The meeting of multiple governmentalities and technologies of participation in protected areas: the case of the Amarakaeri Communal Reserve (Peruvian Amazon)

• The Amarakaeri Communal Reserve (RCA) is a protected area in the Peruvian Amazon co-managed by ECA-Amarakaeri, an indigenous organization representing the communities in its buffer zone, in partnership with Peru’s Protected Areas Service. • The RCA’s co-management is the product of the interactions over the past three decades between three different govermentalities – a conservationist governmentality, an extractive governmentality, and a negotiated eco-governmentality – which build on the priorities and politics of the stakeholders to the RCA. • The negotiated eco-governmentality displays two technologies of participation that are in contradiction: co-management and a management committee – a multi-stakeholder forum composed of the different stakeholders to the RCA. • Despite the work put into the RCA’s co-management and management committee, the emphasis in including the local indigenous population in the Reserve’s management has excluded the local non-indigenous population from participating in it. • This leads to the RCA government being centralized in the co-management partners, empowering indigenous peoples but denying participation to other local stakeholders in the management of RCA.


INTRODUCTION
There are two contradictory processes of territorial governance in the Peruvian Amazon: a longstanding one that has involved the extraction of natural resources and colonization of its territory, and a more recent one whose objective is to conserve the biodiversity of its forests (Álvarez 2012(Álvarez , Larsen 2016. This contradiction is evident in Madre de Dios region in south-eastern Peru. Madre de Dios has historically been considered a frontier space for the expansion of extractive dynamics in the Amazon, most notoriously evident in the current environmental crisis caused by gold mining in the area (Álvarez 2012, Oliart andBiffi 2010). Paradoxically, 44.87% of Madre de Dios' territory is divided into seven natural protected areas, including the Amarakaeri Communal Reserve (RCA, in Spanish) (Rodriguez 2018).
The RCA is a protected natural area under a special co-management regime where Peru's Protected Areas Service (SERNANP, in Spanish) shares the management of the area with the indigenous populations that live in the 10 native communities 1 of its buffer zone. The communities are represented by ECA-Amarakaeri (the Executor of the Administration Contract for the Amarakaeri Communal Reserve), which has a management council elected by their inhabitants every three years. This co-management regime has received various international awards and acknowledgment in recent years, including its addition to the International Union for the Conservation of Nature's Green List in 2018 and the United Nations Equator Prize 2019. These recognitions emphasise the co-management model's achievements, including decreased deforestation as well as the participation of the local indigenous population in the management of the RCA 2 .
Nonetheless, co-management is not the only mechanism of participatory governance in this protected area. Peru's Law for Protected Natural Areas requires that all such areas have a management committee (comité de gestión). These are open, free and voluntary multi-stakeholder forums (MSFs) for the participation of any relevant stakeholders that wish to support protected area management. In the RCA, this MSF includes more than 30 stakeholders -including national and subnational government actors, grassroots organizations, NGOs and indigenous organizations -that intervene directly or indirectly in the RCA and, as per the law, seek to inform the co-management of the protected area. However, in practice, the presence of an indigenous organization as a co-management partner has led to the exclusion of other local actors from its management committee, which affects the ability of the committee to fully achieve its objective.
What socio-historical processes have led to the production and overlap of different governmentalities in the RCA? How have indigenous peoples contested and negotiated these governmentalities? What are the governmentality challenges stemming from the creation of two arenas of participation in apparent competition with each other? These questions demonstrate the relevance of the RCA as a case study, as it reveals a series of intersections and overlaps between different ongoing processes in this territory, including extractive activities and conservation agendas (Orihuela 2017, Pinedo 2017, 2019. In turn, these are juxtaposed to the collective governance institutions of the Harakbut, Matsigenka and Yine indigenous peoples in the area, with the new mechanisms of effective participation promoted by SERNANP for territorial management (Álvarez 2010(Álvarez , Amend et al. 2017, and their own recognized rights as indigenous peoples to territory and self-determination (Gray 2002). Furthermore, the presence of Andean migrants 3 throughout the RCA's buffer zone produces socio-economic dynamics that at times lead to conflict with the local indigenous Amazonian peoples.
This paper seeks to contribute to the discussion on the new participatory governmentalities that are emerging in the Peruvian Amazon (Caruso 2011, Larsen 2016, Orihuela 2017, examining a case where indigenous peoples reappropriate green and participatory government technologies for their benefit and, in doing so, exclude other non-indigenous local actors. We propose that in the history of the RCA, conflicts arise not only over the control and use of the area's resources and how nature and the environment are constructed (Escobar 1999, Bridge, McCarthy, Perrault 2015, but also over the imposition of different forms of government and management for the RCA that form (eco)governmentalities (Foucault 2006, Fletcher 2017, Ulloa 2004.
