Special section : Gorongosa National Park , Mozambique Narrative Fortresses : Crisis Narratives and Conflict in the Conservation of Mount Gorongosa , Mozambique

A single narrative about the Gorongosa Restoration Project (GRP) in Mozambique circulates widely in the popular media. This story characterises the project as an innovative intervention into an ecological crisis situation. The narrative hails the project’s aim to use profits from tourism to address the goals of both human development and conservation of biodiversity, and portrays the park project as widely embraced by long-term residents. This representation helps the project attract broad acclaim, donor funding, and socially conscious visitors, yet it obscures the early emergence of unified opposition to the project’s interventions among long-term residents of Gorongosa Mountain. This article draws on ethnographic research conducted on Gorongosa Mountain between 2006 and 2008 to examine the project’s early activities there. I examine two crisis narratives that led to entrenched conflict between park-based actors and mountain residents. Focusing on the emergence and solidification of divergent narratives—narrative fortresses—about the extension of the park’s activities to Gorongosa Mountain offers insight into the powerful role of crisis narratives in producing and maintaining conflict, leading to outcomes counter to the desires of conservationists. Ultimately, the article points to ways in which narratives of environmental crisis work against aspirations of partnership and collaboration with resident populations in conservation and development schemes.


INTRODUCTION
A single narrative about the Gorongosa Restoration Project (GRP) in Gorongosa National Park (GNP) has gained widespread circulation in the popular media. This story characterises the project as an innovative conservation and development success story in the making-a win-win combination that is celebrated by conservation biologists, Mozambican government officials, foreign tourists, and long term Gorongosa residents. However, this feel-good story belies a much more complex situation on the ground, including animosity and conflict between residents of Gorongosa Mountain and park actors. In order to better understand the nature and emergence of this heated and ongoing conflict, this article analyses how crisis figured in narratives that oriented differently positioned actors. Key to understanding the conflict is how crisis claims in both sets of narratives produced rigid positions-what I will refer to as 'narrative fortresses'where possibilities of compromise or collaboration were foreclosed, ultimately leading to outcomes counter to the desires of both conservationists and mountain residents. Analysis of crisis narratives in this conflict suggests that conservation and development schemes motivated primarily by stories of environmental crisis may be fundamentally incompatible with aspirations for collaboration or partnership with resident populations.
Below, I examine different narratives that circulated from 2005 to 2008, related to conservation and management activities in GNP. During this period, interventions to rehabilitate GNP were managed and conducted through Greg Carr's 1 philanthropic organisation, The Gregory C. Carr Foundation 2 and included significant efforts to extend park management and tourism activities to Gorongosa Mountain which, at the time, lay more than 20 km outside the boundaries of GNP. From the very first interaction between park actors and mountain residents in early 2006, the Carr Foundation's desire to incorporate Gorongosa Mountain into its conservation and development project was met with significant opposition. At this time, two crisis narratives, one figuring mountain residents as a threat to the ecology of the mountain and hence the region, the other figuring the park and its representatives as threatening outsiders hungry for land, came into conflict, and, like oil and water, repelled each other and remained separate.
Numerous scholars of conservation and development projects have observed how narratives emerge in conflicts over land and resources (Peters 1994;Fortmann 1995;Leach and Mearns 1996;Campbell 2002;Hutton et al. 2005;Dressler et al. 2010). As discursive strategies, such narratives often emerge as ways to define and claim resources (Peters 1994;Fortmann 1995). Louise Fortmann has argued that narratives should be seen not only as statements or claims, but also as "a medium through which… events are produced" (Fortmann 1995(Fortmann : 1054. She identifies how, in the context of land disputes, stories serve as discursive strategies that do at least three kinds of work: "to create meaning and validate action; to mobilise action; and to define alternatives" (Fortmann 1995(Fortmann : 1054. This article builds on this work, drawing added attention to how evaluations of crisis and convictions of urgency contribute to a type of intractable conflict that frequently emerges in conservation and development projects. Drawing on Janet Roitman's examination of the power and prevalence of the notion of 'crisis' in both popular and academic thought (Roitman 2013), I focus attention on understanding "the kinds of work the term 'crisis' is or is not doing in the construction of narrative forms" (Roitman 2013: 3). From this perspective, the goal here is not to evaluate the relative truth or falsity of the crisis narratives I will discuss below, but rather to examine how differently positioned actors' evaluations of crisis shaped early interactions between Carr Foundation actors and their supporters and mountain residents, producing an immediate and enduring conflict. As Roitman demonstrates, crisis "regulates narrative constructions… [allowing] certain questions to be asked while others are foreclosed" (Roitman 2013: 94), and "[evokes] a moral demand for a difference between past and the future" (Roitman 2013: 8). Indeed, in the case of Gorongosa Mountain, the moral demands inherent in different narratives of threats to land and resources on the mountain led differently positioned actors to relinquish opportunities for negotiation or compromise, narrowing the range for potential actions and leading actors on either side of the conflict to transgress norms when faced with obstacles.
A central focus in my analysis is examination of the ways in which differently positioned actors' profoundly different understandings of threats to land and resources on Gorongosa Mountain became rigidly divided. For actors on both sides of the conflict that emerged about Gorongosa Mountain, crisis claims intensified feelings of uncertainty about the future of the land on the mountain. The element of urgency central to these narratives led to hardened positions that shaped interpretations of and responses to the other side's words and actions, making positions on either side quite inflexible. My term 'fortress narratives' is shorthand to describe this situation, which plays on the term 'fortress conservation' (Brockington 2002). While, currently, with thousands of human inhabitants, the GNP's extension to include portions of Gorongosa Mountain is not an example of 'fortress conservation,' in fact, tight fortification of information and thick barriers to the affected residents' real involvement in park planning or decision making have created narrative fortresses impermeable to change or influence. My use of the term 'fortress narratives' also points to the connections between divergent understandings of threats to land and resources, and the production of fortress-style management practices in conservation programmes.

