How REDD + Is Performing Communities

Community based approaches are becoming the norm in environmental governance initiatives. One prominent example of this is Reduced Emissions from Deforestation and forest Degradation (REDD+), a climate change mitigation strategy that aims at reducing carbon emissions caused by deforestation and forest degradation in developing countries. REDD+ policies generally evoke communities as both potential beneficiaries of REDD+ and as agents for its implementation. However, it is unclear what REDD+ policies are really referring to when they talk about communities. Drawing on critical social science literature about the idea of community, this article advances a performative perspective to analyze how communities are articulated in international and national REDD+ policy, and reflects on the potential implications of these articulations. Results reveal that international policy documents, including those of the major non-governmental organizations (NGOs) engaged in REDD+, tend to present an interpretation that corresponds to Agrawal and Gibson’s myth of communities as small, localized, and homogenous social units that share social norms. On the other hand, national policy documents reveal enormous variety in the communities that are actually targeted in national REDD+ policies in terms of resources, governance structure, and social cohesion. One conclusion that could be drawn from this is that the dominant uniform interpretation of communities in REDD+ policy, and in much academic and NGO literature, is clearly unrealistic. However, this does not mean that it is inconsequential. We conclude our article by discussing the performative effects of the identified articulations of community.


Introduction
The need to involve communities in the implementation of the international policy initiative on Reduced Emissions from Deforestation and forest Degradation (REDD+) is widely stressed in both policy and academic literature on the subject [1][2][3][4][5][6][7][8].(Indeed, much of the literature that is critical of REDD+ bases its complaints on the failure to adequately involve communities in decision making, or to adequately reward them for their REDD+ efforts [9][10][11].However, what is implied by the term 'community' (or 'local community', as it appears in the United Nations Framework Convention on Climate Change (UNFCCC) sources) is not discussed in any of these texts, which simply assume the presence of entities called 'communities' in or around forested areas.These communities are widely referred to in the context of REDD+ safeguards, i.e., the need to protect rights to use the forest and to receive social benefits (even poverty alleviation).However, they are also frequently portrayed as the stakeholders who will be responsible for management of local REDD+ initiatives.
In this paper, we explore how communities have been imagined in REDD+, and we examine what impact these framings may have had not only on the policy, but also on the communities themselves.
We start by observing that the definition of 'community' was problematized very well twenty years ago by Agrawal and Gibson [12], who observed that proponents of a community approach to natural resource management usually consider 'community' to mean a place-based, small, spatial social unit, with a homogeneous social structure and shared norms.They provided a historico-philosophical explanation for this perception and then criticized it claiming communities tend to be made up of different actor groups with varying interests, the implication being that 'shared norms' may be a myth.Despite this, such myths continue to be evoked by powerful external actors who propagate it to justify interventions based on problematic assumptions about community approaches being pro-poor, efficient because they incorporate local knowledge, and a fulcrum for democratic change amongst others [13,14].However, while there has been vigorous debate on the issue of the efficacy of communities in managing natural resources [15][16][17], there has been relatively little debate on what 'community' actually means in the context of REDD+ specifically.
The notion that communities may be a central pillar in the implementation of REDD+ fundamentally rests on the idea that the rural population, or that part of it that has access to forest resources, is organized into communities which could in principle gear up to carry out activities under the REDD+ banner.This assumes that communities have a number of characteristics, for example leadership with local legitimacy and organizational mandate, as well as the skills needed (which could be bolstered through capacity building), although it is well recognized that community involvement in REDD+ may also require improving the security of land tenure [18][19][20][21].It also seems to imply that where forests are inhabited, there is 'wall-to-wall' presence of communities.However, it may be questioned whether presence of human life necessarily implies presence of communities, in the sense that people consider themselves to be members of a particular place-based 'community' which can decide upon and undertake a series of actions, either communally or through coordination of individual effort.While this may well be true for some communities in some circumstances, it may be questioned whether the entire rural population, particularly in areas that are well integrated into the global economy, feels that it belongs to communities of this kind.Moreover, much human-used forest is very clearly not under any sort of control by communities.Global estimates of the proportion of forest land recognized as owned or managed by communities vary from 22% to 30% [22,23]; other forest areas may be used by the local rural population as an open access resource; often this is officially state or (large scale) private property.
Given the significance attributed to REDD+ as an integrated approach to climate adaptation and mitigation in developing countries [24], it is urgent to examine how communities are envisioned and articulated in REDD+, as well as the implications.We start by examining the role of communities as presented in official policy documents of the UNFCCC on REDD+, and in materials written by the international organizations providing support for REDD+.We then turn to the national documents on REDD+, analyzing in detail eight of the most recent and more advanced REDD+ country programs as submitted to the Forest Carbon Partnership Facility (FCPF) program of the World Bank for financing.This enables us to trace different meanings ascribed to the term 'community' in different contexts, which we believe is important to deepen understanding of how 'communities' can become engaged in REDD+.Before we present the findings, we first discuss our performative perspective and methodological approach.