The paper reveals various forms of governmentalities in the context of the RCA (Ulloa 2004, 2006, Fletcher 2017, Agrawal 2005 which deploy different governance practices that are interrelated, superimposed and/or in conflict. The historical, political and social processes in Madre de Dios have both shaped these different governance practices and configured which of them will prevail over the others, as well as how indigenous peoples have exercised agency to transform them. As the paper shows, the governance of the RCA is concentrated in the co-management model that links indigenous Amazonian peoples and the Peruvian state, creating a negotiated eco-governmentality. As a consequence, other stakeholders with a presence in the territory, such as the Andean migrant population, end up being excluded from the RCA's management committee.
The next section discusses the concepts of governmentality, environmentality and eco-governmentality. The methods through which this research was carried out are subsequently explained, followed by a section on the socio-historical context of the RCA and the area around it. This section is followed by a discussion of the RCA's governmentalities, the co-management model, and its relationship with the Reserve's management committee. The conclusions provide wider lessons for MSFs related to protected natural areas with participatory management models like that of the RCA.

(ECO)GOVERNMENTALITIES: OLD AND NEW GOVERNANCE PRACTICES REGARDING ENVIRONMENTAL ISSUES
Foucault's understanding of power removed it from juridical and economistic models of analysis in order to understand its productive side (Foucault 2008). Foucault argued that power is not a thing that one possesses, or that only represses but is, above all, relations that produce actions with the purpose of shaping behaviour (Foucault 1988). According to Castro-Gomez (2016), the exercise of power for Foucault can be understood through the concept of governmentality, as the "attempt to shape human conduct by calculated means" (Li 2007b: 205). This theoretical framework facilitates engagement with the various governmentalities exercised over people. These are deployed through discourses, tactics and knowledge that configure the different rationalities of governance and are not only bounded to the spheres of the state.
Since the late 1970s, "environment" emerged as a category and a problem to analyse and resolve through diverse disciplines and scientific knowledge -what has been called green thinking (Ulloa 2002). This did not only configure the problem but also how to address it through new governance practices applied to this new subject: the environment. This governance rationality is exercised on people's behaviours towards the environment and the production of the environment itself (Agrawal 2005, Valdivia 2015. Environmentalist discourses define a set of environmental problems for which intervention is required in order to control, regulate and mitigate issues such as climate change, global warming and resource scarcity. But the environment, as a problem, is governed not only by nation-states but by an entire series of multi-scaled global practices that range from small local interventions to the practices of multinational bodies (Ulloa 2005). Luke (1999) was one of the pioneers in thinking about this new set of problems from the governmentality perspective when he coined the term environmentality. Environmentality is a government rationality underpinned by a discursive construction of techno-science that transforms nature into an environmental object that can be managed, controlled, shaped and improved by scientific ecological knowledge. Thus, it is not only "nature", but also "the environment" and "natural resources" that legitimate the intervention of governance practices. Similarly, Rutheford (1999) pointed out that the concerns about ecological problems could be understood as bio-politics, which controls the population and is related to the institutionalisation of a naturalist scientific rationality. This has generated new techniques for managing the environment -ecological governmentality.
Agrawal reinterpreted the concept of environmentality along the lines of how it and its subjectivities are constructed: "the term stands for an approach to studying environmental politics that takes seriously the conceptual building blocks of power/knowledge, institutions, and subjectivities" (Agrawal 2005: 8). The construction and articulation of new subjects and environments are visualised in Ulloa's (2004Ulloa's ( , 2005 work on the production of the 'ecological native' in Colombia through what she calls eco-governmentality. Ulloa explains how the coming together of the environmental and indigenous movements promoted by NGOs and multilateral agencies, led to the discursive formation of the idealized figure of indigenous peoples as the guardians of nature. However, although this discourse became commonplace for environmental NGOs and multilateral agencies, it has also been reconfigured by indigenous peoples themselves. This shows that no governmentality is a finished product as it is constantly being challenged by both powerful actors and historically underrepresented communities. Cepek (2011) presented an important critique and response to the environmental governmentalisation processes. Based on ethnographic work with indigenous Cofán peoples in the Ecuadorian Amazon, Cepek notes the importance of being attentive to the analytical biases and limits of the concept of environmentality. Engaging with Agrawal's reinterpretation of environmentality, he emphasizes that the ability of this form of governance to shape indigenous practices and their qualitative and quantitative approximation of the perspectives of the stakeholders (see Sarmiento Barletti and Larson 2019 for the full research methods). Interviews were carried out with 37 people, including government officers, NGOs actors, researchers, private sector actors, leaders of indigenous federations and native communities, members of Puerto Luz native community and migrant Andean residents of Bajo Pukiri. Interviews sought to understand the different perspectives held by the stakeholders to the RCA over its management and the equity and effectiveness of its management committee. We interviewed both participants and nonparticipants to the RCA's management committee.

MIGRATION AND SETTLEMENT PROCESSES BY INDIGENOUS AND NON-INDIGENOUS POPULATIONS IN THE RCA
The Amarakaeri Communal Reserve is located in Madre de Dios region and has an extension of 402,335.62 hectares. The ten native communities in its buffer zone are inhabited by Harakbut, Yine and Matsiguenka indigenous peoples (see Figure 1). Harakbut people's ancestral territory is within the current boundaries of Madre de Dios region. The area also includes 23 settlements populated by immigrants from, mainly, Peru's southern Andean region (Puno, Cusco and Ayacucho regions). Different kinds of economic activities take place in these areas: livestock raising, timber extraction, agriculture and gold mining.
