Stakeholder Engagement in the Making: IPBES Legitimization Politics

A growing number of expert organizations aim to provide knowledge for global environmental policy-making. Recently, there have also been explicit calls for stakeholder engagement at the global level to make scientific knowledge relevant and usable on the ground. The newly established Intergovernmental Platform on Biodiversity and Ecosystem Services (IPBES) is one of the first international expert organizations to have systematically developed a strategy for stakeholder engagement in its own right. In this article, we analyze the emergence of this strategy. Employing the concept “politics of legitimation,” we examine how and for what reasons stakeholder engagement was introduced, justified, and finally endorsed, as well as its effects. The article explores the process of institutionalizing stakeholder engagement, as well as reconstructing the contestation of the operative norms (membership, tasks, and accountability) regulating the rules for this engagement. We conclude by discussing the broader importance of the findings for IPBES, as well as for international expert organizations in general.

The newly established Intergovernmental Platform for Biodiversity and Ecosystem Services (IPBES) is one of the first international expert organizations to have systematically developed a strategy for stakeholder engagement in its own right. This article explores how IPBES addressed this challenge and finally adopted a formal stakeholder engagement strategy (SES). In doing so, we address a crucial research gap: although there is a growing literature on the opening up of international organizations (Nasiritousi et al. 2016;Tallberg et al. 2013;Zürn 2014) and on the construction of local, situated legitimacy (Connelly 2010;Turnhout et al. 2015), little empirical research has been conducted on such participation in international expert organizations (Lidskog and Sundqvist 2011). One reason for this gap is that stakeholder engagement is almost always conducted as a local, small-scale, and place-based practice, with only few systematic, ambitious efforts to engage stakeholders in global environmental assessments (Saurugger 2010). For example, the Intergovernmental Panel on Climate Change (IPCC) already includes stakeholders, but it has not yet developed a formal strategy (Beck et al. 2014;Gupta et al. 2012). Institutional aspects of stakeholder engagement at the global level "have yet to be investigated sufficiently" (van der Hel 2016, 167).
In this article, we explore how an international expert organization incorporates stakeholder engagement into its operative norms. We apply the concept of the "politics of legitimation" by reconstructing how, why, and with what effects rules and principles guiding stakeholder engagement have been justified and finally adopted (Reus-Smit 2007). We focus in particular on contestation of the operative norms that regulate stakeholder engagement, because they determine membership, the assignment of tasks, and accountability (Reus-Smit 2007).
The emergence of the SES offers an excellent case for studying the politics of legitimation. The idea (simply) of bringing stakeholders to the table was challenged at a very early stage. This moment of uncertainty triggered the politics of legitimation. While there was relatively broad support for the participation of stakeholders in general, implementing it became a matter of justifying who should participate, for what purpose, and with what corresponding rights. Whereas expert organizations such as the IPCC claim to be neutral and are reluctant to openly deal with the politics underlying their activities, 2 IPBES "got its hands dirty," and put "messy" political questions such as participation and representation on the agenda of its intergovernmental negotiations (Klenk and Meehan 2015). Furthermore, the IPBES also faced the challenge of coping with a plurality of stakeholders, all of them defending their own particular claims for engagement; the spectrum ranged from representatives of United Nations organizations, multilateral environmental agreements, and the scientific community, to stakeholders from indigenous organizations and private organizations. Given the broad and diverse spectrum of actors engaged, we reconstruct how IPBES was finally able to adopt a strategy that was accepted by all the actors involved.
Our empirical materials consist, first, of participant observations of smallscale workshops on stakeholder engagement, where for three years we followed the negotiation process, mostly involving representatives of science, conservation, and indigenous NGOs; next, we collected participant observations of the 2013 and 2015 IPBES plenary meetings in Bonn; and finally, we conducted a study of official documents from the IPBES (retrievable from the IPBES website) and reports from multiple stakeholders and intergovernmental meetings (retrievable from the IPBES website and IISD Reporting Services).