METHODOLOGY
This research draws on over five years spent living in Gorongosa District. For three and a half of these years I was conducting ethnographic research split between Gorongosa Mountain, in an area that is inside the newly designated boundary to the GNP, and in the district capital. While the Mozambican government announced their official decision to expand the park's boundaries to include the area above 700 meters on the mountain in 2010, after the initial period of this research, the possibility of this expansion raised the spectre of dispossession for mountain residents throughout the main research period (2006)(2007)(2008) (Schuetze 2010). During this time I conducted countless interviews with differently positioned residents of the Khanda régulado 3 on the mountain and participated in numerous meetings convened by park based actors about the park's planned interventions on Gorongosa Mountain. At each of the meetings, I kept detailed field notes, made and transcribed digital audio recordings, and when possible, 4 made video recordings. Most recently, I have resided on Gorongosa Mountain in June and July of 2011. All of this work has afforded me a long-term view of the area and its transformations.
While my long-term connection to Gorongosa affords me privileged insights into the unfolding of Gorongosa Park's relationship with mountain residents from the time of Greg Carr's initial involvement, I also recognise that my perspective is more richly informed by research conducted on Gorongosa Mountain than it is by deep, long-term ethnographic engagement with members of the park's leadership based primarily in the Chitengo camp inside the park. However, I was in contact with members of the Carr Foundation and the park leadership beginning in 2005 and maintained this communication throughout the course of this research. Although my research focused on the lives of mountain residents, it is my hope that I am able to highlight aspects of both park-based actors' and mountain residents' narratives about the expansion of the park to Gorongosa Mountain that offer greater understandings of both sides of the conflict.

A magic bullet for a crisis: a saviour and a gardener of Eden
The most widely circulating narrative about GNP is a familiar win-win conservation and development success story that has gained popular media attention. From the very start, The Carr Foundation's involvement in GNP has attracted a great deal of media attention. Greg Carr's success in business, his subsequent status as a millionaire, and his philanthropic ventures have served to raise his public prominence and to strengthen the authoritative quality of his efforts. His venture into conservation has tapped into the magic combination of celebrity and environment, allowing him to be counted among a group of people Brockington has termed 'celebrity conservationists,' whose active commitment to environmental causes greatly increases their fame and public profile (2009). The narrative of Carr's philanthropic venture has been featured in widely circulated glossies such as National Geographic Explorer, Forbes, The New Yorker and Outside Magazine. With titles such as 'Gardener of Eden,' 'Greg Carr's Big Gamble,' 'The Saviour of Gorongosa,' and 'Saving a Global Treasure,' these features uniformly present Greg Carr as a heroic protagonist, turning Carr into a rock star of conservation philanthropy. 5 A segment titled 'One Man's Plan to Save a Natural Treasure' that appeared on the popular CBS 6 show, '60 Minutes' on October 26, 2008 is representative of this dominant narrative. In this segment, the focal point of the story is Carr's generosity to use USD 40 million over a period of 20 years to bring about dramatic improvements in both conservation and human development in Gorongosa. The segment in the show opens with the following statement: How much can one man do to save a desperate nation? American entrepreneur Greg Carr is finding out, throwing himself and much of his fortune into one of the poorest places on earth. Mozambique, in East Africa, is a country of spectacular beauty, but it's been laid waste by decades of war, by malaria and by HIV. It takes a lot of vision to see opportunity there, but… Carr thinks he's found it, in a wildlife park called Gorongosa, which he believes could be the salvation of a nation, and maybe a model for the world. 7 The '60 Minutes' narrative portrays Gorongosa as "a tragedy in two parts": the fall of a national park to obscurity from a position as the "most popular national park in all of Africa" after extensive wildlife loss from war-time poaching; and "the suffering of its people whose lives haven't improved much in a few hundred years." 8 With this seemingly hopeless situation as the frame, Greg Carr is figured as a hero whose innovative plan to bring "entrepreneurship to charity" promises to "bring Gorongosa back to what it was." 9 In an interview filmed in the park's Chitengo Camp, Greg Carr states: So, the idea is take the beauty of the park and use that to do human development. Attract the tourists who will spend the money to create the jobs and lift everybody outta poverty. For an entrepreneur, it's kind of a compelling opportunity to, you know, one plus one equals ten. 10 Scott Pelley narrates the segment as he takes a tour of GNP and Vinho, a settlement adjacent to the park's headquarters, which has been the location of most of the Carr Foundation's direct human development initiatives. On this tour, Greg Carr figures prominently as he guides Pelley-with minor parts played by two park employees-the then Park Director Baldeu Chande, and the then Director of Community Relations, Mateus Mutemba. These people, none of whom are long-term residents of Gorongosa, figure as the protagonists in the narrative-actors who implement benevolent projects in the park and in the neighbouring settlement. Their viewpoints from interviews are woven together to contribute to the main plot line.
By contrast, in this portrayal, Gorongosa residents are not active protagonists but are represented as an audience to the unfolding events and the raw material for directed interventions. Footage depicts several shots of malnourished children as Scott Pelley cites striking statistics to illustrate the severity of poverty in Mozambique. Scenes from Vinho feature residents making use of facilities built with Carr Foundation funds: students in assembly at the new USD 100,000 school; people waiting in line at the new USD 200,000 health clinic; and scenes of a community gathering where the narrator notes that, "among the villagers Carr is treated like a rock star." 11 This narrative presents the Carr Foundation's efforts in a single population centre adjacent to the park as emblematic of how the project impacts all residents living in and around the park's territory of 4,067 sq. km and a buffer zone of 3,300 sq. km. It thus uniformly depicts Gorongosa residents as the silent and grateful recipients of Carr's goodwill.
Like this '60 Minutes' feature, the dominant popular narrative of Greg Carr's involvement in GNP gains widespread circulation through the work of the park's public relations staff. The park's PR Department actively transmits this narrative through press releases and a sophisticated website to lure both potential donors and prospective visitors. Once visitors arrive at the park, this narrative is also transmitted in new ways to shape visitors' experiences. Hospitality staff actively emphasise the 'beginning' point of the narrative for park visitors by setting up a television in the main restaurant that screens vintage videos of the colonial heyday when vast herds of water buffalo, zebra, and antelope of all sorts gathered at Lake Urema during the dry season.