The demographic composition of this area has been shaped by various processes; these include the rubber boom at the start of the 20th century, the missionisation process by Dominican priests of the indigenous populations from the mid-20th century onwards, and the expansion of gold mining. The rubber boom had three important consequences for both the Madre de Dios region and the indigenous Amazonian peoples in that area (García 2003). The first was the vast reduction of the populations of Harakbut and other indigenous peoples due to enslaving, killings and the diseases provoked by the rubber interests. The rubber barons also forcibly moved new indigenous populations (such as Yine people) into the Madre de Dios area. The second was the confrontation -the Great War, as the Harakbut call it (Gray 2002) -that took place between Harakbut clans. This conflict was catalysed by the move of two clans, the Toyeri and the Arasaeri, into the territories of other clans as they escaped from the rubber barons. The third was the arrival in great numbers of nonindigenous Andean migrants, following the construction of roads connecting Puerto Maldonado with Cusco to facilitate rubber trade (Valencia 2014).
Harakbut people sought to stay clear from the rubber trade by staying away from the areas most impacted by its expansion, but this isolation ended with the evangelization promoted by the Dominican order of the Catholic Church from 1940 (Gray 2002). One of the main consequences of the missionisation process was the opening of the Harakbut economy towards markets and extractive activities. Dominican relation to nature is overstated. For Cepek, by overvaluing environmentality, various analysts have reduced the critical capacity and agency of the very populations that are the subject of the environmental governance policies. Although there are forms of environmentality that could be a model for the construction of green subjects, or 'ecological natives,' this does not mean that such construction will happen in practice. Similarly, Fletcher (2017) invites analysts to go beyond environmentality or eco-governmentality to include different forms and practices of governance that can be articulated, opposed, overlapped, or on differentiated scales. For example, Orihuela (2017: 62) identified different environmentalities at play in the Tambopata National Reserve, also located in Madre de Dios. Some environmentalities were concerned with the promotion of scientific investigation and keeping people out of the reserve, others sought to develop sustainable agroforestry projects, and others promoted community eco-tourism. Within that mix, indigenous peoples struggled and conducted political action for their rights to land but also linked up with several of those environmentalities.
This paper assumes no fixed governmentalities and engages with them as processes that are yet to consolidate and that intersect with other forms of governance. Furthermore, as shown below, there are different forms of governance exercised within the same environmentality; some that protect nature as a stronghold and impede human access to it, and others that promote the sustainable use of resources. This paper frames the actions of the indigenous populations living in the buffer zone of the RCA within this framework of linkages, struggle and negotiation. These people have re-appropriated the governance of their ancestral territory, and face new challenges and contradictions as they engage with diverse practices of government and non-government actors. This context raises questions about the need and role that MSFs can play in arenas in which historically excluded populations have, through their own agency and struggle, succeeded in recovering participation in the governance of their ancestral territories.

RESEARCH METHODS
This paper is based on research and engagement conducted as part of a multi-country research project on multistakeholder forums carried out by the Center for International Forestry Research (CIFOR). Research on the RCA's multistakeholder management committee included published sources and fieldwork that was carried out over April-June 2018, in four main places: the native community of Puerto Luz (in the RCA's buffer zone), the town of Bajo Pukiri-Delta 1 (close to the RCA), and the cities of Puerto Maldonado (Madre de Dios' capital) and Lima (Peru's capital). The paper is also informed by insights from continuous engagement with ECA-Amarakaeri, SERNANP and the RCA's management committee throughout 2019 and 2020. A mixed methodology was implemented that combined qualitative fieldwork, semi-structured interviews and focus groups, allowing a priests promoted timber extraction for sale and formed debt and peonage relations with the indigenous labour force in exchange for medicines and metal tools that the Harakbut previously did not use (Pinedo 2017).
Another transformative process was spurred by the introduction and expansion of alluvial gold mining. The gold fever starting in the 1970s generated the greatest impact on the indigenous populations of the RCA. Promoted by the Peruvian government through its Mining Bank and the Law for the Promotion of Gold Mining passed in 1978 (Valencia 2014), the government supported Andean migrants in colonizing the forest and granted them extraction rights in Harakbut ancestral territory. This was done so that they could mine gold, which would be later sold to the Mining Bank. The influx of people from the Andes increased throughout time from 12 000 migrant miners in 1972 to 20 000 in 1980 (Moore 2003: 85), reaching a total of 40 000 during the 2000s (Gray 2002: 123).