This article is divided into five sections. The next section provides an introduction to the concept of the politics of legitimation in international relations, which serves as our framework for analysis; we discuss how it can be applied to and operationalized for an expert organization. The third section describes the genesis of IPBES and explains the rationale behind its development of an engagement strategy. The fourth section then analyzes the process of negotiating the design of the SES. Focusing on the politics of legitimation, it examines how operative norms about the membership, tasks, and accountability of stakeholders are justified, included in intergovernmental negotiations, and finally adopted by the plenary. The concluding section discusses what lessons can be learned from the IPBES regarding stakeholder engagement in international expert organizations in general, and regarding the politics of legitimation.

International Expert Organizations and the Politics of Legitimation: A Framework for Analysis
In this article, the concept of the politics of legitimation is used to analyze the negotiation of an institutional design for stakeholder engagement. In line with recent constructivist work, the concept refers to the different ways in which actors demand, challenge, and justify legitimacy (Reus-Smit 2007).
As Reus-Smit (2007) observed, political actors constantly seek to justify their identities, interests, practices, or institutional designs, and in doing so they engage in the politics of legitimation. This concept is in line with constructivist scholarship that claims that a merely procedural view of legitimacy (Bäckstrand et al. 2010) provides an insufficient basis for understanding citizen participation and evaluating the legitimacy of participatory action (Leino and Peltomaa 2012). Legitimacy cannot simply be derived from theoretical concepts or normative principles. As an alternative, constructivist scholarship addresses the important yet little understood issue of how the legitimacy of organizations comes into being and is maintained in a specific, situated context that also shapes an organization's interpretations (Connelly 2010;Leino and Peltomaa 2012;Turnhout et al. 2015). In this view, legitimacy is consequently fragile and open to challenge; it plays a reciprocal and highly political role. Also, these approaches shift the emphasis from "who" to "how" questions, by drawing attention to the practices through which claims of legitimacy acquire status and authority (Hajer 2009;Jasanoff and Martello 2004). Following Reus-Smit, the contestation almost always focuses on what should constitute the operative norms, and how they should be interpreted. When it comes to stakeholder engagement, these are specifically the operative norms that regulate membership, the assignment of tasks, and accountability, as well as the corresponding practices of justification.
In addition, we advance the scholarship by operationalizing the reflexive and a performative dimension of this concept. First, as Reus-Smit (2007, 159) held, politics is essential to the cultivation and maintenance of the legitimacy of an institution. He directed attention to moments where the legitimacy of an institution could no longer be taken for granted, but was challenged and had to be justified (Connelly et al. 2006). This reflexive element characterizes the emergence of an expert organization at the international level that faces a normative and institutional void, a lack of shared norms and rules according to which (legitimate) decisions should be made (Hajer 2009, 34). For example, in the case of IPBES, the operative norms regulating engagement were not given, but needed to be chosen and justified.
Second, the politics-of-legitimation lens focuses our attention on strategies to justify and institutionalize particular operative norms that have a performative dimension; they are by no means neutral, but have substantial practical implications. Following Foucault, constructivist scholarship has helped to show how forms of knowledge-making are not only critical, but also mutually constitutive of world-making (Jasanoff 2012). From such a perspective, the politics of legitimation are not a simple, neutral input into policy, but a set of performative and reiterative practices that change the phenomena they seek to represent. Operative norms, for instance, specify who gets to participate in defining what forms of knowledge matter. They codify the rights and roles of actors within an organization, and determine how norms are put into practice, and thus constrain or enable agency. The definition of categories such as who speaks for science and who for stakeholders largely influences how stakeholders are "invited" and how rules of procedure are set up to guide the work of the organization (Chilvers and Kearnes 2016;Wynne 2007). For example, narrowly framing an environmental topic as a purely technical issue reduces the role of stakeholders to one of trusting or distrusting the experts, leaving little room for engagement (Lidskog and Sundqvist 2015).
To operationalize these insights for the empirical analysis, we examined how and to what end the involved actors justified and adopted norms, and what the effects of the choices were. The three operative norms studied here are as follows: the membership of stakeholders as constituencies as well as addressees; the task(s) delegated to stakeholders, which implicitly delineate lines of responsibility between stakeholders, scientists and governments; and the accountability and governance of stakeholders-how and by whom the work of stakeholders is monitored and regulated. We reconstructed the politics of legitimation-operationalized as the negotiations about these three operative norms-by identifying how, when, and why the spectrum of choices of ways to include stakeholders was opened up, and when it was closed down (Stirling 2008).

