II. GLOBAL EXPERIMENTALIST GOVERNANCE: RATIONALE, ELEMENTS AND CONDITIONS

Since the 1990s, multiple approaches have shaped the language of ‘global governance’.Footnote ¹² The notion of governance, as opposed to the one of government, prominently responds to the limits of States' individual action, and to the growing influence of a wider range of non-State actors in key areas of international relations and law.Footnote ¹³ From this perspective, the unitary idea of a sovereign, hierarchical, regulatory State needs rethinking, as States are now called to deal with issues that do not have obvious solutions and involve multilevel and transnational interactions and cooperation.Footnote ¹⁴ ‘Global governance’ then suggests a system of norms, processes and structures established, and implemented, by a constellation of State and non-State actors (eg private actors, business, NGOs, international organizations) formally and informally operating within the international order.Footnote ¹⁵ This notion points not only to the multilevel architecture of the system, but also to its pluralistic inspiration, calling for greater accountability, transparency and legitimacy. Inevitably, trends and modes of global governance have mutated over time to reflect increasing fragmentation, regime complexity, network orchestration and transnational dynamics.Footnote ¹⁶

Responding to these transformations, experimentalist governance has emerged as a ‘template for governance reform in new areas’.Footnote ¹⁷ This entails ‘an institutionalized process of participatory and multilevel collective problem-solving, in which the problems (and the means to address them) are framed in an open-ended way and subjected to periodic revision by various forms of peer review, in the light of locally generated knowledge’.Footnote ¹⁸ While this process retains a normative dimension, it does not operate within formal command-and-control mechanisms, as it favours ‘regulatory approaches which are less rigid, less prescriptive, less committed to uniform outcomes, and less hierarchical in nature’.Footnote ¹⁹ In exploring its contours, de Búrca, Keohane and Sabel refer to ‘a set of practices involving open participation by a variety of entities (public or private), lack of formal hierarchy within governance arrangements, and extensive deliberation throughout the process of decision making and implementation’.Footnote ²⁰ Experimentalist governance practices can be recognized in various settings (eg private or public initiatives), operate at different levels of decision-making (eg local, national, regional or international) and emerge within and between regulatory regimes.

As a new mode of governance, experimentalist governance prominently responds to two specific challenges: ‘strategic uncertainty’ and ‘polyarchic distribution of power’.Footnote ²¹ ‘Strategic uncertainty’ is said to arise where ‘the parties face urgent problems, but know that their preferred problem-solving strategies fail and therefore are willing to engage in a joint, deliberative (potentially preference-changing) investigation of possible solutions’.Footnote ²² Strictly linked to this paradigm, ‘polyarchic distribution of power’ is the result of situations where horizontal, non-hierarchical, decision-making dynamics between State and non-State actors lead to a conundrum where no actor has the capacity to impose its own preferred solution, without taking into account other views.Footnote ²³

Of course, there is no single correct model of experimentalist governance, but rather an ‘ideal type’ against which evaluating governance practices as ‘a basic default account of the world’.Footnote ²⁴ This ideal type represents a multilevel architecture, based on four ‘deliberation-fostering elements’.Footnote ²⁵ First, the elaboration of framework goals and metrics for evaluating their achievement represents the foundation of the system. The open-ended nature of these goals means that they can have a provisional character, as they adjust to changing circumstances and are shaped by diverse experiences and local contexts, including new technological and scientific knowledge. Second, the implementation of these open-ended goals is left to ‘lower-level’ units—such as national authorities, local communities, or scientists—in coordination with the ‘centre’ and in consultation with civil society. This is where ‘lower level’ actors are accorded considerable discretion in adapting and experimenting problem-solving practices, based on contingent knowledge and decentralized authority. Third, regular reporting obligations are imposed on local units to provide feedback on their performance, through monitoring, peer review and benchmarking. As a result of peer review of implementation experiences, the benchmarking of best practices is particularly important as it provides the basis for mutual learning and accountability.Footnote ²⁶ Finally, as a result of this feedback loop, framework goals and decision-making practices are periodically re-evaluated, and, if necessary, revised.Footnote ²⁷

There is no universal model for experimentalist governance, but for a global experimentalist governance framework to be effective, it is generally thought that four preconditions must also be fulfilled. First, governments must be unable to agree on a comprehensive set of rules and efficiently monitor their compliance. This is a recurrent factor when central actors cannot easily foresee the local effects of rules, and when effective rules are subject to unforeseeable change. This condition is a challenge in climate change governance, where uncertainty and risks, including with respect to distributive impact, combines with a decentralized decision-making.Footnote ²⁸ This is even more striking with respect to the governance of emerging technologies, where the predictability of regulation and its effects are frequently challenged by the need for constant update and revision, in the light of the fast pace of technological development.Footnote ²⁹ This can lead to the problem of ‘regulatory disconnection’ between law and emerging technologies, which occurs where there is a ‘mismatch between the regulation-in-the-books and the latest technology in action’, as the technology emerges, circulates and evolves before regulators are able to reach an agreed position, finalize the terms of the regulation or adapt existing regulations.Footnote ³⁰ Technology then repeatedly leaves the law behind, and often operates in contingent regulatory voids.

