The discursive construction of conflict in participatory forest management : the case of the Agoua forest restoration in Benin

Disclaimer/Complaints regulations If you believe that digital publication of certain material infringes any of your rights or (privacy) interests, please let the Library know, stating your reasons. In case of a legitimate complaint, the Library will make the material inaccessible and/or remove it from the website. Please Ask the Library: http://uba.uva.nl/en/contact, or a letter to: Library of the University of Amsterdam, Secretariat, Singel 425, 1012 WP Amsterdam, The Netherlands. You will be contacted as soon as possible.


INTRODUCTION
The establishment and preservation of nature reserves and protected areas are constrained by a signifi cant challenge, i.e., the distribution of costs and benefi ts related to such an enterprise (Krueger 2009).Although the wider community may benefi t from these protected areas because they are considered as global or national goods, the costs are borne by individuals or groups who live in or around them.In many cases, conservation actions have led to the physical and economic displacement of millions of people who formerly lived, hunted, fi shed, and farmed in areas now protected for wildlife, e.g., watersheds, reefs, forests, or rare ecosystems (Brockington & Igoe 2006;Agrawal & Redford 2009).The implementation of such decisions often results in resistance and confl icts between local communities and the organisations in charge (Connor 2005).
In 2002, the Benin government initiated the restoration of three forest massifs (Agoua, Wari-Maro and Monts Kouffè) because of their importance in Benin's ecosystem and their high level of degradation, due to severe anthropogenic infl uences such as poaching, logging, grazing, and agricultural activities.In the case of the restoration of the Agoua Forest, a confl ict arose between the local communities settled in the forests and the Management Project for the Wari-Maro, Monts Kouffè and Agoua Forest Massifs (PAMF: Projet d'Aménagement des Massifs Forestiers d'Agoua, des Monts Kouffè et de Wari-Maro) charged with the restoration of these forests.
Negotiations between the stakeholders involved ended in an impasse, although formal agreements were signed, and tension still remained after the ending of the project in June 2008.
Since the creation of Benin's protected areas in the colonial period between 1940 and 1960, and subsequently until the beginning of the 1990s, the state managed the forests using force and repression.Communities living adjacent to these forests were kept at a distance as they were considered to be a threat to these natural resources.However, this management system did little to stop or slow down the degradation of these forests.It rather encouraged corruption and illegal logging, creating confl icts between the forest rangers and the local communities whose livelihood depended, at least partly, on the forests (Siebert & Elwert 2004).The objective of sustainable management of natural resources was not reached as there were too few rangers to control its implementation.After the Rio de Janeiro Summit of 1992, forestry reforms were undertaken in Benin.A new forest law was established in 1993.The main reform was the commitment of Benin to participatory management of its natural resources.Since then, many projects have been initiated to stimulate local communities' participation in the sustainable management of forests and parks.However, despite many efforts, the results of these interventions are still questionable (El-Hadj Issa 2001;Tchiwanou 2001;Zoundoh 2001).Several studies have revealed that the forestry reforms of the early 1990s have not been successful due to the failure of participatory management (MDR & PGFTR 1999;Siebert & Elwert 2004).In many cases, confl icts have arisen between the stakeholders involved in these projects.These confl icts stemmed from the opposition between forest department representatives and villagers, due to corruption in collecting and sharing forest management revenues, and illegal logging (Siebert & Elwert 2004), and among the forest department, farmers, and herders, because of the restriction on access to forest lands for farming and grazing (Arouna 2006).
Elsewhere too, studies have revealed confl icts related to natural resource management (Hellström 2001;Yasmi 2003;Hares 2009).Confl icts arise because the stakeholders involved have competing interests, perceptions, and ideas about how natural resources should be managed (Buckles 1999;Castro & Nielson 2003;Yasmi et al. 2006).With the realisation that the traditional top-down forest management was not effi cient in terms of conservation and had more often than not led to confl icts, efforts have been made to involve local people in forest management during the last two decades (Kassa et al. 2009).However, in Benin, confl icts re-emerged between forest administrations and local communities, despite many efforts to involve local communities in forest management (Arouna 2006).In this paper, we aim to improve our understanding of how confl icts emerge, evolve, and either end in resolution or persist in intractable confl ict in the context of participatory management.This, in order to enhance our capability to handle such confl icts and thereby contribute to the sustainable management of natural resources.
