeled Case Name / Case No., Filing / Decision Date, and Main Facts and PP Relevance. The table lists various international investment disputes involving the precautionary principle. Each row provides the case name, case number, filing and decision dates, and a brief description of the main facts and the relevance of the precautionary principle. The cases span from 1999 to 2024 and involve various countries and legal contexts.
Source: Data compiled based on International Centre for Settlement of Investment Disputes (ICSID) and Permanent Court of Arbitration (PCA) official registries, with supplementary cross-referencing via Jus Mundi and Investment Arbitration Reporter.
To preserve analytical focus, the article selects from the above corpus a set of paradigmatic awards in which the reasoning most clearly illustrates the structural dynamics of the PP’s adjudicatory use. This methodological choice reflects a broader tendency in empirical legal scholarship to rely on ‘learning-by-doing’ approaches, whereby conceptual structures are identified through iterative engagement with evolving case practice.Footnote 34 The analysis focuses on whether these treatments generate reviewable standards capable of mediating the interaction among law, science, and politics in adjudication. Not all cases involve substantive engagement with the principle; indeed, several illustrate how it is procedurally acknowledged yet functionally neutralized. This variation is analytically valuable, as it exposes the structural conditions under which the principle becomes marginalized, ignored or transformed into a legitimating facade.
Risk prevention reflects a robust construction of the PP that authorizes protective measures when potential risks are identified under conditions of incomplete or contested scientific evidence. In arbitral settings, this construction is frequently applied in the form of substantive prohibitions without calibrated procedural safeguards, which can result in rigidity and inconsistency. In contrast, risk management embodies a more flexible construction that frames the PP as a balancing instrument, integrating scientific, political, economic, and social considerations. Yet this flexibility often broadens policy discretion and invites deference to administrative judgements, while leaving epistemic uncertainty weakly addressed.
Section 3.1 examines risk prevention and shows that it is frequently applied in ISDS in the form of suspensions, revocations, and outright regulatory prohibitions without calibrated procedural safeguards, resulting in rigidity and inconsistency. Section 3.2 shows that risk management, although more flexible, often broadens policy discretion and invites deference to administrative judgements, while leaving epistemic uncertainty weakly addressed. Section 3.3 identifies nominalization as a recurrent posture in which the PP is acknowledged in the record but not translated into reviewable standards; it does not structure the standard of review or affect the outcome. Among the awards examined in this section, these patterns are roughly evenly distributed, with each category accounting for approximately one-third of the cases. Taken together, these treatments do not supply reviewable criteria for adjudication and indicate a procedural gap within current practice.
3.1. Risk Prevention: The Rigid Edge of the Precautionary Principle
The interface between law and science is contested in adjudication.Footnote 35 Strong forms of precaution, in particular, tend to shift decision-making authority towards scientific expertise and away from legally mediated modes of justification, narrowing routines of participation especially where uncertainty is invoked to support restrictive regulatory action.Footnote 36 This dynamic becomes particularly evident in the application of the PP, which reflects a strong-form construction authorizing pre-emptive suspension or revocation of activities under conditions of scientific uncertainty. Its core orientation is preventive: it treats uncertainty itself as sufficient justification for prohibition, even before concrete harm materializes. While this approach is often celebrated in environmental discourse as a safeguard against irreversible harm,Footnote 37 its transplantation into the adversarial and institutionally fragmented setting of investment arbitration exhibits structural limits. In practice, strong-form precaution tends to be applied in the form of substantive prohibitions that translate uncertainty into indefinite freezes, assign priority to scientific assessment in the absence of calibrated procedural safeguards, and invite arbitral review without offering legal closure. In such contexts, scientific knowledge assumes a dominant role, while legal reasoning narrows to process validation, unable to translate epistemic openness into normative closure.
The Unglaube v. Costa Rica Footnote 38 arbitration exemplifies how strong-form precaution, when untethered from procedural guardrails, can harden into open-ended prohibitions. In August 2005, Costa Rica’s National Technical Secretariat for the Environment (SETENA) issued a sweeping order suspending all ongoing and future environmental assessments within Las Baulas National Park, invoking ‘precautionary’ measures to safeguard the nesting habitat of the endangered leatherback sea turtle.Footnote 39 While the applicable Forestry Law authorized preventive suspensions of environmental procedures, it limited them to one year.Footnote 40 Although the prolonged suspension may have been ecologically justified given the uncertainties at the time, its extended duration effectively transformed a temporary safeguard into a de facto moratorium on development rights, raising questions about the proportional calibration between precaution and legal predictability.Footnote 41
Domestic jurisprudence reinforced this preventive orientation. In March 2005, the Costa Rican Constitutional Chamber directed SETENA to issue guidelines to ensure that no project approval would endanger turtle nesting beaches,Footnote 42 and the Ministry of Environment and Energy (MINAE) subsequently instructed SETENA to halt the processing of permits in the area pending further ecological studies.Footnote 43 This ‘suspend first, study later’ posture exemplified the systemic constraint of strong precaution: rather than weighing proportionality or exploring adaptive alternatives, the administration transformed preliminary uncertainty into binding urgency standards. Unlike risk management approaches that permit conditional approvals or mitigation and leave room for incremental adjustment, the Costa Rican framework represented a preventive model that translated uncertainty into prohibition, thereby foreclosing flexibility.