Communities: A Performative Perspective
As discussed in the previous section, the idea of communities, defined as small spatial units with homogenous social structures and shared social norms, is problematic, and this can be viewed in at least two different dimensions.The first relates to representation, being the way in which ideals of community fail to represent the reality of communities.Agrawal and Gibson [12] highlight the ideological character of the idea of community and of the way in which communities are evoked in current natural resource management policies.Confronted with the failure of the state to conserve forests and nature, and the negative social consequences of many state-led policies, as well as piggy-backing on wider trends to participation and democratization, communities have emerged as an attractive and appropriate target for policies.This reflects a belief that community involvement could "promote desirable collective decisions" thereby enhancing the effectiveness, as well as legitimacy of these policies [12] (p.636).However, this community ideal falls short of capturing the reality of communities and particularly the diverse actors, interests, and norms that constitute them and Agrawal and Gibson suggest that this may result in disappointing effects of community based natural resource management policies.Specifically, it makes these policies vulnerable for elite capture and may result in them reproducing or even enhancing existing power inequalities [25].Whilst much of the REDD+ academic literature, alongside much of the official REDD+ documentation, attaches great importance to communities, this last point is increasingly being recognized.Several recent contributions have highlighted risks of elite capture within communities, and the risk that the lion's share of benefits will go to those who by tradition or by circumstance have a greater share of the resources and power over community decision making [26][27][28][29][30].
This leads us to the second dimension of the problematic of communities, which relates to performativity.Quite apart from the question of whether policy discourses that evoke the myth of communities are realistic, it must be recognized that they are not innocent or inconsequential.As Igoe and Fortwangler [31] (p.66) explain: "while these discourses may not accurately describe the realities of the problems that these interventions are meant to resolve, they are nevertheless effective for mobilizing significant revenues and extending state control over people and natural resources".This suggests that they are performative in the sense that they produce effects that cannot be understood with reference to their representational accuracy [32].This argument has amongst others been applied to the process of nation building.Scott [33] (p. 3) explains that representational devices, such as maps, rework reality to fit with the ideal: "They did not successfully represent the actual activity of the society they depicted, nor were they intended to; they represented only that slice of it that interested the official observer.When allied with state power, they would enable much of the reality they depicted to be remade".Anderson's well-known book 'Imagined Communities' makes a similar point.To Anderson, Nation States can create their constituencies by presenting authoritative and attractive visions, in media, maps, or museums, which people can evoke and make their own [34].
However, as the many examples in Scott [33] show, this constitutive power of visions does not mean that these visions become reality exactly as imagined; they often create unintended effects and they can fail even on their own terms.Performativity means that there will inevitably be overflows [35], that is to say, unpredictable and contingent surprises [32,36].Studies of participation and engagement using the concept of performativity have highlighted how identities and interests of participants do not pre-exist but are formed in the participatory practice, often with unintended and unpredictable outcomes [25,37].Thus, a performative perspective on policy invites us to consider what policies do and how they produce effects.Importantly, this is not the same as asking how effective policies are.Instead of an evaluative perspective that judges the performance of policies using a yardstick, such as the stated objectives of a policy, or criteria for good governance, performativity is tied to a practice-based perspective [38,39] and does not assume the existence of yardsticks or policies outside the practices they judge or intervene in.This interpretation of performativity finds its roots in Science and Technology Studies [40,41], and recognizes that policies and their outcomes are entangled and that they are both constitutive of and constituted in practice.Applied to the notion of community, a performative perspective suggests that communities do not simply exist, passively waiting for REDD+ to engage them.Instead, REDD+ performs communities, it brings them into being in practice in specific, contingent, and context specific ways.According to Marres [42], this argument can be traced back to pragmatist philosophy and particularly to Dewey's 'The Public and its Problems' [43].Marres [42] suggests that the formation of publics (communities in the case of REDD+) is not just about will formation (articulating the interests of the communities), but also issue formation (shaping the meaning of REDD+ and the problems it addresses).
In other words, a performative perspective offers insights into the workings of a policy, which go beyond evaluating its representational quality (does the policy adequately reflect reality) and its effectiveness (does it achieve stated objectives) and allows for critical examination of the intended and unintended political implications of policy prescriptions.Consequently, this paper asks how REDD+ performs communities, that is, what visions of community do REDD+ policies at various levels articulate and with what potential consequences?

Method
To understand how communities are envisaged in the international policy on REDD+, we examined three sets of literature; Firstly, all the official policy decisions, as well as submissions made by parties and observers, which were made during the negotiations on REDD+.These are available on the UNFCCC website [44].This represents a relatively concise body of centralized literature, which can readily be overseen.Secondly, literature from NGOs and support organizations, which in contrast is vast and widely dispersed.Although we could not make an exhaustive or all-inclusive study, we attempted to distil the main lines of discussion in this material in as far as they refer to communities in REDD+.