Harakbut people deployed two main strategies to cope with this pressure on their territory. First, they sought legal recognition and titling as native communities, which were recognized throughout the 1980s. Second, to reduce conflicts, they allowed the Andean settlers to extract gold in some areas of their territory by renting out areas within their native communities. Moreover, they asked their Andean tenants for access to their gold extraction knowledge and technology so they could extract gold themselves for an income (Pinedo 2019). Although mining has become an important threat to their territories given the river pollution and deforestation it causes and the overlap of concession rights granted to Andean migrants and their own territories, it has also become one of the main ways to generate income in a fast and constant manner for some of the inhabitants of the native communities in the RCA's buffer zone.
Thus, the ancestral indigenous territory of the Harakbut people is, in reality, a "hybrid of different forms of territoriality that are superimposed, combined and have ambiguous effects" (Pinedo 2019: 30). It is the product of different forms of governing the territory based on the intervention on natural resources and relations between Amazonian and Andean populations, which would become even more complex with the creation of the Amarakaeri Communal Reserve.

THE PRODUCTION OF THE AMARAKAERI COMMUNAL RESERVE: A CROSSROAD OF MULTIPLE GOVERNMENTALITIES
Protected natural areas are territorialized forms of state government practices over a delimited space that, through various technologies, produce the idea of a specific and differentiated environment to be controlled and conserved (West, Igoe andBrockington 2006, Valdivia 2015). Protected areas in display different forms of governmentality -fortress conservation, co-management, limited resource use areas, etc. (Orihuela 2014, Rodriguez 2018. Legally, Communal Reserves are types of protected areas that allow the participation of indigenous peoples in the government of the territory (Protected Areas Law 1997 Nº 26834, art. 22). For Madre de Dios' indigenous movement, led by the Native Federation of the Madre de Dios River and Affluents (FENAMAD in Spanish), the recognition of native communities was limited and proved to be insufficient for the defence and protection of the entire Harakbut territory; there was an overlap of the territorial rights of native communities, miners' extraction rights, and an expansion of mining concessions around their communities. Interviews with indigenous leaders and their NGO allies noted that FENAMAD proposed the creation of the Amarakaeri Communal Reserve as a defence mechanism for those communities (see also Alvarez et al. 2008). Communal Reserves were promulgated in Peru's national legislation in 1975 as a way to ensure the territorial defence of the, then, newly titled native communities in the Peruvian Amazon (Amend et al. 2017).
The process of creating the RCA had two phases. The first phase (1990)(1991)(1992)(1993)(1994)(1995)(1996)(1997)(1998)(1999)(2000)(2001)(2002)(2003)(2004)(2005)(2006) started with enthusiasm, supported by indigenous peoples and their NGO allies. FENAMAD, together with the Harakbut people, promoted the creation of the RCA to defend their ancestral territory from the extractive interests including gold mining and the extraction of timber and hydrocarbons (Moore 2015). The original idea was for the RCA to become a legally-recognized indigenous territory that would allow indigenous peoples to re-appropriate the protected area -a green technology of governance -in order to strengthen their own forms of self-government and territorial rights. In 2000 a section of Harakbut people's ancestral territory was recognized as a conservation area. The indigenous movement considered this recognition as a victory in terms of its ongoing struggles for territory. Hector Sueyo (2002), a senior Harakbut leader and in charge of Indigenous Affairs in Madre de Dios' regional government at the time of research, noted that their objective was for the RCA to be managed by indigenous peoples. Through this, they sought to achieve the state's recognition of their territorial rights. The indigenous plans on how they would manage the RCA included a proposal to set up a council of elders to support its management (Sueyo 2002).
This first phase ended with disillusionment for indigenous people. Once the RCA was created in 2002, SERNANP 4 imposed a conservationist governmentality, impeding indigenous peoples' access to their ancestral territories. It proposed that the management of Communal Reserves would be done jointly by SERNANP and the indigenous communities in the reserve's buffer zone. In practice, SERNANP was in charge of exercising control and surveillance to conserve the biodiversity of the entire protected area, and restricted access to most of the area outside its buffer zone 5 (see Figure 1). This co-management approach was made into law in 2005 when the Special Regime for the Administration of Communal Reserves was approved, creating a new participatory technology of governance for the Amazon in which the Peruvian state would manage an area together with the indigenous populations of its buffer zone, represented by the Executor of the Administration Contract (Amend et al. 2017). However, our interviewees noted that this new governance approach did not consider the diverse ways in which indigenous peoples understand, engage with, and live in their territories (see also Gray 2002). Government practices in the RCA imposed from above a green governmentality that prohibited the entry, use and extraction of resources by indigenous peoples living in its buffer zone. Indigenous leaders told us that this generated uncertainty, annoyance and disappointment among the communities in the reserve's buffer zone. They were banned from going into the RCA to conduct their traditional practices and other income generating activities such as extracting timber or mining for alluvial gold.
