The trade-offs of win–win conservation rhetoric: exploring place meanings in community conservation on the Wild Coast, South Africa

In attempts to reconcile conservation and development for poverty alleviation by establishing protected areas, economic values of nature and compensation for loss of access to resources are often prioritized over cultural and personal values. Additionally, conservation interventions in local communities are often hindered by contested visions of sustainability. We explore the utility of place meanings to unpack diverse local interests by examining an intervention that proposed to establish a fenced protected area in a community on the Wild Coast, South Africa. We describe the narratives that argue for or against the project and how they make use of the place meanings attributed to parts of the landscape, including forest, communal grazing land and plantations. We then examine the coalitions behind narratives: groups of actors who share the meanings and constructs of the problem and who employ these for a particular strategy. This allows us to map the negotiation process, and understand how community dissent influences the project. We find that a focus on economic benefits from protected areas neglects alternative meanings, e.g. cultural and spiritual value of forests as well as potential alternative pathways for development such as investing in small-scale agriculture. Our analysis reveals the tension that exists in the ‘win–win’ discourse of conservation between the rhetoric of sustainable resource use and co-management as well as a trend back towards ‘fortress conservation’. A community counter-narrative is successful in stalling the project which illustrates the importance of considering the plurality of meanings for interventions to be sustainable in the long term.


Introduction
There is a long-standing debate about how to align conservation of biodiversity with poverty alleviation and development in communities (Adams et al. 2004;Wells and McShane 2004). Interventions to establish reserves or parks for biodiversity and collaborate with local communities to manage these often attempt to reconcile conservation objectives with local livelihoods and economic concerns, for example, through economic compensation for loss of access to natural resources or investments in income-generating activities associated with ecotourism (see Mangome and Fabricius 2004;Ramutsindela 2007). These community conservation interventions are often framed as 'win-win' opportunities with ecological and socio-economic benefits (Christensen 2004;Chaigneau and Brown 2016). However, they seldom take into account the cultural, spiritual and personal values of nature for local people Bologna and Spierenburg 2015). In this context, the economic value of 1 3 nature is often emphasized over other local values and meanings (Büscher and Dressler 2007).
Conservation and development interventions are also often hindered by contested visions of sustainability within a community (Kepe 2008). Interventions may fail when the diversity of interests of local people and their lived experiences and desires for the future of a place are not taken into account (Bologna and Spierenburg 2015). Even in the literature that is quite critical about protected areas as the prevalent model for conservation interventions in southern Africa (Brockington 2002;Kepe 2008;Ntshona et al. 2010), the diversity of community interests and perspectives is often mentioned, but seldom demonstrated or discussed in detail.
Sense of place has been suggested as a conceptual tool through which to engage with conflicting visions of sustainability and to unpack local concerns in, for example, conservation interventions (Yung et al. 2003;Chapin et al. 2012;Chapin and Knapp 2015). In this article, we explore this suggestion by focusing on the competing meanings attributed to place by different groups of people (Davenport andAnderson 2005, Brehm et al. 2013). Studies of connections to place have captivated scholars since the 1970s, and place-related research has increased dramatically since then (Lewicka 2011;Manzo and Devine-Wright 2014), with interest growing recently in the environmental management and sustainability sciences (Stedman 2016;Masterson et al. 2017a). This rich literature has examined relationships between people and place through many traditions and using multiple constructs (for different accounts of this see Gustafson 2001;Scannell and Gifford 2010;Lewicka 2011). Much sense of place research has focused on the strength of attachment to place using positivist or hypothesis-testing quantitative tools, but such metrics have limits in answering why we depend on a place or what it means to our identity (Stedman 2002;Davenport and Anderson 2005;Brehm et al. 2013;Masterson et al. 2017a). A complementary part of the literature emphasizes qualitative and often phenomenological investigations of place. Drawing on foundational place writings, a focus on place meanings highlights how people's lived experiences in a place create meaning (Relph 1976;Tuan 1977). As set out by Stedman (2016) and Masterson et al. (2017a), we view place attachment and place meanings as two subconcepts under the umbrella of sense of place that relate to each other, i.e. place attachment rests on the meanings attributed to place. Therefore, a focus on place meanings may help to interrogate conflicts between groups equally strongly attached to a place but who are attached to different symbolic and emotional meanings that they attribute to this place (Yung et al. 2003;Jacquet and Stedman 2013;Masterson et al. 2017a).
We view the construct of place meanings broadly as the descriptive answer to the questions 'what kind of place is this?' or 'what does this place mean to you?' A holistic treatment of place meanings requires consideration of both experiences of belonging and experiences of alienation, and thus place meanings can be positive but also can reflect negative and ambivalent feelings about place (Relph 1976;Manzo 2005). We recognize that place meanings can be created through interaction with the biophysical environment (Stedman 2003), i.e. the experience of landscape attributes contribute to the meanings that people hold for a place (Masterson et al. 2017a). However, we also recognize that these place meanings are constrained and constructed through social interactions and discursive practices, towards particular political ends (Stokowski 2002). In other words, we also pay attention to the discursive construction of place meanings through language which can reveal the ideas of legitimate use and management of a place and the future desires that people have for a place (Yung et al. 2003;Di Masso et al. 2014). Therefore, meanings do not compete equally but rather in the midst of hegemonic discourses, structures and power relations and a focus on the discursive presentation of place meanings may help to reveal some of these place shaping structures and relations (Stedman and Ingalls 2013;Ingalls and Stedman 2016).
Dominant narratives shape ideas about the condition and meanings of nature and places. Such ideas held by powerful actors with vested interests are carried through in conservation policies and interventions and can have immense and persistent power (Leach and Mearns 1996;Kepe 2001a). Conservation narratives enlist support for conservation measures by defining the nature of environmental problems (von Heland and Clifton 2015; Berdej et al. 2015). In this paper, we use narratives as accounts that frame environmental problems as well as what solutions and interventions are, therefore, possible (Ernstson and Sörlin 2009;Lidström et al. 2015;von Heland and Clifton 2015). As such, conservation narratives are often based on and make use of specific preconceived notions of place (Hope Alkon 2004;Bell and York 2010), even if these notions are constructed by powerful people far removed from the interests and experiences of a place (Hajer 1995;Neumann 1997;Brockington et al. 2008). Therefore, in this paper, we look into narratives about a conservation intervention in a particular place, as a vehicle through which place meanings are articulated and mobilized.
The picturesque but contested landscape of the Wild Coast in the Eastern Cape province of South Africa is the subject of development and conservation interests with the aims of protecting biodiversity as well as redressing economic neglect under the Apartheid regime. 1 Narratives 1 The Apartheid political system of inequality was broadly based on a racial hierarchy where those who were classified as 'White' were socially superior to those classified as 'Coloured' or 'Black'. This system precluded Africans from all rights normally associated with being a citizen of South Africa, e.g. the right to own land. These laws and government were officially dismantled in 1994, but the legacies of the benefits of community conservation for seeking to redress the historical exclusions of people from their land have been influential in the region (for a discussion of these dominant broader narratives about the region that have influence in local places but originate elsewhere, see the supplementary material). These interventions have largely taken the form of co-management 2 of protected areas and ecotourism, so far with mixed success (Kepe 2001b(Kepe , 2008Ntshona et al. 2010). To examine how such narratives play out in competing claims to local places, we selected one such proposed co-management intervention to establish a protected area where the management would be shared by the provincial government and the local community. During our work in this area, the declaration of the nature reserve was delayed by conflicts among the community, until funding for the initiative ran out. These conflicts are reflected in the emergence of different narratives about the project, which illustrate the different meanings that people associate with particular places in the landscape.
As authors we are sympathetic to community conservation and the possibilities that such interventions may align financial resources of international conservation agencies, with the interests of poor and in this case even transfer legal ownership of forest land to communities from whom it was previously dispossessed. However, we posit that this is only possible when local people and their ideas about nature are a meaningful part of such arrangements. Co-management is often enacted in ways that are forced by top-down interests, and only symbolically 'involve' local actors and restrict their access to customary land (Büscher and Dressler 2012). Therefore, in this paper, we aim to use discursive place meanings to deconstruct co-management and the idea of 'win-win' conservation, as well as debunk the idea of a homogenous community that is often implicit in many socalled co-management projects. In previous work, we have shown how multiple meanings of particular types of places in the landscape (ecotopes) are evoked and created in the narratives shared about the area in response to the project plans (see Masterson et al. 2017b). In this paper, we seek to understand how these storylines make use of place meanings and how this plays out in the negotiations of the project. We are inspired by Hajer's (1995) mode of discourse analysis which looks into narratives, storylines and particular linguistic devices employed by coalitions of actors to strategically align with powerful interests. Therefore, we ask: How do coalitions of actors make use of place meanings and how do these fit within the broader narratives of conservation in South Africa? What tensions arise within the negotiations around this intervention and what are the implications for community conservation?
The article unfolds as follows: first, we describe the conservation intervention, 'the project', under study and how we recorded narratives about this intervention through interviews and how these were analysed. In the first part of the results, we summarize the different local narratives about the proposed conservation intervention and the storylines used to attribute meaning to places, with focus on how they were used to negotiate the project [for further details, see Masterson et al. (2017b) which analyses the heterogeneity of place meanings and their links to landscape complexity]. Then we present our analysis in which we identify coalitions of actors that reveal underlying processes of negotiation of the project and ultimately shed light on how the counternarrative stalls the project. We give particular attention to place meanings in narratives and highlight how groups may have multiple meanings, hold these in tension and how this may change over time. We end with a discussion of the implications of these tensions and competing meanings for pathways towards sustainability. What can the strategic coalitions of actors around particular narratives tell us about the meaning of conservation interventions in these affected communities on the Wild Coast? This begins to address the research gap in the intersection of place research and sustainability science identified by Masterson et al. (2017a) to specifically examine whose place meanings are favoured and the implications of these power dynamics for social-ecological systems.