Having depicted the colonial past as a glorious one, the dominant GNP narrative then highlights a present situation of crisis (the combined challenges of depleted wildlife and the poverty of area residents), and the imagined future restoration of the park's former glory through the salvation of a "restoration project" which will both rejuvenate tourism and serve as an "engine of economic growth for the entire region." 12 This narrative has wide appeal because of the apparent plausibility of a simple solution, the absence of conflict, and its heart-warming content.
In these popular stories, crisis functions to garner praise and support for a worthy and benevolent cause and creates a sense of 'history in the making'-a crucial moment in time which prompts potential visitors to desire to see the progress first hand. Conflict and complicated realities are edited out of this popular version of the story-allowing it to retain its wide appeal. But, such popular narratives that figure Carr as a hero bringing uniformly positive benefits to eager recipients belie a much more complicated situation on the ground.

Where is the 'crisis'?: the persuasive power of narratives of environmental degradation
The '60 Minutes' segment and other features on the GRP offer an abbreviated view into what has been a much more expansive effort to rehabilitate GNP. From the very start, a significant focus of the Carr Foundation's involvement in GNP has been to extend the boundaries of the park to include the upper elevations of Gorongosa Mountain. In this context, the dominant narrative is a powerful depiction of an environmental crisis that was circulated to generate financial support and public approval for the park's expansion. Examining this narrative in greater detail reveals the deeply persuasive power of crisis narratives in conservation and development projects.
After an aerial tour of GNP that also included Gorongosa Mountain in September 2005, Greg Carr and American ecologist Rich Beilfuss became alarmed at fires and deforested areas on the mountain's slopes. After this helicopter tour, Carr concluded urgently, "colleagues of mine who are scientists believe that the mountain top will be mostly destroyed in a very few years if something is not done." 13 Thus, in one brief sentence, a narrative was put forward, legitimated by the authoritative knowledge of 'scientists,' and creating a sense of urgency for something 'to be done.' Two days later Beilfuss, the then head of the Carr Foundation's new Gorongosa Park Research Station, 14 stated that their "strategy to deal with the problem" would "involve efforts to entice people off the mountain with some sort of settled, irrigated household schemes in the lowlands below 2,000 feet." 15 The contours of this narrative quickly gained more detail, more solid grounding in authoritative knowledge, and clearer definition for paths of action to take. Beilfuss and other Carr Foundation staff soon undertook "preliminary studies" including six helicopter transects, a "nine-day ground visit with residents and land users on the mountain," and "analysis of change using satellite imagery" (Beilfuss et al. 2005: 2). A report of findings from this rapid assessment was released in December 2005, and declared that the mountain was "under immediate threat" (Beilfuss et al. 2005: 1) and that "slash-and-burn encroachment and uncontrolled fire escaping from hunting or clearing, are the two main destructive forces behind the loss of the forest ecosystem" (Beilfuss et al. 2005: 5). Less than four months from the initial observations, the authors presented their findings, arguing that, "the rate and extent of land clearing on the Gorongosa Mountain is accelerating rapidly" and that "it will take no more than three to five years before the ecosystem is degraded to a point from which it is unlikely to recover" (Carr Foundation 2006: 10). At this time, the Carr Foundation funded the production and circulation of a sophisticated video titled 'Save the Mountain' in order to generate awareness of and support for their efforts.
After receiving the reports myself, I was anxious to return and see the situation first hand. Having lived on the upper elevations of the mountain in 2004, I was concerned about what sounded like an alarming situation for people I knew personally. Given the inflammatory language of the Carr Foundation's 2005 'Save the Mountain' public campaign, when I returned to stay on the Khanda portion of the mountain in 2006, I expected to now be able to see in person what the video and reports described as 'denuded hillsides' above my host family's fields. But as I gazed up at the top of the mountain that loomed over my host family's home, the dark edge of the montane rainforest looked just as it had two years before. There was no sign of crisis, no visible clear cuts, and no barren hillsides as Greg Carr's emails had led me to imagine. The fields here were just as they had been since I began visiting the area in 1999.
The surprise I felt on my return resulted from one of the effects of prevalent and globally circulating environmental crisis narratives. Though I knew of residents' complex agricultural, fire control, and land management systems from previous stays on the mountain, the characterisation of crisis in authoritative technical language and through high quality media appealed to powerful universal ideas (Tsing 2005)-ideas so potent that I was led to expect that the situation on the mountain had made a sudden change for the worse. The Carr Foundation's reports of the situation in Gorongosa had overwhelmed my own grounded knowledge of the people and place of Gorongosa that came from more than three years of living in the district.
Part of the power of environmental crisis narratives comes from their tendency to simplify complex situations into familiar and recognizable stories. Examining the emergence of the park's crisis narrative of ecological collapse on the mountain illustrates this aspect well. When Greg Carr and other members of his foundation's team first made aerial visits to the mountain in 2005, they incorporated their observations into pre-existing narrative frameworks, linking brief observations into a complete vision of the nature of the situation. As has been noted elsewhere in Africa regarding the creation of land and resource policy, Carr and his team of scientists inferred sweeping crisis from casual 'snapshot' observations (Fairhead and Leach 1996). Aerial views that were literally detached from both social context and time depth perspective became the basis for formulating a vision of the scale and speed of deforestation despite weak empirical evidence. 16 Carr Foundation scientists' readings of the Gorongosa Mountain landscape are likely to have been further skewed since they made their primary observations during the end of the dry season (September-October) when many people set fires as part of the regular agricultural cycle and to create fire breaks around homesteads surrounded by dense, tall grasses. As Leach and Mearns have noted, a common methodological error in scientific assessments of environmental degradation in Africa has been "to take short-run observations as evidence of a secular long-run trend, when they may simply describe one phase in a cycle" (Leach and Mearns 1996: 15).