The Making of the IPBES and Its Rationale for Engagement
Throughout the 1990s, leading scientists as well as policy actors pushed for independent global scientific assessment in the field of biodiversity (Mauz and Granjou 2010). These efforts resulted in the Millennium Ecosystem Assessment (MA), which published its results in 2005 (Reid et al. 2006). Established in 2012 under the auspices of United Nations Environmental Programme (UNEP), the IPBES builds on the achievements of the MA and IPCC. After the MA, a series of intensive multistakeholder negotiations were held under the auspices of the UNEP, in which the representatives of governments, intergovernmental agencies, and science and conservationist NGOs were included. A major milestones occurred in 2010, at a meeting in Busan, Korea, where the main features of the future organization were adopted. Interestingly, in this multistakeholder setting it was decided that the new panel should-like the IPCC-be an intergovernmental organization. As result, it was governments that, in Panama in 2012, founded the IPBES as an independent intergovernmental organization open to all member countries of the United Nations. So far, the organization's plenary sessions have been held in Bonn, Germany (January 2013); in Antalya, Turkey (December 2013); and again in Bonn (January 2015). A small secretariat in Bonn assists the plenary and its two subsidiary bodies: the Bureau, which is in charge of the administrative functions, and the Multidisciplinary Expert Panel (MEP), which provides advice on scientific and technical matters. 3 In the IPBES negotiations, we have reconstructed three interlinked rationales for why the actors called for and justified stakeholder engagement. First, during the scoping process it was recognized that the IPBES would need diverse forms of knowledge; producing yet another global assessment report on the state of biodiversity and ecosystem services science would not meet the needs of its audiences and would fail to support local action on the ground. Instead, in contrast to the IPCC's globalism, and informed by the experiences of the MA (Larigauderie 2015, 73), IPBES decided to pursue a global but multiscale assessment, and to complement this process by setting up three additional working areas: namely, knowledge generation, capacity building, and policy support. To perform these additional tasks, input from local stakeholders was required and representatives of "local" and "indigenous" knowledge were invited to the process (Borie and Hulme 2015). Stakeholders have access to knowledge about both the specific regional and local conditions and impacts of biodiversity loss, and are key to the implementation of management strategies. Thus, IPBES sought, at least in words, to take seriously the claim that global environmental knowledge-making requires diverse voices and openness to diverse forms of knowledge.
Second, as leading involved scientists claimed, stakeholder participation was seen as a means to enhance the sense of ownership within and commitment to the IPBES process (see Watson 2005). These depended in part on "whether it has been produced through a participatory and inclusive process" (Gupta et al. 2012, 70). It was therefore seen as a way of ensuring buy-in regarding the final knowledge product (Nasiritousi et al. 2016). The rationale is that engaging stakeholders reduces skepticism toward research results, thus enhancing the likelihood of scientific knowledge having an impact (van der Hel 2016, 169).
Third, inviting those most affected to participate in knowledge-making and implementation can lead to better governance, since the active involvement of stakeholders is required if science is to provide useful products (see van der Hel 2016). If an expert organization on biodiversity includes those directly affected by ecosystem degradation and loss of biodiversity, most notably local and indigenous peoples, then the chances are high that tailored solutions will be developed. 4 The rationale for incorporating stakeholders into the processes of knowledge production was thus to enhance the effectiveness of the governance regime. In sum, the rationales for engagement were mainly instrumental, and had a strategic orientation toward promoting effective participation as a way to enhance the knowledge base and increase its relevance for decision-making, as well as to enforce global governance.

Stakeholder Engagement: Tracing the Trajectory
Engaging stakeholders has remained a constant quest for IPBES (Opgenoorth et al. 2014), in both content and form. The negotiations turned out to involve dynamic processes in which the content of the agenda, the style of negotiation, and its context were substantially transformed. Throughout this process, various actor groups developed, justified, and constantly rewrote the engagement strategies. To account for this dynamic, we distinguish three phases: experimentation and opening up, translation into operative norms and conflict, and institutionalization and closing down. In each phase, new categories and operative norms emerged, and these norms were contested within and between the groups who commented (states and stakeholders) and those engaged in translating the often diverging comments into a new draft of the SES (the IPBES bodies). In what follows, we examine each phase by elaborating on the three operative norms of membership, tasks, and accountability and governance of engagement.