Case study
The case study takes place in a contested social and ecological landscape, namely the overlapping regions of the Transkei and the Wild Coast. The Transkei is a former homeland or Bantustan: a region created and set aside by the Apartheid state for indigenous South Africans on the basis of an assumed 'ethnic' homogeneity. 3 The Transkei was one of the two Bantustan areas designated to Xhosa-speaking people, and largely designated for subsistence agriculture. This undeveloped, mostly rural area has high levels of poverty and unemployment and low service delivery due to historical systematic neglect under the Apartheid policy of separate development. The broader area known as 'emaXhoseni' continues to be an area of cultural belonging for isiXhosa-speaking groups. The Wild Coast refers to a coastal area of rugged land-and seascape which includes the Pondo-Albany Thicket biodiversity hotspot, a national biodiversity priority area (DWAF 2004). The iconic Wild Coast has been enrolled in ecotourism branding aimed at 'White' 4 middle and upper income tourists (see supplementary material for further detail). These names begin to reveal the competing place claims which play out in an area suffering from the legacies of a complex ethnic and racial history. Conservation in the region has historically excluded local people by dispossessing people of their land, denigrating local knowledge and management practices, and vilifying traditional resource use as 'poaching'. This has meant that conservation is largely seen as a typically White concern (Carruthers 2006;Graham 2017). Contemporary conservation agendas have sought to redress these injustices of Apartheid, through the inclusion of previously disadvantaged (Black) communities in community conservation and co-management which poses significant challenges in the context of persistent racialized relations in these areas so deeply scarred by Apartheid policies. In the supplementary material, we present a short review of the narratives of conservation and development in the broader region.
The empirical study was conducted in the three villages 5 adjacent to a state forest in the Mnquma Municipality in the Eastern Cape Province. These three villages are made up of a total of approximately 4000 people (Statistics SA 2012) and form one traditional authority. The state forest is managed by the provincial forestry authority. The community land around this forest is a grassland-savanna-forest matrix. This land is under customary land tenure, which means that user rights to land are allocated by the traditional authority. 6 In 2009, an internationally funded conservation project aiming to extend the protected area network of the Eastern Cape earmarked this particular state forest and surrounding areas, among a number of other sites, for establishment of protected areas. At the study site, this entailed the transfer of ownership and management of the state forest land from the provincial forestry authority to a co-management agreement between the provincial conservation authority, the provincial forestry authority and the three local communities. The proposed nature reserve would comprise the existing forest reserve as well as areas of community land (including patches of forest, grassland, abandoned fields, and pasture). The project included the fencing of this area, and the possibility of importing large game, particularly buffalo, which would be supplied by the conservation authorities. The proposed reserve presents the opportunity to breed disease-free buffalo which fetch very high prices at wildlife auctions as they are popular for tourism and hunting farms. Diseasefree buffalo are more valuable as they may be moved to other places once sold (Hunt 2010). The project also had the explicit aim to stimulate development of local communities at nodes along the Wild Coast, including providing employment opportunities within the operation of the nature reserve.
As will be discussed below, reservations and disagreement in the community delayed the co-management agreement and the declaration of the nature reserve and in August 2013, the project's 5 years of funding ended terminating the employment of the project staff and local villagers responsible for clearing alien invasive plants, and the local rangers.

