Governing the ungovernable: contesting and reworking REDD+ in Indonesia

Reducing Emissions from Deforestation and Forest Degradation plus the role of conservation, sustainable forest management, and enhancement of forest carbon stocks in developing countries (REDD+) has rapidly become a dominant approach in mitigating climate change. Building on the Foucauldian governmentality literature and drawing on a case study of Ulu Masen Project in Aceh, Indonesia, this article examines the practices of subject making through which REDD+ seeks to enroll local actors, a research area that remains relatively underexplored. It interrogates the ways in which local actors react, resist or maneuver within these efforts, as they negotiate multiple subject positions. Interviews and focus group discussions combined with an analysis of documents show that the subject making processes proceed at a complex conjuncture constituted and shaped by political, economic and ecological conditions within the context of Aceh. The findings also suggest that the agency of communities in engaging, negotiating and even contesting the REDD+ initiative is closely linked to the history of their prior engagement in conservation and development initiatives. Communities are empowered by their participation in REDD+, although not always in the ways expected by project implementers and conservation and development actors. Furthermore, communities' political agency cannot be understood by simply examining their resistance toward the initiative; these communities have also been skillful in playing multiple roles and negotiating different subjectivities depending on the situations they encounter.


Introduction
In late October 2011, the car I had rented moved slowly as it navigated through the difficult terrain of the dense forest on the way to my research site in Pidie District in the Aceh Province, Indonesia. We were heading towards the resource-rich district, which was also known as the heartland base of support for the Free Aceh Movement (GAM/Gerakan Aceh Merdeka) and the hometown of Teungku Hasan Di Tiro, a founder of GAM. The journey was long, taking around seven hours from the provincial capital city of Banda Aceh. It was nearly dusk when I heard a loud sound coming from the back of our car. The source of the noise was unclear, whether it was from a flat tire or something else. "You know, that was just a regular shooting, and if we stopped here, some thugs could just rob us. That happens a lot in this area", explained the driver calmly.
Notwithstanding the signing of a peace agreement between GAM and the Government of Indonesia in 2005, Aceh still struggles to maintain political stability. Given the endurance of this post-conflict situation, how does a green project like Reducing Emissions from Deforestation and Forest Degradation (REDD+) manage to establish itself at the village level? What does it mean for the local population? How do local communities respond to the project?
Since its inception in 2005, REDD+ has gradually become a dominant approach in global climate mitigation efforts. The rise of REDD+ has prompted growing academic debates. Informed by Foucault's governmentality framework, a growing number of political ecology scholars analyze REDD+ as a new form of environmental rule that govern human-environment relationships and examine strategies and tactics to govern neoliberal subjects and the disciplinary effects of the initiative (Bumpus and Liverman 2011;Lövbrand and Stripple 2011;McGregor et al. 2015). More recent studies productively use this framework to examine resistance and subjectivities in REDD+ (Asiyanbi et al. 2019;Astuti and McGregor 2017;Benjaminsen 2014). This article contributes to and advances this body of literature by examining the production of new subjectivities through REDD+ and highlighting the agency of local actors, an area which remains relatively underexplored (Fletcher 2017;Singh 2013). In particular, it follows a recent call (Asiyanbi et al. 2019) to unpack the multiplicity and even contradictory forms of local engagement in REDD+ and their dynamics over time to illuminate the emancipatory potential of such engagements.
Drawing on an ethnographic research of the first REDD+ project in Indonesia, the Ulu Masen Project, this article interrogates the processes to produce new subjectivities and how communities have reacted, resisted and creatively accommodated strategies implemented by the project. In particular, I examine three governmental strategies to produce new subjectivities implemented in the project: first, awareness raising activities on conservation and REDD+ to enhance communities' commitment to support the carbon emission reduction goals; second, increasing forest surveillance to reduce illegal logging and wildlife trafficking through recruiting and training forest rangers and community rangers; finally, engaging communities in spatial land use planning and participatory mapping to clarify forest tenure. Moreover, I investigate the ways in which subject-making processes intermingle and are shaped by complex political, economic and ecological conditions. In doing so, the research was carried out within two village sites that reflect rather contrasting economic, social and ecological conditions.
In this article, I argue that the communities' diverse reactions are influenced by the history of their prior engagement with conservation and development initiatives, their experience of land dispossession, and the prolonged civil conflicts in Aceh, Indonesia. Communities' political agency cannot be understood by simply examining their resistance toward the initiative. As the findings show, communities in the two case studies have also been skillful in playing multiple roles and negotiating different subjectivities depending on the situations that they encounter. They are empowered by their participation in REDD+, although not always in the ways expected by project implementers and conservation and development actors.
The next section outlines scholarly debates on REDD+ governmentalities and agency and discusses how a local practice of muslihat (Siapno 2002) could help us to understand Acehnese political agency and illuminate local responses to the initiative. This is followed by a brief description of research methods, broader context on REDD+ in Indonesia, and case studies. This sets the stage for elaborating and analyzing three key governmental strategies as described above. The article concludes with the broader relevance of the findings.