Second, although they are unable to agree on specific rules, governments must not disagree on basic principles, and allow discussion and deliberation on the goals and benchmarks.Footnote ³¹ Rather than a call for universal agreement, this condition requires a thin consensus. With respect to global climate governance, the fulfilment of this condition is less obvious. As polarized positions among actors persist, they are possibly combined with a thin, high-level convergence towards the scale of the problem and the urgency of a new international climate agreement.Footnote ³² The Intergovernmental Panel on Climate Change (IPCC) has played a primary, and yet much criticized, function in thickening such consensus based on claims of objective, expert knowledge.Footnote ³³

Third, cooperation among decision-makers is pivotal, as experimentalist governance will not effectively thrive where (formal and informal) veto powers or obstructionism are exercised to block consensus and push forward hidden agendas and strategic interests. To drive such cooperation, this model relies upon destabilization mechanisms, such as ‘penalty defaults’, established by the central authority as a disincentive for refusal to joining the proposed governance system. These are ‘mechanisms for unblocking impasses in framework rule-making and revision by rendering the current situation untenable while suggesting—or causing the parties to suggest—plausible and superior alternatives’.Footnote ³⁴ They are structured by imposing unfavourable default rules to push parties to cooperate by contributing to the information-sharing regime in order to avoid such rules.Footnote ³⁵ At an international level, there are examples of these mechanisms being established, especially in international trade, to induce environmentally-oriented collaboration.Footnote ³⁶ In global climate governance, where multi-actor cooperation is highly problematic, mostly due to political and economic interests, the threat of penalty defaults is potentially very powerful. The EU has, controversially, started to act in this direction by using (the threat of) unilateral action to catalyse adequate international or third countries climate regulation, through the contingent imposition of the rules governing the EU Emission Trading Scheme to emissions of greenhouse gases generated abroad.Footnote ³⁷

Finally, non-State actors (eg industry, the scientific community, consumers and NGOs) play an increasingly influential role in global governance.Footnote ³⁸ As a result, to fill the information gaps in areas of uncertainty and provide a form of legitimacy through the involvement of different interests, a key role is attributed to participation of civil society. Compared to other global environmental issues, climate change has certainly engaged a wider range of actors, including NGOs and epistemic communities.Footnote ³⁹ But cooperation with, and participation of, civil society and NGOs at an international level is not straightforward, due to—inter alia—their legal status under international law, the debate on the democratic legitimacy of their participation, and the challenge of identifying the relevant public(s).Footnote ⁴⁰

Although experimentalist governance is far from a panacea, it has clear advantages. First, the focus on iterative processes based upon open-ended goals, benchmarking of best practices, reporting and periodical review allows decision-making to be based on much richer knowledge and information about alternatives than is traditionally available. This emphasizes cooperation between different actors and learning as a way to support accountability and transparency. As a result, some see experimentalist governance as a way to routinize dynamic accountability and transparency.Footnote ⁴¹ Second, experimentalist governance pursues deliberative participation in decision-making, where ‘actors’ initial preferences are transformed through discussion by the force of the better argument’.Footnote ⁴² In this context, de Búrca, Keohane and Sabel argue that ‘we should often establish processes that help us generate unimagined alternatives and improve our ability to choose among these alternatives by rigorously exposing each to criticism in light of the others’.Footnote ⁴³ As explained later in the context of emerging climate change technologies, achieving deliberative participation is problematic as opportunities for upstream, substantive, participation are often limited, especially in international fora.Footnote ⁴⁴ This also brings up the issue of deference toward scientific expertise at the expenses of non-scientific inputs from the lay public in the decision-making process, affecting the legitimacy of its outcome.Footnote ⁴⁵ Finally, global experimentalist governance may have the merit of catalysing innovation in international law practices and processes. Its constructive engagement with strategic uncertainty and polyarchic distribution of power allows the decision-making process to remain flexible and adaptable to changing circumstances and new practices towards effective and innovative regulation and institutions. As noted by Cottrell and Trubek, this approach suggests a transformative view of international law, which goes beyond a merely legalist vision of international law, as a rigid system of precise and enforceable rules, to unleash its expanded problem-solving potential.Footnote ⁴⁶

De Búrca, Keohane and Sabel note that ‘[t]he concept of experimentalist governance is more demanding than the broader category of pluralistic governance processes […]’.Footnote ⁴⁷ The main difference lies in its ability to regularize pluralist practices that mostly emerge spontaneously and are undertaken on an ad hoc basis. While pluralist modes of governance run mostly in parallel with hierarchical international regimes, experimentalist governance constructively engages with them, using some of their features and procedures in a recursive and participatory fashion.Footnote ⁴⁸ In so doing, experimentalist practices can get the best from both experiences: on the one hand, they do not reject legally binding norms, but use them as penalty defaults, which are characteristic of traditional modes of governance; on the other hand, they institutionalize ad hoc forms of decentralization, consultation, discretion and cooperation, which are key features of transnational networks and regime complex.

III. AN EXPERIMENTALIST GOVERNANCE PERSPECTIVE OF CLIMATE CHANGE

The debate on experimentalist governance has flourished in the US and EU social science and legal doctrine since the 1990s.Footnote ⁴⁹ Its resonance is now increasingly evident in international law and politics. In environmental policy, examples of global experimentalist governance have been identified most notably in forestry initiatives, the international agreement for the protection of dolphins against tuna fishing practices and the Montreal Protocol to the Vienna Convention on the Protection of the Ozone Layer.Footnote ⁵⁰ But, in the context of climate change, experimentalist practices remain largely unexplored.Footnote ⁵¹ This is surprising as climate change has increasingly become a crucial global governance issue, which presents linkages with the paradigm of strategic uncertainty and polyarchic distribution of power, constituting the scope conditions for experimentalist governance.

On the one hand, climate change is seen as a wicked problem: that is, a problem to which solutions cannot readily be found in existing mechanisms and which persists even after the application of the best-available practices.Footnote ⁵² Von Homeyer refers to climate change as a ‘persistent environmental problem’ that could drive the emergence of experimentalist governance, based on relatively close links between the problem and the operating logic of the economic sectors causing it; high complexity; low visibility and a global dimension.Footnote ⁵³ These characteristics point towards the experimentalist idea of strategic uncertainty, where traditional problem-solving strategies fail.Footnote ⁵⁴

On the other hand, climate change certainly constitutes a collective action problem. In Cole's definition, this is ‘one that cannot be solved by a single individual or member of a group, but requires the cooperation of others who often have disparate interests and incentives, raising the costs of transacting or negotiating a cooperative solution’.Footnote ⁵⁵ In an experimentalist governance context, this could be framed as a question of ‘polyarchic distribution of power’, where no actor is able to operate in isolation, nor impose its preferred solution, without cooperating with others. Under this reading then, experimentalist experiences could make a valuable contribution to climate change governance, providing a new framework for dealing with the failure of traditional problem-solving strategies and multilevel cooperation in decision-making.