Thirdly, for the analysis of how communities are framed in national level REDD+ planning, we used as examples the eight countries which had submitted so-called Emissions Reduction Programs (ERPs) to the World Bank's Forest Carbon Partnership Facility (FCPF) in final or draft form by the time this study was carried out (end of 2017).These are the most complete expressions of the rationale of national REDD+ thinking available and can be compared since the reports follow a common format requiring among other things, the identification of stakeholders and an explanation of the causes of deforestation and degradation, as well as how the interventions planned are consistent with this.The countries that had submitted final ERPs at the time of this research were: Chile, Mexico, Vietnam, the Democratic Republic of Congo (DRC), and Costa Rica, whilst Ghana, Mozambique, and the Republic of Congo (R. Congo) had entered draft ERPs.In all cases, except for Costa Rica, these plans were not for the entire country, but for selected geographic areas only.
The ERPs were studied using the textual analysis software Atlas-ti.A full document search was made for all mentions of the word 'community' or 'communities' (including 'local communities' and 'indigenous communities').Uses of terms that were irrelevant for our purposes (e.g., international community, donor community, plant community, and proper names) were ignored, as were those parts of the text that were provided by the World Bank (i.e., embedded in the format) and sections that were copies of standard texts from other sources, such as the UN Declaration on Rights of Indigenous Peoples.The reference lists at the end of the reports were also excluded from the analysis.The remaining uses of the term 'community' were coded using a set of alternative signifiers, which are explained in the results section of this paper.Words which could have similar meanings to 'community' (e.g., village, indigenous group, commune, agrarian nucleus) were also checked and compared.All the reports were in English, but where a version in an alternative language (e.g., French, Spanish) was available, the equivalent terms were also identified to ensure consistency.

Communities as Articulated in UNFCCC REDD+ Policy
The Paris Climate Agreement [45] includes in Article 5 explicit reference to 'policy approaches and positive incentives for activities relating to reducing emissions from deforestation and forest degradation, and the role of conservation, sustainable management of forests, and enhancement of forest carbon stocks in developing countries'.The initiative is generally known as REDD+.In this context, no direct mention of communities is included, though the preamble to the Agreement lists the need to consider obligations, such as support for the rights of indigenous peoples and of local communities, and Article [8], on adaptation, refers among other items to support for the resilience of communities.However, the need for engagement and participation of 'local communities' and 'indigenous peoples' (IPs) in various ways is clearly expressed in earlier UNFCCC policy decisions on REDD+.In these decisions, the term 'communities' is mentioned ten times, particularly in the context of recognizing their needs [1].The documents specify it will be important that REDD+ ensures communities' sustainable livelihoods, develops safeguard information systems that show respect for their rights and their knowledge, provides options for adaption to climate change which take into account their traditional and indigenous knowledge, and engages them in monitoring and reporting (Table 1).There is strong emphasis on consultation and the use of community knowledge to arrive at better REDD+ interventions, which, at minimum, will do no harm to their livelihoods.More recent decisions also recognize the potential of communities to be involved in the implementation of these interventions.The influence of IP lobby groups, which joined with civic environmental movements to form a front for issues relating to social justice during the climate change negotiations [46][47][48], is evident in the repeated use of the combo 'local communities and indigenous peoples' and undoubtedly served to increase the prominence of 'communities' in these policy texts.
The views of the UNFCCC negotiating Parties on the engagement of communities in REDD+ can also be traced through the submissions that have been entered on different topics throughout the process.What is most noticeable is that, as with the policy developments outlined above, there has been a progression in how the role of communities has been seen.Whilst the call for benefits to flow to communities has been present in broad and unspecific terms throughout the negotiation process, quite quickly contributions turned to focus on the need to engage communities in the generation of information (monitoring) of carbon stocks and of the effectiveness of REDD+ efforts locally.This was followed by calls for community monitoring of safeguards, including those that refer to respect for knowledge and rights of IPs and members of local communities, promotion and support for effective participation of communities (e.g., through Free, Prior and Informed Consent (FPIC)), and protection and conservation of natural forests, whilst enhancing social and environmental benefits considering the need for sustainable livelihoods.Although there have been sporadic references to direct incentive payments to communities (e.g., by Colombia, [49]), most refer to community benefits in general, without committing to cash payments as such, and the most recent submissions have focused on the need for non-carbon benefits to communities, possibly financed in addition to carbon.Overall, whilst there are clear calls for the participation of communities in the sense of providing information and receiving some benefits (and above all, not being disadvantaged by REDD+), there is no clear picture given in either the national submissions or in the UNFCCC policy texts on whether and how communities would be engaged in the actual implementation of REDD+ post planning phase, presumably in part because this would interfere with national sovereignty on the matter.