The investors challenged these measures precisely on the ground that a nominal one-year precautionary suspension had been unlawfully prolonged into an indefinite moratorium, undermining their development rights and upsetting the balance between substantive protection and procedural fairness.Footnote 44 While the Tribunal acknowledged that Costa Rica’s actions were consistent with its domestic legal framework, it refrained from scrutinizing how the vaguely defined ‘ecological vulnerability assessments’ had been transmuted into binding urgency standards justifying the moratorium.Footnote 45 To be clear, this article does not dispute that when scientific evidence is determinate and enjoys wide consensus, legal reasoning should respect that certainty. The challenge this article addresses, however, arises specifically when scientific knowledge itself remains contested or ambiguous. In these specific contexts, this judicial reticence reflects a broader challenge: because environmental science is technical, uncertain, and politically malleable, tribunals risk defaulting into limited scrutiny.Footnote 46 In the absence of procedural standards to test scientific claims, such as proportionality thresholds or temporal limits, the law cedes its mediating authority – a concern echoed in scholarship on the growing reliance of international tribunals on expert-driven forms of scientific assessment.Footnote 47 The Unglaube case thus illustrates the ‘administrative layer’ of strong precaution: temporary regulatory tools are hardened into indefinite freezes when procedural triggers are absent.
If the Unglaube case demonstrates the administrative rigidity of strong precaution, the Infinito Gold arbitration deepens this structural limitation by constitutionalizing precaution into a technocratic framework, thereby embedding expert authority into the very foundations of legality.
In Infinito Gold Ltd v. Costa Rica,Footnote 48 the PP was elevated to a mandatory threshold condition. In its 2004 Murillo Amparo decision, the Constitutional Chamber interpreted Article 50 of the Costa Rican Constitution, which guarantees the right to a ‘healthy and ecologically balanced environment’, as incorporating the preventive/precautionary principle and requiring that an environmental impact assessment (EIA) be conducted before any exploitation concession is granted.Footnote 49 By elevating precaution to a constitutional command, Costa Rica transformed scientific uncertainty into a dispositive legal criterion, sufficient to revoke or deny concessions without recourse to proportionality or alternatives. Here the preventive orientation hardened further: environmental science became a decisive constitutional constraint, leaving no role for adaptive calibration.
This framework shifted authority from political balancing to technocratic expertise, illustrating a broader concern that the reception of scientific knowledge may overwhelm rather than be channelled into autonomous legal reasoning.Footnote 50 When SETENA approved the EIA in December 2005, it was treated not as one factor in a proportional calculus but as the decisive gatekeeper of legality.Footnote 51 The investor even sought to ‘cure’ its concession while leaving its validity contingent on the outcome of the EIA,Footnote 52 language the Tribunal read as the most accurate interpretation of the Chamber’s ruling.Footnote 53 In this way, uncertainty was translated into a categorical legal condition, foreclosing incremental adjustment.
At the international level, the Tribunal reinforced this dynamic by narrowing its review to procedural legality rather than substantive reassessment of science. It emphasized that an International Centre for Settlement of Investment Disputes (ICSID) tribunal does not act as an appellate court with regard to national decisions and that annulment of domestic judgments is possible only under the high threshold of denial of justice.Footnote 54 This posture effectively confined arbitral scrutiny to whether Costa Rica had followed its own legal framework, leaving untouched the deeper issue of how probabilistic ecological models displaced proportionality review.
The bilateral investment treaty (BIT) analysis further entrenched this deference. While the majority found a breach of fair and equitable treatment linked to Costa Rica’s 2011 legislative mining ban and its 2012 implementation, it immediately turned to the environmental exception in Annex I, section III(1) of the BIT, which permits environmentally motivated measures only in so far as they are ‘otherwise consistent with this Agreement’.Footnote 55 The focus thus shifted from whether the EIA-based determinations were proportionate to whether such determinations fit within treaty exceptions – again insulating expert-driven uncertainty from direct scrutiny.
Placed within the analytical framework of the systemic limitations of risk prevention, Infinito illustrates the ‘technocratic layer’ of strong precaution. Scientific models became constitutional law; tribunals deferred to expert determinations without proportionality review; and legal authority focused on validating process rather than closing uncertainty. In this sense, environmental law risks ‘too much science and not enough law’,Footnote 56 as strong-form precaution, when embedded without procedural triggers or stop rules, narrows the mediating capacity of law in ways that ultimately escape legal scrutiny.
Taken together, the Unglaube and Infinito cases expose the rigid edge of strong-form precaution in investment arbitration. In Unglaube, precaution settled into an indefinite administrative freeze; in Infinito, it was constitutionalized into a technocratic threshold. Despite these institutional differences, both cases reveal the same structural chain: scientific uncertainty dominates, law yields its mediating role, proportionality and alternatives vanish, and tribunals are left with little capacity for normative closure. The systemic limitation of risk prevention thus lies not only in its substantive severity but also in its institutional design: once scientific models are allowed to dictate prohibitions without procedural triggers or stop rules, legal authority contracts into a passive validator of expertise. This diagnosis sets the stage for the next section. If risk prevention falters as a result of being too rigid, collapsing uncertainty into indefinite bans and constitutionalized thresholds, then the risk management variant promises greater flexibility. Yet, as seen in Section 3.2, this flexibility is often underspecified: when precaution is absorbed into the discretionary logic of state policy, it risks becoming a legitimating device for political expansion rather than a structured tool of adjudication.
3.2. Risk Management: The Open-Ended Flexibility of the Precautionary Principle
When strong-form precaution prioritizes scientific assessment over legal balancing, the next question is whether a management-oriented framing improves adjudicative structure; the following cases suggest that flexibility can broaden discretion without clarifying review. The uncertain political foundations of the liberal project behind international law have made it susceptible to becoming a tool that can justify radically opposing outcomes, depending on the political motivations pursued by international lawyers in each case.Footnote 57 Bureaucratic administration, in this context, often means administrative control through expertise.Footnote 58 While science plays an important role in identifying risk, determining the appropriate response to this risk remains an intrinsically political decision, which legitimately can be taken only by the state, which is accountable to its own citizens. In many ways, the core foundation of environmental law is precautionary because it is ‘grounded in a political judgment that the environment demands and deserves proper protection to preserve life and enhance well-being’.Footnote 59
If the previous section exposed how strong-form precaution prioritizes scientific assessment over legal balancing, this section brings the inverse dynamic into view: framed as public-regarding, precaution was in practice redeployed to expand sovereign discretion; in the Bilcon v. Canada case,Footnote 60 it was invoked to safeguard community values, and in the Windstream v. Canada case,Footnote 61 to ensure science-based regulation. Rather than mitigating uncertainty, the PP was instrumentalized to manage it politically: it justified shifting criteria, open-ended moratoria, and the avoidance of substantive justification.