Data collection
This study is based on empirical data collected through qualitative interviews, and participant observation over 6 weeks in February and March 2014, 5 months after the project funding had ended. The first author conducted semistructured interviews with 27 actors in the local communities and organizations involved in the project [see Table 1 in Masterson et al. (2017bMasterson et al. ( , p 1446]. A part of this dataset was also used by Masterson et al. (2017b) to analyze place meanings in relation to ecological landscape heterogeneity. Here we summarize the data collection and emphasize aspects 5 The names of which we have not revealed to protect the anonymity of informants. 6 This system of informal rights arrangements and customary law for land are governed by chiefs (an inherited leadership position reserved for men) and their headmen at the village level in the former Bantustans. These 'traditional authority' institutions were heavily interfered with and influenced by colonial and apartheid administrations. As a result, land is under insecure tenure arrangements and the jurisdiction and interaction of chiefs with elected municipal ward councilors serving on local government institutions established in post-apartheid South Africa is often a source of contention (Oomen 1999;Ntsebeza 2004). 4 While contemporary international audiences may be uncomfortable with the use of the terms 'white' and 'black' as racial categories, we use these terms as they have contextual relevance to the analysis. The Apartheid regime imposed a classification of its subjects into racial categories, and a legacy of this is that such categories have become part of the lived experience of South Africans (Posel 2001). However ambiguous the ontological and moral underpinnings of such a racial classification are, to reject these categories would mean overlooking how respondents in this context construct their identity and relationships to others, and the contemporary reality of the lived experience of South Africans (Brandt 2013). Respondents use the English or isiXhosa terms for 'black' and 'white' and we, therefore, are compelled to use these terms in our analysis as terms of relevance for respondents in their making sense of this conflict. of the data collection that are particularly relevant for the analysis of coalitions and negotiations of the project.
We employed a purposive sampling strategy, first selecting interviewees to represent the different stakeholder groups who through pilot work we identified had roles within or interacting with the project such as community groups, local people employed by the project, external people working with the forestry and conservation agencies and representatives of the traditional authorities. Additionally, we employed purposive snowball sampling (Kvale 1996) to also sample a range of views of the project within and outside of these stakeholder groups. We interviewed three central actors in the provincial conservation department (non-local to the village, two White men and one Black isiXhosa-speaking man); four (out of ten) non-local amaXhosa employees of the provincial Forestry Department who were stationed at or visited the reserve; three of the twelve community forestry guards employed by the project; three project staff engaged with the community primarily through an elected community committee; and six of the 21 members of the community Participatory Forestry Management Committee (PFMC). We also interviewed the traditional authorities in the area including the chief, one of the village headman, and two sub-headmen. We also undertook seven formal interviews with community members who do not fall into any of the other categories. All of the local community members interviewed were isiXhosa-speaking people who identified themselves as local. Through snowball sampling we interviewed individuals who other interviewees told us had views on the project opposing their own. At first, a few community members opposing the project did not want to be interviewed as they did not want to be associated with the project in any form, and there was some distrust of the first author as a white non-local. Every effort was made to build trust with community members with the first author spending 6 months in the community talking with people about other non-project-related research and attending community events. Additionally, time constraints meant that we could not interview everyone with an opinion on the project, but we saw an acceptable level of thematic saturation within interview responses indicating that we interviewed a good range of views, and we supplemented these data with informal interactions in the community.
Interviews first engaged the interviewee's perceptions of the landscape and particular places, and particular emotions or uses attributed to these to get at place meanings (see Masterson et al. 2017b supplementary material for interview schedule). Interviews particularly focussed on opinions about the project through a short narration about the interviewee's involvement with the project and his or her understanding of its aims and impact, including perceptions of how the project was perceived and played out in the community. During these interviews we asked interviewees whether (and how) their perceptions of the project and of the landscape changed during the project. Four interviews were conducted in English, and the rest of the interviews and discussions were conducted in isiXhosa through an interpreter. Communicating around ideas of 'sense of place' can be difficult in one's own language, let alone in a cross-cultural context, and so the first author and interpreter made use of previous research with the same communities to find appropriate isiXhosa words with which to discuss the landscape and people's relationship to it. Through this previous research, we had spent many months walking and driving through the local landscape with local informants, which helped us to describe a local endogenous biocultural classification of types of place which was used as a basis for asking the interviewees about what emotions they associate with types of place (Masterson 2016). We have retroactively termed this classification as 'ecotopes' for the scientific audience (see Masterson et al. 2017b for further details).
These data were supplemented with observations from shorter and informal interactions with community members as well as attendance at four community meetings. Participant observation during these meetings as well as informal discussions on the street or at community events, allowed the first author to observe relations of trust and distrust within the community as well as the ways in which community members, project employees and the PFMC presented the project and the landscape in informal discussions. The majority of these meetings, and informal discussions occurred in the local language isiXhosa and were interpreted afterwards for the first author by our local interpreter, but notably, one official meeting between the PFMC and the project staff was held in English despite the fact that many local PFMC members were not fluent in English. These data were recorded in a notebook, and both used to inform the choice of interviewees, as well as used to triangulate the interview data in the analysis of coalitions of narratives.