This paradigm of uncertainty and distribution of decision-making power is particularly relevant when it comes to the governance of climate change technologies. Over the last decade, these issues have increasingly been debated with respect to the potential governance of geoengineering research. While mechanisms are in place for dealing with some aspects of their potential control (eg environmental impact assessment), these techniques present huge challenges, raising the recurrent question of how to deal with uncertainties, risks and asymmetric distribution of impact across regions and communities, while ensuring effective participation, accountability, transparency and regulatory flexibility.

IV. GEOENGINEERING TECHNOLOGIES: RISKS AND UNCERTAINTIES

Geoengineering methods are generally divided into two categories: Carbon Dioxide Removal (CDR) and Solar Radiation Management (SRM) techniques.Footnote ⁵⁶ CDR techniques ‘address the root cause of climate change by removing greenhouse gases from the atmosphere’.Footnote ⁵⁷ These include land-based methods, such as afforestation, reforestation, or direct air capture; as well as ocean-based methods, such as ocean iron fertilization. The latter aims at increasing the rate of CO₂ transfer into the deep sea by manipulating the ocean carbon cycle through addition of nutrients, such as iron.Footnote ⁵⁸ SRM techniques, instead, ‘attempt to offset the effects of increased greenhouse gas concentrations by causing the Earth to absorb less solar radiation’.Footnote ⁵⁹ These refer to a series of untested methods that would increase the reflectivity of the Earth by making its surface brighter; cool the Earth through injection of cloud-condensing particles into the atmosphere; scatter sunlight back to space, through injection of sulphate aerosols into the stratosphere; or reflect solar radiation by positioning sun-shields into space. Most of these techniques are merely technological concepts and are not even at the stage of development, as ‘much more research on the feasibility, effectiveness, cost, social and environmental impacts and possible unintended consequences is required to understand the potential benefits and drawbacks, before these methods can be properly evaluated’.Footnote ⁶⁰

However, all these methods assume that we can successfully manipulate the climate to counteract dangerous climate change, and present significant risks and uncertainties.Footnote ⁶¹ Ocean fertilization, for instance, is likely to increase ocean acidification and lead to ecosystem shifts and loss.Footnote ⁶² This would result in huge distributive impacts, especially upon those communities relying on traditional uses of marine resources for their livelihood.Footnote ⁶³ SRM might adversely change and reduce the precipitation pathways, disrupting large scale food production,Footnote ⁶⁴ causing droughts and water resources scarcity.Footnote ⁶⁵ The ultimate effects of SRM and some CDR, such as ocean fertilization, remain largely unknown and unpredictable, as we remain ignorant of some of the potential impacts, including with respect to the scale and consequences of the ‘rebound effect’. This refers to the circumstance by which, upon cessation of geoengineering, the climate would warm more rapidly than if no geoengineering had been conducted, making the impact of climate change much more disruptive than what we experience now.Footnote ⁶⁶ These techniques are also characterized by uncertainties with respect to their potential effectiveness. According to the IPCC, the ability of CDR techniques to control the climate system is likely to be limited, as there is insufficient knowledge to calculate how much CO₂ could be offset by these techniques on a century timescale.Footnote ⁶⁷ While SRM methods, if realizable, have the potential to substantially offset a global temperature rise, they would also modify the global water cycle, and would not reduce ocean acidification.Footnote ⁶⁸ The insufficient knowledge about the potentially huge intended and unintended consequences of these techniques, and their probabilities, clearly constitutes a barrier to effective governance and regulation.Footnote ⁶⁹

These challenges are aggravated by social and ethical concerns, including issues of public engagement in, and consent to, a potential geoengineering scenario.Footnote ⁷⁰ The side effects of these techniques are potentially going to be felt more by the poor, who are more vulnerable to climatic changes and less able to commit resources to adapt, raising substantive intra- and intergenerational equity dilemmas.Footnote ⁷¹ This makes the question of social consent and participation in the decision-making even more urgent, and yet problematic. Whose consent is necessary when it comes to manipulating the global climate? How can we control the eventuality of unilateral action by States, or individuals outside government oversight, on climate geoengineering, and reduce the risk of global conflicts and security threats?Footnote ⁷² These are all questions that should be addressed by allowing upstream public participation in the decision-making process.Footnote ⁷³

These risks and uncertainties expose the decision-making about these technologies to Collingridge's ‘control dilemma’. This claims that until a technology is sufficiently developed, its impacts cannot be sufficiently understood in order to assess, regulate and control its deployment; but, at the same time, early regulation is necessary to control technological development and avoid negative impacts.Footnote ⁷⁴ In a geoengineering context, this impasse is unavoidable. As most of these techniques are still in a conceptual stage, more research is needed to reduce risks and uncertainties, as a fundamental step to design effective governance and regulation. At the same time, ‘global, transparent and effective control and regulatory mechanisms’ are increasingly viewed as a precondition for any geoengineering research activity.Footnote ⁷⁵

While an unconditional prohibition is probably undesirable, international governance mechanisms are crucial to control and manage unintended, potentially dangerous, effects and the risk of unilateral action.Footnote ⁷⁶ As existing international treaties and rules were not conceived with this technology in mind, they do not provide direct control of geoengineering techniques, nor bespoke mechanism to ensure participation, flexibility and transparency for the governance of these techniques. Although this is not to suggest that potential geoengineering activities would emerge in a legal vacuum, as customary international law principles would still apply, experimentalist governance theories could contribute to support and redirect international law towards effective governance of geoengineering research.