Additionally, is it noticeable that there is no definition at all in these policy texts of what 'community' means, barring one remark from COMIFAC (the Commission of Central African Forests) which suggests that the definition should be left to individual countries according to their national circumstances [50].Clearly, social institutions vary greatly between countries and trying to impose a single definition would be impossible.However, it is apparent that the term is intended to refer to groups of people resident in or near forests who are dependent on these forests for part of their livelihoods, whether or not they have formal tenure, rather than to individuals who own forest land as private property.However, in reality many rural communities are made up of smallholders (operating essentially as private land owners) or at least include some quasi-private situations.Farmers within communities usually operate as individuals even where there is no formal tenure (i.e., usufructure systems are common).Therefore, whilst use of the term 'community' implies some kind of group or communal decision-making system, there is no implication in the texts that a community necessarily possesses and/or manages areas of communal land as this may or may not be the case; nor is any particular size of community or form of community governance suggested.The phrase 'local communities' as noted above is almost always paired with, but clearly seen as different from, 'indigenous peoples'.The latter term is well recognized, e.g., in the UN Declaration on the Rights of Indigenous Peoples in relation to territories owned by or claimed by groups who identify themselves as being of a specific local ethnicity, and it is safe to assume that the term 'local communities' is considered to have a broader, more catch-all meaning.To conclude, the UNFCCC does not provide a definition of communities but does envision them as being able to benefit from REDD+, and as potential agents in the planning, implementation, and monitoring of REDD+ without a clear specification of how exactly the communities are to accomplish this.This formulation appears to have been chosen to ensure that REDD+ is directed at the poorer, more traditional, users of rural land and forests, rather than profiting large land owners.This was partly in response to concerns that arose (among others, from international NGOs and Indigenous Peoples' (IP) organizations, see below) when REDD+ was proposed, with a view to ensuring that REDD+ would be socially equitable.However, the formulation makes no reference to possible inequalities within communities and how individuals within communities would be engaged.

Preamble
Recognizing the need for full and effective engagement of indigenous peoples and local communities in, and the potential contribution of their knowledge to, monitoring and reporting of activities relating to decision 1/CP.13,paragraph 1 (b) (iii); Also requests developing country Parties, when developing and implementing their national strategies or action plans, to address, inter alia, the drivers of deforestation and forest degradation, land tenure issues, forest governance issues, gender considerations and the safeguards identified in paragraph 2 of appendix I to this decision, ensuring the full and effective participation of relevant stakeholders, inter alia indigenous peoples and local communities;

Communities as Articulated by International Support Agencies and NGOs
The position of intergovernmental support organizations and NGOs on the role of communities in REDD+ is, in comparison to the UNFCCC and its Parties, much more detailed.In a formal submission to the UNFCCC [49], the UN Convention on Biological Diversity calls for the development of self-sufficiency in REDD+ of indigenous and local communities.Similarly, the UN Forum on Forests calls for community forestry under REDD+, for the involvement of communities in meaningful decisions on state forests, and as one of the first, for attention to the role and importance of tenure.Crucially, a community approach to implementation appears to be at the heart of the UN REDD+ program (a support and funding program jointly managed by United Nations Development Programme, the Food and Agriculture Organization and the United Nations Environmental Programme, which has a specific community based sub-program co-financed by the Global Environmental Fund Small Grants Program.The World Bank's Forest Carbon Partnership Facility (FCPF), one of the major funders of REDD+, follows a charter which requires testing ways to sustain or enhance livelihoods of local communities and conserve biodiversity, whilst its Forest Investment Program (FIP) has a dedicated mechanism that provides grant funding for indigenous people and local communities to engage in the implementation of REDD+ processes.Conservation International states that the participation of indigenous peoples and local communities is crucial to the successful development and implementation of REDD+ mechanisms, and has developed a community manual to support such efforts [51].However, perhaps more than any other agencies, the CGIAR organizations have been influential in shifting the nature of much discourse on REDD+ from its earlier focus on international results-based finance to national governments, to one in which the central issue is what communities can do at the local level.This can be seen in the work of the Centre for International Forestry Research (CIFOR), which championed the idea of the '3Es' (effectiveness, efficiency and equity) as the standard against which REDD+ implementation should be held [52], and followed this up with in its global comparative study and global database of community level REDD+ projects [53].This type of thinking is also present in the work of The World Agroforestry Centre, ICRAF, with its Local Perspectives on REDD+ [54] and arguments that REDD+ projects which involve communities are more likely to save forests [55].For: "Who can manage forests better than those living within or beside them? with REDD+ redefining the forest management and conservation landscape, community forest management can contribute to reduced forest emissions and increased forest carbon stocks [6] (pp.201-202)".