This example also reveals another key aspect of the power of crisis narratives in this context-their influence on how one interprets evidence and experience. The fact that such a rapid assessment led so quickly to definitive explanations and prescriptions for action, including ideas for massive 'voluntary' resettlement schemes for mountain residents, reveals how, beginning in 2005, Carr Foundation staff made sense of the situation in Gorongosa drawing on wider discourses or what Leach and Mearns refer to as "received wisdom about environmental change" (Leach and Mearns 1996: 3). Indeed, Carr Foundation scientists' narrative of crisis depends on and perpetuates a conventional view, that African agriculturalists are "incapable of acting as resource custodians" (Leach and Mearns 1996: 20). The choice of the pejorative language of 'slash-and-burn' to describe observed fires and cleared forest patches also reveals an implicit value judgement of mountain residents' land use as irrational (Guha 1997), which gains even more explicit expression when early Carr Foundation reports described fire and farming techniques as 'destructive forces.' Linking into prevalent conservation and development narratives, assumptions about the nature and speed of ecological change on the mountain seemed solid and legitimate enough to serve as the basis to propose policies as drastic as 'enticing' people to move off of their land.
Perhaps most significantly, the extreme degree of urgency that characterises this narrative of crisis reveals more about the interests of Carr Foundation leaders than it does about the nature of deforestation on the mountain. Emery Roe's observation about the links between crisis narratives and claims to land and resources applies directly here: Crisis narratives are the primary means whereby development experts and the institutions for which they work claim rights to stewardship over land and resources they do not own. By generating and appealing to crisis narratives, technical experts and managers assert rights as 'stakeholders' in the land and resources they say are under crisis. (Roe 1995(Roe : 1066 New on the scene in Gorongosa in 2005, the Carr Foundation team had numerous interests at stake in highlighting the situation on Gorongosa Mountain as a crisis. Not only did the mountain offer an additional area for intervention that made the nascent project seem more important and necessary, but it also presented the possibility of enhancing the park's attraction for tourists. A park without much wildlife would need special attractions to draw tourists far from the beaten path of eastern and southern African safari travel. In fact, the December 2005 report on Gorongosa Mountain includes "loss of tourism potential" along with "degradation of forest cover" and "degradation of water supply" in a list of the "main threats to the mountain system" (Beilfuss et al. 2005: 7-8). Accordingly, the Carr Foundation began promotions and preparations for tourist visits to the mountain as early as 2006, even before they gained official permission from the area's residents ( Figure 1).
My experiences living on the southwest side of Gorongosa Mountain from 2006-2008 revealed a situation, which, (while not a utopia), was much more stable and much less dire than the Carr Foundation's portrayal of crisis. Over time, even representations of the 'crisis' on the park's website have required modification. When I left the area in 2008, around the time that the Carr Foundation had predicted the ecosystem on the mountain would have "degraded to a point from which it is unlikely to recover" (Carr Foundation 2006: 10), little had changed in the forest on Khanda's portion of the mountain. In 2009, the Gorongosa website altered the timeline to place the end point of disaster at 2011. Now, the ominous time-oriented predictions of immanent collapse are absent from the website, but a tone of urgent crisis remains. 17 Environmental crisis on Gorongosa Mountain, from the GNP perspective, has become a protracted condition. The persuasive power of environmental crisis narratives is here visible: despite the inaccuracies of assessments of the timeline for ecosystem collapse on the mountain, the determination of the situation as one requiring outside intervention remains unchanged and unquestioned.

A different kind of crisis: narratives of threats to land tenure on the mountain
While the Carr Foundation made much of the 'crisis' of deforestation on Mount Gorongosa, most of its residents generated a narrative of an ominous predicament of a much different sort. This sense of present crisis struck me as overwhelming when I first returned to stay with my host family on Gorongosa Mountain in 2006. My first excited conversations with old friends were strikingly different than on my previous return visits. Those who knew me well told me their latest news, as was always the case. But rather than telling me about a child who was ill, the latest new-born, or the results of this year's harvest, people pulled me aside to tell me with urgent concern that 'Fundação Carr' 18 had come and was trying to take their land away. Residents of all stripes shared this sentiment: women, men, youth, elders, church leaders, and traditional leaders alike. This atmosphere revealed another aspect of crisis narratives. Situations deemed to be urgent or pressing move to the forefront as a primary concern, superseding all others. Just as demonstrated by the Carr Foundation's 'Save the Mountain' public campaign, crisis narratives evoke a desire for transmission-a desire that was also apparent when I visited Khanda's régulo, 19 Eugénio Almeida, shortly after my arrival. He too expressed an urgent sense of crisis. He spoke in a hushed tone about the Carr Foundation and then asked me if I had brought my voice recorder. Revealing a strong desire to transmit this narrative of crisis and his awareness of having limited access to channels for its spread, he said, "Turn it on… this is important." The régulo picked up a stick and began drawing a map of Gorongosa district at our feet. Pointing to the oval representing the mountain and tracing it all the way to the edge of the Pungue River he said, "From here all the way to there-all of this belonged to the régulos. Each régulo had their population. And when the Portuguese came, people were living here." He pointed to the area inside the present borders of the park, saying: When they arrived, they encountered animals of all kinds. And they expelled the people… That area they call 'park'-there in the park they threw out Régulo Chikale, Régulo Nyanguwo, Régulo Tambarara: 'Get out! Get out!' 20 The park's history depicted in this narrative stands in stark contrast to the popular media's narrative, which begins with the park's 'glorious past'. In the regulo's narrative vision, restoring the colonial past of the park is not a positive goal. While the park was once splendid for wealthy leisure tourists, it was not so for Gorongosa's residents. The human costs involved in the park's creation were extremely high. Elder Gorongosans recall how colonial agents used the brutal system of taxation and conscripted labour to force the very residents who had been displaced from the present-day park to return to their former territory to build roads and lodging for tourists. Thus, to many long-term Gorongosa residents, the project to rehabilitate the park is re-creating, repeating and extending the negative legacy of the park into the present.