REDD subjectivity and political agency
A growing body of scholarly works has interrogated REDD+ using critical literature on neoliberal environmental governance. REDD+ has been described as part of an increasing trend toward neoliberalization of environmental governance that promotes neoliberal strategies such as deregulation, privatization, decentralization and market reliance to govern the environment (Bumpus and Liverman 2011;McAfee 2016;McGregor et al. 2015). Larner (2003) posits that neoliberalism is not merely a set of policies but is instead a new form of political-economic governance that shapes individuals and institutions to comply with market norms. Thus, neoliberal policies can form self-governing and autonomous neoliberal subjects. This proposition has prompted a growing literature that employs the Foucauldian concept of "governmentality" (Foucault 1991(Foucault , 1995, or sometimes reframed as "environmentality" (Agarwal 2005a(Agarwal , 2005b or "green governmentality" (Rutheford 2007), to understand how the creation of market incentives could alter human behavior toward the environment.
Foucault defines governmentality as "conduct of conduct" that embraces the exercise of discipline over bodies, 'policed' supervision of the inhabitants of the sovereign territory, the bio-political regulation of 'species life' of a population, and self-formation through ethical care of the self (Foucault 1991(Foucault , 1995Huxley, 2007). He argues that power is not repressive but 'productive' through social construction of subjects rendering the governed governable. It produces subjects, forms their character, and 'normalizes' them, making them willing to, and capable of, complying with norms of various forms of propriety (Foucault 1980). Drawing on this framework, Agarwal (2005aAgarwal ( , 2005b interrogates strategies used by the government to create how people care about environment. However, Fletcher (2017) suggests that Agarwal's conceptualization is merely one of a number of disciplinary modes of environmentality, which include also sovereign environmentality (e.g. fortress conservation, command and control conservation) and governmentality 'according to truth' (e.g. traditional ecological knowledge). He also contends Agarwal's approach, that overemphasizes top down manipulation exercised by the state through "intimate government", could overlook the ability of local communities to self-mobilize in managing natural resources. Accordingly, Fletcher (2017) Journal of Political Ecology Vol. 27, 2020 459 calls for studies to examine how people self-mobilize and exercise their agency in environmental governance. Indeed, Foucault suggests the inherent possibilities of resistance closely linked to power as he notes, "Where there is power, there is resistance, and yet or rather consequently, this resistance is never in a position of exteriority in relation to power" (Foucault 1980: 95). Earlier studies have documented the ways communities contest development and conservation interventions (Peluso 1992;Scott 1985Scott , 1992. In his landmark study in Kedah, Malaysia, for example, Scott (1985) shows how peasants' resistance toward the Green Revolution manifests in mundane activities such as food dragging, dissimulation, desertion, and false compliance, which he coins as "everyday forms of resistance" (Scott 1985: xvi). More recent studies have emphasized a multiplicity of local forms of political agency which include not only resistance but also compromise and creative accommodation of such interventions (Asiyanbi et al. 2019;Li 2007;Osborne and Shapiro-Garza 2018;Wilshusen 2010). Indeed, many governmental schemes entail not only contestation and negotiation but also compromise among actors that ultimately transform peoples' identities, spatial practices, and the interventions themselves (Li 2007). Li (2007) suggests that to understand how resistance emerge we should closely examine diverse positions that people occupy, and the various forms of power that they encounter. To summarize, this article productively combines multiple governmentality frameworks and literature on the technology of resistance and political agency to provide a nuanced understanding of diverse reactions of Acehnese toward REDD+.
To illuminate Acehnese political agency, it is important to understand Aceh's historical and political backdrop that shapes the ways people react, respond or resist to particular development interventions. Aceh Province, located in the northern part of Sumatra, has been a site of civil conflicts between the GAM and the Indonesian military forces. A separatist movement emerged in the early 1950s as a response to the failure of the newly independent Indonesian Republic to fulfill its promise to grant Aceh a special autonomy status to govern its region based on Islamic Law (McCarthy 2007). The movement gained more traction in the early 1970s after the discovery of a liquefied natural gas field in Aceh by the American oil company, Mobil, which led to a forced resettlement of Acehnese, the arrival of non-Acehnese skilled workers, and an increasing military presence and violence in the area. The military repression lasted for over three decades. After the tsunami disaster swept the region in 2004, the GAM and the government of Indonesia signed a peace agreement in 2005 ending decades of civil conflict.
After decades of the Indonesian military's total war policy where frontal resistance against them was too risky, the Acehnese become skillful at playing dual performative roles of loyal/disloyal citizens, while working around repressive structures of power and turning them into productive mechanisms to serve their interests (Siapno 2002). This strategy has been used in a range of areas, such as silently working within the government or a multinational business and publicly vowing loyalty to Pancasila 2 state ideology and Indonesian citizenship, while at the same time supporting GAM. In this context, Siapno (2002) argues that Acehnese political agency cannot be understood as a "pure, uncontaminated opposition", but is in many ways produced in and through mechanisms put in place by the occupying culture. To understand Acehnese political agency, Siapno productively uses a local practice known as muslihat. Tiwon (1985as cited in Siapno 2002 describes muslihat as follows: The word "muslihat" is not easily translated into the western scheme of values without using words or phrases that carry the unfortunate connotation of dishonesty. In the Malay world, however, muslihat can be construed to mean "using indirect means to attain the goal." It is a neutral quality available to both hero and villain. The hero, always triumphant, is he who makes the most efficient use of muslihat. This can be done by hiding one's physical presence, by disguising one's true identity, by indirect movement, or by the indirect use of language.

Journal of Political Ecology
Vol. 27, 2020 460 In Aceh, the words muslihat, tipee-muslihat or peunget are known in "Acehnese tales and local practices as a positive strategy used by small animals or common people in forging alternative power relations vis a vis bigger animals or rulers" (Siapno 2002: 11). In the wider context in Indonesia, such seemingly duplicitous behavior could be considered negative behavior, as it is sometimes referred to as licik or sly and tricky. Indeed, when I conducted research in Aceh, some people in Jakarta warned me about Acehnese tipu muslihat. Siapno (2002) further argues that there is no clear line between power and resistance. Thus, we should consider multiple configurations of power relations that are intertwined with each other at different levels and "…seriously examine, not moralistically dismiss, co-optation, collusion and contamination, forms of political contestation that are not pure opposition…" (p. 16). Using these analytical tools, this paper examines the multiple responses expressed by Acehnese in responding to REDD+ interventions.