Yet despite this focus on the importance of communities in REDD+, these texts also fail to explain or explore what they mean or imply by 'community'.They use the term as if it were self-evident (e.g., the otherwise very extensive REDD+-related glossary at the end of Angelsen et al. [51], for example, does not include an entry for 'community').This follows earlier practice in highly influential publications on communities and forests, such as White and Martin [22], which also talked about communities without critical analysis of what this means.It is quite difficult to read between the lines to determine how communities are actually being imagined here.In terms of organizing natural resource management and REDD+, involvement of communities is seen as a 'third way' (other than regular government channels and the private sector).There are underlying pragmatic efficiency reasons for this, as the quote from Agrawal and Angelsen above indicates, but also ideological/political reasons related to the idea that the rights of poorer, small scale users to control their local resources need to be protected and bolstered to ensure that REDD+ does not give advantage to the more powerful elements, such as larger landowners and corporations.Along with this, it seems that the whole rural population is implicitly conceived as being made up of communities, and secondly that these communities are imagined largely as Agrawal and Gibson [12] suggest, although the list might be broadened.Specifically, communities appear to be envisaged as relatively small territorial and placed-based social groupings (1) that are positioned outside the normal hierarchy of government administration, (2) that have some form of internal, communal decision making procedure which may be based on traditional/indigenous rules or on other norms and procedures, (3) that may or may not have formal rights to land and forest and may or may not have communal forest land, and that (4) under favorable conditions, are expected to be able to act as a unit to manage the forest to meet common goals.
Whilst this last expectation does imply the ability to work towards common goals, it is less clear whether communities are understood in these texts to have homogeneous social structures and shared norms [12,56].The shared norms which supposedly help communities to manage common properties are usually understood to be rooted not just in shared ethnicity, religion, and history, but also in the fact that members have similar livelihoods strategies, seen as contributing to internal harmony.The way that the REDD+ discussion on equity has developed offers interesting insight into this issue.Although there have been a number of interesting discussions on the meaning of equity in REDD+ in academic literature [57][58][59][60], the main focus at the policy level has been on safeguards and benefit sharing (see Table 2 for the case of UN-REDD).FCPF requires countries to provide and justify a benefits distribution plan that respects customary rights to lands and territories and that reflects broad community support "so that incentives are applied in an effective and equitable manner, and shared with those stakeholders most critical to protecting forests (the benefit sharing plan will provide details on the benefit-sharing process, related distribution criteria and timelines and types of beneficiaries)" [61].Moreover, there is emphasis in these texts on the need to be gender inclusive and to protect the interests of marginalized groups (i.e., usually seen as marginalized communities), thus a generally pro-poor approach.However, they do not really peek into questions of equality (for example, as regards to size of landholdings) within communities and there is no requirement to carry out a class analysis at community level.This failure to attend to heterogeneity within communities, suggests that these texts do assume some degree of social homogeneity.This also applies to NGO literature, which critiques the social inclusivity of REDD+ programs.The Rights and Resources Institute [62] for example, deals with generalized issues such as community rights and community tenure, but does not broach questions of equity and fairness within communities.To our knowledge, this has been addressed, and only in the context of benefit sharing, in just two countries, Nepal [63] and Vietnam [64,65], where a formula to differentiate needs of the family (Nepal) or their individual contribution to increasing carbon stocks (Vietnam) has been used.We note that these benefit distribution systems relate to rewards for increases in carbon storage, rather than for decreased deforestation, which is much more difficult to link with individual behavior or position (see Skutsch et al., [60] for a detailed explanation).We also note in passing that lack of equality within communities does not necessarily mean there are no shared norms.In many communities, inequality is the norm and is accepted as such.Protect and enhance economic and social well-being of relevant stakeholders, with special attention to the most vulnerable and marginalized groups

Communities as Articulated in National Policy Documents
The eight Emissions Reductions Program documents revealed a range of uses of the term 'community', which in most cases was applied in several different senses even within one country document (Table 3).The term is used much more often in some countries (e.g., Mozambique) than in others (e.g., Costa Rica), even taking into account the length of the reports, and that six distinct meanings can be identified, as explained below.The first important finding is that 'community' is very frequently (and some cases, predominantly) used not to refer to a group of people, their organizational form or governance, but simply to refer to the rural population.Examples are 'community awareness raising' and 'increasing community income'-where what is meant here is not community (in the sense of shared or communally decided), but the rural people in general, and this use is repeated in many reports, e.g., 'the resilience of communities to climate change'; 'access of communities to forest resources'; 'develop viable alternative livelihoods for local communities'; 'sustainable community livelihoods'; 'bushmeat is a primary source of income for communities', etc.In all such cases, terms such as 'rural families' or 'the local population' could equally well be used instead of 'community', since the reference is not to any specific organization, institution, or social grouping.We see here 'community' being used as shorthand for 'people'.
A second meaning of the term community is found in reference to indigenous peoples or indigenous cultures in general (i.e., not to specific communities such as local organizations).Statements such as ´Indigenous Law No. 6172 set forth that the indigenous reserves remain the property of indigenous communities´Costa Rica ERP is a typical example of this.Thirdly, it is often used to refer to 'villages/settlements', i.e., the physical areas in which groups of people live (e.g., 'forest-fringe communities in the ER program area are ethnically diverse' (Ghana ERP)).In this sense, it implies a presence of people living together in groups, but without reference to any particular form of governance; however, we mostly used the term 'villages' to classify references to communities where there does not appear to be clear authority present that would enable the community to take charge of REDD+ implementation (e.g., 'Communities/villages are not legal administrative entities in Vietnam so are without decision making powers' (Vietnam ERP)).