The Régulo paused, looked up from the map he had been drawing, and looked me right in the eyes. Lowering his gaze, he slowly moved the stick to point again at his map and said: Now this here is Gorongosa Mountain. [T he Portuguese] weren't able to claim it-they weren't able to claim the area all the way to here. But now, our government, because they are poor, now they want to put this mountain in, for all this to be included with Chitengo. 21 But we don't want this! We don't want it! We don't want it, we don't want it, we don't want it. 22 On the mountain, where nearly all residents make their living from small-scale farming, land is central to political, economic, and social life. Political power and rights over land and resources are all linked to ancestral tenure-those who can claim to be descendants of the first settlers on Gorongosa Mountain are considered the present owners of the land (Isaacman 1972;Shipton 1994). Thus, mountain residents' dominant narrative is centred around the view of long-term tenure and autochthony as the legitimate basis for claims to land and resources.
This ancestral claim is embodied in the explicitly political figure of mhondoro spirits. Mhondoro-the spirits of the first political leaders to rule in the area-reside on the mountaintop, and exert ongoing political authority through well known regulations and prohibitions. In Gorongosa, mhondoro are seen as the owners of the land who, when well respected, look after the well-being and prosperity of their descendants (Lan 1985;Shoko 2007). Mhondoro, and other spirits residing in an area must be accorded due respect, or they may show their disapproval by causing misfortune. Mountain residents recall historical moments, stretching back to pre-colonial times, when mhondoro spirits have intervened in times of crisis to protect mountain residents from outside threats.
To the residents of Gorongosa Mountain, the arrival of Greg Carr's heavily financed project to rehabilitate GNP raised the familiar spectre of the possibility of dispossession and thereby also the anger of mhondoro spirits. The unified sense of urgency that emerged among long-term residents of the mountain derives from both recent and deep historical experience with struggles over the most important shared source of and political and economic security-land. For over a century, Gorongosa Mountain has been a site of struggle. The mountain has been a place of refuge and a stronghold for groups opposing various forms of centralised outside governments 23 (French 2009;Schuetze 2010). In the late colonial period, the Portuguese government granted a large portion of land on the Khanda side of the mountain to a German man to start a dairy farm. An enormous swath of land high on the mountain was turned to pasture for his cattle. Many of those displaced later worked on the dairy farm, but this employment and the 'economic benefits' did not compensate for the loss of land and led to deep-seated anger.
Following independence, displaced residents were able to return to their land, but the war between Frelimo and Renamo 24 soon brought land tenure to the centre of a new struggle. During the 16 year conflict, mountain residents were the focal point in a struggle over control of the population. Because Renamo troops had a base on the mountain and relied on residents to supply them with labour and basic foodstuffs, Frelimo strategy aimed to cut off this important support base by forcing residents into aldeias or 'communal villages' in the lowlands. Mountain residents were targeted in a scorched earth campaign where homes, fields and granaries were burnt, and goats and chickens slaughtered, in order to leave them with no choice but to seek refuge and food aid in the Frelimo protectorates. But, for many mountain residents, remaining on their own land was so important that they preferred to find a way to eke out a living on wild foods and by planting crops in hidden valleys, risking punishment from Frelimo or capture by Renamo troops who periodically raided homesteads to conscript people into the army.
These past threats to land tenure coalesced into a powerful, shared narrative that fostered defiant attitudes against the Carr Foundation that were so widely shared as to be almost universal among mountain residents. This narrative, also informed by previous experiences with GNP, painted a homogenous picture of anyone assumed to be Carr Foundation staff as villainous outsiders with only nefarious intentions and a singular desire to force residents off their land. Colouring mountain residents' interpretations of the Carr Foundation's planned interventions, this narrative thus ignited resistance to park initiatives as soon as they were first introduced.

Community meetings: narratives like oil and water
Unified and intense opposition to the Carr Foundation's early initiatives on Gorongosa Mountain began even before programmed interventions were put in place. The Carr Foundation began its project on the mountain by convening several 'community meetings' held in the Khanda Regulado on the southwest slope of Gorongosa Mountain. Such meetings were a critical aspect of what the Carr Foundation heralded as "the full involvement of the communities in all aspects and stages of development" of interventions on the mountain (Beilfuss et al. 2005: 8). Following legal guidelines in the national land law, some of these meetings were also required in order to obtain the approval of the area's residents for proposed land use. In the context of these meetings, the two crisis narratives came into conflict, but, like oil and water, they repelled each other and remained separate.
At the first 'community meeting,' held on January 13, 2006, Greg Carr and other leaders of the Carr Foundation, with district government representatives and the then Director of GNP, Roberto Zolho, met with a crowd of residents of the Khanda region of the mountain. The meeting focused on a proposal to form a 'partnership' with the park to manage the land at the top of the mountain. In their speeches, park and government leaders shared precious few details about what such a partnership would entail. District government officials explained that the Carr Foundation had generously offered to manage the land on the mountain. Park representatives characterised the Carr Foundation's interests as: the "exploration of Gorongosa Mountain" for a "joint-venture in ecotourism" (GTZ, PRODER 2006). They explained that tourists would contribute money to the Khanda community so their lives would improve.
After this brief presentation, attendees responded by expressing their concerns. The official meeting report summarised mountain residents' statements in the following way: …all were unanimous in affirming that they had no intention of forming a partnership [with the park] because… they had already had negative experiences in the past or during the colonial period, when many Portuguese settled in an area and then shortly thereafter began to impede and expulse people… (GTZ, PRODER 2006).
Mountain residents' narratives, grounded in historical experience, shaped powerful resistance to the idea of park management of the land and resources, and quickly fomented a sense of unity in opposition to an outside threat. The Carr Foundation's narrative and vision for intervention on the mountain, which made great sense from a social and physical distance, thus encountered an obstacle when placed before area residents for approval.
Rather than withdraw or work towards a modified, collaborative arrangement, project leaders concluded that the problem lay in local residents' 'misplaced' fear of land loss and ignorance of the goals and values of conservation. The goal for Carr Foundation leaders then became to bring mountain residents' understanding of the situation and goals for intervention into conformity with that of project leaders. A note appearing at the end of the official report of this initial meeting reads: "A weak capacity to understand complex issues such as this is common in the communities" (GTZ, PRODER 2006). 25 And, in a bullet list of recommendations it was suggested that park staff "study other ways to present the issue in order to generate more interest in the community" (GTZ, PRODER 2006). Revealing how, 'full involvement of communities' in the project, in practice, referred merely to a process of consultations to gain approval for inflexible and predetermined schemes, the report made no recommendations to modify or drop the original proposal. Rather, a second meeting was called, one month after the first.