The second phase (2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014) was marked by conflict and crisis in the governance of the RCA, as a neoliberal extractive form of government was imposed from above, in what appeared to contradict the conservationist governmentality and indigenous government practices (Pinedo 2017). The indigenous population's disillusionment turned into protest when the Peruvian government signed an oil and natural gas exploration contract with the transnational corporation Hunt Oil for Lot 76, which overlaps most of the RCA and the communities in its buffer zone, and even part of the neighbouring Manu National Park. This generated more uncertainty and anger among the local indigenous population, as they were banned from engaging in activities they had always carried out in their own territories, while international companies were being provided with extrative concessions in them (Álvarez et al. 2008).
As Pinedo (2017) notes, Hunt Oil worked to form clientelist relationships with indigenous communities, gaining support from some of them in exchange for economic and social support ranging from bribing leaders to health campaigns. This generated divisions within the indigenous movement, as clientelism disciplined indigenous groups through the construction of moral debt relationships (Graeber 2014). This form of extractive governmentality produced unequal power relations and conflicts within the indigenous movement, and proposed a way of understanding and governing nature as something to be exploited for private benefit.
In theory, extractive governmentality contradicted other forms of conservationist governmentality. However, in these neoliberal post-frontier arenas (Larsen 2016, Orbegoso 2018), it is not only possible for them to co-exist, but also to articulate with each other. In 2007, ECA-Amarakaeri and SERNANP created the Master Plan for the 2007-2009 period for the RCA, in which a Strict Protection Zone was created to ban hydrocarbon exploration in and extraction from the area of the headwaters of the Madre de Dios River. Months later, however, SERNANP changed its position and approved a new Master Plan, with ECA-Amarakaeri's support, that eliminated the Strict Protection Zone and included a new technical opinion on hydrocarbon activities, considering that they were not a threat for the RCA. On the contrary, SERNANP noted that the extraction of hydrocarbons could provide funds for a better management of the RCA (FENAMAD 2009). Conservationist and extractive governmentalities not only coexisted but the former, while deploying technologies to restrict indigenous people's access to the RCA, allowed Hunt Oil's extraction of resources. This allowed for the introduction of governance practices that sought to produce different kinds of subjects -morally indebted indigenous communities -and environments -extraction zones in the RCA -that would allow extractive activities to be carried out. The regional indigenous movement challenged this intersection of governmentalities as FENAMAD and the communities it represents responded with protests and demanded the resignation of SERNANP's Head of the RCA and of ECA-Amarakaeri's President. In 2008 the first Master Plan (2008-2012) for the RCA was approved, in which the changes made to include Hunt Oil were maintained. This Master Plan became a legal planning document once it was signed by the Head of the RCA and ECA-Amarakaeri's President. The plan had never been consulted with the communities of the RCA's buffer zone or with FENAMAD. FENAMAD refused to recognize ECA-Amarakaeri and formed its own Executor of the Contract of Administration, which opposed Hunt Oil's activities in the RCA.
The conflict between the native communities, both ECA-Amarakaeri, and SERNANP went on for a few years until it stabilized with the election of a new executive committee for ECA-Amarakaeri in 2011 and the replacement of SERNANP's Head of the RCA around the same time. The new Head implemented a new strategy of working closely with ECA-Amarakaeri's executive committee and developing the technical capacities of its members to be able to participate more effectively in the RCA's co-management. Interviews revealed that this renewed attempt at collaboration was well received by Harakmbut leaders, who responded by setting aside their mistrust of SERNANP after it had supported Hunt Oil.
This reconciliation, negotiation and adaptation of governance practices between ECA-Amarakaeri and SERNANP could be described as a negotiated eco-governmentality (2014-present) -the formation of new governance practices through a new model of participatory governance for the RCA. This is a co-management (cogestión) model in which the Peruvian state (SERNANP) and the indigenous populations (ECA-Amarakaeri) are responsible for the administration, vigilance and control of the area. The positive experience with this new modality of governance for the RCA set the ground for a new ideal management of Communal Reserves: the co-management model with an intercultural focus. This model consists of transforming the classic, conservationist governmentality model where protected areas are managed solely by SERNANP, to one in which ECAs and SERNANP are equal partners in the conservation, management, and development actions in Communal Reserves. As SERNANP's Head of the RCA noted in an interview:

"The (…) duo of an indigenous leader and the Head of the RCA [is] not easy because it is a question of whether they are able to speak and understand what co-management is. The topic of co-management in protected natural areas is a challenge which I would say that SERNANP is still learning about. […] I've repeated it many times and I'm going to continue saying it, for me the issue of the individuals involved [in co-management] is crucial. People have to come with an open mind. An openness regarding [their] own actions. That you, as the Head of a [protected] area, can talk to an ECA president and reach consensus about how things are going to be done, that it is effective co-management, that's key to it."
The negotiated eco-governmentality was crafted in conversations and agreements between SERNANP and ECA-Amarakaeri through which they committed to conserve the RCA's biodiversity without challenging the relationship between indigenous peoples and their territory. In part, this negotiation was successful due to the constant work of the indigenous movement, which overcame previous internal conflicts, that at its worst point had led to there being two ECAs at the same time, to retake the government of the territory. It was also successful due to SERNANP's internal transformations, leading to the adoption of new strategies for local populations to participate in the government of the Communal Reserves. Thus, an unprecedented new partnership was established to govern the RCA: the participatory co-management between the state and the ten native communities. Co-management, a participatory technology, allowed for the introduction of negotiated eco-governmentality to the RCA. 6 However, there is a second participatory technology at play: the multi-stakeholder management committee.