However, in many cases, the term 'community' clearly relates to a form of governance, when it refers to communities in the sense of having a defined leadership system and mandate which could, for example, enable it to take charge of local implementation of REDD+.Here we discerned three more meanings of community.Firstly, there are communities that are independent of the government hierarchy, in the form of autonomous decision making units (e.g., 'internal organization among ejidos and communities, resulting from the weakening of their internal governance mechanisms, such as their general assembly, and their internal bodies, such as the ejido commissioner and the board of protection, are one of the main barriers to ensuring a successful development of activities' (Mexico ERP)).Then there are communities that specifically occupy indigenous peoples' reserves (i.e., 'different' from the majority of the rural population).For example, 'This diminishes the land use rights of local communities and leaves notably indigenous communities-those outside dedicated areas within concessions-in limbo' (ERP Republic of Congo).Finally, there are communities that belong to the (lowest tier of) the government hierarchy (e.g., 'The administrative organization comprises two former districts, 8 territories, 23 sectors, 66 communities, 4 towns and thousands of villages' (ERP DRC)).In allocating each mention of 'community' to these different categories, we have endeavored to be as consistent as possible, and we included a 'not clear' category for cases where there was significant doubt.
The six different meanings associated with the term 'community' are shown in Table 3, in terms of the frequency with which they occur in each report (%).In Mexico, the term is exclusively used to refer to ejidos and comunidades indigenas, two different forms of communally organized agrarian nuclei which have legal land tenure over defined territories and clearly defined internal decision-making systems; they are independent of the government administration.Farmers and forest owners not living within these areas, but who might be engaged in REDD+, are in contrast referred to as smallholders or 'other landowners'.In Chile, 'community' is used both to refer to the lowest level of the formal government hierarchy, i.e., the commune (meaning municipality or county) and to indigenous communities living in areas over which they have communal tenure by law.In Costa Rica, it refers mainly to communities in indigenous territories, although other types of communities ('peasant communities', 'other rural communities') are sporadically mentioned in the report without further specification, as well as small landowners and farmers, who as in Mexico, are not considered to belong to communities.In Vietnam, 'community' most often refers to informal or traditional groups of people living in villages within the communes, which represent the lowest tier of the formal administrative system.Although there are traditional leaders in these villages, they do not enjoy formal powers, for example, the Civil Code does not permit these communities to own land as a group and they are not legal entities, so they cannot have bank accounts.The situation in the African countries is less clearly defined.In the DRC, R. Congo, and Ghana, communities are framed in the context of the traditional and customary system of chiefs.These traditional governance systems may fuse with the formal government system, for example in DRC, chiefs are usually appointed by the government as administrators at the village level.The DRC REDD+ document identifies the county or municipal level as the 'community' and smaller population centers nested within this, sometimes as 'villages' and sometimes (confusingly) also as 'communities'.In Ghana and the R. Congo, there are autonomous local level communities whose local leaders have some authority to make decisions about land use etc., although much of the land is in private hands through usufructure rights.Ghana already developed overarching umbrella institutions called Community Resource Management Areas (CREMAs) and envisages similar organizations for REDD+ called Hotspot Interventions Areas (HIA), which will be governed by a local governance board of land owners, land users, local authority entities, and community leaders (including leaders of minority groups), as the key players in the REDD+ strategy.This means that the 'community' is not seen as the focus in quite the same sense as in many other countries.Moreover, the target of the entire strategy is the individual cocoa farmer, not the community as a group.In Mozambique, the term 'Local Communities' (usually, but not always, capitalized) is specified in the Land Law to mean groups of people with communal title to the land and authority over its use.
In summary, we see that there exist quite different representations of communities in these different segments of policy making.These are all implicit rather than explicit, and while the national reports are more specific about the form their communities take, none of the sources really examine the nature of communities and their internal workings.There is an unquestioned assumption that a community approach will be efficient and will produce benefits for the rural people, with an underlying coda that suggests that all rural people are equally poor.