During this second meeting, park representatives again made general speeches about their proposal but added no details. Khanda residents again responded with impassioned speeches of refusal. Seeing that their answer had been ignored twice in a row, all the attendees stood up and walked out in protest before the meeting ended. This display of unity and defiance in the face of authority illustrates an effect of mountain residents' crisis narrative. Building on habits ingrained from generations of living under colonial rule, rural residents of Gorongosa typically display a supremely diplomatic and almost submissive stance in formal meetings with authority figures.
This surprisingly showy and unified protest was animated by narratives that figured the Carr Foundation as foreign interlopers bent on taking land on the mountain, leading the meeting attendees to collectively abandon social conventions of deference in order to demonstrate their unified resistance. As one man put it: "we have known for a long time what it's like to have our land taken from us. First, they speak sweet words… they will act like our friends today, but tomorrow, they will kick us to the side." 26 In meetings held in Khanda from 2006-2008, the narrative fortresses on either side were not breached, but rather, the persuasive qualities of crisis narratives shaped interpretations of and hardened responses to the other side's words and actions, making positions quite rigid, further illustrating the power of divergent narratives of crisis to shape and maintain conflict.

Mhondoro narratives, unity, and resistance
Despite the opposition they encountered in these meetings, the Carr Foundation pushed forward with its agenda to extend conservation and tourism activities to the mountain. In response, mountain residents continued to present a unified front of resistance. In the eyes of mountain residents, from 2006 to 2008, the Carr Foundation's actions were characterised by empty promises and disrespect. Slotting them into a single character position in their narrative, residents of Khanda saw Carr Foundation employees and associates as a homogeneous group, and felt that they were treating the mountain as their own. Despite their overt refusal to approve the park's proposal to initiate conservation efforts on the mountain, residents watched as Carr Foundation employees moved forward with conservation and tourism schemes on their terms, ignoring traditional regulations governing access to sacred areas. Some park projects, such as the actual and planned construction of buildings and structures to serve tourists violate specific rules ordained by the mhondoro including strong prohibitions against placing buildings above certain elevations. These actions intensified residents' sense of the threat to their land and fomented a powerful sense of unity and defiance that would build over time.
In 2007, the Carr Foundation hired a team of dozens of fiscais (fiscais=rangers) to patrol the mountainside in Khanda.
These fiscais visited fields above 700m, imposing fines for cutting certain tree species and pressuring residents to plant tree seedlings in fields on steep slopes, immediately adjacent to streams, or in areas otherwise deemed to be 'unsustainable.' Mountain residents regarded the deployment of forestry rangers as a major affront and responded with outspoken resistance, leading to overt conflicts with residents of the lowlands who were employed as fiscais. As one man explained: "four fiscais came and wanted to plant trees in our garden. We refused, [and said to them] that if you want to plant [the trees] why don't you plant them in your field?… If you want to plant them, go ahead, but as soon as you leave, we will pull them out." 27 Mountain residents also watched as park staff buzzed around the mountainside in helicopters to travel to meetings, take important guests for aerial tours, or shuttle biologists back and forth between the mountaintop and their accommodations in the park. For many, the frequent roar of helicopters stirred up the terror of the recent war when Frelimo made aerial attacks on home sites in a last-ditch effort to remove any families who remained in hiding (Schuetze 2010). These events thus reinforced mountain residents' crisis narrative, grounding the narrative's early predictions in tangible signs that their land was under immediate threat.
The Carr Foundation's offences touched off a flurry of stories of the protective actions of Mhondoro spirits-narratives that further strengthened mountain residents' sense of unity and righteousness of autochthonous land claims. In the face of the Carr Foundation's offences, one story spread like wildfire as proof that these spirits continue to have force. In late 2006, the Carr Foundation hired a 'community liaison' to manage relationships with mountain residents in Khanda. While visiting different mfumos (mfumos=subchiefs) near Nyankhukhu, this young British man set his sights on an impressively tall dome-shaped granite outcrop known as Bango Moliro, which, in Chigorongosi translates as 'fire mountain.' Seeing the landmark as an exciting place to bring tourists, he ignored the explicit warnings of the mfumos against visiting this prohibited area and climbed up to explore. Soon after, while driving back to the park in one of the Carr Foundation's new 4x4 vehicles, he noticed smoke flowing out from under the hood and stopped to investigate. Before long, the car was consumed in flames and was destroyed completely.
How a brand new truck with no previous engine problems could catch fire baffled park staff, but made complete sense to mountain residents, who spread the story with glee. After all, the young man had climbed Fire Mountain in defiance of the mhondoro's prohibitions. Offended spirits were actively at work.
This and other stories depicting the protective and retributive actions of mhondoro reveal the power of narrative to foster unity and defiance and a sense of moral righteousness. They also illustrate how Gorongosa Mountain residents' sense of patrimony is much more strongly based on their land and ancestral political claims than with the nation or the national government of Mozambique. It was not long before it became clear to mountain residents that the national government was in support of the Carr Foundation's project, and many were not surprised. As one leader put it: "I've told you that the government wants money. They've been bought. Now, we, the people, are familiar with how our government works. The people here have no power." 28 In stark contrast to Mozambique's central government, narratives figured mhondoro spirits as responsive and active in protecting mountain residents' land tenure. Resistance to the park's interventions on the mountain, then, also gave expression to resistance to the authority and legitimacy of the national government. The park's expansion to Gorongosa Mountain was therefore deeply political, rekindling long-standing opposition to external centres of political power. The park's project reaffirmed mountain residents' assessment of the national government as an illegitimate power which acts not to protect the interests of the people but only to generate wealth for a limited few.