In Bilcon v. Canada, the Joint Review Panel (JRP) was mandated by the Canadian Environmental Assessment Act (CEAA) to determine whether ‘significant adverse environmental effects’ would remain after mitigation.Footnote 62 It did not apply that statutory test. Instead, it introduced a new criterion – namely, consistency with the ‘community core values’ of Digby Neck and its islands – and, on that basis, concluded that the quarry would cause ‘major and irreversible’ social transformations that warranted rejection.Footnote 63 The Tribunal held that the JRP had ‘in effect created a new standard’ unauthorized by law and undisclosed to investors, thereby failing to conduct the mitigation-then-significance analysis required by the CEAA.Footnote 64 What was presented as precaution became a vehicle for blocking investment through indeterminate and unannounced socio-political criteria, illustrating a broader pattern in which precautionary language absorbs policy preferences that resist transparent legal review.Footnote 65
The procedure reinforced this drift. Across nearly 90 hours of hearings, the claimants’ experts had only about 90 minutes in which to present technical evidence.Footnote 66 At one point, the Chair remarked that the process was ‘in a small way’ ‘a kind of a referendum’.Footnote 67 The JRP also broadened ‘traditional knowledge’ beyond its usual Indigenous scope to include all local residents.Footnote 68 Technical appraisal thus ceded ground to a referendum-like process. Ultimately, the Tribunal emphasized that the term ‘community core values’ was open to multiple readings, none consistent with the statute or the record.Footnote 69 What appeared as a safeguard for communities functioned as an unreviewable political veto, an outcome that sits uneasily with the reminder that precaution in environmental law rests on a political judgement to protect the environment rather than to displace statute-bound, evidence-responsive review. These procedural distortions also underpin the concern raised in Professor Donald McRae’s dissenting opinion, which warned that the Tribunal’s intervention risked substituting ex post judicial ‘second-guessing’ for domestic environmental review processes and thereby ‘chilling’ their future operation.Footnote 70
A different facet of this structural tension emerges in Windstream v. Canada, which illustrates a complementary face of precaution: governance by deferral. On 11 February 2011, the Government of Ontario announced that offshore wind would not proceed ‘until the necessary scientific research is completed and an adequately informed policy framework can be developed’, suspended intake under the Feed-in-Tariff (FIT) programme, and cancelled Crown-land applications lacking a FIT contract.Footnote 71 This was framed as precaution pending science, not as a prohibition grounded in proven hazard. Months earlier, the Ministry of the Environment had acknowledged that Renewable Energy Approval standards for offshore projects were incomplete; it floated a tentative five-kilometre shoreline exclusion, subject to revision, and launched consultations.Footnote 72 The process drew 1,403 submissions, most questioning the sufficiency of the five-kilometre line and calling for further study.Footnote 73 Rather than resolving the uncertainty said to justify the moratorium, these inputs gave the Province further latitude to postpone a definitive policy response.
The Ontario filings stressed that any framework must be ‘firmly supported by sound science and the precautionary principle’.Footnote 74 Yet, after invoking that rationale, follow-through remained thin: research plans stalled, no offshore-specific rules materialized, and the freeze persisted. In the meantime, Ontario centralized control by cancelling non-FIT Crown-land files and holding Windstream’s FIT contract in suspension, without articulating reviewable standards. Without institutional criteria for determining when scientific inquiry is sufficient to support a regulatory decision, appeals to sound science can become a basis for delay rather than a framework for decision.Footnote 75 The Tribunal did not specify what would constitute sufficient science.Footnote 76 It found a breach of Article 1105(1) of the North American Free Trade Agreement (NAFTA)Footnote 77 – which requires each party to accord investment treatment in accordance with international law, including fair and equitable treatment – on narrower grounds: Ontario left the investor in prolonged limbo despite having the tools either to complete the framework and reactivate the contract or to amend the regime and terminate it.Footnote 78 The deeper logic went unaddressed: uncertainty, once mediated through expert-led administration, became a basis for continued regulatory control without generating clear criteria for legal review.Footnote 79
Read together, the Bilcon and Windstream cases show how precaution, outwardly a commitment to public good, can in practice cloak discretionary state action. In Bilcon, the statutory, mitigation-sensitive test was displaced by a socio-political veto; in Windstream, ‘science-based’ prudence translated into an indefinite deferral. In neither case did adjudication construct procedural triggers to discipline this drift, such as advance notice of evaluative standards, transparent thresholds for the sufficiency of science, or proportionality backstops. Under a stewardship rationale, precaution becomes a technique for governing through uncertainty: it enables the state to defer, redefine, and insulate choices from justificatory review. That the PP is rooted in a political judgement to protect the environment helps to explain its rhetorical force, but without procedural constraints, that very judgement furnishes the vocabulary through which sovereign authority asserts itself amid the discipline of law.
3.3. Nominal Acknowledgement: Invocation as Compliance Assurance
Beyond the widely debated substantive framings of risk prevention and risk management, a third recurrent position appears in ISDS: nominal acknowledgement. In this pattern, the PP is invoked by the parties as compliance assurance – for example, as evidence of project diligence or of domestic preventive control – but it is not translated into reviewable standards that structure the reasoning or affect the outcome. This is not a third substantive method; it is a mode of non-application that illustrates how both prevention and management can prove ineffective in practice. In such instances, precaution remains a record-level citation rather than part of the articulated standard of review, which in turn motivates the design developed in Section 4.