V. THE LIMITS OF INTERNATIONAL LAW AND THE POTENTIAL FOR EXPERIMENTALIST GOVERNANCE IN GEOENGINEERING RESEARCH

The relationship between law and governance is unavoidably complex.Footnote ⁷⁷ De Búrca and Scott have described such interaction based on three models.Footnote ⁷⁸ In some instances law resists and inhibits governance, or is confronted with a general reduction of its capacity (‘gap thesis’). In others, they are ‘mutually interdependent and mutually sustaining’, mainly through the interaction of soft and hard law (‘hybridity thesis’). But there are also examples where both governance and law are reshaped and transformed, as a result of their reciprocal interaction (‘transformation thesis’). Yet, governance practices appear to always question the role of law in its ability to reflectively adapt to uncertainty, changing circumstances and multipolar actor dynamics.Footnote ⁷⁹ By conceiving a global experimentalist governance approach to the international regulation of geoengineering research, the relationship between governance and law might principally be viewed as transformative.

As most geoengineering techniques are largely untested and pervaded with unknowns, the debate on their governance and regulation remains mostly speculative. However, norms and institutions that can potentially control field-experiments and small-scale research are being analysed. Domestic law and institutions would theoretically have the capacity to control encapsulated techniques through existing regulatory tools, such as existing environmental and planning law mechanisms within their jurisdiction.Footnote ⁸⁰ However, they show a limited ability to address unencapsulated methods, entailing a higher probability of transboundary effects or impact upon areas beyond national jurisdiction, which require international control.

With respect to their international regulation, there is no international treaty specific enough, nor international institution with a sufficiently clear mandate, to provide a direct governance and regulatory framework for all geoengineering methods or all aspects of individual methods. As a result, there is virtually no dedicated transnational control on climate engineering research or, potentially, large-scale deployment. This means that, as a matter of international law, ‘[a]ny State may legally conduct [geoengineering], on or over its own territory, or that of other consenting states, or over the high seas’.Footnote ⁸¹

However, this apparently permissive approach is limited by customary law principles of general application based on State practice.Footnote ⁸² Here, the no harm principle, the duty to consult and notify of potential transboundary harm, and the obligation to conduct an environmental impact assessment are of specific interest.Footnote ⁸³ First, the ‘no harm’ principle imposes an obligation on States to prevent, reduce and control pollution and significant transboundary environmental harm to the territory of other States or to areas beyond national jurisdiction arising from activities within their jurisdiction and control. This principle has been enunciated in soft-law declarations,Footnote ⁸⁴ endorsed by the UN General Assembly and the International Law Commission,Footnote ⁸⁵ recognized in important judicial decisions,Footnote ⁸⁶ and codified in most environmental law treaties through more specific treaty obligations, such as the obligation to protect and preserve the marine environment or to conserve biological diversity.Footnote ⁸⁷ Fulfilling these obligations is likely to be problematic with respect to some geoengineering technologies. As this principle primarily operates in relation to shared natural resources and hazardous activities, the extent to which geoengineering techniques would affect shared natural resources and/or constitute hazardous activities will need to be clarified. Moreover the principle not only includes activities conducted within the territory of a State, but also under its jurisdiction and control, thus including State's flagged vessels and registered aircraft wherever they may be. This clearly extends its applicability to most geoengineering-related activities, including new methods.Footnote ⁸⁸

Second, the obligation to consult and give notice of potential transboundary harm, as well as to conduct an environmental impact assessment (EIA) will apply to geoengineering activities presenting a risk of ‘significant adverse impact in a transboundary context, in particular, on a shared resource’.Footnote ⁸⁹ Although the scope and content of the assessment is to be determined by national law through the establishment of a national procedure for the ex-ante and regular assessment of the impact, international law requires consideration of the ‘nature and magnitude of the proposed development and its likely adverse impact on the environment’,Footnote ⁹⁰ as well as the inclusion of ‘the effects of the activity not only on persons and property, but also on the environment of other States’.Footnote ⁹¹ The discretion left to national authorities in establishing the procedure is subject to the exercise of due diligence, in the absence of general international rules or specific treaty provisions.Footnote ⁹² In this context, treaty assessment frameworks, such as the one established for scientific research on ocean iron fertilization under the London Protocol—which are analysed below—can ‘play an important role in ensuring both harmonization of national measures and the application of appropriate standards and thresholds for assessment of geoengineering activities’.Footnote ⁹³ These obligations, however, do not entail a duty to obtain the consent of neighbouring States to the activity, as they do not provide them with a veto power over the conduct of the activity.

Against the backdrop of customary international law rules and soft-law principles (eg the precautionary principle), most of the international law literature has focused on how these techniques, particularly ocean fertilization and sulphate aerosol injection, could—directly and indirectly—be governed by existing international treaties and institutions.Footnote ⁹⁴ This assessment has been based on their subject matter or mandate (eg climate change, protection of biological diversity, environmental modification techniques), geographic scope and impact (eg marine environment, atmosphere, outer space) and regulated substances (eg ozone depleting substances). It has then become clear that, in most cases, adjustments were needed to effectively regulate these techniques.Footnote ⁹⁵

Although existing international law is a necessary starting point, this approach risks underestimating the fundamental question of how decisions about these technologies should be made and who might legitimately make them.Footnote ⁹⁶ In other words, the broader governance question about the legitimate forms and actors of the decision-making on these techniques becomes an afterthought. Using an experimentalist paradigm, these might be described as questions surrounding the ways to deal with strategic uncertainties related to successful problem-solving techniques, and polyarchic distribution of power across a multitude of States and non-State actors legitimately involved in the decision-making process.