Data analysis
In the process of conducting interviews and attending meetings and informal discussions, we inductively identified three broad narratives about the project and the landscape: two narratives in support of the project, and one in opposition (Masterson et al. 2017b). In our analysis, narratives include storylines and arguments that make use of place meanings. In fact, narratives both make use of existing place meanings and are implicated in their construction and negotiation as actors make claims about place and what they think places should be. Sections of the transcribed interviews were deductively coded (Saldaña 2015) by the first author (with co-authors validating coding through regular discussions) for storylines used to argue for or against the intervention, as well as how storylines made use of particular meanings of landscape units or 'ecotopes' that are found within the proposed nature reserve boundaries. Place meanings were identified from statements about the emotions, uses and symbolism attributed to the landscape and particular ecotopes through interviews and informal discussions. In this paper, we present the narratives emphasizing how they are used by different groups in the process of the project development. For the purpose of the analysis of this paper, we focus on three categories of biocultural landscape types (simplified from the six ecotope categories from an endogenous local isiXhosa classification of the landscape presented in Masterson et al. 2017b We then analyse the narratives to understand the alliances and process of negotiation of the project outcomes in the community, with a view to exploring what meanings may reveal about underlying intentions and views in communities with heterogeneous interests. To do this, we conducted a second round of open coding of the interview data, to identify coalitions and tensions in these coalitions. Based on Hajer's (1995) mode of discourse analysis we look to the storylines used to attribute meaning to places, events and the project. Hajer (1995) used these as markers to reveal discourse coalitions of actors: groups of actors who share the meanings, and constructs of the problem and employ these for a particular strategy. In this view, individuals or groups make choices about the storylines and meanings that they use to present why things are the way they are and what needs to be done. In the competition for the problem definition by different narratives, individuals or groups may have to strategically adopt storylines and particular language dissonant with their own values, to argue for their cause. Here we pay particular attention to the ways in which actors represent others' storylines, and position themselves in relation to community conflicts. We also code the data for instances where multiple meanings of place or storylines are expressed by interviewees or in informal discussions and whether or not these are compatible, or where these illustrate potential cognitive dissonances. We also compare the place meanings elicited about the ecotopes in the first part of the interview, with the place meanings espoused in the narratives about the project. Informal interviews and observation at the field site also inform this analysis.

Narratives about the Wild Coast Project
Here we synopsise from Masterson et al. (2017b) the three main narratives that could be distinguished from each other in the data. We present here the storylines and the meanings of place which are employed in each narrative either in support of (two narratives) or in opposition to the project (one narrative). While the narratives (also summarized in Table 1) are not mutually exclusive, each narrative has main proponents and supporters as told to us through the interviews.

Narrative 1: an opportunity for restoration of unique biodiversity and wilderness
The first narrative focuses on the need for protecting unique forest biodiversity and wilderness, which conservation and forestry officials (some of whom are White and non-local) as well as community forest guards claim is threatened by illegal resource extraction and alien-invasive species. To set up the urgency for protecting the forest, this narrative references the increased rate of harvesting and sale of indigenous plants and animals, particularly by outsiders and younger generations. The threats to the forest as a unique biodiversity refuge are described with the use of technical understanding of ecological processes such as concern over the removal of understory forest layer, and specific concern for slowgrowing species with episodic recruitment in forest.
The proponents argue that co-management of the reserve is a solution to overharvesting by including communities in the benefits of conservation. Proponents reference the uniqueness and "irreplaceability" of this place by pointing to its biodiversity and that the forest patches represent a priority area for national and provincial protected area expansion under the National Environmental Management: Protected Areas Act (No. 57 of 2003). Many community members praise the project's work in removal of invasive alien species such as Lantana camara. They refer to people employed to remove invasive vegetation as "cleaners" which builds on a meaning of the forest as pristine and wild. This reifies the meaning of the forest as having been mismanaged and in need of saving, and illustrates how the project contributes to restoration and purification of the forest. The project plan to reintroduce indigenous antelope and buffalo is also used in this narrative to show how the project contributes to "restoration of the natural ecosystem" and the place meaning of wilderness.
Through this narrative community land is presented as suitable habitat for grazing game species, and this is mobilized in arguing for the inclusion of community land in the reserve. This is also based on meanings attributed to this communal land that it is degraded due to previous cultivation and overgrazing.