In Tethyan Copper v. Pakistan,Footnote 80 the investors invoked the PP, the only published instance in our review in which the principle was advanced by an investor. After the provincial government of Balochistan refused a mining lease under the Balochistan Minerals Rules 2002,Footnote 81 effectively terminating exploration rights,Footnote 82 the investors emphasized that their environmental and social impact assessment incorporated a ‘worst-case scenario’ approach to water demand and sludge management.Footnote 83 They argued that this modelling reflected compliance with precaution by anticipating potential risks of wastewater toxicity and designing safeguards accordingly.Footnote 84 The PP was thus invoked not as a shield against development but as a legitimating device to demonstrate that scientific uncertainty had been addressed in advance. Yet, the Tribunal treated these submissions as procedural detail, noting them without interrogating the relevance of the worst-case modelling or requesting further data.Footnote 85 The arbitral inquiry narrowed to whether provincial authorities had acted consistently with domestic licensing procedures and contractual expectations, leaving precaution marginalized. Here, the investor’s attempt to transform science into law, casting precaution as risk prevention, was absorbed into the procedural record without shaping the Tribunal’s analysis,Footnote 86 illustrating how arbitral review can acknowledge ecological diligence, yet render it irrelevant.
The Gaspar v. Costa Rica Footnote 87 arbitration reflects the mirror image: here it was the host state that invoked domestic environmental regulations, including precautionary environmental controls grounded in Costa Rica’s Ley Orgánica del Ambiente (Organic Environmental Law),Footnote 88 to justify temporary closure measures against a poultry processing plant.Footnote 89 These measures were implemented under SETENA’s ‘preventive control’ authority, which mandates environmental viability assessments for activities that generate hazardous waste.Footnote 90 The Tribunal acknowledged that the SETENA regulatory framework granted administrative discretion to impose sanctions under Article 99 of the same law, including partial or total facility closures, as part of its preventive enforcement functions.Footnote 91 However, the Tribunal focused narrowly on procedural compliance with domestic law, avoiding substantive scrutiny of the proportionality of the measures.Footnote 92 For example, when the investor argued that the one-day compliance deadline was operationally impossible and that the closure disregarded the plant’s actual wastewater treatment capacity, the Tribunal noted that subsequent judicial suspension of the measures had remedied procedural defects.Footnote 93 This reasoning prioritized procedural validation over assessing whether the precautionary measures (i) were scientifically supported or (ii) constituted the least restrictive means available.Footnote 94 The Tribunal concluded that compliance with domestic procedural requirements was sufficient for international legality,Footnote 95 effectively reducing the PP to a formal procedural step. Even if the risk assessment had gaps or inadequacies, mere adherence to procedural norms was deemed sufficient. The PP thus functioned merely as a legal citation to justify SETENA’s enforcement authority, not as a principle shaping proportionality or evidentiary reasoning.
These cases do not turn on who may wield precaution; they show a recurrent posture. Whether invoked by investors in Tethyan Copper v. Pakistan or by the state in Gaspar v. Costa Rica, the PP is acknowledged but not operationalized. In Tethyan, it appears as project-level diligence; in Gaspar v. Costa Rica, it appears as a statutory preventive requirement. The Tribunals in both cases place precaution in the procedural background rather than treating it as a rule of review, so it is sufficient to enter the record yet insufficient to shape outcomes. This approach is distinct from those discussed in Section 3.1 and Section 3.2. It reflects neither the dominance of science nor politics; rather, it amounts to nominalization, a procedural absorption that supplies no threshold for evidentiary sufficiency, no proportionality constraint, and no allocation of the burden of proof under conditions of uncertainty. The result is a principle that persists in name while offering legitimacy without legal discipline. The gap is procedural: without triggers to translate citation into adjudicatory standards that specify how much science suffices, when a measure is proportionate, and who bears the burden, the PP cannot perform legal work.
Across prevention, management, and nominalization, the PP is acknowledged but not translated into reviewable criteria, confirming a procedural vacuum. Section 4 specifies the procedural architecture – proportionality review and reallocation of the burden of proof – through which the PP can perform adjudicatory work.
4. Precaution as a Procedural Trigger: From Cognitive Openness to Normative Closure
Section 3 traced three distinct trajectories through which the PP falters in investment arbitration. These instances reveal not isolated missteps but a consistent trend: tribunals situated at the intersection of science, politics, and law consistently translate precaution into procedural form, while leaving absent the criteria that might discipline such translation.
Crucially, this persistent proceduralization is not a contingent weakness or a downgrading of the principle; it reflects, instead, the legal system’s self-protective logic under conditions of cognitive uncertainty. Once it is recognized that arbitral tribunals’ reduction of PP to procedural considerations is not an aberration but an institutionally embedded practice, it is possible to begin to locate the normative foundation for designing procedural trigger mechanisms capable of structuring this inevitability.
This section therefore reconceives the role of the PP at the international level, specifically as a procedural tool available for investor–state tribunals in adjudicating disputes under conditions of scientific uncertainty. Rather than treating it as a substantive obligation to prevent or manage risk, it should function as a justiciable procedural trigger that determines how reasons must be given under conditions of uncertainty and where adjudicatory contestation must provisionally stop. The point is not to eliminate uncertainty but to transform cognitive openness into legitimate closure – that is, into a form of closure that rests on explicit and reviewable decision rules, as illustrated by the proportionality review and burden-of-proof mechanisms discussed in Sections 4.2 and 4.3 below.
To that end, this section frames the solution as a procedural governance spectrum: a family of institutional techniques capable of channelling scientific uncertainty into structured and reviewable legal reasoning. These may include calibrated proportionality review, burden-shifting mechanisms, ongoing data-reporting duties, periodic reassessment requirements, interim measures, and declaratory remedies. While the spectrum is wide, the section focuses on two mechanisms that most clearly illustrate how tribunals operationalize stop rules in practice – namely, proportionality review and the reallocation of the burden of proof – as these are the doctrinal sites where the jurisprudence reveals the clearest structural pressures and where procedural closure is produced with the greatest transparency. Proportionality review dynamically calibrates scientific evidence to balance environmental protection and investment protection (as seen in Methanex v. United States),Footnote 96 and the reallocation of the burden of proof institutionalizes reflexive governance by channelling uncertainty into procedural duties (as seen in David Aven v. Costa Rica).Footnote 97 Together they supply the stop rules absent in the jurisprudence examined in Section 3.