In the RCA, this second arena of participatory governance is framed by an intricate socio-political panorama. The management committee is an MSF that, in theory, should close the representation gap of the non-indigenous populations living in and developing livelihoods from the RCA's buffer zone. In terms of the structure through which the RCA is managed, SERNANP deals with the vigilance and control of the RCA, and its co-management partner, ECA-Amarakaeri, is in charge of the area's management and representing the indigenous population. The management committee is a third actor, which should function as a voluntary multistakeholder participation mechanism aimed at creating an arena of dialogue, coordination and collaboration with local populations, civil society and sectors of the subnational and local governments. Legally, all stakeholders that in some way or another hold some kind of relationship with the RCA (cultural, social, economic, political or research) can participate in this MSF.
Although the RCA has existed since 2002, its first management committee was not created until 2014. Ideally, the RCA management structure should be shared among SERNANP, ECA-Amarakaeri and the management committee. These MSFs have two internal bodies of representation and execution: the assembly of all the members to the committee and its executive commission. The assembly is responsible for approving the reserve's Master Plan and electing the committee's executive commission. It is convened at least once a year, although research revealed that this is not always possible due to the scant funds that are set aside for the functioning of the management committee. The MSF's executive commission, which meets at least twice a year, is in charge of both drafting an annual work plan to support the comanagement activities of the RCA and to evaluate ECA-Amarakaeri's work. To that end, interest groups have been established that function as micro-arenas of work on specific issues and meet apart and more often than the two meetings of the executive commission 7 .
The legal requirement of having a multi-stakeholder management committee complicates the governance practices of the negotiated eco-governmentality between SERNANP and ECA-Amarakaeri, as it allows for the introduction of other stakeholders with different interests that may be contrary to those of indigenous peoples in the RCA's buffer zone or SERNANP's conservationist agenda. Although the RCA's management committee should support the co-management partners and be their liaison arena with the stakeholders external to the reserve's co-management, in practice these actions are carried out directly by SERNANP and ECA-Amarakaeri together with their NGO allies and regional indigenous organizations.
Some contradictions arise between these participatory technologies on two levels. At a first level, in seeking to widen participation, the RCA's government becomes entangled, as it is unclear what space should make the reserve's management and administration decisions. At a second level, there is a "spatializing of social and ethnic differences" (Pinedo 2019: 24) regarding access to the RCA and its resources between the indigenous and non-indigenous populations in its buffer zone. Indeed, the local indigenous Amazonian peoples have territorial, political, and technical representation through their communities and ECA-Amarakaeri, while the nonindigenous population have political participation and representation through local governments. As explained below, these contradictions are resolved through the centralization of governance in SERNANP and ECA-Amarakaeri's comanagement partnership.

NEGOTIATED ECO-GOVERNMENTALITY: PARTICIPATORY GOVERNANCE TECHNOLOGIES AND THEIR EXCLUSIONS
Interviews and continuous engagement with the RCA's management committee by the authors demonstrated that the actors involved in the reserve's governance -indigenous leaders, national and subnational government officials and NGOs -are constantly thinking about and reflecting on how to manage the area given its special status. Given the RCA's status as a Communal Reserve, they consistently noted that it is co-managed by indigenous peoples themselves.
A SERNANP official noted that it is important that ECA-Amarakaeri does not compete with the RCA's management committee, as the ECA holds the co-management of the protected area with SERNANP: "The management committee is a public arena open for the participation of anyone who wants to support the management of the protected natural area (…). In a Communal Reserve [the management committee] is to some degree opaqued, if you want to call it that, by the ECA because many institutions, despite being members of the management committee, are already collaborating and coordinating through the ECA. Thus, the ECA ends up being both, the management and the support to the management of the area".
The interviewee continued reflecting on the work of ECA-Amarakaeri by pointing out that in a Communal Reserve the organization acting as the Executor of the Contract of Administration has specific executive functions, including being part of the area's co-management and administration and thus must constantly coordinate with SERNANP. He said that ECAs are "so active in a Communal Reserve because apart from having a contract with the state, they have a commitment to fulfil [to the communities they represent]. So, the management of Communal Reserves is a little more complex. It is co-management; that's the difference, where the decisions are often made between the state and indigenous peoples". Importantly, ECA-Amarakaeri is seen as a participatory avenue for the local populations and already works closely with the NGOs that work in the area. Thus, the management committee loses importance as a space for coordination and collaboration, which it holds in most of the other protected areas in Peru where there is no ECA. Interviews revealed that the members of the management committee preferred to coordinate their activities directly with the comanagement partners rather than through the RCA's management committee. In fact, research showed that there is little discussion between committee members about any collaborative work or activities. Therefore, the collaboration and coordination with the stakeholders to the RCA fall to the co-management partners rather than to the management committee, whose impact as a space for multi-stakeholder participation is minimal, except for when it has to periodically approve the RCA's Master Plan. The leadership and coordination success of ECA-Amarakaeri is also due to the fact that indigenous leaders themselves are managing the RCA, so that projects are implemented in the native communities. An important indigenous leader in Madre de Dios noted that: "We have worked to design our own strategy to manage an area. This was not the work of the regional government or NGOs. We, the indigenous organizations, changed the whole conversation about Communal Reserves. We influenced the creation of the co-management [model] more than the municipalities, more than the NGOs. We, indigenous organizations, are the ones that have led and have made, let's call it, a road map. It has not been so much the NGOs, but our own work [as part of the co-management model] that has done it".