Methods
This research is based on fieldwork conducted over approximately 13 months during several visits in 2010-2013, with follow-up visits in 2017-2018. In collecting the data, I carried out multi-site ethnography in Jakarta, Banda Aceh (the capital city of Aceh Province), and two selected REDD+ pilot villages. Semistructured interviews were conducted with 65 key informants representing government agencies at the national and provincial levels, environmental non-government organizations (NGOs), indigenous activists, donor agencies and communities in selected village sites (see Figure 1). I also conducted six focus group discussions with men's and women's groups representing different social-economic status and village leaders, two focus group discussions with community and forest rangers, and household surveys in the selected villages.
Two village sites selected for the research are pilot sites of the Ulu Masen Project with rather contrasting conditions: one reflecting the REDD+ conservation goal (an area with intact forest needing protection) (Village 2), and the other reflecting its sustainable development goal (an area with scattered forest cover needing sustainable development (Village 1). In the villages, household surveys were conducted to examine household income levels, asset distribution including land ownership, and data on forest-based livelihood strategies. I also explored different uses of forest resources by men and women, the amount and value of forest resources harvested and used, and communities' understanding of and participation in REDD+ project activities. To deepen the information and triangulate the data, I carried out in-depth interviews and focus group discussions on the history and pattern of forest resource access and use; concerns and issues being raised and negotiated by the villagers; and the villagers' strategies to respond and negotiate terms concerning the REDD+ project.

Broader context in Indonesia
Indonesia is an important country in the development of REDD+ initiatives. It is the third largest greenhouse gasses emitter due to deforestation and forest degradation and the second largest recipient of REDD+ finance in the world (Wolosin et al. 2016). The country has aggressively pursued REDD+ objectives, especially to fulfill its intended nationally determined contribution (INDC) to voluntarily reduce its carbon emissions 29 percent by 2030 and up to 41 percent with international support (GoI 2017).
The REDD+ initiatives are implemented in the context of deep structural change in Indonesia through the enactment of decentralization policies, which delegate substantial authority on forest governance to provincial and district governments. The initiatives are also implemented in the context of unclear forest tenure arrangements. Notwithstanding recent social forestry policies that aim to clarify forest tenure arrangements and address long standing forest tenure conflict, the forest areas remains heavily controlled by the state, and approximately 85 per cent of forest areas are still administered by the state (RRI 2013). This has left villagers living in 31,957 villages located in and around the areas classified as forest with insecure access to forests and livelihood. Hence, forest tenure conflicts are widespread.

Journal of Political Ecology
Vol. 27, 2020 461 Despite regulatory complexity and unclear institutional arrangements, many multilateral and international agencies have invested in readiness and demonstration activities in Indonesia for nearly a decade, which manifested in a rapid proliferation of REDD+ demonstration activities from 2009-2014. In 2014, a national REDD+ agency, an ad-hoc ministerial agency, was established to oversee the implementation. However, the agency was short-lived, having been dissolved in early 2015. Subsequently, policy initiatives and demonstration activities have slowed down. Recent data suggests there are 86 REDD+ projects in the country, yet only 20 of them are considered active (Simonet et al. 2016). Although REDD+ initiatives are dwindling, the government of Indonesia continues to actively seek avenues to sustain the initiative, including accessing climate finance. 3 Figure 1: Research Sites in Aceh, Indonesia.

Ulu Masen REDD+
In many respects, the Ulu Masen Project is well suited to study how neoliberalism-inspired environmental initiatives and environmental subject-making are manifested in an area with a complex historical and political backdrop. It is implemented in Aceh Province, which was once known as a resourcerich and rebellious region due to prolonged civil conflicts. With the cessation of conflict, Aceh was granted a special autonomy status through the stipulation of the Law on Governing Aceh (LOGA) in 2006, which delegated significant authority to the Acehnese government including over forest governance (Setyowati 2020). When Irwandi Yusuf, a former GAM combatant and environmental activist, was elected Governor of Aceh in 2007, he launched Aceh Green Vision with the World Bank's support.
The Aceh Green Vision is an ambitious vision to improve Aceh's economic situation. It integrates themes of climate change via renewable energy and land use management, community development, commerce, and conservation. The Ulu Masen Project was established as a part of this vision to solve two main challenges for Aceh: ensuring sustainable forest management, while at the same time generating a source of financial revenue for development (PDD 2007;Setyowati, 2014). In developing the project, the Government of Aceh collaborated with Fauna and Flora International (FFI) and a carbon broker company, Carbon Conservation (CC), herewith referred to as 'the project proponents.' The Government of Aceh implements the project with FFI providing technical assistance and facilitating capacity building and community engagement activities, and Carbon Conservation acts as the sole seller of the carbon credits generated from the project.
The project covers 738,000 ha of forests and potentially affects the livelihoods of approximately 130,000 rural Acehnese living in and in adjacent to the Ulu Masen area, shown in Figure 1 above. It aims to reduce carbon emissions through reducing the deforestation rate by 85 percent in 30 years, and to make carbon financing a means to improve community livelihoods and provide incentives to protect forests (Clarke 2010). To ensure the communities conform to the emission reduction goals, the project proponents carry out capacity building activities, enhance law enforcement, develop agreements with communities, create employment for local people, and conduct forest monitoring and patrols (PDD 2007). The proponents also acknowledge the role of mukim, a traditional governance/administrative unit that was originally part of the Aceh Sultanate system (17 th century), in forest resources governance through engaging mukim in participatory land use planning and forest management at the local level.
The initial project development was rapid, and involved the elite inner circle of the Governor's office in regular communication with FFI and Carbon Conservation, partly due to the desire to build on momentum created during the United Nations Framework Convention on Climate Change (UNFCCC) meeting in Bali in December 2007. As shown in Figure 2 below, most Ulu Masen project activities took place from 2007-2012 but dwindled due to the failure of the carbon market to flourish over the course of the project and the provincial administration change after Governor Yusuf failed to secure his second term in 2012. He was reelected as Governor of Aceh in 2017, reviving hopes for many of the continuity of the Aceh Green Vision. 4 However, in April 2019, Governor Yusuf was sentenced to seven years in prison for corruption. The future of Aceh Green Vision remains to be seen (see Figure 2).