Phase 1: Experimentation and Partial Opening Up
The IPBES plenary invited the International Union for the Conservation of Nature (IUCN) and the International Council for Science (ICSU)-both of which have been outspoken advocates of IPBES for many years-to work with relevant stakeholders, including indigenous peoples and local communities, and the private sector, and with the secretariat, to prepare, in consultation with the Bureau and the Multidisciplinary Expert Panel, a draft stakeholder engagement strategy for supporting the implementation of the work programme. (IPBES 2013a, 1) Following this request, IUCN and ICSU issued a call for a broad set of stakeholders (with 314 submissions received in response to 1,500 invitations) in spring 2013. A meeting then took place in Paris (2013) to produce a preliminary draft of the SES. This draft was reviewed by all invited participants, including those who did not attend (180 comments in total). The IPBES Bureau and the MEP revised the document at a meeting in June 2013. 5 The IPBES secretariat made this draft available online for an open review between June 17 and July 28, 2013. During that time, stakeholders met at the regional and subregional levels to discuss and comment on the document. In total, stakeholders as well as state representatives produced 81 comments that were then displayed on the IPBES website. In cooperation with the Bureau and the MEP, the IPBES secretariat incorporated the comments into the document in August before putting it online as one of the working documents for the IPBES plenary session in Antalya, Turkey, in December 2013 (IPBES-2). 6 While stakeholders and state representatives articulated the importance of this draft produced during the first phase, a concluding discussion on the SES was postponed, apparently "due to a lack of time" (IISD 2013).
Task: IPBES invited stakeholders to take part in "supporting the implementation of the work programme" (IPBES 2013a, 1, emphasis added). Instead of having stakeholders contribute to, or even take over, a broad range of tasks (such as scoping the agenda and policy support), their task, and thus their scope of engagement, was restricted to assisting with the implementation. However, a number of stakeholders challenged this narrow definition of their task. As an alternative, they demanded that the scope of stakeholder engagement be enlarged to encompass "developing the work programme and [participating] in other decision-making processes" (IPBES 2013b). The WWF, for instance, called for a broader set of tasks (including more rights), enabling them to make contributions "to decisions made by the plenary, Bureau and MEP." However, this statement was not included in the draft document prepared for the 2013 plenary session in Antalya. 5. The result was a document called "IPBES Intersessional Process, Online Review, Draft Stakeholder Engagement Strategy for Supporting Implementation of the IPBES Work Programme." 6. This document appeared as IPBES (2013b). Accountability and governance: During this phase, stakeholders were invited to develop an SES, while operative norms defining their accountability and governance were excluded from the negotiations. In accordance with the intergovernmental status of the platform, states decided on the final wording of the strategy and developed it in close coordination with other IPBES bodies. The secretariat and the MEP deleted statements in the draft that did not support the frame for the SES set by the plenary. This frame included a facilitator role for the two major science and conservationist NGOs, ICSU and IUCN. Their role was welcomed in general. However, some indigenous organizations stressed in small stakeholder meetings, as well as in the plenary, that ICSU and IUCN cannot speak for all stakeholders. Indigenous representatives underlined the need to "recognize the diversity of [stakeholder] groups collaborating with the Platform." 7 Along these lines, the issue of the representation and underrepresentation of stakeholder groups was raised frequently in the comments on earlier drafts of the strategy; this theme included the question of whether private firms should be involved. 8 Among the states that commented on the draft, China argued that "fullscale stakeholder engaging actions will add considerable burden to IPBES, distract limited resources from the priority" (China 2013). In contrast, the Philippines requested the "establishment of institutional mechanism within IPBES to enable and facilitate stakeholder engagement in all activities" and to "be inclusive: enable access and effective participation of different stakeholders in IPBES processes and ensure that the diverse views of stakeholders are adequately presented and given due consideration" (Philippines 2013). While other states welcomed the engagement of stakeholders in general, they also stressed the need to deal with potential conflicts-such as asymmetries in capacity and financial resources. The problem of fair and balanced representation was explicitly addressed in the draft. As a result, the need to ensure "balanced representativeness in participation among the different stakeholders" (IPBES 2013b, 3) was taken up as a guiding principle.