The indigenous leader expresses that although comanagement is now under the law for Communal Reserves, it is an achievement of the struggles of the regional indigenous movement for a structure that allows indigenous peoples to be part of the formal governance of the RCA with SERNANP. In other words, the indigenous communities of the RCA's buffer zone have succeeded in transforming the various forms of governmentality into one that places indigenous peoples at the center of governance.
For many of the indigenous leaders we interviewed, the latest achievements obtained by the RCA have been thanks to indigenous people's direct participation in the area's management and administration. This sense of indigenous empowerment in the governance of the RCA is also expressed in the following quote from an interview with an NGO representative, which portrays the current type of relationship between ECA-Amarakaeri and SERNANP: "the line of work is very oriented to conserving the area. It is all based on that. But most of SERNANP's activities have to do with meetings with the ECA. That's it. There have been many moments when SERNANP decided about something on its own, mostly on budgets, for example, and when the ECA found out, it drew the line. It told them, 'Why are you making that decision about a given amount of money; are we not a co-management [partnership]?'".
Indigenous people's political agency in the production of the negotiated eco-governmentality allows them to propose certain ways of governing the territory through the participatory technology of co-management. But it also allows them to question, criticize and censor (at least internally) SERNANP's actions, shifting their actions closer to their own interests and objectives.
What are the consequences of centring governance on the indigenous peoples of the RCA? The RCA's structure of governance produces participatory arenas in favour of the indigenous populations, but it excludes a large group of Andean migrants and non-indigenous Amazonian people living and working in the buffer zone from the political and technical governance of the protected area. The governance practices of negotiated eco-governmentality between SERNANP and ECA-Amarakaeri aim to provide benefits as well as control and shape the behaviours of the indigenous Amazonian people living in the RCA's buffer zone. Andean migrants are only engaged with when indigenous communities or the goals of the co-management partnership are threatened by the extractive activities which many of them carry out, and only to expel them from the areas where they have settled.
At the time of research, there were no sustainable development projects implemented by the co-management partners for the benefit of the non-indigenous population of the RCA's buffer zone. There were no participation incentives for this population in the RCA's management committee either nor were they able to lay out their interests or initiatives to the management committee. When they tried to, indigenous organizations blocked or limited their participation. As a nonindigenous person who participated once in the management committee pointed out during an interview: The concentration of governance in the co-management partners and the exclusion of non-indigenous populations from participation in the wider governance of the RCA is due to internal and external pressures from gold mining. On the one hand, there is pressure from the migrant miners on the RCA's buffer zone. These miners are not part of the alliance that has been created among SERNANP, NGOs and ECA-Amarakaeri -the main stakeholders of the negotiated eco-governmentality -as their interests are linked to gold extraction and not to conservation or the sustainable management of the territory. Moreover, the mining groups are represented by the local and regional governments as they reflect their development interests, so they have political access to Madre de Dios' government. On the other hand, as noted earlier, some indigenous communities are also linked to that same extractive activity as they still hold rental agreements with migrant miners, or mine for gold in their own communal territories.
There are divergent objectives, interests and desires among the stakeholders directly or indirectly involved in the RCA's governance. Thus, the practices of eco-governmentality are in dispute and possibly in transformation, revealing ecogovernmentality as a process that needs constant reaffirmation. In this context, the limits of action and influence over the participatory management for an MSF like the RCA's management committee becomes clearer. Rather than being an arena that provides more strength to historically excluded populations, such as indigenous peoples, they can turn into a threat to the territorial rights they have achieved. It could be argued that indigenous Amazonian peoples fought and negotiated for their power current position in the governance of the RCA and are reluctant to let other stakeholders take over the management their territories as has happened throughout their history of interactions with the Peruvian government and other powerful actors such as the Catholic church and multinational extractive companies. Yet, at the same time, although they are central to the governance of the RCA, indigenous peoples do not have legal ownership rights to the RCA, which are held by the Peruvian state and have historically benefited other stakeholders.