To this end, we analysed the confl ict that arose in relation to the participatory restoration of the Agoua Forest by the PAMF.Because we wanted to study the manifestation of this confl ict over time, we used a framing perspective (Aarts et al. forthcoming).According to Kretsedemas (2000: 639), "Frame analysis has been used to examine the ways in which movement groups articulate their goals, recruit participants, and respond to the counterframes of their opponents."Frame analysis is therefore useful in confl ict situations to investigate how frames emerge and their applications after they have been developed (Kretsedemas 2000).

Research Approach
Frame and framing concepts are particularly relevant for researchers studying confl ict, negotiation and inter-group interactions (Gray 2003;Dewulf et al. 2009).The notion of frame is rooted in cognitive psychology (Bartlett 1932) and anthropology (Bateson 1954).It is often associated with Erwing Goffman (1974), who described how to use the concept of frame to understand human thought and interactions.Since then, the concept of frame has evolved.Current research on framing distinguishes two main approaches: the cognitive approach and the interactional approach (Dewulf et al. 2009).
The cognitive approach in framing research was explicitly formulated by Minsky (1975) in the field of artificial intelligence.This research tradition has its roots in Bartlett's (1932) schema theory of memory.The cognitive approach focuses on cognitive frames or mental structures that help us organise and interpret incoming perceptual information by fi tting it into pre-existing categories about reality (Minsky 1975;Dewulf et al. 2009).In research using the cognitive framing approach, frames are considered as stocks of knowledge used by individuals to assess new information.The defi nitions of frame by Goffman, and Gitlin fi t in this framing approach.Goffman (1974) defi ned frames as schemas of interpretation that enable individuals to locate, perceive, identify and label phenomena.In the same way, Gitlin (1980: 6) said that, "Frames are principles of selection, emphasis and presentation composed of little tacit theories about what exists, what happens, and what matters."They are knowledge schemas or structures of expectations about people, objects, events and settings (Tannen & Wallat 1987;Dewulf et al. 2009).
The interactional approach to framing research is linked to the early work of Bateson (1954) on meta-communication, in which framing is defi ned as exchanging cues that indicate how ongoing interaction should be understood (Dewulf et al. 2009).In this approach, the defi nition of framing corresponds to what Entman (1993: 52) said: "To frame is to select some aspects of a perceived reality and to make them more salient in a communicating text, in such a way as to promote a particular problem defi nition, causal interpretation, moral evaluation, and/or treatment recommendation".Frames are alignments or co-constructions produced and negotiated in interactions (Dewulf et al. 2009).Interactional frames are thus communication devices used by participants in interaction to negotiate meanings and alignments.The interactional approach to framing thus enables us to understand how participants in a confl ict co-construct meanings and negotiate alignments while interacting.
Both framing research traditions are useful to gain insight into confl ict dynamics, but each gives a different kind of understanding of how and why frames change.The cognitive approach notes that stakeholders hold multiple frames as knowledge schemas and shift from one to other when they get new information (Minsky 1975).The interactional approach links frame shifting by stakeholders to what is going on during interaction (Dewulf et al. 2009).In this research, we opted for the interactional approach to get more insight on how and why stakeholders' frames evolve in confl ict situations.In the interactional approach, frames are considered as agency used to act on the world (Marullo et al. 1996;Pellow 1999;Benford & Snow 2000).People use frames to perform actions.For example, in social movements, (e.g., peace, civil rights, environment, women's movements, etc.), where groups of people engage in collective action, frames are used to activate and motivate the greatest number of potential adherents (Marullo et al. 1996).Social movement actors are signifying agents who actively engage in producing and maintaining frames for constituents, antagonists, and bystanders and observers (Benford & Snow 2000).In efforts to change policy, policy makers use frames to gain broader support, by linking the preferred issue framing of a group to the core values of an external group (Marichal 2009).According to Gray (2003: 15), "Frames are used to 1) defi ne issues, 2) shape what action should be taken and by whom, 3) protect oneself, 4) justify the stand we are taking on an issue, and 5) mobilise people to take or refrain from action on issues."Thus, frames are constructed in interaction and used strategically to persuade others to our point of view, gain advantage in negotiations, and rally likeminded people to our causes (Shmueli et al. 2006).Thus, in the interactional framing approach, people co-construct social realities in conversations (Pearce & Cronen 1980;Ford et al. 2002;Dewulf et al. 2009).Since constructed realities provide the context in which people act and interact, the nature of these realities establishes the opportunities for action, how people see the world, what actions to take, etc. (Ford et al. 2002).Frames are then iterative; this means they are constructed in a particular reality and infl uence this reality as well (Ford 1999;Aarts & van Woerkum 2006;Aarts et al. 2011).In this framing perspective, a confl ict is neither a state of the world nor a state of mind, but a reality that resides in the social interaction among disputants (Ford et al. 2002;Dewulf et al. 2009).So, confl icts arise only because of how people co-construct issues, relationships, and interactions (Dewulf et al. 2009).