Membership: The inclusion of the issue of legitimate representation on the agenda also opened up a discussion on who would actually count as a stakeholder. The Phase 1 draft states: In the context of the implementation of the work programme, stakeholders are institutions, organizations or groups that: As a result, the category of stakeholders is framed in an encompassing and inclusive way (by including the tasks of contributing, using, and encouraging (a-c). In fact, participants wondered how to identify potential "unknown stakeholders" that have not yet heard of IPBES. Along these lines, the Annex to the draft document provides a chaotic assemblage of criteria for the inclusion of stakeholders. It states that they should have the "relevant rights, knowledge, skills, experiences and qualifications" or the "capacity to take action" (IPBES 2013b, 6). An alphabetical list of potential stakeholders is then presented; this contains overlapping categories beginning with "academic education communities" and ending with "young people and children," with twenty others in between (IPBES 2013b, 6). Thus, we can observe a phase of opening up for negotiations on membership and its operative norm.
A number of critical issues also emerged during this phase. For instance, participants criticized the definition of a stakeholder and its lack of specificity. At the plenary sessions in Antalya (IPBES-2), the conflict over the categorical attributions of stakeholder status was explicitly expressed; ICSU and IUCN argued that "stakeholders should be seen as a relationship of partners and not just unspecified generic 'stakeholders'" (ICSU and IUCN 2013). Similarly, at a stakeholders' day prior to the meeting in Antalya, an indigenous peoples' representative articulated the concern that indigenous peoples "do not feel encompassed by the word 'stakeholders'" (IPBES Stakeholders 2013). Later, during the plenary, she said that "IPBES … must recognize … indigenous peoples and local communities, as knowledge-holders, rights-holders, and partners, and recognize our distinct contribution to the platform" (Carino 2013).
The notions of rights-holders and partners refer to the principle of being affected (d) by the strategy, quoted above. This principle is often cited in normative debates on the legitimacy of global governance. It states that "those who are affected by a decision (the impactees) should be entitled to have a say" in decisions (Marchetti 2012, 32). This far-reaching definition of stakeholders appeared in earlier drafts of the SES but was deleted when the secretariat, together with the Bureau and the MEP, prepared the document for the plenary meeting in Antalya. In this document, any mention of being affected that might result in a rights-based terminology had vanished.
This first phase can be characterized by a mode of experimentation (see Chilvers and Kearnes 2016), in which some operative norms were made open to influence by stakeholders while others were made closed to inputs through conventional framings of the task and accountability.

Phase 2: Translation into Operative Norms and Conflict
The IPBES plenary session in Antalya ended without a working session on the draft of the engagement strategy being held, and an interlude-the second phase-followed. The plenary requested that the IPBES secretariat provide a revised version of the SES draft without giving any specifications. Published online for comments by the secretariat in summer 2014, the revised version of the SES included an implementation plan and stated: In drawing up this revised strategy, the secretariat also consulted the Stakeholder Engagement Standard (AA1000SES) developed by the Institute of Social and Ethical Accountability, which provides a framework to help organisations ensure that stakeholder engagement processes are purpose driven, robust and deliver results. (IPBES 2014b, 2) Task: What is most striking about this new draft is its attempt to formalize operative norms for stakeholder engagement. In this context, the aim of formalization was to make sure that the SES procedures are "purpose driven, robust, and deliver results" (IPBES 2014b, 2). However, this type of formalization certainly contributed to restricting the available space for bottom-up and inclusive process (see Sundqvist et al. 2015).
Governance and accountability: This intervention by the secretariat shows that negotiating the content of stakeholder engagement can also lead to conflicts about who is actually in charge of the process-that is, about the accountability and governance of the strategy. Given the discussion above, one might expect the secretariat's formalized new draft simply to become the basis for negotiations in the plenary. However, in fall 2014, IUCN, ICSU, and the Germanybased Network NEFO held a Pan-European workshop and produced a response. It states, "we believe that the document IPBES/2CRP.6 [the older draft from Antalya] should be the basis for the development of a revised strategy" (IUCN, ICSU, and NEFO 2014). The document then further comments on the Antalya draft, basically ignoring the draft produced by the secretariat. It justifies this action with reference to the "wide stakeholder consultative process" and the review by members of the Bureau and the MEP that resulted in the Antalya draft (IUCN, ICSU, and NEFO 2014). This combined statement had an effect. The final draft for the negotiations presented at the 2015 plenary session in Bonn mainly built on the Antalya draft. Thus, the secretariat's attempt to impose formalization from above failed to make the draft more robust. As a consequence, it faced resistance from stakeholder groups and mobilized bottom-up processes.