Narrative 2: hope and development for future prosperity
The proponents of this narrative describe the need for development initiatives by representing communities as suffering under poverty with little access to infrastructure or job opportunities. The project and the protected area are portrayed by project staff as a way to develop the community and as "a catalyst for socio-economic growth". The potential for breeding and sale of buffalo is a very important storyline here. Disease-free buffalo fetch very high prices on the market, and have the potential to attract tourism. Most of the project staff 7 use this as a motivation for the economic sustainability of the project: If [the community] goes into a co-management agreement which has a legal basis, it provides them with some sort of equity which they wouldn't have had before. If you say we're in a deal with [the provincial conservation authority], and they are going to give us 20 buffalo, and in five years' time there are going to be 30 and we're going to auction off six. Government's going to take 3, and community will take 3; it's on a 50/50 basis.
[…] They're equal partners in the reserve itself; there's this kind of equity.
After the project staff took the PFMC (including some of the traditional authority representatives) to a buffalo auction, the high value of buffalo became a very important and often repeated storyline for community proponents of the project too. It involves the specific retelling of the hypothetical example above and how proceeds from the auction of buffalo would be invested in infrastructural assets for the villages. In this way again, conservation and the protected areas also sustain employment for local game guards and justify fencing off the forest to community members.
Fencing the forest also resonates with community members who want to protect the forest as an asset belonging to the whole community. To this end, this narrative makes use of the meaning of the forest and landscape as a scenic and aesthetically beautiful place with the potential to generate income from ecotourism. Some community proponents mention their own experiences of tranquillity in the forest to motivate why tourists would want to visit.
In this narrative, there is a focus on customary tenure of community land. Community members refer to the sadness and regret about abandoned fields that have been encroached by acacias and which people no longer have the financial means to plough. 8 However, they argue that these areas now represent the chance to be part of the project because they are required to support grazing buffalo. In this way, the community land, the plantations and the forest are emphasized as communal assets.

Narrative 3: exclusion and encroachment on ancestral land
The community members who argue against the project claim that the nature reserve threatens ownership and use of community land for grazing and building. This narrative makes use of stories of dispossession of their land during Apartheid to underscore the importance of maintaining ownership of land. Opponents of the project claim that their fathers were successful in resisting Apartheid era schemes to "steal" their land, 9 and, therefore, they have a duty to retain sovereignty over community land. This also draws on the meaning of the plantations which are referred to as ihlathi abelungu, literally translated to "White man's forest", as this resident explains: "Our fathers were cheated. Their land was taken away from them and gum trees were planted there.
[…] Had they known before, they would have said no, because the trees suck all the water to the last drop. It's very dry where they are". This is also employed to demonstrate the meaning of the forest as appropriated land. Opponents describe how they have been excluded for many years from the indigenous forest which has sacred value as the location of benevolent ancestral spirits, denoted by ihlathi lesiXhosa or "Xhosa forest". The opponents claim that the forest remains a place of exclusion since the PFMC captures many of the benefits of the project, especially the investment into micro-enterprise activities. One community member speculates: "seemingly those people were not striving for the good of the whole community but for their own stomachs".
The narrative also draws a parallel with this Apartheid dispossession of land and points to the threat that the project poses to the agricultural meanings attributed to this community land. Each abandoned field is referred to by the 8 Title to arable land is passed down from father to son. However, under the current customary tenure system these de facto land owners have no official title to land and have to plough land to demonstrate ownership. But, amidst a larger process of deagrarianization, it is difficult for many people to plough these fields due to financial constraints. 9 In this area, government created a forest reserve on local and indigenous peoples' land for the purposes of conservation in the 1930s, and planted woodlots in the same forest patch in the 1950s (see Masterson et al. 2017b for further detail). household name and the sacred and symbolic significance of the ancestral heritage and de facto ownership of this land is emphasized. Community land also carries meanings of being suited for a small-holder traditional agricultural lifestyle and precious grazing land for livestock. There is also emphasis on how the buffalo would endanger community members (the unpredictable nature of buffalo makes them very dangerous to humans).

Coalitions amongst actors
Here we examine in detail the coalitions of groups and individuals who use the same meanings and storylines within the narratives above, to argue for or against the project in these villages. We examined how actors who strategically make use of a particular storyline may also carry alternative meanings, dissonant with the dominant narrative for and against the project.