Here, existing international law instruments appear ill-suited to answer these questions in isolation. This is because potentially applicable international treaties have not been conceived with geoengineering in mind, and could therefore merely perform a default, simply passive function with respect to the governance of these technologies, until and unless specific amendments are actively adopted. As it stands, potentially applicable international treaties are not flexible enough to take strategic uncertainties into account, nor reflect the multilevel distribution of power that we experience in this area. Moreover there is a concrete possibility of regulatory disconnection in geoengineering. As with other technological developments (eg biotechnology, information technologies), the technology here might develop faster than its regulatory framework, reducing international law to a merely reactive function and requiring constant readjustment to new scientific and technological development.Footnote ⁹⁷ These factors make most existing international treaties ill-equipped to constitute an effective governance framework for geoengineering, in the absence of specific mechanisms for building flexibility and adaptability.Footnote ⁹⁸ More flexibility, collaborative participation and adaptability should be built into their framework to adequately reflect uncertainties, involve actors at different levels of decision-making and ensure that regulation constantly connects with its regulatory target. Some have already started suggesting alternative paradigms to international treaties, focusing on the potential for regional collaborations, adaptive management and a stronger reliance on soft-law mechanisms and bottom-up initiatives for governing geoengineering research.Footnote ⁹⁹

On this basis, it is argued that lessons from experimentalist governance could help transform and redirect international law instruments towards a more adaptive and proactive approach to governing these technologies. An iterative reflection on the mechanisms for decision-making and participation, based on provisional goals, recursive learning and cooperation could enable a more effective governance framework. This might also allow the regulatory process to be more responsive to changes and therefore reduce the potential for regulatory disconnection between law and technology. Furthermore, as Cottrell and Trubek convincingly argue with respect to new governance-type mechanisms within transnational regulatory contexts (ie WTO and EU), the recourse to open-ended standards, benchmarking of best practices, deliberation and negotiation, networks for coordinating multiple levels of governance, and the use of soft and hard law suggests an ‘expanded vision’ of international law as a framework for problem-solving, to deal with strategic uncertainty and multilevel distribution of power.Footnote ¹⁰⁰

Certainly, this is not to suggest that the existing international law rules and institutions are irrelevant. On the contrary, international law provides indirect legal constraints on States’ conduct associated with geoengineering through customary international law and any treaty provisions potentially applicable, or at least adaptable, to individual activities or their effects. It therefore serves an essential backstop function in constraining behaviour and restraining unilateral action; helping structure international and national discussion; and directing geoengineering governance to specific international institutions.Footnote ¹⁰¹ As a result, international law provides the necessary context for global governance practices to emerge. Experimentalist governance structures cannot thrive in isolation, but will need to interact with, and rely upon, existing international law and institutions, with the result of eventually being influenced by these rules and processes. But using an experimentalist model for geoengineering governance would not be straightforward. As discussed later, its practical implementation within the new international marine geoengineering governance framework under the London Protocol remains difficult, due the barriers to cooperation and public participation.

VI. MARINE GEOENGINEERING UNDER THE LONDON PROTOCOL ON DUMPING OF WASTES AND OTHER MATTER AT SEA

Among the variety of geoengineering concepts, ocean fertilization has received greatest attention in the international legal literature.Footnote ¹⁰² This is due to its alleged technical viability and the fact that some small-scale experiments concerning ocean fertilization have already taken place, steering the debate on their effective international regulation and control.Footnote ¹⁰³ Here I analyse the 2013 marine geoengineering amendments to the Protocol to the London Convention on Dumping of Wastes and Other Matter at Sea. Subject to its entry into force, these would constitute the first internationally binding control mechanism for marine geoengineering activities.Footnote ¹⁰⁴

The London Convention regime's objective is to promote effective control of all sources of pollution of the marine environment, by preventing, reducing and, where practical, eliminating dumping of wastes and other matter at sea.Footnote ¹⁰⁵ Dumping is defined as ‘any deliberate disposal at sea of wastes or other matter from (and of) vessels, aircrafts, platforms or other man-made structures at sea’ and is prohibited under this regime.Footnote ¹⁰⁶ Placement of wastes for the purpose other than disposal is not considered dumping, and is permitted, insofar as this is not contrary to the aims of the Convention. The London Protocol, which is intended to replace the Convention for Parties who have ratified it, applies a restrictive, ‘prohibited unless permitted’ approach to dumping, based on the precautionary principle.Footnote ¹⁰⁷

Parties started addressing marine geoengineering in 2007, issuing a Statement of Concern with respect to ocean fertilization experiments.Footnote ¹⁰⁸ In line with a non-binding decision adopted by the Conference of the Parties to the Convention on Biological Diversity (CBD),Footnote ¹⁰⁹ in 2008 they adopted a non-binding resolution, clarifying that ocean fertilization activities were not to be allowed as these would violate the aims of the Convention and Protocol, and qualify as dumping.Footnote ¹¹⁰ However, ‘legitimate scientific research’ on ocean fertilization was considered ‘placement for a purpose other than disposal’, and therefore permitted, subject to a national permit and provided that it is not contrary to the aims of the Convention and Protocol.Footnote ¹¹¹ Legitimate scientific research was defined as ‘those proposals that have been assessed and found acceptable under the assessment framework’, on a case-by-case basis.Footnote ¹¹² Ocean fertilization activities other than legitimate scientific research were not to be allowed.Footnote ¹¹³ In 2010, a Specific Assessment Framework for Scientific Research (AFSR) involving Ocean Fertilization was adopted to guide the evaluation of ‘legitimate scientific research’.Footnote ¹¹⁴ This constitutes an iterative process to support national authorities in issuing the permit, intended to be reviewed at appropriate intervals, in light of new and relevant scientific knowledge and experience in applying the framework.Footnote ¹¹⁵ Finally, in 2013 the Parties translated these non-binding resolutions into binding amendments to the Protocol to regulate marine geoengineering activities.Footnote ¹¹⁶ This has been presented as ‘a mark of true leadership in global standard-setting’.Footnote ¹¹⁷