Discussion
At first sight, the different sets of REDD+ documents seem to represent communities in rather different ways.The policy texts developed at the level of the UNFCCC through national submissions and negotiations are not very specific about what defines a community, although they represent them as groups of people living in forest areas, which may or not have legal tenure over these resources.These texts appear to have been quite strongly influenced by lobby groups representing IPs, and most of the references to communities relate to safeguarding IP and community rights, including the right to consultation, and emphasize their role as beneficiaries, whilst there is little reference to how they may be involved in the direct implementation of the policy.Given the lack of specificity, one could question why terms such as 'rural people' or 'small farmers and forest users' were not used instead of 'communities'.Evidently, 'communities' implies a level of organization and a means of communication which could be used directly by REDD+.The fact that 'communities' do not exist everywhere is simply bypassed, for convenience.The agencies that are promoting strong community involvement in the actual implementation of REDD+, including FCPF, UN-REDD+, and many international NGOs and research organizations do not provide a clear definition or critical discussion of communities either.However, as we have shown, these organizations also appear to imagine the entire rural population as made up of local territorially coherent communities, which are not part of the governmental administrative system, but are independent entities with their own decision-making systems.In these texts, a community approach appears in some ways to indicate a 'pro-poor' approach, since communities are considered to be 'poor'.However, it is noticeable that very few cases actually consider inequality within the community.This may reflect the dilemma mentioned already at the end of Section 5, i.e., the development industry in general supports the idea of self-determination at community level but at the same time demands equality, and these may be incompatible objectives.Interestingly, while these texts attribute great significance to communities in relation to REDD+, they also acknowledge that communities often do not have formal rights to forest or tenure over land and that they lack legal and technical skills for participation in REDD+, which means that the capacity of these communities to contribute to REDD+ may be limited.
In contrast, the individual country reports revealed quite a range of interpretations of 'community', reflecting their own national circumstances.Firstly, it was evident that in very many cases the term 'community' is being used not in the sense of a type of organization, but simply to refer to rural people in general, as a form of shorthand.Secondly, in at least two cases (Chile and DRC) 'communities' is a term, which, among other uses, is routinely used to mean municipalities or counties, i.e., the lowest tier of the government administration, which is very different from the view held by NGOs and support agencies, which clearly see communities as outside the normal government system.Thirdly, although in some countries 'communities' does imply place-based social groups that have the mandate to make their own decisions as territorial units, in others, this is not so clear.The role envisaged for communities varies, in some there is more emphasis on communities as beneficiaries, whilst others contain at least some practical suggestions for how they may be engaged in implementation.However, there is almost no reference to how existing social inequalities within communities might influence outcomes.In these national reports, the inequalities that are mentioned are only in terms of ethnicity or gender.
Thus, we may conclude that the first two sets of literature on REDD+ policy articulate a vision of communities that resembles Agrawal and Gibson's myth.We see an implicit assumption in the documents that all rural people are members of communities, which in principle could have the characteristics necessary for participation in REDD+.Community level REDD+ is seen as being for the general benefit of everyone without any reference to existing social inequalities within communities, other than gender and sometimes ethnicity (the euphemism 'marginalized groups' is often used in this context).Moreover, the engagement of communities is clearly seen as contributing to the effectiveness and legitimacy of REDD+.Importantly, this is done without clearly specifying how exactly, given their limitations, these communities will be able to live up to these expectations.
The interesting point is that the country reports display a whole range of different visions of community.In some cases, a wall-to-wall concept is employed, in others, only specific groups of people in specific areas are targeted; the forms of community governance vary and are more explicit in these reports.Aside from this, the function of communities in REDD+ is represented in terms rather similar to those used in the global texts.In that literature, there are expectations that communities can be capacitated to carry out REDD+ activities, and that this will promote an equitable, efficient, and legitimate approach to REDD+.
These observations raise questions about why, given the evident lack of community capacity in reality, and the social inequalities that quite clearly exist within communities, the idea that engaging communities in REDD+ will not only be effective but also egalitarian, continues to be evoked.While we agree to some extent with Pasgaard [66], that this can reflect pragmatic use of rhetoric to further the positions of actors at all levels, we suggest that in fact this formulation has a strong ideological appeal, which stems from particular political positions and social and environmental justice principles.It is also strengthened by lack of faith of many observers in government efficiency and sincerity regarding rural poverty alleviation, as well as by the general trend, also found many official government rural development policies, towards the idea that local participation in decision making is positive and even necessary.
Apart from the concerns discussed above related to the realities and capacities of communities, this leaves the legitimacy of community-led decision making and implementation unaddressed.Following the majority of published literature on participation (for overviews and critical discussion see [25,37,67]), there is an implicit belief in REDD+ documentation on the global level that strengthening the relative bargaining power of rural people in the modern-day situation is unquestionably benign.Whilst we recognize that in some places there are strong alternative (traditional) forms of governance at the local level, particularly in the case of communities that identify themselves as indigenous people, taking this for granted overlooks the fact that in many cases community empowerment in REDD+ is likely to end up being inegalitarian or even undemocratic (see [68] for the case of REDD+ in Nigeria).