Coalitions in support of the project
Narrative 1 (conservation and restoration) brings together the project staff, conservation and forestry officials and community forest guards who argue for conservation of the forest particularly. The project staff and conservation officials are educated White men, who do not reside in the area. As part of the project operations, they represent the authority on the project aims and implementation and are dominant in shaping this narrative. Additionally, they each have a role in providing access to the project funding for this community. These actors employ the meanings of unique biodiversity and vulnerability of the forest to exploitation by local (Black) people in poverty, mirroring mid-century race-biased narratives of degradation of land. Part of their narrations focus on sustainable use and the need to include local people in the management of local resources through the project, mirroring the community-based conservation rhetoric that has gained popularity in South Africa in the democratic era (see supplementary material) and this is particularly important to the conservation and forestry authority officials. Community members who are employed as forest guards (the majority of whom are women who were unemployed before the project) and some other community members are convinced by this narrative. This traditional healer who made use of the forest extensively for medicinal plant harvesting was trained by the project as a community forestry guard and describes how her meaning of the forest changed: How I viewed the forest changed after I was employed. At first I thought the forest was just there for me to do with whatever I felt like. For example, if I wanted a bundle of firewood, I used to just go into the forest and get it. But now it's different because I have been taught a lot about the forest. […] Now I know that it's wrong to go there and do whatever you feel like, debarking the tree the whole way around the trunk and killing it. I was unaware of that fact.
The project staff and forest guards also express the desire to protect "wilderness" and restore the natural system and biodiversity. These meanings reveal traces of fortress conservation notions in the urgency of protecting biodiversity resources and not allowing use of threatened forest resources. This reveals a tension in this narrative coalition: for project staff, there is acknowledgement of local communities' need to use the forest for medicinal plants, and cultural rituals, but this is eclipsed by the notion (and old racist narratives of degradation) that the unique forest and particular species are under threat from poachers and illegal harvesting of medicinal plant resources for sale in markets in the cities. Interestingly, the exclusion of communities from the protected area is also supported by the community forest rangers. These traces of fortress conservation are also reflected in the way this community forestry guard privileges technical and expert knowledge in management: "What could the community do with the forest? They would only destroy it. They can't manage this place. They must go to school and learn more about nature. Otherwise they can't manage it". This group of forestry guards thus set themselves apart from the rest of the community who they view as implicated in destruction of forest resources, by referring to their own education and awareness and effectively arguing for the importance of their own jobs guarding the forest.
The project is presented by project staff as a win-win solution for conservation and development of communities with an emphasis on the protected area as a catalyst for development and "spinoffs from conservation" (as one project staff member described it). Through the storylines of narrative 2 (development), the project staff makes use of the storyline that the forest and abandoned fields will be used as equity in the project and for the communities' benefit, which convinces many community members. Harnessing the dominant and powerful broader discourse of win-win interventions means that proponents of narrative 2 find themselves in alliance with proponents of narrative 1, and a conservation agenda. To argue for any development benefits, they must argue for a protected area with buffalo, which limits access to land for the surrounding communities. On the surface, narratives 1 and 2 appear to be compatible, for example, the desire to protect wilderness and biodiversity (conservation narrative 1) could be compatible with the desire to protect the forest as a community asset that could be used for ecotourism. However, this win-win discourse coalition also presents a tension for community members who use the power of both narratives to argue for the project in that their meanings of scenic beauty of the forest are aggregated with the ideas of threatened wilderness which consider Black communities a threat to the landscape. This mirrors a colonial discourse still prevalent in conservation which Neumann (1997) refers to as 'good natives' and 'bad natives: the closer local people are to nature, the 'better' the more deserving they are of economically benefiting from such initiatives, but the more 'modern', the more they pose a threat to conservation.
Another tension is particularly evident in the stories told by the elected community group, the PFMC, who were the community liaison for the project (and, therefore, also the gatekeepers of the information about the project to the community), and were benefiting from the micro-enterprise activities. In describing the benefits of the project for development, this group presented the forest and the plantations as a development asset and repeated the storyline that buffalo breeding and sale would benefit the communities. However, this is held in tension with the view of the forest as critical to the community and their well-being through the provision of forest products such as poles, medicinal plants, and the cultural and aesthetic importance of the forest for rituals and spiritual well-being. For example, this PFMC member walks in the forest almost everyday and says, "That forest is so important to me. My life. My health. Even the breeze there makes me feel relaxed". But when asked how the forest should be used, he responded, "In order to develop the community through the forest, we should not allow people to go as they please, and do what ever they feel like there, damaging the forest". There was little attempt by this group to reconcile the ability to walk and experience the forest (also required for hiking based tourism) with the presence of dangerous buffalo. 10 To argue for the development benefits of the project, these community members strategically align with the narratives of the project staff and the economic benefits of contributing and giving up community-owned land for conservation. This is despite their cultural attachment to the forest. This is a significant tension in this coalition mired further by assumptions about the superiority of a Western worldview, as illustrated by this White project staff member who attempts to discredit the significance of cultural use of the forest by stating that the importance of cultural practices would always be second to the need for employment: I do believe that there are strong cultural connections to the forest which are real, but I think that the need to generate income overrides any bloody thing they like. If you were to say you've got to clean that [ancestral] grave away here because we want to build a hotel which can employ 40 people, they'll put up a bit of a stink but they'll move the grave. They want the hotel, you know.

Changing allegiances: opposing the project
Interestingly, the traditional authorities and some powerful community elites first made use of narrative 2 and its place meanings to argue in favour of the project, describing the forest as "a place that can help communities develop" through jobs and projects. However, by the end of the project, the traditional authorities had turned against the project. The traditional authorities explained this change of position as a loyalty to their constituents who opposed the project. However, the project staff and the PFMC cast this in a light of political interference as a project staff member explains here: The opposition to the forest being expanded onto communal land comes from a politician who has an influence over [the traditional authority]. There are also a number of plans proposed by a developer, which include building a mall and low cost housing, as well as a hotel on the land which would have fallen within the proposed reserve and the Coastal Conservation Area.
Here, place meanings of the coastal forest indicate a tension and consequent shift in allegiance. The traditional authority presents the coastal forest as aesthetically beautiful and with tourism and development potential. However, this conflicted with the idea of the coastal forest as protected as a Coastal Conservation Area 11 (over which the traditional authority does not have jurisdiction). The traditional authorities who were invested in the plans for the mall (colloquially known as 'the mall project'), began to emphasize the place meanings of narrative 3 of the coastal forest and community land in that area as belonging to the community.
With the support of the traditional authorities, the counter-narrative could no longer be ignored. It is this narrative that eventually succeeded by stalling negotiations around the project. 12 However, in this counter-narrative too, actors do not necessarily have the same agendas, despite using some similar storylines to argue against the project. Many community members align with and benefit from the support of the traditional authority, but do not support the mall project. Many of the more vociferous community members behind the third narrative are cattle owners. Instead of the mall, these community members argue for importance of agriculture and grazing land to the community, as well as for the freedom of use and access to community land under insecure tenure.
But there is also tension in the narrative opposing the project which argues against restricted access to land that would be part of the protected area. While they refer to the sacredness of the forest and a desire to use the forest, they also express concern that people are destroying the forest through overharvesting. This respondent, Mr M, who argues against the project, concedes that the state forest portion of the land belongs to the state now and that the community has no power to do anything about the access that they would lose if a fence were to be erected: That forest is a good place, to me it's a sacred place. There are species there that they need for customary rituals and the species there in the forest are used when addressing the spirits. [Fencing the forest for a reserve] will affect people because if it's fenced, people won't get access to the forest. We just agree because we can't do otherwise. If you take heed of each thing that will affect people, that forest would never be fenced.