Producing destruction: narrative fortresses
When I returned to Gorongosa Mountain in June and July 2011, I was officially entering a national park. While no signs indicating this park status had been posted, the park designation had a heavy impact on the area. What struck me most was the way in which communication between park leadership and mountain residents was so infrequent as to be nearly absent. Early aspirations for 'full involvement' of resident populations in the implementation of the project had been completely abandoned. Instead there was even tighter fortification of information and thicker barriers to affected residents' real involvement in park planning or decision-making. Thus, narrative fortresses became more apparent, and more impermeable to change or influence.
Throughout the early years of the project, rumours about park officials' plan and actions in regards to the mountain circulated widely. Excluded from the planning table and barred from influencing the design and vision of the project's initiatives from the start, mountain residents were left to fear the worst about the intentions of outsiders. Information shared at 'community meetings' had been overly vague and general, fostering fearful speculation about the park's actual plans. This situation has not changed over time. Tellingly, I learned of the official inclusion of the upper elevations of the mountain in the park via an internet news article from my home in the US in the summer of 2010, long before the news reached mountain residents. It wasn't until March 2011 that park authorities convened an official 'community meeting' in Khanda to make the announcement.
By the time I arrived in June 2011 it seemed that communication problems had worsened. Or at least the stakes were now higher. Upon my return to stay with my host family I was shocked to find that everyone seemed to be talking about immanent dispossession. And, I watched exchanges where residents living inside the new park boundaries received clear information from people in official positions that they should prepare to move their homesteads to the lowlands. These reports came from sources including the District Director of Education, the Governor of Sofala Province, and fiscais employed by the GRP.
When I shared this information with leaders of the park's new division set up to manage the mountain, they were dismayed and explained that there were no such plans in the works. The park project leaders' failure to involve mountain residents in planning and design and, at a minimum, their failure to open clear lines of communication about the implications of the park designation, opened the way for narratives of impending dispossession to dominate. Such dramatic failures to communicate have had negative effects for all parties. Fearing immanent dispossession, mountain residents were altering plans for the immediate future, including questioning whether to invest in the upkeep of their property. Here, a tragic impact of these narrative fortresses is revealed. Without means for effective communication, park actors and mountain residents came to live in separate realities-with harmful consequences for mountain residents whose actions in the present and plans for the future were shaped by the rising fears of forced removal.
Most disturbingly, the animosity and fear of the situation contributed to a visible increase in deforested areas-exactly what the park project on the mountain aimed to halt. My return visit allowed me to witness how a 'crisis' of deforestation had become more palpable. In early July, I hiked around the mountain summit and saw large areas cleared for fields-areas that had been solid forest during my last visit in 2008. In interviews with numerous residents, I found everyone I talked to was well aware of the situation. Most mountain residents I interviewed expressed deep dismay about the people who were suddenly clearing fields on the mountain's summit. The mountain's summit has long been understood to be territory off limits to cultivation-an area where mhondoro and other land spirits reside. Spiritually mandated prohibitions against cutting trees on the summit were still strictly adhered to while I lived on the mountain from 2006 to 2008.
While I encountered different interpretations of the reasons for the transgressions, all of them were linked to a profound sense of powerlessness, anger, and immanent crisis. Some attributed these people's actions to last-ditch efforts to accumulate money from potato farming to buffer the material devastation that would accompany removal to the lowlands. Others explained that transgressions were based in a sense that if the forests were the focus of outsiders' interest in controlling the land, destroying forests might drive them away. Others saw it as sabotage-as a form of retribution where destroying forest was aimed explicitly against the park's clear interest in protecting the forests-to frustrate park actors and make them also feel a sense of powerlessness. As one man put it, all this started when the park came to claim control of the land: "it's almost a competition. If you want to see [deforestation], then that's what we'll do." 29

CONCLUSION
Between 2006 and 2008, the Carr Foundation's initiatives on Gorongosa Mountain touched off a conflict that led to a hardening of positions-the development of narrative fortresses-rather than creating any sort of genuinely collaborative efforts with area residents. As the community meetings and their aftermath illustrate, crisis narratives played a powerful role in the creation and perpetuation of this conflict. These narratives shaped different actors' sense of legitimacy and purpose, their interpretation of events, and their actions, thereby serving as "a medium through which events [were] produced" (Fortmann 1995(Fortmann : 1054. Fundamentally different apprehensions of crisis in narratives about threats to the land and resources on Gorongosa Mountain and a failure to breach them with respectful communication or genuine collaboration has had negative effects on the well-being of residents. Locked in rigid fortresses, these divergent narratives of crisis intensified over time, leading actors on both sides to transgress moral and legal norms. By 2011, the deepening and intensifying conflict with GNP had catalysed dramatic transgressions of long-held prohibitions that give expression to a local conservation ethic. Crisis narratives lent actors who felt otherwise powerless a sense of righteousness in these actions. Thus, the project's environmental crisis narratives ironically produced outcomes counter to the goals of conservationists. Expanding beyond a focus on narratives to examine the centrality of contrasting evaluations of 'crisis' offers several important insights. First, it reveals the deeply persuasive power of environmental crisis claims in conservation and development projects. Linking into broader circulating narratives, these claims quickly generate authority and legitimacy for rapid action despite obstacles or opposition. Narrating particular contexts in terms of environmental crisis creates the sense of necessity for intervention. Crisis-claims about threatened ecosystems motivate desires for urgent action and, drawing on a growing sense of global patrimony of biodiversity, they legitimate the authority for external intervention. Further, as Roitman notes, "crisis is posited as an a priori; the grounds for knowledge of crisis are neither questioned nor made explicit" (2013: 10). Environmental crisis narratives thus have a powerful and persuasive effect on a broader public-tapping into the power of already existing narratives of global environmental crisis, such narrative portrayals of particular contexts can lend any proposed interventions legitimacy without the need for extensive empirical evidence to substantiate claims. In the case of the Carr Foundation's early efforts to extend park activities and governance to Gorongosa Mountain, narratives of environmental crisis were created in the space of a few months, gaining expansive 'communicability' (Briggs 2005) through the authoritative language of ecological science.