To understand the dynamic of the confl ict between the PAMF staff members and the farmers in relation to Agoua's restoration, we investigated how these stakeholders framed the problem, both their own and the other stakeholders' role in the process, and the participatory process that was designed to manage the forest.We analysed the frames that stakeholders brought to the fore in different conversation contexts over time (see Aarts et al. 2011;Dewulf et al. 2009;Gray 2003).
Issue or problem frames deal with what the confl ict is about.Disputants often start talking about the confl ict by giving a brief summary of what they believe the confl ict is about.Gray (2003) called this 'whole story frames'.These frames shed light on which aspects of the confl ict are important to each party.Issue frames also include cause and solution frames as the disputants, by describing the issue, tend to highlight their meanings of the causes and their desired solutions.
Relationship frames include identity, characterisation and power frames.Identity frames refer to how stakeholders involved in a confl ict present themselves.Parties in confl ict view themselves as having a particular identity in a specifi c confl ict situation (Kaufman et al. 2003).Identity frames are the different answers to the question: 'Who am I?' (Hoare 1994;Gray 2003).Answers to that question may vary from one stakeholder to another depending on their demographic characteristics (e.g., race, gender, and ethnicity), location (e.g., their work place, where they are living, and their origin), their role (e.g., as a farmer or a facilitator), the institution with which they work (e.g., a project staff member or a government offi cer) or their interests (e.g., whether or not they agree with the legislation) (Gray 2003).Characterisation frames are the mirror of identity frames as they concern how a group of actors present the others involved in a confl ict.They are the answer to the question: 'Who are they?' (Gray 2003).In confl icts, parties tend to stereotype and portray opponents negatively or positively.They construct characterisation frames that are often different from the identity frames of the other parties.In this case, such characterisations often undermine the others' legitimacy, cast doubt on their motivations, or exploit their sensitivity (Kaufman et al. 2003).Characterisation frames are also used by parties in confl ict to strengthen their own identity and justify their actions towards the others.Parties in confl ict describe their relation with each other using power frames.Power frames are related to power resources as they are the way actors involved in a confl ict evaluate their own resources and those of the others to infl uence each other (Marfo 2006).Disputants use power frames to characterise the forms of power that are legitimised in the confl ict as well as the forms of power that they preferred to comfort their own position (Kaufman et al. 2003).
Process management frames refer to the way parties judge the implementation of the process, and their preferred management process (Gray 2003;Kaufman et al. 2003).When the process concerns a confl ict, we talk about confl ict management frames.Depending on parties' identity frames, their characterisation frames, and their interests, they may hope for or prefer a particular type of management process.Confl ict management frames in environment confl ict studies may vary from avoidance or passivity to struggle, sabotage or violence (Gray 2003).
Identity frames, characterisation frames, and power frames are important for understanding confl ict, as confl ict almost inevitably arises when people feel their identities have been threatened (Aarts et al. 2011;Blok 2001).These frames also infl uence people's feelings of whether or not there is a problem and the way they defi ne the issue at stake (Gray 2003).People are always negotiating identities and the problem at stake in the presence of others in interaction.Thus, the above frames have been investigated and analysed in different interaction contexts, varying both over time and in the composition of the stakeholders involved in the conversations.

Research Methodology
The interpretive approach was used to investigate how the confl ict in Agoua's restoration emerged, evolved and was managed (see Yanow 2000;Bommel 2008).This approach suggests that we live in a world that is variously understood.It is based on the assumption that there are multiple possible interpretations of a social situation (Yanow 2000).We observed what the stakeholders involved in the confl ict were saying and doing, and the contexts in which these talks and actions took place (Silverman 2001).