Case studies
Village 1 As shown in Figure 1 above, village 1 is administratively under Aceh Besar District and located in close proximity to the province capital. As Table 1 shows, the village's ecological condition is relatively dry, often resulting in famine and drought during the dry season. Due to limited water availability and poor irrigation systems, villagers cultivate dryland rice once a year and supplement their livelihood with various commercial crops such as corn, candlenut and more recently oil palm. Logging activities and land pressures in this village are considerably lower than other communities in the Ulu Masen area, due to the lower Journal of Political Ecology Vol. 27, 2020 463 availability of high value commercial timber, poor vehicle access and less fertile farmlands. During the conflict period, Village 1 was labelled a 'red area' as the rebel combatants often passed through the village territory to flee and find sanctuary in the Ulu Masen forests. As a consequence, villagers were subjected to frequent military surveillance and their mobility was rather limited. Village 1 has been subjected to various development interventions. 5 In the late 1970s, the government resettled local populations who lived scattered in the forests to accessible places to ease the delivery of 'development' programs, removing them from the forest area which was later classified as a state forest reserve. 6 The villagers also have relatively high exposure to environmental issues as the village has been a site of various conservation projects since the mid 1990s. Exposure to the state's coercive power that displaced villagers and the subsequent engagement in development projects have increased communities' social cohesiveness, and they act collectively in response to development interventions. These engagements have equipped the villagers with new skills and the language needed to demand better services from the government and to be strategic in accessing various forms of development assistance. For example, the mukim leader describes how the villagers have made their best effort to improve village conditions by proving that they are obedient citizens by participating in the state's development projects. 7 They have also established a watershed forum to protect the watershed area covering a 2,500 ha forest area.

Village 2
Village 2 is administratively under Pidie District, a heartland base of GAM located in the west part of Aceh Province, as shown in Figure 1 above. The village has fertile agricultural land and is rich in minerals, especially gold. This led to widespread artisanal mining in the area, and as the village profile in Table 1 below shows, members of the community have long been integrated into the market through trading rice, coffee and more recently oil palm. Forest resources have significantly contributed to local livelihoods. They benefitted in the 1980s from the boom of agarwood (Aquilaria trees that become infected with an aromatic mold) locally known as gaharu, and various high value hardwood timber (PDD 2007). During the 1980s and 1990s, several 5 FGD with men (15/10/2011), women (14/10/2011) and village leaders (13/10/2011). 6 Resettlement projects started to be implemented throughout Indonesia during Suharto's administration by the Department of Social Affairs, usually targeted to hill farmers labeled as masyarakat terasing (estranged/isolated people) to help them to be "ordinary average Indonesians" (Li 2007 logging companies were active in the district, and villagers frequently found work as low paid laborers (EoA 2009). Illegal logging was also rampant in this area. The availability of high valued natural resources and fertile lands, high market demand and easier vehicle access have attracted newcomers, middlemen and other actors seeking to acquire local farmland and participate in the commercialization of local products, hence increasing pressure on land and forest resources. The village was a 'no go' area during the conflict due to its close proximity to the GAM headquarters location. The long period of conflict had profound impacts for the villagers' lives and livelihoods. In addition to frequent torture and killings, the villagers also lost their source of livelihood. During the height of conflicts, villagers' mobility was strictly limited by the military. They could only move around a 100 meter range from their homes; otherwise, they would be suspected GAM combatants. 8 As a result, large agricultural areas were abandoned. Despite its rich natural resources, the economic status of the community is considerably lower than Village 1 due to the long period of intense conflict. Military repression caused the relative isolation of Village 2 from various development and conservation interventions. Only after the signing of the peace agreement in 2005 was a conservation project implemented in the area by FFI to monitor and mitigate humanwildlife conflicts.
In sum, both villages have contrasting economic, social, and ecological conditions and histories that shape their responses toward REDD+ interventions (see Table 1). Regardless of their differences, they receive similar interventions from the project, as described in the following section.

Social and Economic Context
 67 Households  Acehnese majority  Mixed shifting cultivation and permanent agriculture  Commercial rice production (once a year), commercial crops (e.g. corn, legumes, coconut, areca nut/betel nut)  Poor vehicle access  264 households  Acehnese majority, around 50 percent moved in 1980s to work in logging companies or transmigration project  Mixed shifting cultivation and permanent agriculture  Commercial rice production (twice a year), commercial crops (e.g. cocoa, coffee, areca nut/betel nut)  Better vehicle access

Ecological Condition
 Less fertile soil  Dry area, prone to drought and famine  Lower availability of timber and other commercial forest products  Fertile soil  Wet area  High availability of high valued timber, gold, wildlife products (e.g. ivory)

Access to Forest
 Non-timber forest products  Timber, mostly for household needs  Non-timber forest products  Timber, gold

History of Environmental Encounters
Involvement with environmental activism since the mid-1990s More recent encounters with environmental NGOs

Findings and discussion: inside the production of REDD+ subjectivity
The production of model communities is central in REDD+, in which the proponents produce ideal assumptions and expectations about communities (Skutsch and Turnhout 2018). Communities are expected to act as 'rational ecosystem service providers' who understand the monetary value of their services and are willing to give up forest-based livelihoods in exchange for performance-based payments. By forming communities as rational ecosystem providers, the REDD+ interventions "harness and direct existing community dynamics", including their willingness to participate in markets (Milne and Adams 2012: 153).
In the same vein, the Ulu Masen Project proponents assume that the availability of carbon financing and potential employment generated from the project (e.g. the recruitment of community rangers) will entice the communities to support forest conservation and forgo opportunities from logging activities (Project Design Document/PDD 2007). The proponents expect communities to engage actively in the project, understand its conservation objectives, seek alternative 'environmentally friendly' commercial activities, and apply entrepreneurial skills to map market demands. Technical assistance and funding will be provided for communities "that commit to protecting forest" (PDD 2007: 24). They also seek to render the communities responsible for conserving the forests and to find alternative sustainable livelihoods. Such characteristics are promoted through setting conditions that encourage the communities to 'conduct their conduct' to support the carbon emissions reduction goal. This vision is eloquently described by the director of Carbon Conservation: 9 Ultimately, we have been trying to think about ways where it is not hand-outs. So, the money will get to the people by the forest, but we feel that it's actually damaging to give them money without recourse, without strings, because if it seems as though it's money for free then people won't appreciate it, they won't invest it, they won't learn the skills required to actually create a meaningful life.
The statement echoes discourses employed in neoliberal social policies that position the community as a self-governing formation that can direct local livelihoods -both their successes and their failures (Castree 2010). Rose (1999) calls this practice "government through community" insofar as communities are encouraged to be responsible for their own improvement by participating in markets, learning how to be competitive, and making rational choices. In this manner, the role of government is transformed, not to plan but to "enable, animate and facilitate" community action (Li 2007: 234). To analyze the production of environmental subjects, in the following sections I examine three governmental technologies implemented in the Ulu Masen project.