‘Marine geoengineering’ is defined under the amendments as ‘a deliberate intervention in the marine environment to manipulate natural processes, including to counteract anthropogenic climate change and/or its impacts, and that has the potential to result in deleterious effects, especially where those effects may be widespread, long-lasting or severe’.Footnote ¹¹⁸ Based on this definition, a new Article 6bis states that ‘Contracting Parties shall not allow the placement of matter into the sea from vessels, aircraft, platforms or other man-made structures at sea for marine geoengineering activities listed in Annex 4, unless the listing provides that the activity or the sub-category of an activity may be authorized under a permit.’Footnote ¹¹⁹ Ocean fertilization is the only type of marine geoengineering currently listed under a new Annex 4, which states that ‘ocean iron fertilization may only be considered for a permit if it is assessed as constituting legitimate scientific research taking into account any specific placement assessment framework’ (ie AFSR for ocean iron fertilization).Footnote ¹²⁰ A binding Generic Assessment Framework established in a new Annex 5 is to guide the evaluation of ‘matters that might be considered for placement under Annex 4’.Footnote ¹²¹ In the case of ocean iron fertilization, this Generic Framework is to be complemented with the Specific Assessment Framework for Scientific Research (AFSR) involving Ocean Iron Fertilization, which ‘shall meet the requirements of [Annex 5] and may provide further guidance for assessing and issuing permits’.Footnote ¹²²

Other marine geoengineering activities could be included in Annex 4 in the future, based on a ‘recommended non-binding procedure’.Footnote ¹²³ This process enables the Protocol to remain flexible and adaptable to future developments in this field, while ensuring that the primary objective to protect the marine environment is pursued, in the light of the precautionary principle. Contextually, a procedure for the appointment of independent international experts was adopted. They should provide advice on listed activities or consideration for listing, under the oversight of the Secretariat, and can be nominated by both Parties and observers (eg NGOs). While the initial proposal envisaged a standing body, the final agreement was to appoint a roster of experts leaving discretion to the Parties as to when and how to seek their advice.Footnote ¹²⁴

Should these amendments enter into force, Parties would need to adapt their national permit system to ocean fertilization research activities. For instance, an iron fertilization activity, when conducted in the UK marine areas (although unlikely), or anywhere in the sea from a British vessel or a vessel loaded in the UK, would constitute a licensable activity under the Marine and Coastal Access Act (2009) and the Marine (Scotland) Act (2010).Footnote ¹²⁵ However, as of March 2015, no ratification has yet been notified to the IMO Secretariat. Some predict a period between five and ten years for it to come into force, during which the non-binding 2008 and 2010 resolutions will still apply, including the specific AFSR.Footnote ¹²⁶ However, subject to its entry into force, this amendment could drive controlled marine geoengineering research and provide initial governance for these activities.

VII. A CASE FOR EXPERIMENTALIST GOVERNANCE IN CLIMATE CHANGE TECHNOLOGIES?

The new regime for marine geoengineering research under the London Protocol represents a significant development in international environmental law. It combines a positive attitude towards scientific research and multilevel cooperation with a desire to accommodate the regulatory disconnection between law and technology. These amendments reflect a number of features of experimentalist governance. This connection, however, remains loose as some fundamental barriers prevent them from embodying an ideal type of this mode of governance. Section II outlined four deliberation-fostering elements. As set out above, these elements are: provisional framework goals, and metrics for evaluation; implementation by ‘lower level’ units; regular reporting and peer-review obligations; and periodical re-evaluation and review of the goals and decision-making practices. These aspects of experimentalist governance are, at least loosely, reflected in the rationale of the amendments and the assessment framework(s), and operate within the normative backstop of the rules and institutions established under the Protocol.

First, the amendments' objective is to ensure that ocean fertilization activities are compatible with the protection and preservation of the marine environment. Based on a precautionary approach, this objective can be viewed as a provisional framework goal. Its scope is open-textured as it broadly relates to the protection of the marine environment, and the possible extension of the regulation to other marine geoengineering types suggests that its current regulatory target (ie ocean fertilization) might be merely provisional.Footnote ¹²⁷ This approach then provides flexibility by ‘future-proofing’ the Protocol to allow for ‘quick regulatory responses to new techniques’.Footnote ¹²⁸ This factor is significant as it directly responds to the issue of regulatory disconnection between law and technology. However, such flexibility seems pre-framed in favour of technological development, as there is no express indication that this objective can be provisional in the sense that a marine geoengineering type should be delisted from—rather than added to—the permitted activities, as a result of new knowledge.Footnote ¹²⁹

The Generic Assessment Framework, in combination with the AFSR involving ocean iron fertilization, represents the metric to evaluate the implementation of this framework goal. This is a central tool of the architecture established under the London Protocol and is modelled on the London Protocol Annex II of Assessment of Wastes and Other Matters that May be Considered for Dumping. This template is to be used by national authorities to judge whether the proposed activity amounts to legitimate scientific research, and can therefore be permitted, in compliance with the framework goal. The assessment is an iterative process of collection of information, peer review, and comparison of data, experiences and monitoring methodologies. The evaluation focuses in particular on the scientific questions at the basis of the field experiment; research methodology; and potential conflict with economic interests. Project proposals must provide information for the assessment of the placement site, substances used, potential effects, risk management and monitoring. If the project is likely to have an impact on another area of the sea or in areas beyond national jurisdiction, consultation must be carried out with the State exercising jurisdiction or with any other potentially affected State. The consultation should engage with any other stakeholders and involve advice from independent international experts.Footnote ¹³⁰

Second, national authorities retain discretion in the implementation of the Protocol's obligations, including with respect to the application of the Generic and Specific Assessment Frameworks and other guidelines, and the selection of experts. This amounts to an opportunity to contribute to the identification and definition of the problems, through different solutions. National authorities are therefore the centre of gravity of the regime as they engage with both project proposers and the IMO Secretariat, while retaining clear autonomy in decision-making and implementation.