Membership: This version (summer 2014) of the SES, published online by the secretariat, reframed the very notion of the stakeholder. Instead of defining stakeholders in line with the practice of contributing, encouraging, and using, as we quoted above, this version spoke of "three broad groups that need to be involved: policymakers … including the government officials of the 119 Member states of the IPBES, scientists and knowledge holders, … [and] practitioners (or implementers)" (IPBES 2014b). This recategorization is indicative of the constant struggle to come to terms with the contested concept of the stakeholder. This version would have undone the clear distinction between governments, on the one hand, and stakeholders, on the other, disturbing the allocation of power within the organization. That is why the notion of membership was once again up for negotiation in Bonn in 2015.

Phase 3: Institutionalization and Closing Down
Task: The final phase of the strategy spans the negotiations at the IPBES plenary session in Bonn, 2015. For the first time, state representatives negotiated the SES draft in the plenary, as well as in a contact group, to remove the remaining brackets indicating disagreement from the document. The contact group, which was open to observers during its three-day meeting, essentially dealt with two major areas of concern: membership issues (eligibility criteria), and accountability and governance issues (oversight mechanisms).
Membership: The first concern emerged when the wording "balanced representation" of stakeholders was deleted from the document, because some states wanted to avoid funding requests that might result from use of the term. When it was excluded, a number of developing countries objected that stakeholder participation was and might continue to be dominated by Western countries. They insisted that the IPBES must ensure balanced participation. The negotiations paused; microphones were switched off and informal chatting in small groups of three or four dominated the scene. Delegates, especially from the USA and a number of developing countries, headed by China and Egypt, discussed alternative wordings that would satisfy the requests. As a result, questions about stakeholder membership became contested issues. With some delegates already suggesting that work on the document be postponed for another year, a solution was announced: IPBES "encourages all stakeholders representing, inter alia, their regional, disciplinary and knowledge systems in their diversity to collaborate with the Platform" (IISD 2015). "In their diversity" was the wording that finally resolved the crisis, expressing concerns about balanced representation without granting any specific rights. Furthermore, stakeholders were defined as "(a) Contributors: scientists, knowledge holders and practitioners and (b) End users: policymakers," merging the categorization of practice with that of actors (IPBES 2014a, 3).
Accountability and governance: To understand the dynamics of this debate, it is worthwhile to stress that informal meetings took place not only between state delegates, 9 but also between delegates and stakeholders. To give an example of this dynamic, each morning stakeholders would meet up to discuss the agenda of the day and how they would lobby delegates on stakeholder-related issues. One day near the end of the negotiations, part of the Chinese delegation, including its head, showed up at the morning gathering of the stakeholders. The participants at the stakeholder meeting briefly introduced themselves, stressing their diversity and the networks to which they were connected. Through an interpreter, a few cordial words were exchanged, after which the head of the Chinese delegation encouraged the stakeholders to continue their meeting while he remained in the room. The stakeholders discussed a few items 9. Due to their expertise in these matters, some (former) stakeholders have been nominated by states to become part of the state delegations.