Educating communities on REDD+
The production of new knowledge is essential to inform the regulation and the shaping of practices and human subjectivities toward the environment (Agarwal 2005b). As such, ensuring that local communities become REDD+ literate is crucial to instill new behavior that conforms to climate mitigation goals. In the Ulu Masen Project, capacity-building and awareness-raising activities involving community representatives are carried out by distributing information materials and holding informational workshops on REDD+ and climate change. These activities aim to enhance community understanding on the forest's role in carbon sequestration, to ultimately shape their motivation to better conserve the forest.
Ideally, the REDD+ consultation process occurs prior to project implementation. Following Free, Prior and Informed Consent (FPIC) principles, the consent to engage in the REDD+ project should be sought after the communities have full information about the scope and impact of any proposed activities Mahanty and McDermott 2013). However, the consultation in the Ulu Masen REDD+ occurred much later in response to increasing pressure from civil society, and the requirements of the project's validation process for the Voluntary Carbon Standard (VCS). 10 The proponents assume primae facie that local communities have agreed with the project design and will be consulted as the project progresses (Clarke 2010).

Journal of Political Ecology
Vol. 27, 2020 466 The approaches taken in informational workshops proved to be problematic. First, during the workshops, the proponents only invited village and mukim representatives, with the expectation that the information shared would trickle down to other villagers in the Ulu Masen area. 11 This stems from a set of assumptions that resemble the 'myths of community' as described by Agarwal and Gibson (1999) -small localized homogenous social units that share common norms. Similar assumptions have also been seen in REDD+ projects elsewhere (Skutsch and Turnhout 2018), leading to a failure to recognize power relations embedded in the culture and social structures of local 'communities.' This could result in the exclusion and marginalization of certain groups, such as women, and potentially lead to elite capture.
Second, the proponents tended to simplify the information about REDD+ and downplay the prospect of earning fast cash from the project. While they described how REDD+ is connected to global efforts to mitigate climate change, the potential benefits and risks of the initiative were rarely explained. The information delivered generally highlights citizens' responsibility to conserve the forest. A government officer argued that such an approach is taken to avoid unrealistic community expectations about the project due to looming uncertainties, particularly with regard to when the project's financial benefit would accrue and how it would be distributed. 12 The proponents often provided oversimplified information to villagers. For instance, during a village meeting, a project field staff described: 13 Simply put, REDD+ is [selling] the 'wind' program. We breathe in oxygen and breathe out carbon dioxide. Carbon dioxide is poisonous and can absorbed by the trees, which then produce oxygen. So the logic is that the developed countries have a limited amount of trees and their carbon dioxide circulates around the globe…Therefore, they need to pay us because we have many trees.
As a result, most villagers interviewed labeled REDD+ as a project to "sell the wind" (jualan angin), a metaphor used by the proponents to simplify the abstract and complex nature of the initiative. The metaphor is rooted in the cultural perception on the wind in Aceh that the villagers consider similar to the nature of REDD+: invisible, abstract, and highly unpredictable.
Notwithstanding the proponents' efforts to conceal detailed information about potential monetary benefits, communities could access information through various means such as newspapers and television. Exposure to such information without clear transparency from the project generated expectations and speculation about the potential financial stream from the project amongst the communities. Indeed, expectation plays a performative role to mobilize actors and resources amid uncertainties in REDD+ projects (Marasella et al. 2018). In Village 1, despite initial confusion during the early stage of project implementation, the village leaders became eloquent in expressing their expectation of monetary compensation and other project benefits, which was partly due to their broader exposure to the issues. The village, considered by the project proponents as an exemplary site for the Ulu Masen Project, has regularly received national and international attention and visits from those who want to learn about REDD+. Such interaction has enabled the villagers to access further information about REDD+ and build broader networks.
The villagers have quickly learned compelling frames to articulate their expectations. Despite the differences in the understanding of REDD+ and expectations on forms of benefits, most villagers expected that it would bring certain kinds of development (pembangunan) which might manifest as governmental provision of public services such as infrastructure, schools, and income-generating activities. However, some