Potential common ground
The tensions in place meanings presented above, reveal some potential common ground for these opposing coalitions. As we have shown, individuals and groups hold more than one meaning of each part of the landscape (Table 1), and often hold these in tension. Underlying and sometimes despite the meanings co-opted and created through each of the opposing narratives, we observed some commonalities in personal significance of places within the landscape, particularly for the forest as a tranquil and soothing place restoring well-being. Interestingly, these meanings of tranquillity were reported by actors across agendas, genders and race. For example, this isiXhosa-speaking female forest ranger describes how: "the forest is important to me, and how I feel. My health changed due to the atmosphere there. I learnt about animals. The different scents inhaled here make your body invigorated". Mr. S, the male isiXhosa-speaking PFMC member mentioned earlier, ascribes great personal significance to the forest and the relaxation of walking through the forest, a meaning echoed by one of the female traditional leaders: "The forest is a different world, a different atmosphere-listening to the birds. I go there when I'm stressed". An opponent of the project, Mr M, also mentions the unique species and desire to protect to the forest. And, the English-speaking conservationists also spoke of the personal significance of the forest and a sense of well-being that it brought as well as how it was "unsettling to see the forest in decline". We identify this overlap in meanings across actors, as a potential leverage point for future negotiations.

Discussion
The discourse coalition of win-win conservation: internal tensions Through our exploration of this process of the negotiation of the project, we have shown how meanings may change over time, in strategic ways for particular agendas. By examining coalitions in this process we have also identified instances where groups had to adopt a storyline or meaning, dissonant with their own values, to argue for a desired outcome. Actors arguing for the project and nature reserve declaration, make use of the win-win discourse of conservation and development (Chaigneau and Brown 2016). This is first described through involving the communities in co-management of a protected area, reflecting the popular discourse of community-based conservation in post-Apartheid South Africa (e.g. Fabricius 2004). However, this rhetoric of eco-modernist sustainable use ideals and co-management (cf. Hajer 1995), is betrayed by proponents who desire the protection of wilderness by a fence. We have shown the prevalence of fortress conservation ideas used by project staff and forest rangers alike in arguing for the preservation of wilderness and the exclusion of local (Black) people. This mirrors a trend in the broader conservation discourse "back to the barriers" of colonial fortress conservation (Hutton et al. 2005;Büscher and Dressler 2012). Despite an emphasis on participation and benefit sharing, many community conservation interventions replicate old forms of coercive conservation (Neumann 1997). Advocates of this neoprotectionism argue for the immediacy of the need to protect the intrinsic value of biodiversity through protected areas as safe havens, and use this to justify the exclusion of local people for the good of the environment and biodiversity (Hutton et al. 2005;Büscher and Dressler 2007). Interestingly, here, the project proposes to bring in buffalo to restore "wilderness" which reinforces the need for a fence. This also reflects the notion that wilderness must include large game (and thus requires large fences) which pervades the conservation sector in (South) Africa (Brockington et al. 2008).
Here the influence of neoliberal ideologies in the realm of conservation in southern Africa is also evident. Economic benefits are particularly important in motivating protected areas in the vicinity of the poverty and underdevelopment attributed to the former homeland areas (Kepe 2008). The proponents of the project (both project staff, officials and community members) refer often to the poverty and underdevelopment of local communities in this area. But, as we allude to, the win-win discourse of protected areas prevalent in southern Africa focuses on the profitability of protected areas through ecotourism and investment in wildlife (Ramutsindela 2007;Büscher and Dressler 2012;Holmes and Cavanagh 2016). Importing buffalo to this proposed reserve fits the bill of win-win first by restoring biodiversity and second by providing economic benefits to communities outside of the protected area. The rhetoric of economic benefit of buffaloes through ecotourism potential and breeding 13 replicates the development paradigm of the private wildlife industry, which excludes communities from land (Brooks et al. 2011;Spierenburg and Brooks 2014) replacing local livelihood needs through other means (Ramutsindela 2007). However, our results demonstrate that for many people who argue for the project, these economic benefits are attractive, but the lack of access to the forest in the protected area would be detrimental to their spiritual and cultural well-being.

Community resistance to win-win conservation: implications for sustainable futures
Examining the process of negotiation of the project proposal through the lens of narrative coalitions and meanings allowed us to see how the heterogeneous interests of the local communities played out and how the project failed to take place. As we have shown, the win-win discourse of conservation through this project had broad support within the communities (including the traditional authorities). However, the counter-narrative gained momentum after the traditional authorities aligned with it, and succeeded in stalling the project negotiations. This is particularly interesting because this counter-narrative competes against the internationally popular win-win conservation discourse as well as powerful actors backed by the state. The beneficiary communities are not in a position of power in terms of the control of funds (a situation that many communities in the vicinity of protected areas face, for example, see Bologna 2008). What can be learned from the counter-narrative for both future community conservation interventions, as well as alternative pathways of development?

Focus on economic benefits of protected areas obscures alternative pathways of development such as agriculture
First, this resistance demonstrates that the dominant model of economic benefits from protected areas obscures alternative framings of development. The pro-project narratives framed the community lands as degraded due to overgrazing and encroachment on abandoned cultivation sites, referencing the large decline in cropping and cattle ownership in the area. However, this neglects the importance of an agricultural lifestyle and identity to these communities evident in narrative 3, where community land and abandoned fields represent hope for an agricultural future (see also Shackleton and Hebinck 2018). This view to retaining abandoned fields in the hope of future agricultural endeavours is discredited as "backwards" (traditionalist) or as "sentimentality" by the proponents of the project, and these proponents represent community land as equity for the community in the development project. This is an illustration of the neoliberal pressures on community interventions that Büscher and Dressler (2012) identify as a shift away from local constructions and meanings of nature, to what the environment could mean in terms of capital and global markets.
Interestingly here, both this development-through-protected-area discourse as well as the dominant agricultural discourse in South Africa which favours commercial agricultural ventures (Hall 2009) neglect small-holder agricultural development as a strategy for poverty alleviation. In contrast, the counter-narrative contributes to and draws on a less powerful discourse that supports small-holder agriculture for food security (Hebinck and Lent 2007;Ntsebeza and Hall 2007), which makes it all the more interesting that the counter-narrative in this community was successful. The counter-narrative also makes reference to a hotel and mall development in an area that overlaps with the proposed protected area. While this seems to be the agenda of only a few community members (particularly the traditional authorities and an advisor), it does offer an alternative mode of development based on and maintaining the sovereignty and control of customary lands by local and indigenous people, problematising the dominant model of fortress-protected areas for development.