Third, and strictly linked to lower units' discretion, an institutionalized dialogue between key actors is established through regular reporting and peer review. National decisions and experiences must feed back into the central international process as ‘the outcome of any assessment and documentation of any permit issued shall be reported to the Secretariat and shall be made publicly available at or shortly after the time the decision is made’.Footnote ¹³¹ This reporting of information on the permitting supports the learning process and allows monitoring and peer review to feed back to the Conference of the Parties, as the regime's ultimate decision-making body, and inform the benchmarking of best practices. In connection with this requirement, a web-based repository of references relating to the application of the AFSR is being developed, which would be accessible to all London Convention/London Protocol Parties, in cooperation with the CBD, UNESCO-IOC and other forums.Footnote ¹³² Although the terms of reference indicated that this will be open to all Parties, there is no indication that information will be confidential and that non-Parties or non-State actors, including NGOs, would be denied access.Footnote ¹³³ Although this repository does not amount to an institutional clearing-house mechanism, such as the one under the Montreal Protocol's financial mechanism or the CBD, it is nevertheless important to centralize technical expertise and local decision-making practices, and enable exchange of experiences and best practices.Footnote ¹³⁴ This iterative dialogue between project proposers, national authorities and international bodies reflects the attention to cooperation between local and central units, as we see in democratic experimentalism.Footnote ¹³⁵ Following a recursive review of scientific information and permitting procedures, data from local units are regularly refined and shared with the Secretariat, as the global standard-setter.

This leads the reporting and peer-review process to finally trigger periodical revision and review as part of a reflexive exercise.Footnote ¹³⁶ The learning process through the Assessment Frameworks and the central reporting could result in adjustments to be required to the research proposal, the national permit or the list of activities permitted under the rubric of marine geoengineering. This is to ensure that regulation and governance remain flexible and adaptable to evolving knowledge, while fulfilling the framework objective of protecting and preserving the marine environment. In particular, monitoring and peer review of data under the Generic Framework ‘will indicate whether field programmes need to be continued, revised or terminated and will inform decisions regarding the permits’.Footnote ¹³⁷ K Scott stresses how this ‘risk assessment framework is a model for precautionary and adaptive management, and compares favorably with instruments providing for environmental impact assessment such as the 1991 Environmental Protocol to the 1959 Antarctic Treaty’.Footnote ¹³⁸ The emphasis on the need to regularly review decision-making regarding marine geoengineering has been central to the adoption of the 2014 Guidance for Revision of Permitted Marine Geoengineering Activities.Footnote ¹³⁹

At first glance, then, the presence of these elements suggests that we might fairly consider the governance regime for ocean iron fertilization as an example of experimentalist governance in the international regulation of climate change technologies. However, as also discussed in section II, four conditions are generally considered necessary for effective experimentalist governance. These are: the inability of governments to agree on a comprehensive set of rules and efficiently monitor their compliance; a general agreement on basic principles to allow discussion and deliberation on the goals and benchmarks; cooperation among decision-makers; and civil society participation.

The new amendments build on the first two conditions. On the one hand, they are limited to an ocean fertilization context as Parties were not able to agree on a ‘one size fits all’ governance mechanism for all marine geoengineering techniques. Such an approach would have fallen short of addressing the characteristics of each technique and prevented effective monitoring of compliance with the regulation. On the other hand, Parties were able to agree on a control mechanism for legitimate scientific research, establishing basic principles and procedures. The characterization of the main attributes of legitimate scientific research is a central aspect of the framework, and supports other international mechanisms for the control of marine scientific research (ie LOSC and the 1991 Environmental Protocol to the 1959 Antarctic Treaty).Footnote ¹⁴⁰ But, despite these two factors, the overall architecture of the international regime for ocean iron fertilization demonstrates a limited scope for cooperation between decision-makers, aggravated by the absence of penalty defaults, and restricted opportunities for deliberative participation of civil society.

Cooperation between Parties is a necessary precondition for experimentalist governance to thrive.Footnote ¹⁴¹ But the space for such cooperation under the London Protocol amendments appears narrowly framed, effectively acting as a barrier to experimentalist governance. Here, cooperation has taken different routes. With respect to the negotiations of the amendments, cooperation among Parties (including non-State actors) has been high, and widely encouraged, both in plenaries and in working groups. This is important as, while the Protocol enjoys limited participation,Footnote ¹⁴² non-Parties, such as the US, have been active in the debate and suggestion of possible options for the amendments.Footnote ¹⁴³

But, with respect to the marine geoengineering framework as operated, cooperation in information sharing essentially occurs between national authorities, project proposers and the Secretariat, with the potential support of nationally-designated experts. As discussed further below, there is limited space for external participation. Furthermore, even the scope for cooperation between ‘insiders’ appears restricted, as the exchange and learning process happens within the confined boundary of the specific proposal under discussion, taking a vertical approach to cooperation. Little is done to encourage horizontal cooperation among different national authorities and proponents to drive comparability and learning across experiences in implementation. Horizontal cooperation between national authorities is not uncommon within international environmental law treaties and institutions. Examples range from coordination between national Management Authorities under the Convention on International Trade of Endangered Species (CITES); through the IMO-supported Memoranda of Understanding for Port State Control (PSC) and cooperation under MARPOL, and multiple fora for cooperation between national nuclear energy regulators, including in the event of nuclear emergencies.Footnote ¹⁴⁴ In this context, then, the lack of specific mechanism to encourage such horizontal cooperation under the new marine geoengineering regime appears short-sighted, given the potential transboundary, and even global, impact of these activities. Even the web-based repository of information under the Secretariat is a rather weak tool, as no formal obligation is imposed on Parties to take this information into account to drive recursive learning and inform revision of framework objectives and decision-making practices.