4.1. From Functionalist Aspirations to Reflexive Closure
Section 3 showed that whether framed as risk prevention or risk management, the PP has not supplied tribunals with reviewable standards under conditions of scientific uncertainty. The result is repeated acknowledgement without operative thresholds for evidence, proportionality or burden allocation. To move beyond this impasse, this article shifts the theoretical lens: the PP is not a substantive adjudicatory norm but a procedural trigger, a device that activates proportionality review and burden-of-proof reallocation as structured stop rules when epistemic openness threatens to overwhelm legal reasoning.
This reframing is not without antecedents. The broader literature on law and uncertainty has long debated how legal systems can remain cognitively open to external knowledge while maintaining normative coherence. Several models have emerged, each offering distinct insights into law’s relationship with science and complexity. One prominent strand, represented by Steven Lierman, conceptualizes law as a complex adaptive system: a dynamic institutional arrangement that gradually converges with the epistemic practices of science and the normative aims of governance.Footnote 98 This convergence perspective suggests that legal systems can evolve alongside other subsystems to produce coordinated risk regulation. Another perspective, developed by Suryapratim Roy, draws on Latourian insights to present law as a reflexive mediator, emphasizing testimonial practices that filter expert knowledge through iterative discursive processes.Footnote 99 Both approaches offer important tools for understanding how legal systems can respond to uncertainty – not by retreating from it, but by evolving alongside or actively mediating external epistemic claims.
However, these adaptive strategies face distinctive constraints in international investment arbitration. Unlike domestic environmental regimes, where precautionary measures operate within robust administrative frameworks capable of iterative learning, ISDS tribunals function within a formalized and adversarial architecture designed for one-off dispute resolution rather than recursive adjustment. Limited institutional resources, procedural rigidity, and pronounced information asymmetries prevent tribunals from engaging in the iterative monitoring, feedback, and revision that adaptive governance presupposes.Footnote 100 These structural tensions give rise to what may be termed an ‘adaptive crisis’: tribunals are asked to decide cases under conditions that call for institutional learning; yet, the procedural design of ISDS affords no mechanism for cumulative updating, intertemporal coordination or reflexive correction. As a result, adjudication becomes vulnerable to deformalization (erosion of procedural discipline through ad hoc pragmatism) and materialization (substantive domination of legal judgment by scientific or political claims).
It might be possible to rethink the interactions among science, policy, and law so as to avoid the dangers of deformalization and materialization.Footnote 101 Recognizing this constraint, the article proposes a procedural response, grounded in Ladeur’s concept of stop rules, that enables tribunals to manage scientific and political complexity through disciplined, reviewable reasoning rather than substantive improvisation. Such an approach preserves cognitive openness to external knowledge while safeguarding the normative autonomy of law, turning the adaptive crisis into an opportunity for procedural renewal.
One point of departure for developing a workable framework is Hey’s observation that the PP emerged not as a static rule but as a dynamic process of legal evolution responding to the epistemic challenges of ecological uncertainty.Footnote 102 This insight underscores the core premise of this section: the PP is best understood not as a substantive yardstick but as a structural response by legal systems to endemic uncertainty.
This recognition invites a systemic approach: one that not only acknowledges the limits of institutional adaptation but also conceptualizes legal decision-making under epistemic complexity as an autonomous normative process. Building on Niklas Luhmann’s theory of law as a functionally differentiated subsystem,Footnote 103 further developed by Ladeur,Footnote 104 this part advances the view that law must remain cognitively open to scientific and political inputs while preserving its normative closure in justification. In this frame, the role of the PP is reconceived in procedural terms to mediate complexity. As noted in Section 1, this procedural reading of the PP finds a theoretical analogue in Ladeur’s work on stop rules (Stopp-Regeln), which function not as formal procedures or normative tests but as cognitive boundaries that shield legal judgment from the paralysis of endless scientific debate and to enable normative closure under conditions of epistemic contingency.Footnote 105
Stop rules operate by fixing certain distinctions and prestructuring the horizon of admissible options, thereby combining variation with selective closure. They are not aimed at eliminating uncertainty but at making further exploration temporarily suspendable, especially when a particular pattern of reasoning or knowledge appears to be provisionally effective. In doing so, they prevent infinite regress and shield decision-making from paralysis caused by the endless search for alternative possibilities.Footnote 106 They do not establish truth; rather, they suspend doubt. The continued possibility of legal decision-making across all social domains depends on a shared – if implicit – commitment to such stop rules: without stable frameworks for limiting recursive contestation, judgement becomes impossible under conditions of limited time and incomplete knowledge.Footnote 107
In this light, stop rules should be understood as operating at the level of adjudicatory method rather than substantive law. They guide how tribunals structure reasoning when facing epistemic complexity, by defining thresholds for the sufficiency of evidence, delineating when additional scientific inquiry becomes excessive, or specifying procedural burdens without displacing governmental prerogatives. Positioned in this way, stop rules help tribunals in maintaining coherence and accountability within their limited jurisdiction, while respecting states’ sovereign authority to determine broader policy directions. Respect for state choices therefore forms the default position. Procedural intervention becomes warranted only where concrete structural shortcomings compromise the possibility of reviewable adjudication. Such shortcomings include, for instance, open-ended administrative inaction, the absence of genuine adversarial testing, or credible indications of selective use or withholding of scientific information. In such circumstances, procedural triggers serve to safeguard the integrity of legal reasoning while maintaining appropriate deference to sovereign regulatory autonomy.