The focus on economic benefits from protected areas neglects other meanings of the landscape
Second, the success of the counter-narrative demonstrates that the project intervention has neglected alternative meanings of the landscape to the communities. The dominant win-win conservation and development paradigm focuses heavily on the need for conservation and economic compensation or benefits from loss of access to land. Büscher and Dressler (2007) attribute this commodification of nature through protected areas to the widespread assumption that rural people do not want to use or should not be allowed to use resources; reasoning which resonates strongly with storylines from the project staff in this case. This mirrors an often debunked but still lingering colonial degradation discourse of generalized degradation of biodiversity at the hands of poor resource-dependent Black communities (Maddox 2002;Hajdu 2009; and see supplementary material). Interestingly, these conservation narratives which we have shown have such power in this local place are not formulated at the local level but rather by powerful non-local conservation interests (Hajer 1995;Brockington et al. 2008). The dominance of western images of conservation and indeed of White conservationists in conservation rhetoric in southern Africa, and the power of these ideas to pervade community conversations are evidenced by the pro-project narratives actively discounting the alternative meanings of the forest and land to local people (see also Masterson et al. 2017b).
The counter-narrative on the other hand employs the cultural and spiritual significance of ihlathi lesiXhosa to local people. The view of forests as ihlathi lesiXhosa as sacred spaces where the ancestors reside and a source of medicinal plants and resources is widely valued by amaXhosa in the Eastern Cape Cocks et al. 2016). The discrediting of this cultural significance of the forest to motivate the exclusion of people from the forest reserve is based on the assumption that economic benefits will always trump cultural values. However, as we have demonstrated, the proponents of the project (including project staff and authorities, and community members) also have strong cultural and personal values for the forest beyond economic value that motivate their desire to protect and experience the forest. Cocks et al. (2012) argue that the importance of cultural relationships with land and nature can be a strong motivation for conservation of forest and thicket in the Eastern Cape. The same argument for the value of cultural and personal values in motivating stewardship of indigenous lands has been made many times (Tengö et al. 2007;von Heland and Folke 2014;Comberti et al. 2015;Bologna and Spierenburg 2015).
Strikingly, the win-win narrative of the proposal for a protected area neglects the historical context of these communities, who were dispossessed of land when colonial and Apartheid governments seized indigenous forests. The project actors simplify the complexity of the informal system of tenure by arguing for the inclusion of community land into a fenced-off protected area. But as narrative 3 argues, this neglects the importance of de facto ownership and heritage of abandoned fields ploughed and maintained by individual families. The project claims to provide economic benefits for the whole community, but as narrative 3 shows, this does not take into account the significance of loss of land to individual families, and the struggle they face to retain sovereignty of land under diverse informal land rights arrangements governed by traditional authorities (Fay 2005). Therefore, the use of "community" in the project narrative that argues for co-management and community benefits replicates coercive forms of conservation and pays no attention to the equity of distribution of these benefits or losses to individuals in the community (common in other community conservation interventions in southern Africa (Neumann 1997;Bologna and Spierenburg 2015)). Additionally, the failure of this intervention complements other studies of co-management in the former Transkei that illustrate conflicts and ineffective institutional arrangements around successful land claims managed as protected areas (Palmer et al. 2002;Kepe 2008;Ntshona et al. 2010).

Conclusion
It has been suggested that conflicts about land-use interventions can be understood as conflicts about the meanings of place to which opposing parties are strongly attached. Here we have demonstrated how meanings of place are mobilized through narratives and in the context of broader discourses. By examining coalitions and the way in which this reserve negotiation played out, we were also able to demonstrate the heterogeneous interests of the communities involved and the meanings they attribute to land. Although not the main focus of the article, the results also demonstrate the ways in which issues of race permeate community conservation in South Africa, through the hegemony of degradation narratives and an insidious move back to fortress conservation. Place meanings and the narratives told about this project also reveal how deeply racialized relations between communities and conservators are in the region. Issues of race are still alive in the ways that people think and talk about each other and the meanings that they attribute to place, and this can pose significant challenges to co-management conservation arrangements (Graham 2017).
The success of the counter-narrative suggests that failure to recognize local meanings of nature and place can risk interventions, as well as obscure the legitimate claims to sovereign land and alternative pathways of stewardship, through agriculture, for example. Engagement with the full range of meanings of place may help to disentangle the trade-offs between conservation and development, particularly in the popular win-win conservation discourse. We suggest that attention should be paid to the agricultural heritage values to which people are attached: these underlie a desire to continue to practice agriculture for lifestyle reasons. We also demonstrate how some of these meanings, particularly around the tranquillity and uniqueness of the forest, are shared amongst the diversity of actors both local and from the provincial conservation and forestry authorities. This suggests that in the future, transparent and respectful dialogue that is encouraging of a plurality of meanings and innovative development solutions, offers hope for the development of a more equitable and sustainable stewardship of this land. However, achieving conservation and development goals will require dealing with historically racialized relations and claims to place that intersect in the management of this contested land.