Examining the nature and effects of crisis narratives in this case also helps to shed light on why the Carr Foundation's initial aspirations to foster amicable relations and create plans for the project with the 'full involvement' of resident populations were quickly replaced by a top-down governing structure. This initial aspiration for partnership with Gorongosa residents was shaped by a larger set of structures, including Mozambique's current neoliberal economic development strategies that encourage tourism ventures. Ecotourism is bolstered by narratives of 'neoliberal conservation' (Büscher and Whande 2007), which offer the satisfying illusion of parks as progressive forces-as correcting mistakes of harsh fortress conservation practices of the past. In Mozambique, this vision for parks is clearly expressed by Mozambique's Ministry of Tourism which states its goal is to move beyond "past" practices when conservation areas were "planned and managed against people" towards a "future" when conservation areas will be "run with, for, and in some cases by local people" (MITUR 2004: 19). The centrality of tourism to neoliberal conservation strategies shapes the kinds of win-win conservation and development success story narratives that GNP generates. Heart warming, progressive narratives proliferate in order to draw socially conscious visitors, who are driven by a neoliberal logic of "consumptive activity" as environmental action (Igoe et al. 2010: 504).
With such powerful forces promoting partnership relationships and joint management of protected areas with resident populations, what led Carr Foundation staff to so quickly abandon this goal? The sense of urgency, conviction, and righteousness that crisis evokes showed itself to be much more powerful than visions of undoing past wrongs and running conservation areas "with, for, and in some cases by local people" (MITUR 2004). By their very nature, crisis claims are judgements of a situation that lend themselves to rigid, fixed positions. As Roitman (2013: 3) has revealed, "crisis is mobilised in narrative constructions to mark out...moments of truth," bolstering a sense of authoritative knowledge of the state of a problem and the direction needed for change. For the early staff of the Carr Foundation, facing what was deemed a crisis situation, there appeared to be no time or space for compromise. Thus, 'community meetings' on Gorongosa Mountain almost immediately became information sessions with no room for negotiation or joint decision making. Plans set before mountain residents for approval were carried out despite vocal opposition to them. Mountain residents' refusal to grant the park permission for land use was subverted by taking the request to the national government, which issued a decree that officially annexed the higher elevations of the mountain to the existing park.
Meanwhile, relationships between park actors, government representatives, and residents of affected areas of the mountain deteriorated, communication channels became confused and broken, and fears of mountain residents became heightenedopening the way for retaliatory acts. Thus, the sense of crisisboth for conservationists and for mountain residents alike intensified. This situation fuelled a spiral of conflict that later threatened to move toward violent outcomes, as beginning in 2011, park staff responsible for management of the mountain began contemplating the potential for employing armed guards to protect the forests. 30 The early phase of the Carr Foundation's efforts in Gorongosa reveals the dangers of conservation schemes in a context of severe power imbalances. It reveals how, in a postcolonial context, interventions that impact basic rights to control land and resources gain external legitimacy through crisis claims and calls for urgent action as well as through familiar development narratives that highlight the needs of recipients and the benevolent intent of interventions. The familiarity of such apolitical narratives contributes to their popularity and the power that they have to cover over complex realities and painful conflicts. This narrative easily overpowers and erases the deeply political and historically grounded narratives of Gorongosa residents who are cut off from the networks of influence, power, and control enjoyed by park-based actors, government leaders, and conservation advocates. The severe power imbalance that allows only one narrative to circulate beyond Gorongosa further consolidates the legitimacy of international conservation actors to intervene in the area. Apolitical narratives obscure the ongoing and heavily political conflict that is at play in Gorongosa where those who have the most at stake-their land, their lives, their livelihood-are given the least amount of control or authority in park affairs.
Where severe power imbalances foster fundamentally different interests in land and resources, divergent crisis narratives are bound to emerge. As examination of this case has revealed, in such circumstances, even when conservation and development projects aim explicitly to create collaborative partnerships with resident communities, environmental crisis narratives tend to push project leaders away from genuine attainment of such goals. With this in mind, it is the responsibility of those in positions of greater power to forge ways to deliberately work against dominant narratives and rigid assessments of crisis. Focusing on the power and consequences of crisis claims to generate narrative fortresses highlights the need for conservation actors to not simply consult with residents affected by conservation schemes, but to create structures of genuinely collaborative governance. It points to the need to think differently about conservation so that residents of a protected area can be respected as partners in strengthening existing land management practices rather than as the source of destruction or as objects of interventions. As the Gorongosa case reveals, creating, circulating, and maintaining a 'single story' of conservation and development visions through tight control of park governance and public image does not, in the end, lead to greater control over complex situations. Opening to a diversity of experiences, positions, and forms of knowledge has greater potential for breaking down narrative fortresses and generating more effective and more just approaches to land management.
15. Rich Beilfuss, pers. comm.;September 21, 2005 16. Data continues to be scant and based on little more than casual observations (see Walker, this issue).
17. In a recent version of the website, a phrase under a section titled 'YOU CAN HELP!' states, "We need your help in this important campaign to protect Mount Gorongosa-one of Mozambique's most treasured natural wonders-before it's too late." See: http:// www.gorongosa.net/en/page/save_the_mountain/restore-themountain; Accessed on April 20, 2012.
18. 'Fundação Carr' is the Portuguese phrasing of 'Carr Foundation' 19. Régulo is the Portuguese term used to refer to an area's primary 'traditional leader' or 'chief.' 24. Frelimo was the ruling party at independence, and Renamo was an opposition army funded largely by neighboring white minority-ruled governments in order to destabilise the newly formed socialist state.
25. 'Communities,' here, is a term commonly used in the language of staff and leaders of development organisations in Mozambique to refer to residents of rural settlements. As Hughes has noted, the term has come to replace Frelimo's use of the word 'peasantry' (Hughes 2005). Thus, the term 'community' is a vague concept whose referent invokes notions of harmony and homogeneity that 'works to disguise differential abilities to access power' (West 2006: 36). It is also a political and 'spatialised' concept that locates people in particular geographical areas (Smith 1992