In clarifying the architecture of these procedural mechanisms, two additional points warrant brief explanation. Firstly, the normative grounding of stop rules lies not in any judicial ambition to reshape substantive international law, but in the modest domain of adjudicatory method. Stop rules are not judicial inventions of new obligations, nor domestic administrative standards imported into ISDS. Their legitimacy is derivative: they arise from the tribunal’s duty to decide disputes under conditions of epistemic complexity, from treaty-interpretation principles such as good faith and reason-giving, and from the procedural fairness requirements embedded in arbitral practice. Positioned in this way, stop rules function as rational constraints on legal reasoning rather than as new normative commitments imposed on states. They do not displace domestic authority over environmental policy (which remains with governments) nor the allocation of substantive obligations under treaty law (which remains with states collectively). Rather, they guide tribunals only in the narrow domain where adjudicators must interpret indeterminate treaty provisions and produce reviewable decisions. This cumulative, case-based evolution, captured in Ladeur’s account of law’s iterative self-adjustment, explains how procedural boundaries can stabilize without purporting to redefine the content of substantive international law.Footnote 108
A similar clarification concerns the temporality of these mechanisms. The ‘closure’ produced by stop rules is epistemic and procedural, not political or permanent. ISDS awards remain final and enforceable, but the reasoning they employ must, at some point, suspend further scientific contestation in order to permit adjudication. This does not foreclose future legal or policy development: subsequent cases may recalibrate evidentiary thresholds; treaty reforms may clarify standards; and states retain full prerogative to revise environmental measures. ‘Legitimate closure’, in this sense, refers to the structural requirement that legal justification must reach a point of decision under uncertainty – while remaining open to later systemic evolution. Far from entrenching arbitrary judicial discretion, stop rules discipline it by ensuring that the suspension of doubt occurs through transparent and legally reviewable procedures.
Understanding the PP through the lens of systems theory reframes its role; on this view, the principle operates as a trigger that activates internal stop rules: a device that activates internal stop-rule mechanisms when legal systems face cognitive overload.Footnote 109 Under conditions of epistemic uncertainty, law must maintain its normative closure despite the openness and instability of scientific knowledge. This approach initiates legal processes, such as proportionality review or the reallocation of the burden of proof, that enable closure without demanding certainty. This reframing accords with public-law accounts of the standard of review in natural-resources disputes: tribunals should calibrate both the nature and the intensity of review; treaty texts seldom specify this standard; and any calibration engages questions of deference to domestic decision-making.Footnote 110
The following sections turn to practice. Two procedural architectures illustrate how tribunals can operationalize the structural role of the PP. Proportionality review, as developed in Methanex v. US, shows how tribunals can dynamically calibrate scientific evidence, acknowledge uncertainty while ensuring a rational connection with credible risks. Reallocation of the burden of proof, exemplified in David Aven v. Costa Rica, demonstrates how tribunals can reconfigure evidentiary duties to prevent uncertainty from being endlessly deferred or strategically manipulated. Together, these mechanisms supply the stop rules that Section 3 identified as missing and enable the PP to function as an operative procedural device capable of structuring review, rather than remaining a nominal acknowledgement.
4.2. Proportionality Review: Dynamic Calibration of Scientific Evidence
The substantive application of the PP, conceived here not as a substantive mandate but as a procedural trigger, requires legal systems to establish flexible safeguards that balance scientific uncertainty with regulatory legitimacy. At its core, proportionality review functions as the key mechanism, systematically evaluating scientific evidence while ensuring that environmental policies align with actual risks. This delicate balance prevents two extremes: the law becoming blindly dependent on scientific consensus, or political agendas using precaution as a tool for legitimacy.
This article does not, however, endorse a fully fledged proportionality stricto sensu review that would delegate substantive policy decisions to arbitral tribunals. Rather, it advocates a more calibrated, deferential form of proportionality analysis, focused on the rational connection between regulatory means and legitimate environmental objectives. As Caroline Foster notes, proportionality in ISDS should remain a ‘light-touch’ standard, emphasizing reasonableness and regulatory coherence rather than importing the intrusive balancing characteristic of European public law.Footnote 111
The Methanex v. US case exemplifies this approach. The Tribunal applied the rational-connection test to reconstruct the relationship between law and science. It recognized the host state’s right to set the level of protection while preventing the PP from being used as a mere tool for political legitimacy by dynamically selecting the appropriate level of evidence. This case was significant not only for its application of the PP but also because it was among the first of the NAFTA cases to accept and structure amicus curiae participation. The case arose from California’s complete ban of methyl tertiary-butyl ether (MTBE) as a gasoline additive, a move that caused significant economic loss to Methanex, the world’s largest methanol producer. The amicus curiae brief emphasized that the PP allows countries to take preventive regulatory action, even with incomplete scientific evidence, to protect human health or the environment.Footnote 112 While amicus curiae briefs are rarely used to introduce scientific evidence, they have become a crucial tool for non-governmental organizations and stakeholders, especially in climate change cases, to present scientific data in legal disputes.Footnote 113
The amicus curiae submissions emphasized that states have the right to adopt preventive measures based on ‘conflicting scientific evidence’ and that the ‘political nature’ of setting the level of protection should be respected under international law.Footnote 114 Methanex argued that California’s ban was not scientifically grounded but rather a form of trade protectionism in relation to the domestic ethanol industry.Footnote 115 The Tribunal applied the rational-connection test within a proportionality-type framework and did not require California to prove a direct causal link between MTBE pollution and harm. Instead, it focused on the ‘rational basis’ for the decision, evaluating the connection between the precautionary measure and ‘credible scientific concerns’.Footnote 116 Despite conflicting conclusions from the World Health Organization and Methanex, the Tribunal found that a report from the University of California, detailing the migration model of MTBE and its groundwater contamination data, provided a ‘rational basis’ for risk inference.Footnote 117
In response to the investors’ claim that the ban was driven by environmental populism, the Tribunal scrutinized the scientific arguments presented in the administrative record, stripping away political narratives. It concluded that California’s decision-making process was focused primarily on protecting water resources, not on harming foreign methanol producers.Footnote 118 This approach effectively reframed the scientific controversy into a legal question centred on the rational basis for risk recognition, avoiding the influence of political discourse.