A second reason for the persistence of this myth is more pragmatic.The prior experience of the agencies that are promoting REDD+ is largely confined to projects such as community forest management and PES.These programs have been carried out in communities which possess at least some of the idealized characteristics described above, and which see themselves as management units.They are communities which were identified by government agencies or NGOs as potential partners or candidates for this kind of intervention, precisely because they possess these attributes.Although they may not represent the more general condition of the rural population, it has been expedient for some of these agencies to continue their programs in these kinds of communities and expand them with REDD+ objectives, and moreover it clearly influences the way these agencies conceive the situation and the possibilities for intervention in rural areas.However, there are many rural settlements where this kind of social organization is missing and where such initiatives have never been attempted because there is no mechanism for communication and negotiation, or where the label 'community' has been applied by external agents, but does not reflect local realities [14].Therefore, all the literature on community level natural resource management is based on an inherently biased sample and this bias is being reproduced in REDD+.CIFOR's Global Comparative Study on REDD+, which used existing community forest management and payment for environmental services projects as proxies for REDD+ is a clear example of this.It is a bias which is extremely difficult to overcome, since programs such as REDD+ depend on being able to communicate both about and with local forest users.If these are not arranged into clusters with a recognizable contact point, i.e., in the form of 'communities', it is difficult to visualize how communication could be managed.
This brings us to our final point.The community myth persists because it performs, i.e., it does work.Pasgaard and Nielsen [14], for example, show in a case in Cambodia how conservation rhetoric is used at all levels, from local leaders to donor organizations, essentially to further own interests, and Blaikie [13] has a similar story for Botswana and Malawi.The propagation of the idea that community is central to REDD+ could be seen as an example of this.However, at the same time, the ideological attractiveness of performing communities in this way has enabled consensus about the direction of REDD+ (as unspecific as this direction may be).It has also worked to reinforce the legitimacy of NGOs and other agencies that collaborate with communities, where these do in fact exist and to advocate for their rights, resulting in the potential for new project funding for collaborative work under the banner of REDD+.Therefore, although the visions of community in REDD+ may not be real in a literal or material sense, they are clearly a reality that actors have chosen to believe in and from which they stand to benefit.This means that they are real in their consequences [69].Whether or not REDD+ will be able to shape the reality of actual communities according to its vision is an open question.Past experiences in community based natural resource management does suggest that this is possible.The initiation of community-based arrangements and institutions can re-order social relations and remake communities to fit with this myth.This is in fact the basis of the Ostrom tradition in institutional analysis and natural resource management.Moreover, in a recent analysis of the Payment for Environmental Services program in Mexico, McCall et al. [70] find evidence that the presence of this program, which provides relatively easy money to communities that can justify a case for it, may in fact have served as an opportunity and incentive for communities to create, or revive, the necessary decision making apparatus which may have earlier fallen into abeyance due to historical circumstances, i.e., it has formed the modern 'community glue'.Newton et al. [7] have also made the case for how REDD+ is strengthening existing CFM initiatives in Nepal.
Policies are often built on myths, which does not always invalidate them or even necessarily result in failure.Policies built on myths can sometimes transform the situation they are designed to address.However, the power of myths lies not only in the ideals they imagine, but also in what they obscure, and so there will always be unpredictable outcomes.The strong attachment to the community myth could be the fuel through which REDD+ becomes a widespread catalyst for positive change, but it may also cause blindness in terms of the practical implications for communities, which will ultimately do them a disservice.

Appendix 1 , 1 ( 3 Further
para.2c (Safeguard C) Respect for the knowledge and rights of indigenous peoples and members of local communities, by taking into account relevant international obligations, national circumstances and laws, and noting that the United Nations General Assembly has adopted the United Nations Declaration on the Rights of Indigenous Peoples; Appendix 1, para.2d (Safeguard D) The full and effective participation of relevant stakeholders, in particular indigenous peoples and local communities, in the actions referred to in paragraphs 70 and 72; Appendix Footnote to Safeguard D) Taking into account the need for sustainable livelihoods of indigenous peoples and local communities and their interdependence on forests in most countries, reflected in the United Nations Declaration on the Rights of Indigenous Peoples, as well as the International Mother Earth Day; agrees that enhanced action on adaptation should tak(e) into consideration vulnerable groups, communities and ecosystems, and should be based on and guided by the best available science and, as appropriate, traditional and indigenous knowledge; Invites all Parties to enhance action on addressing loss and damage associated with the adverse effects of climate change by undertaking, inter alia (f) Involving vulnerable communities and populations, and civil society, the private sector and other relevant stakeholders, in the assessment of and response to loss and damage; 2015, Paris FCCC15/21 Preamble, ToR for Doha WP review of Article 6 of the Convention Reaffirming the importance of taking into account indigenous peoples, local communities and non-governmental organizations in activities related to Article 6 of the Convention.

Table 1 .
References to communities and indigenous peoples in the United Nations Framework Convention on Climate Change (UNFCCC) decisions on Reduced Emissions from Deforestation and Forest Degradation (REDD+).

Table 2 .
Criteria relating to social equity established by the UN-REDD program on Social and Environmental Principles (SEP).

Table 3 .
Meanings of the term 'community' in the Emissions Reduction Programs (ERP) documents submitted to the Forest Carbon Partnership Facility FCPF (frequency of occurrence in %).N-number of relevant references to the term 'community' in the national report; PP-pages; N/PP average occurrence of the term 'community' per page of the document.