Journal of Political Ecology
Vol. 27, 2020 467 villagers felt that they preferred monetary compensation from REDD+. As elaborated by a village leader in 2011: 14 We are expecting fair compensation based on efforts that we have contributed to conserve the forest. We have been working voluntarily to protect the environment, but we are concerned whether or not we can continue protecting the forest on a voluntary basis. We hope that all of our voluntary work can be fairly compensated, which means that the [monetary] compensation should cover our daily needs so that we can just focus our energy to protect the forest. Maybe some people say that [REDD+] compensation will be used for improving the quality of village infrastructure…but we prefer to have monetary compensation.
In Village 1, the villagers' engagement with environmentalism and relatively high familiarity with environmental discourses enable them to internalize conservation values and articulate their commitment to conservation. Their experience of drought and difficult access to water seems to provide a crucial base for them to be more willing subjects for the adoption of conservation values. Introducing conservation values through the water issue, which is deemed crucial locally, has helped the villagers to make sense of their situation and enhance their willingness to conserve the forest resources. Despite their active roles in conservation activities and familiarity with environmental narratives, Village 1 residents refused to be restricted from accessing various forest resources. Their engagement in the conservation initiative and REDD+ was strongly motivated by their desire to obtain rights over mukim forest and have their water catchment area legally recognized by the state. As described by a mukim leader, 15 Our focus right now is protecting 2,500 hectares of water catchment area. It will be helpful to get help to map the territorial boundaries of our village so that we know exactly our territory, and in the future, we can sell it [carbon credit] for REDD+…If we don't delineate our territory boundaries, it could be easily encroached by outsiders. We need to get our rights to manage the area recognized by the government…to ensure that people will respect our rights.
The experience of land dispossession by resettlement projects decades earlier, and subsequent unfulfilled promises of land certificates from a failed oil palm plantation establishment, have generated villagers' awareness of the importance of secure tenure over their forestlands. Through their engagements in REDD+ activities, villagers became familiar with the discourse of conservation embedded in REDD+ documents as well as customary (adat) 16 rights and they were increasingly outspoken in expressing their opinions during various consultation meetings. When discussing REDD+, I heard some villagers talk more about the discourse of rights than conservation. They saw REDD+ as an avenue to get their rights over customary forest recognized and to obtain compensation that could improve local livelihoods. They had a different vision than the proponents on how the customary forest should be managed. The mukim leader indicated that the villagers wanted to maintain some access to timber products, while only limiting access to these resources in areas considered pivotal for conservation, whereas the proponents sought a general ban on timber harvesting.
In contrast, the residents of Village 2 suffered enormous loss of life and livelihood during the years of conflict, and have struggled to get back to normal ever since. With the availability of highly marketable natural resources, such as gold and timber, they turned to logging and artisanal gold mining to obtain fast

Journal of Political Ecology
Vol. 27, 2020 468 cash. This situation, coupled with easier access to markets and elevated land pressures, challenged the efforts of outside agencies to produce environment subjects. In Village 2, villagers had less exposure to, and a more limited understanding of, REDD+. Only a few villagers, especially village and mukim leaders and community forest rangers, had even heard about the initiative. Furthermore, most of them were more interested in cultivating commercial crops while expecting the development agents to provide technical and financial assistance for their endeavors. Even the mukim leader, who frequently participated in meetings on REDD+ both at district and provincial levels, had a somewhat vague understanding about the initiative. In 2011, he was much more enthusiastic about talking about how mukim will utilize the carbon fund: If we receive the carbon fund, we want to allocate it for local communities; we will train them how to log the trees properly, how to manage forest properly. We could cultivate commercial trees, like sengon [Albizia chinensis]. We will make a local regulation though which every household will have to plant ten trees for one tree they cut for their household needs. 17 The mukim leader's statement suggests that he, indeed, did not envision REDD+ as something that could restrict access to timber. Rather, he saw it as a program promising economic benefits from forestlands through the planting of commercial trees. Despite a comprehensive conservation campaign program aiming at getting villagers to embrace forest conservation and abandon shifting cultivation, many villagers continue their cultivation practices and refuse to comply with conservation regulations. For the inhabitants of Village 2, environmentalism simply remained irrelevant to their situation. REDD+ also seemed to be a distant concept for most of them.

Surveillance over forest and people
The creation of disciplined villagers who are willing to internalize conservation values and practices is central to the success of REDD+. The project attempted to discipline people to comply with emission reduction goals through intensifying forest surveillance and policing activities. The proponents hired nearly 2,300 forest rangers as contract staff stationed across Ulu Masen forests, and recruited and trained community members as community rangers to collaborate with state agents for policing forest areas. This approach is closely aligned with growing securitization of forest landscapes to ensure REDD+ permanence and to combat activities that are considered illegal (Asiyanbi 2016;Cavanagh et al. 2015).
Most forest and community rangers were former GAM combatants, illegal loggers, and poachers. The project proponents considered that engaging those perceived as 'threats' to the Ulu Masen forest was pivotal to "reduce pressure on the forest" by providing an alternative livelihood for those likely to place the protected forest at risk (FFI 2011). The hiring of forest rangers not only served the proponents' interest in securing the forest areas, but also provided a 'quick fix' for unemployment problems experienced by former GAM combatants (PDD 2007). The proponents also assumed that involving the former combatants in REDD+ would reduce the likelihood of them returning to the conflict and prevent them from turning to illegal logging activities, hence fostering successful reconciliation processes in the region (EoA 2009).
Community rangers were recruited from communities living in and around the Ulu Masen area to allow them to take 'ownership' and 'feel responsible' for securing forest areas (FFI 2011). This method is what Neumann (2001) calls "self-surveillance" through which the villagers became involved in conservation activities by delegating the burden of monitoring and securing the forest area. The community rangersmostly of them ex-combatants, illegal loggers and poachers are expected to become the project's 'eyes and ears' on the ground who closely monitor forests and people. A key informant described: 18