This approach aligns with Ladeur’s concept of stop rules, which argues that the law must establish cognitive boundaries in decision-making to prevent the indefinite delay of scientific disputes.Footnote 119 In the Methanex case, the Tribunal achieved a form of ‘stopping’ by dynamically weighing the evidence. It treated the University of California’s empirical data as core evidence, while secondary evidence, such as animal studies, provided supplementary evidence.Footnote 120 This method created a structured evidentiary spectrum for inferring risk under uncertainty. Moreover, the Tribunal noted that California was required, under Executive Order D-5-99, to conduct ongoing analyses and to evaluate potential substitutes for MTBE. This requirement acknowledged the provisional nature of the ban while ensuring that regulatory discretion remained subject to periodic re-evaluation, rather than drifting into open-ended deferral.Footnote 121
This review mechanism prevents the legal system from being hijacked by scientific disputes and guards against the misuse of political power under the guise of precaution. Methanex demonstrates how proportionality review insulates the legal system from being overwhelmed by unresolved scientific controversy. When scientific uncertainty threatens to dominate legal judgments, proportionality review filters the evidence, transforming scientific conclusions into normative propositions that the legal system can manage. California’s ban was justified not by scientific consensus but by the logical coherence of the risk inferences. Even with scientific disagreements, the Tribunal found that California had shown a sufficient connection between the measure and ‘credible scientific concerns’, thus meeting requirements of international law.
Furthermore, while respecting California’s right to choose a high level of protection, the Tribunal implicitly required the measure to be proportionate to the severity of the risk posed by MTBE contamination. The Tribunal’s evaluation of the University of California’s evidence on groundwater contamination established a rational basis for California’s total ban, ensuring that the scope of the measure was justified by the significant environmental threat identified.Footnote 122 Although the University of California Report recommended a phased transition rather than an immediate ban, the Tribunal ultimately upheld California’s complete ban on MTBE and found no evidence of improper motive.Footnote 123 Through procedural review, the Tribunal ensured that the decision-making process was transparent and free from capture by specific interest groups, such as Archer Daniels Midland Company (ADM), thereby safeguarding the legitimacy of the regulatory action.Footnote 124 This approach reinforced proportionality review as a cognitive firewall, aligning with this article’s emphasis on dynamically calibrating scientific evidence to support normative legal outcomes.
While the Methanex case successfully navigated the balance between scientific evidence and political influence, the Bilcon case illustrates how the legal system can be colonized by political discourse. In Bilcon, it was the JRP that failed to justify how its novel ‘community core values’ test translated environmental evidence into a legally cognizable standard. The JRP neither applied the mitigation-then-significance analysis required under the Canadian Environmental Assessment ActFootnote 125 nor provided adequate notice that it intended to employ this new evaluative framework.Footnote 126 In contrast, Methanex demonstrated how proportionality review, as a procedural instantiation of reflexive governance, both preserved the host state’s ecological sovereignty and defended the legal system’s autonomy in managing risk allocation. The Tribunal refrained from making direct judgments on scientific disputes, such as the carcinogenicity of MTBE, and instead focused on the rational foundation of California’s decision-making process. This aligns with the core principle of reflexive governance: law should facilitate a dynamic dialogue between systems through procedural rules, rather than replacing scientific or political expertise. By incorporating temporary measures and regular reviews, the legal system acknowledges the evolving nature of scientific knowledge while preventing the PP from becoming a tool for indefinitely suspending policy.
This dynamic balance underscores a key proposition in functionalist jurisprudence: the adaptability of the legal system does not lie in abandoning formal rationality but in using cognitive techniques to transform complexity into resources for constructing normative order.
4.3. Reflexive Governance: Reconfiguring the Burden of Proof
Before examining how tribunals have operationalized evidentiary innovation, it is useful to situate the discussion within Foster’s analytical framework. Foster identifies several procedural techniques by which international courts and tribunals manage scientific uncertainty – these include limited presumptions, calibrated adjustments to evidentiary standards, and iterative reassessment mechanisms – which preserve adjudicative review without demanding conclusive scientific demonstration.Footnote 127 These techniques do not displace the default rule that each party bears the burden of proof but they enable tribunals to respond to epistemic complexity in ways that are disciplined, time-bound and contestable. Introducing this framework clarifies how the developments discussed below fit within a recognized repertoire of methods for managing evidentiary complexity.
The substantive application of the PP goes beyond merely adapting law to scientific uncertainty; it entails a structural reconfiguration of legal epistemology. By recasting rigid substantive mandates as procedural mechanisms embedded in goal-oriented ambiguity, law shifts from a deterministic rule-enforcing role towards an adaptive medium for risk communication and institutional learning.Footnote 128 This transformation is most visible in procedural devices that reconfigure how evidentiary burdens operate: a change with significant implications for international economic law.