CONCLUSIONS
The diverse dynamics and processes that have structured the indigenous and non-indigenous local populations must be considered in order to grasp the interactions of the three governmentalities and their governance practices in the recent history of the RCA. The first is a conservationist governmentality (1990)(1991)(1992)(1993)(1994)(1995)(1996)(1997)(1998)(1999)(2000)(2001)(2002)(2003)(2004)(2005)(2006) promoted by the Peruvian state through the figure of the protected natural area under the modality of the Communal Reserve and its various technologies of fortress conservation in order to preserve biodiversity. The second is an extractive governmentality (2006-2014) that employed neoliberal clientelist governance practices to build debt relationships that facilitated their objective of resource extraction. Both governmentalities intersected when the first generated various mechanisms so that the second could be inserted in the RCA territory.
However, these two forms of governmentality have been disputed and transformed by the struggles of the regional indigenous movement, as it sought self-government and the recognition of indigenous peoples' rights over their territories. The contestation of governmentalities from below has led to the formation of a third and current form of government of the RCA: the negotiated eco-governmentality (2014-present) between ECA-Amarakaeri and SERNANP, with its own green and participatory technologies of governance (comanagement and management committee) that seek to govern the RCA in an articulated way in favour of biodiversity conservation and the sustainable development of indigenous communities. Orihuela (2014: 58) noted that the original enthusiasm of the indigenous organizations for the Communal Reserves faded sometime after they were created because they did not end up being part of their government. This idea coincides with the first period that we have narrated in the RCA's history, in which the original enthusiasm for its creation disappeared before the imposition of the conservationist and extractive governmentalities. However, we propose that it is necessary to have a long-term analytical approach on protected areas in order to understand their change processes.
Like in the Ashaninka Communal Reserve in Peru's central Amazon, where the indigenous population has managed to transform the asymmetries of power (Caruso 2011), this paper has shown that the indigenous peoples of Amarakaeri have politically reappropriated the government of the RCA, which is expressed in the historical process of the construction of the co-management partnership between SERNANP and ECA-Amarakaeri. This process has allowed for the political repositioning of indigenous peoples in the selfgovernment of their territories and in doing so has created a new way of relating to the Peruvian state.
Certainly, during recent years, ECA-Amarakaeri, with the support of SERNANP and allied NGOs, has succeeded in erecting itself as the liaison organization between the stakeholders external to the RCA and the ten native communities in its buffer zone. That indigenous peoples are central to the RCA's management is a product of the struggles for selfdetermination by the regional indigenous movement, as well as of the internal transformations of the Peruvian state (particularly of SERNANP), which, given the pressure from below, sought to create another form of governance for the RCA and distanced itself from the classic forms of green governmentality.
Nonetheless, as West, Igoe and Brockington (2006) point out, indigenous peoples are not necessarily the groups that end up worst in terms of dispossession or economic impoverishment by the creation of a natural protected area. In the RCA, the co-management partners promote their interests and those of the organizations they represent. By not having any representative within this co-management, the different stakeholders that are also present in the RCA -the nonindigenous population and the different local governments -are excluded from informing and participating in its management. Normatively, this could be resolved in the multi-stakeholder arena of the management committee, which should be opened to the participation of the non-indigenous stakeholders who also live in the RCA's buffer zone. However, in reality, the co-management partners (and especially ECA-Amarakaeri) displace the management committee as a space for collaborative planning and coordination, superimposing its functions and blocking the possibility that the latter may be an arena of more effective and wider coordination and collaboration with other civil society stakeholders and with other levels of the Peruvian state. Therefore, two mutually exclusive macro-spaces of governance have been constructed. On the one hand, the RCA's governance is dominated by the indigenous Amazonian population and SERNANP, which excludes the non-indigenous population from the management of the territory. On the other, there is the state's classical arena of political representation and participation -at play through the actions of the local and subnational governments. This space is dominated by and focused on the extractive interests of the non-indigenous populations, and at the same time excludes the indigenous Amazonian people in its jurisdictions, as there is little or no representation of Harakbut, Yine and Matsiguenka peoples in local governments.
This separation of interests and spheres of governance is the final challenge the management committee, an MSF, must deal with to find opportunities to bring different stakeholders together in coordination and collaboration, or even conversation. Ideally, the main objective of Communal Reserves in Peru -as established by the state -is that local populations themselves participate in the management of their territories in partnership with the state. Yet, who does the Peruvian state understand as the population with the right to manage (or co-manage) a protected natural area? If the populations that were historically excluded by that same state have now succeeded in becoming part of the co-management of an area like the RCA, is it necessary to further expand this governance through the management committee?
There is a need for dialogue and negotiation among the different stakeholders related to the RCA to continue supporting its participatory and sustainable governance. However, that dialogue and negotiation does not have to grow out of an official MSF such as a management committee. A different arena of participation is needed, one that is co-created by all the stakeholders involved, in which they mutually recognize not only their differentiated interests but also the importance of collaborating and finding common grounds. This is far from being a participatory utopia, since it is possible that a new participatory sphere brings new problems, tensions and conflicts between the stakeholders. Nevertheless, these reflections are beginning to emerge among the organizations in the Amarakaeri Communal Reserve and will possibly make way for transformations of new (eco) governmentalities in the future.