Journal of Political Ecology
Vol. 27, 2020 469 It is amazing to see that among these hardened men, these guys are going from outcasts and criminals to heroes. They are becoming our eyes and ears. They let us know what is going on in very remote parts of the jungle, places that are normally very difficult to monitor.
Community rangers are generally more active than forest rangers in taking active measures to mitigate human-wildlife conflicts. In Village 1, for instance, community rangers' credibility in mitigating humanwildlife conflict seemed to be widely acknowledged beyond the village. Oftentimes, they had to travel beyond their assigned territory in response to communities' requests, especially to mitigate human-wild elephant conflicts that are rampant in this area. 19 In Village 2, however, the forest and community rangers found it much harder to curtail rampant illegal logging, which was deeply entrenched in networks involving various actors such as military officers, the police, GAM combatants, and local communities. Many GAM combatants have intricate network ties with logging business actors who often supplied logistical support to GAM. GAM's involvement in illegal logging is an open secret in Aceh (McCarthy 2006). Those networks have successfully adapted to the changing political situation in Aceh. Furthermore, illegal logging has been an important part of the local economy around the Ulu Masen area for decades (PDD 2007). It is estimated that around 2-3,000 villagers in the Ulu Masen area were involved in illegal logging in the 2000s (FFI as cited in PDD 2007). The number might be higher considering the scale of illegal logging operations. Illegal loggers often collaborate with forest rangers to avoid getting arrested. A former illegal logger turned community ranger recalled that a close collaboration with forest rangers is crucial in carrying out illicit activities: They [forest rangers] should actually patrol and catch illegal loggers. But we never met them. Some of them in fact were involved in illegal logging. [When I was still logging timber illegally], I used to collaborate with them. They would let us know if the inspection would happen so we could stop the operation and avoid being arrested. They would say, "Do not bring the timbers tomorrow and hide them." From them we had information on the best time to log the trees. 20 While some forest rangers took on their new roles and responsibilities with a sense of pride and enjoy an elevated social status, many face a dilemma in performing their duties. Their private interests are sometimes incompatible with the roles they have been hired to perform. With limited income as forest rangers, they could hardly make ends meet. Some turn to illegal logging activities and are subsequently fired. A forest ranger also suggested that it is often difficult for them to directly oppose logging players who once supported them during the conflict. Moreover, during a group interview, the rangers explained difficulties in gaining the trust of communities considering their past identities. In some cases, villagers resented the rangers' attempts to educate the broader public about conservation and to prohibit the villagers from conducting illicit activities in the forest area, something that the rangers themselves did in the past. To summarize, the effort to discipline people through hiring forest and community rangers proved to be challenging due to their intricate relationship with illegal logging networks, which contradicted the role that they were expected to perform. This, in turn, has hampered their ability to discipline and police the conduct of communities in support of REDD+ goals.

Participatory mapping and land use planning
Spatial ordering and classification, such as through participatory mapping and land use planning, are crucial governing mechanisms because they allow increased control over space, resources and the movement of people (Scott 1998). In the context of REDD+, creating maps serves as a tool to regulate activities in Journal of Political Ecology Vol. 27, 2020 470 different forest zones. It can also support forest tenure clarification, an essential element to ensure successful REDD+ project implementation, because it allows the proponents to decide who gets project benefits, devises benefit sharing mechanisms, determines incentives for behavior change, and ensures long term security for the project (Corbera and Schroeder 2011;Larson et al. 2013;Sunderlin et al. 2013). Further, the mapping exercise in REDD+ is crucial to frame the space in order to stabilize the 'forest-as-carbon sink' and create a sphere of action and intelligibility (Lansing 2012). Like other forests in Indonesia, the Ulu Masen forests are classified as state forests regardless of de facto customary forest tenure arrangements that continue to be practiced by local communities. Most villagers do not have formal property rights over forests. However, some of them obtain locally-recognized documents signed by the village head or the sub-district head (Dunlop 2009). This situation is complicated by the fact that after the tsunami in 2004, many communities who lost their lands were evacuated to other areas, including areas in close proximity to or inside the forests (Fan 2006). In some cases, tsunami survivors were relocated to the Ulu Masen forest area, which fueled tension with local communities. 21 An integral part of the Ulu Masen Project involved mukim based participatory mapping and spatial planning processes. The mukim traditionally plays an important role in governing natural resources and territory, usually encompassing several villages (Syarif 2007). In the early 1970s, however, the Indonesian government implemented the Village Governance Act (Law 5/1979) to better govern the population by creating a standardized government structure in the form of village (desa) administration. In Aceh, the policy significantly erodes the authority and legitimacy of mukim and other traditional institutions, including local forest governance. However, the Law of Governing Aceh (LOGA) of 2016 formally recognizes mukim as a government structure that covers several villages. To strengthen mukim's role in forest management, the Ulu Masen project proponents carried out mukim based participation to enhance community awareness of mukim institutions and their role in forest management, engaging them in developing mukim land use maps that would support local conservation and development goals. The map that included mukim based land-use plans was incorporated into the district spatial planning. It was hoped that this approach would resolve widespread forest tenure conflicts in the Ulu Masen area (PDD 2007;FFI 2011).
The mapping exercises took place in a series of phases, beginning with community workshops to increase awareness of mukim and indigenous rights, developing sketch maps of current land use, and discussing how the community envisioned future land use aligned with the REDD+ goals. 22 It is followed by ground checks in the field to clarify existing land uses and identify mukim boundaries as pointed out by communities. The map was later digitized using GIS, revised by the participants, edited and eventually designated into a final map that was presented to the community in a public meeting. The final map was used as a reference for the Ulu Masen Project and planned to be integrated in the district's spatial planning.
However, translating communities' conception of space into the map grid is not a simple process because they have different conceptions of space compared to that of the project proponents. Most mukim do not have a definite marker of their territorial boundaries but using a traditional method known as 'si uro jak wo' (literally translated as "a one-day return journey") to measure the boundaries of customary forest areas. Such a fluid conception of boundaries is commonly shared by most of the mukim in Aceh, including in the two study communities. A villager expressed similar views during my initial visit to Village 1 in 2010: "When we have tired of walking in the forest, that's the boundary of our forest." 23 It reveals that community members would like to continue to have access to areas within walking distance. Farther away from their village areas, and generally at higher elevations, the villagers believe that a lower intensity of forest use is appropriate.
A key NGO informant, who facilitated the mukim-based spatial planning, noted that many mukim leaders extend their mukim territorial claims as far as possible, thereby causing territorial overlaps among different mukim. This situation could potentially cause territorial disputes. When the project proponents proposed territorial boundaries of mukim for the principal map described above, most of mukim leaders were Journal of Political Ecology Vol. 27, 2020 471 reluctant to accept it. Having witnessed land dispossession, especially during the heightened conflicts, they were concerned that by fixing their territory and borders on a map, the fluid nature of access to their mukim territory and the forest area would be hampered. Although they finally agreed with the principal map, mukim leaders rejected the term 'mukim border.' They were aware that once they agreed to delimitation of a border, their access to the forest would be restricted. The impact of participatory mapping was more profound in Village 1, where villagers' sense of territorial boundaries and entitlement has been increased. They understand that they needed to assert their customary rights over their area and gain the state's recognition to secure their territorial claims. They initially proposed that the district head issue a decree recognizing the community-managed water catchment area, but they were unsuccessful. When the study was carried out, villagers explored another possible avenue to obtain the government's recognition of their rights by trying to obtain a permit from the Ministry of Forestry and Environment through its social forestry schemes. 24 In Village 2, even though the village was an early starter for the participatory mapping exercise, only a few village elites understood the process. The villagers had no clear plan for using the map created through the participatory process to claim their rights to the forest. However, the village and mukim leaders plan to integrate FPIC principles in a village regulation that requires all development initiatives coming to a village to obtain free, prior and informed consent from the community.