halfheartedly for a while, and then the Chinese delegation explained that they had to leave to attend their own morning meeting. The situation remained unclear, and stakeholders were left wondering what this was all about. Why could this sequence of events matter? Recall that it was China that had stated in Phase 1 that "full-scale stakeholder engaging actions will add considerable burden to IPBES," and suggested that it would instead be better to "engage a small group of key stakeholders" identified by the secretariat (China 2013). This position directly addressed the second concern, namely that of governance and accountability. The draft for the negotiations proposed two options for the proposed oversight (IPBES 2014a). In option 1, the development and operation of an engagement strategy remained essentially the responsibility of the secretariat. By contrast, option 2 suggested that the process be "overseen by an inclusive, open-ended forum of stakeholders … in collaboration with the secretariat" (IPBES 2014a, 2). These two options are indicative of the politics of legitimation, since they reflect the struggle over the meaning of engagement and how it should be governed. Option 1 presented a rather top-down model, whereas option 2 emphasized ownership and self-governance by stakeholders. However, with the proposal that notions of ownership and self-governance be handled by a stakeholder forum, conflicts about representation came to the fore again. 10 These concerned different regions (North and South) as well as an increase in the participation of powerful actors (not least from the industrial sector). Delegates explored different versions of these two options, came up with a third option, discarded it again, and continuously communicated new wordings back and forth with the heads of their delegations. China, in particular, was very reluctant to agree on a final text. Again, a small change of wording -from "forum" to "network"-made a consensus possible, since "network" indicates a lesser degree of institutionalization. In the adopted text, the responsibility mainly remains with the secretariat (similar to option 1). It requests "the secretariat, under the supervision of the Bureau and the Plenary and in collaboration with an open-ended network of stakeholders, to undertake the activities, as appropriate, set out in the initial implementation plan" (IPBES 2015, 98). The relationship between this open-ended network and IPBES remains to be specified. In other words, instead of creating an institutional sub-body of the IPBES, in which stakeholders would oversee their contributions to the organization, state representatives kept stakeholder engagement on a tight leash and did not integrate the open-ended network into IPBES's institutional design.

Conclusions: Context, Reflexivity, and Power
In this article we have highlighted how the politics of legitimation shaped both the dynamics and the design of the IPBES stakeholder strategy. In this concluding section, we first draw some conclusions that are closely related to our empirical findings, and then discuss their wider importance for IPBES as well as for other international expert organization.
IPBES has included a broad spectrum of actors, and as a result it has had to respond to a diverse spectrum of legitimacy claims. By understanding the dynamics of the politics of legitimation in terms of opening up and closing down design choices for these operative norms, we can trace three phases through which the politics of legitimation has evolved.
In the first phase, we observed an opening up of the negotiation processes to a plurality of voices, and thus to a diversity of legitimacy claims. The politics-of-legitimation lens focused our attention on reflexive moments; during this phase, the prevalent instrumental mode of participation was challenged, and alternatives were put on the agenda by actors who justified participation on alternative, democratic grounds. Their rationale was to give stakeholders a substantial voice and to empower them. During this phase, we observed an opening up in relation to the definition of the rights and responsibilities of stakeholders.
In the second phase, we saw how crucial categories such as "stakeholders" (and their subcategories) themselves became sites of political contestation. Our analysis also demonstrated how and why operative norms determining stakeholder tasks, membership, and accountability affected the allocation of decisionmaking authority within the IPBES. The performative character of these operative norms offers a partial explanation of why stakeholder engagement and the norms regulating it became contested issues. In relation to this, conflicts over the allocation of authority are often rooted in more fundamental differences in values, such as competing rationales for engagement.
The third phase was characterized by attempts to close down the negotiation processes during the subsequent formal adoption of the SES. Our analysis also showed that calls for empowering a diversity of actors by giving them a voice were marginalized during the negotiation processes. Although terms such as "rights-holders" and "partners," which imply putting states and stakeholders on an equal footing, were frequently put forward, they were excluded from the final version of the SES. This resulted in an ambivalent situation. On the one hand, the adoption of the strategy was perceived as a crucial signal to stakeholders. 11 On the other hand, the increasing contributions of stakeholders inside the IPBES are not reflected in operative norms (such as the status of their membership and voting rights). As we have shown in this case study, while stakeholders were welcomed as knowledge holders, little room was afforded for their empowerment as rights holders or partners. The instrumental manner in which dominant actors justified participation also led to the exclusion of alternative voices or democratic visions of the engagement.
What more general lessons can be learned from the IPBES about stakeholder engagement in international expert organizations in general, and about the politics of legitimation? We will stress three aspects that are of wider relevance: namely, lack of reflexivity, the institutional setting, and power asymmetries.