In international arbitration, the default evidentiary position is that each party proves the facts on which it relies.Footnote 129 This baseline, however, admits important doctrinal qualifications. Most notably, the necessity defence, understood as a circumstance precluding wrongfulness, carries an exceptional evidentiary burden: a state invoking necessity must establish the existence of a grave and imminent peril and that the measures taken were the only reasonable or feasible means to address that peril.Footnote 130 Awards such as Total S.A. v. Argentina apply this high threshold in the narrow context of necessity, but these holdings should not be taken to establish a universal rule that states must in every case prove the reasonableness or necessity of their measures.Footnote 131
Accordingly, necessity jurisprudence should not be taken to establish a general rule that, in every investment dispute, the host state bears the burden of proving the reasonableness and legitimacy of its decision-making process. The Von Pezold case, for example, articulates stringent evidentiary requirements in the specific frame of necessity and related defences; it is better read as context-specific jurisprudence than as authority for an across-the-board evidentiary principle.Footnote 132 At the same time, doctrinal and policy literature recognizes that, under conditions of epistemic complexity, tribunals may legitimately deploy narrow, procedurally bounded burden-reallocation devices – prima facie inferences, lowered standards on discrete questions, targeted presumptions, and reassessment duties – so long as these measures are accompanied by due-process safeguards (early notice, opportunity to rebut, and time limits). Foster’s taxonomy provides a practicable template for these context-sensitive interventions.Footnote 133
David Aven v. Costa Rica illustrates how such devices can operate in practice. Costa Rica’s domestic regime (notably Article 109 of the Biodiversity LawFootnote 134 ) imposed disclosure and procedural duties on developers; the Tribunal treated persistent procedural omissions and failures of disclosure as legally significant triggers that carried epistemic consequences – that is, as factors that could properly affect the allocation of evidentiary risk rather than as invitations to pronounce finally on contested ecological facts.Footnote 135 In particular, while the Tribunal relied on procedural omissions (for example, failure to produce the Protti report, a hydrogeological study that had identified potential wetlands on the project site but was omitted from the environmental viability application) as a basis for finding breaches of domestic law and for justifying supervisory or suspensive measures, it did not in all respects translate those findings into a successful monetary award against the state; the Tribunal dismissed certain counterclaims for damages on quantification grounds.Footnote 136 Thus, the ‘liability’ to which the Tribunal referred is best read as (i) a finding that investors had failed to comply with domestic legal duties – an outcome that removed treaty protection or justified regulatory restraint – and (ii) a procedural basis for imposing supervisory or compliance obligations, rather than a straightforward condemnation that automatically produced treaty-level monetary liability.Footnote 137
Through this remedial and procedural approach, the Tribunal effectively compelled developers to internalize the costs of precautionary inquiry and to engage in iterative compliance and reassessment.Footnote 138 Viewed through that lens, David Aven substitutes dynamic procedural duties for static evidentiary thresholds: plausible technical indicators (aerial imagery and soil analyses) together with process-focused obligations justified precautionary suspension and subsequent supervisory measures.Footnote 139 This pattern closely corresponds to Foster’s account of prima facie inferences and reassessment duties as techniques for managing scientific ambiguity.Footnote 140 Rather than enacting a substantive, across-the-board reversal of the burden of proof, the Tribunal in David Aven treated procedural omissions as legally meaningful triggers that allocated the epistemic consequences of uncertainty to the party best placed to produce information. In this way, the Tribunal operationalized a recognized evidentiary technique: converting unresolved scientific questions into structured procedural obligations while avoiding requiring the Tribunal to decide contentious ecological facts on a definitive substantive basis.
The transformation maps onto Ladeur’s theory of stop rules: legal systems must develop cognitive boundaries that permit decisional finality under conditions of scientific contestation.Footnote 141 Traditional stop rules, such as burdens of proof or evidentiary thresholds, frequently founder when confronted with probabilistic or disputed scientific claims; David Aven demonstrates how tribunals can answer that challenge by turning epistemic gaps into procedural duties and presumptions, thereby institutionalizing preventive self-discipline.Footnote 142
The evidentiary implications are significant. The tort-derived maxim actori incumbit probatio presumes a degree of causal or probabilistic certainty that is often unattainable in complex environmental disputes. Where achieving such certainty would produce paralyzing indeterminacy, limited procedural presumptions and reassessment mechanisms can preserve the rule of law by rendering disputes reviewable and maintaining accountability. Such measures, however, must be strictly circumscribed so as not to usurp the primary policy role of states or to effect covert, generalized rule-making by tribunals. Foster’s work offers operational guidance on how to design these safeguards.Footnote 143
In sum, reflexive governance reframes the evidentiary question: it does not seek to abolish the presumption of legality or to establish blanket reversals of burdens of proof, but to permit context-sensitive, procedurally regulated reallocations that convert unresolved scientific questions into duties of verification, reporting, and reassessment. This architecture allows tribunals to manage uncertainty while preserving normative coherence, democratic space, and institutional legitimacy.
5. Conclusion
This article has shown that the PP falters in investment arbitration not because tribunals misunderstand it, but because treating the principle as a substantive mandate sits uneasily with adjudicating scientific uncertainty in a bilateral, adversarial forum. Reframing the PP as a procedural trigger offers a more workable and institutionally realistic path. When understood as activating structured techniques, such as calibrated proportionality review (focused on rational connection) and targeted burden-shifting, the principle helps tribunals to translate scientific uncertainty into transparent and reviewable legal reasoning. These mechanisms, grounded in Ladeur’s concept of stop rules, do not displace sovereign discretion. Rather, they provide a disciplined way for tribunals to navigate uncertainty while preserving legal authority, without lapsing into arbitrary choice or unexamined deference. In this sense, the PP functions less as a substantive command than as a framework for structuring adjudication under uncertainty.
Admittedly, the procedural tools proposed here operate as a pragmatic interim framework rather than a substitute for state-driven treaty reform, particularly since they cannot eliminate the tension between epistemic uncertainty and the finality of monetary awards. Future research, therefore, may examine how remedial design – such as reversible or conditional forms of relief – could help to align the finality of adjudication with the inherently provisional nature of precautionary science.
Acknowledgements
The authors are grateful to the anonymous TEL reviewers for their thoughtful comments and suggestions. All errors remain our own.
Funding statement
The authors acknowledge financial support from the Fundamental Research Funds for the Central Universities, the National Social Science Foundation of China (Grant No. 20BFX200), the Zhejiang Provincial Soft Science General Project (Grant No. 2025C35031), and the Zhejiang Provincial Philosophy and Social Science Planning Project (Grant No. 25BMHZ086YB, in collaboration with the Zhejiang Law Society).
Competing interests
The authors declare none.
Open access