Conclusions
Current efforts to drive and sustain REDD+ policies and initiatives have been manifested in a variety of strategies (Asiyanbi and Lund 2020;Milne et al. 2019). This article focused on examining three key governmental strategies to enroll local communities in the Ulu Masen REDD+ and produce new subjectivities, which have included educating communities about REDD+ and conservation; increasing forest surveillance and policing through recruiting forest rangers and community rangers; and spatial ordering through mukim-based participatory mapping and spatial planning. The governmentality framework, and the local practice of muslihat help to illuminate how local actors resist, adopt and creatively accommodate these interventions based on the multiple subject positions that they occupy.
The process of enrollment in REDD+ involved complex conjunctures. The findings suggest that different social, economic and ecological contexts and different experiences with development/conservation encounters produce different outcomes for subject-making efforts. Despite different socio-economic and ecological conditions, the project proponents implemented similar strategies in the two villages, which generated different outcomes. In Village 1, villagers' long engagement in environmentalism and relatively high awareness of environmental discourses enables them to actively engage in state-led development agendas, including REDD+, while insisting that their needs and aspirations are met. Here, villagers skillfully appropriate narratives of conservation and indigenous rights in order to renegotiate their rights over forestlands.
In contrast, the people of Village 2 have more limited awareness on REDD+ and there are high opportunity costs in shifting their livelihood practices to be aligned with REDD+ goals. This has resulted in weak enrollment of communities in REDD+ on the ground. Despite REDD+ awareness-raising and capacity building activities implemented in the village, most villagers continue their local practices such as shifting cultivation, artisanal gold mining, and illegal logging. This situation, combined with easy access to markets, challenges the proponents of REDD+ to create new and willing subjects.
In both villages, villagers have been quite skillful in playing multiple roles depending on the situations that they encounter. Their engagement in a variety of capacity building activities has increased their ability to think critically about the importance of territory. In Village 1, for instance, the participatory mapping exercise and awareness raising activities about REDD+ and indigenous rights have increased community awareness of

Journal of Political Ecology
Vol. 27, 2020 472 their territorial boundaries and the importance of having their rights recognized by the state. Furthermore, community participation in REDD+ is strongly motivated by their desire to get secured tenure over their forest areas. In Village 2, the idea of FPIC has inspired the mukim leader to initiate the development of a local regulation that requires the application of FPIC for development initiatives to be implemented in the area. Moreover, attempts to fix state forest boundaries and zoning systems face multiple obstacles because of complex tenure arrangements, local politics, and different conceptions of spatiality between the proponents and local communities. In Aceh, many mukim have a rather fluid conception of mukim territories. Therefore, they negotiated during the participatory mapping exercise to allow them to have continuous access to the forest. Moreover, their participation in the mapping exercise was strongly motivated by their desire to have their mukim territory recognized by the state.
To conclude, the analysis in this article offers at least three important insights: first, through revealing diverse local responses to REDD+, this article elucidates that the neoliberalization of nature is incomplete, and far from smooth. Rather than creating mindless environmental subjects, REDD+ has opened up spaces for local actors to contest and creatively accommodate these interventions to better suit their aspirations and needs, as also seen elsewhere (Asiyanbi et al. 2019;Astuti and McGregor, 2017). Examining communities' engagement with the REDD+ project and the multiple positions they adopt towards it not only provides insights into their diverse aspirations and needs, but also into their critical engagement with the project. Indeed, as suggested by Li (2007), community engagement with governmental interventions could provoke their practice of politics and increase their capacity to resist and utilize them to better suit their needs. Channeling in this way this to broader policy making processes could unleash the emancipatory potential of such engagements to shift how the policies are formulated (Asiyanbi et al. 2019).
Second, in this article, I also show that REDD+ outcomes are shaped by social relations and the political and ecological context where REDD+ is implemented. These findings suggest that not all contexts are equally favorable to REDD+ implementation. Clear and secure tenure over forestlands, low opportunity costs, low commercial interest in forest resources, less access to markets, and local social cohesiveness seem to favor REDD+. This finding resonates other studies Milne 2012). However, such conditions seem to be idealistic and rare to find on the ground. Rather than looking for such elusive, ideal conditions, these findings, perhaps, suggest a more modest role for REDD+ than is often envisaged. If REDD+ can only thrive under certain conditions, and such conditions are difficult to produce, then REDD+ cannot be seen as a 'cheap' and 'fast' route to massive emissions reductions (Asiyanbi and Lund 2020). Furthermore, it is vital to pay attention to the problems considered important by local residents and tailor forest policies accordingly, to produce environmental subjects (Agarwal 2005a(Agarwal , 2005b. Finally, the article has revealed that an ideal community portrayed in REDD+ projects and policies, as also described by Skutsch and Turnhout (2018), is problematic because it obscures not only complex social relations, but also aspirations and needs. While such a portrayal could help policy makers, project developers and the like to render communities governable, it could constrain them from benefitting from REDD+ and further marginalize certain social groups in the communities. Integrating a more nuanced understanding of communities is essential to ensure just outcomes of REDD+ projects and broader conservation and development initiatives, and enable communities to get fuller benefit as citizens.