This approach potentially reduces the role of cooperation from a catalyst for learning to a narrow procedural task within the permit regime. To confirm this point, the new regime does not foresee a system of penalty defaults to induce cooperation. As illustrated above, experimentalist governance often works in the shadow of penalty defaults, as a threat of less favourable default rules to be applied in the event that Parties do not cooperate by signing up to a destabilization regime. This is to encourage collaborative participation in the new system, by making alternative rules undesirable and inducing a re-evaluation of the benefit of collective action. In a global governance context, penalty defaults have effectively been designed as threats of trade sanctions, such as those established to penalize non-cooperation within the Inter-American Tropical Tuna Commission to minimize the death of dolphin by-catch, or of trade restrictions, such as those applicable to Ozone Depleting Substances under the Montreal Protocol regime, or those applied on an ad hoc basis by the COP under CITES.Footnote ¹⁴⁵ But in the case of the London Protocol amendments on marine geoengineering, penalty default rules are absent, or at least unclear. While it is true that penalty defaults do not constitute a foundational element of experimentalist governance, they represent a valuable deterrent to reduce the chances of parties ‘to translate reluctance to participate in new arrangements into overt or covert obstructionism’, and support cooperation.Footnote ¹⁴⁶ As geoengineering remains deeply controversial, both nationally and internationally, the likelihood that veto powers and obstructionism inhibit collaboration in this new system is certainly high.

With respect to participation of civil society, some international NGOs and scientific groups have contributed, in the plenary and the working groups, to express concerns, share technical and scientific expertise, and submit draft texts.Footnote ¹⁴⁷ The question, however, remains as to whether such participation opportunities will be retained as during the negotiation process, once (and if) the amendments come into force. Although lacking decision-making power, non-State actors’ participation increasingly plays an important role in global governance of technologies, from providing technical and scientific expertise to inform the learning exercise, to claiming more substantive participation in the decision-making process.Footnote ¹⁴⁸ It is precisely with respect to the space for public participation in the deliberative process that the new London Protocol mechanism appears limited, giving rise to the second barrier to experimentalist governance. Active participation by a broad range of stakeholders is a key feature of experimentalist governance, as a contribution to transparency and deliberative democracy.Footnote ¹⁴⁹ But the mechanisms of public participation in environmental decision-making are multifaceted, ranging from minimal consultation on technical matters to more deliberative approaches.Footnote ¹⁵⁰

Participation in the London Protocol regime takes place mainly within the boundaries of the impact assessment under the Assessment Framework(s). In that context, the exchange of information to support decision-making occurs exclusively between project proponents, national authorities in charge of the permitting process, and possibly independent international experts. The specific AFSR includes (minimal) avenues for consultation with stakeholders, mostly during the mandatory Environmental Impact Assessment for field experiments. Clearly, the choice of regulatory mechanisms for public participation in the permitting and implementation process for geoengineering research rests with the national authorities, leaving them considerable discretion. The AFSR procedure is highly technical and there are limited avenues for civil society to intercept it and engage. But even when consulted, they have no official voice outside this precise realm, resulting in a limited ability to influence the international law decision-making process.

More importantly, the substantive grounds for participation are largely pre-framed within the limits of technical and scientific reasons under the Assessment Framework(s) umbrella. The evaluation of a project relies almost exclusively upon its risk assessment and management, while disregarding potential social, cultural and ethical concerns associated with it. For instance, while ‘social and economic factors such as whether the activity, or regulation of the activity, could have important social and economic effects, including distributional effects, eg affecting certain countries or population groups’ must be considered, they are likely to be interpreted as merely measurable impacts.Footnote ¹⁵¹

In this context, cultural values, concerns beyond risk and alternative discourses about the technology may be equally important to reach ‘good decisions’ on the regulation and governance of technologies.Footnote ¹⁵² However, under these amendments, their legitimacy is overshadowed by a clear emphasis on merely technical, quantitative risk assessments, where other rationalities are not captured. This is unfortunate, but not overly surprising, given the enormous and perennial challenges of addressing socio-cultural values beyond risk in the governance of technological change, which are surely magnified at the international level.Footnote ¹⁵³ This shows that public participation is conceived as an institutionalized process between predetermined actors (ie national authorities, project developers and experts, affected local communities), insofar as it is to contribute to a defined and technical procedure to assess new proposals or new techniques. Although some of the London Protocol Parties have obligations under the UNECE Aarhus Convention to provide effective mechanisms for public participation in the decision-making, access to information and access to justice on environmental matters, this instrument has a mainly regional scope and focuses on the national decision-making, rather than on opportunities for deliberative participation of civil society in international decisions-making and governance mechanisms.Footnote ¹⁵⁴

This analysis suggests that there is an unavoidable obstacle to qualify the new marine geoengineering regime as an example of global experimentalist governance, as this mode of governance is deeply focused on wide participation, especially from civil society groups, to increase dynamic accountability and transparency. Indeed, as Sabel et al note, experimentalist governance intends to reduce the gap between overall responsiveness of the governance system and democratic participation broadly conceived. Opening the decision-making to a variety of actors and a multiplicity of rationalities would then be a necessary condition to achieve deliberative outcomes under an experimentalist governance architecture. This would be of particular importance in the context of international regulation of controversial technologies, such as geoengineering, if we are to seriously address social and ethical concerns beyond risk.

It now seems clear that, while some experimentalist elements are present, the new international regime for marine geoengineering falls short of embracing an experimentalist approach to problem solving and multilevel distribution of power. Here, the two important features of cooperation between decision-makers and participation of civil society are confined, and there is no system of penalty default to provide an incentive towards their widening. This reduces the potential of the new London Protocol amendment to represent an ideal type of experimentalist governance, to a much more modest example of loose experimentalist governance. Under these circumstances then, the advantages of flexibility and adaptability provided under this new global governance mechanism for climate change technologies are outweighed by a limited scope for effectively scrutinizing the legitimacy of the decision-making process through cooperation and participation. From a global governance point of view, it seems that the key question about how decisions on controversial climate change technologies are made and who might legitimately make them is yet to be properly disentangled.