First, whereas engagement was used to enhance the relevance of IPBES's work, our case study demonstrates that increased engagement does not automatically bring about legitimacy, reflexivity, or relevance. This empirical outcome is in no way unique to IPBES, but characterizes a variety of cases in which norms of participation are invoked to address the needs of stakeholders. In these cases, although the process was opened up, the norm of effectiveness continued to dominate the construction of legitimacy, and there was limited scope for participation (Connelly et al. 2006;Turnhout et al. 2015).
Second, our case study shows that the institutional setting in which engagement practices are embedded has a critical impact on the dynamics of negotiations. Whereas the first phase of negotiations was conducted in informal settings such as workshops, making it possible to experiment with design options, the third phase was shaped by the intergovernmental setting, where innovative elements were systematically excluded. This empirical finding points to the decisive role that the secretariat and individual states can play in determining the rights of stakeholders within the IPBES. The intergovernmental status of the plenary and its consensus-based procedures contributed to a narrowing down of agendas. As result, they also limited the range of options put forward for debate in the negotiations. Alternatives to existing institutionalized procedures (such as elevating the official status of observer organizations) have been virtually excluded from the debate. These efforts have ultimately served to limit stakeholders' agency, and thus to protect the prevailing allocation of power between state and nonstate actors. This constraining effect of the institutional setting has also been observed in the case of the IPCC (Beck 2015; see also Haas and Stevens 2011).
Third, these empirical findings highlight the importance of taking into account the relationship between the politics of legitimation and power. The IPBES case study reveals how different actors tactically exploit the politics of legitimation to influence the process and its outcomes and to enhance their own influence. In this way, actors seek to influence the distribution of power and authority in the configuration of this new platform. To safeguard their influence, governments influenced negotiations within the IPBES plenary by controlling the processes and agendas (see Haas and Stevens 2011 for the case of the IPCC).
The prevailing asymmetrical allocation of decision-making power within the platform partly explains why stakeholders did not receive a voice. Governments are the only formal members of the IPBES with decision-making power in the plenary. As a result, they were in the position to formally adopt the SES and make final decisions about norms and rules of procedure. This can explain why options that might lead to a power shift, such as providing stakeholders with voting rights, have been excluded from the agenda for intergovernmental negotiations. This asymmetry, enforced by the rules of procedure, has led to a growing mismatch between the growing tasks of stakeholders and their marginal representation in the main decision-making body of the platform (which once again may lead to a situation characterized by the politics of legitimation).
Also in line with other case studies, stakeholder engagement was finally shaped by the narrow intergovernmental setting, where scientists and governments only pay lip service to the importance of divergent voices, perspectives, and knowledges (for the case of the MA, see also Filer 2009). In many cases, this plea for participation was of a symbolic kind, with policy-makers and scientists publicly embracing stakeholder involvement while internally (backstage) following a narrowly scientific agenda (Chilvers and Kearnes 2016).
Stakeholder engagement has commonly been restricted to procedural questions such as how participation processes can improve the legitimacy of their outcome (see Mitchell et al. 2006). In this perspective, broader and more balanced participation is often assumed to lend greater neutrality, and thus legitimacy, to the end-product of the assessment. These procedural approaches, however, underestimate the extent to which, in a variety of global assessments, knowledge and power can be unequally available and differentially composed (Klenk and Meehan 2015). Power imbalances, biased representation, and the lack of access and capacities are challenges faced not only by IPBES but also by IPCC and Future Earth (van der Hel 2016).
Given that knowledge-making about the global environment is not simply a neutral input into policy, but a set of performative practices, constructivist scholarship has long been concerned with questions of representation and power in scientific advisory processes (Jasanoff 2012;Turnhout et al. 2016). This scholarship has called for renewed attention to avowedly normative questions about whose visions are being accorded legitimacy, at the expense of what other ways of envisioning engagement? Are the people making the legitimacy claim the right kind of people to speak for the phenomena they represent? These persistent problems highlight the importance of openly addressing questions of representation, rather than defending the ideal of neutral expertise. The task for the future is therefore to analyze the politics of legitimation also in terms of how social and political asymmetries lead to inequalities in the "justification power" that enables persons or groups to contest given justifications and create new ones. Given the conditional, situated, and open nature of engagement, there is no general recipe for how to justify and maintain legitimacy. But even if, as we suggest, there is no simple or unitary model, ways of achieving a more self-aware and effective mode of engagement remain an important topic for exploration.