Data were collected from February 2007 to October 2008, a timeframe that coincided with the negotiation period between the PAMF and the local communities.Data were gathered through semi-structured interviews and informal conversations with the stakeholders, participatory observation of meetings held in the framework of the management of the project and the confl ict, and analysis of documents concerning the project and the confl ict.In total, 19 farmers with farms in the forest, six members of restoration committees, seven staff members of the PAMF, six staff members of the Bantè municipality, and the Director of the Centre for Agriculture Promotion of Bantè Municipality (CeCPA: Centre Communal de Promotion Agricole) were interviewed.The individual interviews were supplemented by six focus group discussions with the farmers.These focus group discussions gathered farmers who responded favourably to our invitation to talk about the confl ict.Two negotiation meetings held during the confl ict management process between the stakeholders involved in the confl ict were attended and data were recorded.
The interviews and conversations were tape-recorded, transcribed and coded.We analysed the transcribed texts and documents using notions and techniques from discourse analysis methods.Discourse analysis focuses attention on the way language is used, what it is used for, and the social context in which it is used, including its effects (Punch 2005).Another reason for this choice was that discourse analysis studies use not only transcripts of talks like conversation analysis, but also other sources based on transcripts of open-ended interviews, or on documents of some kind (Silverman 2001;Wooffi tt 2005).Through discourse analysis, we investigated how confl ict emerged, evolved and was managed during the participatory restoration of the Agoua Forest.

Research Setting
This research focuses on the Agoua Forest confl ict in Benin.The Agoua Forest is a state forest, which was declared a protected area by Decree No. 8 104 SE of 4 November 1953.At that time, the forest covered 75,300 ha (Akpado 1996).In 2002, it was reduced to 68,848.43 ha (PAMF 2006).The forest is located in the western part of central Benin and is entirely in Bantè Municipality (see Figure 1).
Before its classifi cation as state owned, the Agoua Forest was managed by local communities whose traditional authorities had taken care of the conservation of this ecosystem for many centuries (PAMF 2006).This traditional institution was steered by the king and his court, who are present even today.
After the conversion of the Agoua Forest into public domain, its management was taken over by the forest administration, which adopted a coercion and repression management approach but had only a few forest rangers to enforce the implementation.Local communities, who were consequently removed from the management and use of the forest, adopted a rebel attitude towards the natural resources on which their livelihood depended.During the 1960s and 1970s, the Agoua Forest was illegally occupied by these local communities in search of land for agriculture.According to the 2002 census, the local population consisted of about 58,594 persons or 8,194 households living in and around the forest of Agoua (RGPH 2003).Most of them depended on the forest for their livelihood; agriculture, the main source of livelihood, was practised by about 80% of the villagers, followed by hunting and fi shing (PAMF 2006).The introduction of cashew plantations increased demand for land in the region and pushed the farmers to settle in the forest.The PAMF estimated that 5,889 ha of the forest were occupied by local communities' farms, plantations, and fallows, and 50 ha occupied by houses scattered in the forest.Most of the farmers settled in the forest had migrated from other regions of Benin and abroad in search of land, and were installed in the forest by the local traditional authorities who managed the land tenure system.To date, in each village, there is a spiritual chief who is the keeper of all the lands inherited from the ancestors.These lands are divided and each part belongs to a collectivity that shares it among its members.Until the end of the 1980s, land was lent to the newcomers without any compensation.Since the beginning of the 1990s, the local traditional authorities have been receiving money or drinks before giving land to any migrant because of the growing scarcity (PAMF 2006).
According to the PAMF ( 2006), the main causes of Agoua's degradation were the illegal and excessive logging and poaching, weak control system in the forest, the slash and burn technique in agriculture resulting in excessive and ineffi cient use of land in the forest, late bush fi res provoking wildfi re, and overgrazing in the forest by the herders' cattle.Illegal logging was organised by small-scale entrepreneurs and loggers at local and regional levels.They used other loggers and unskilled labour recruited in the villages closer to the forests, and they were often in complicity with local authorities and/or local forest administration representatives (Siebert & Elwert 2004).The PAMF was initiated in 2002 to start the restoration of the Agoua Forest.The PAMF was a fi ve-year project implemented by the Ministry of Environment and Nature Conservation to restore and design the management plan of Wari-Maro, Monts Kouffè and Agoua Forest Massifs.

CASE STUDY The Participatory Management Plan for the Agoua Forest
In Benin, participatory forest management is carried out through the establishment and implementation of the participatory management plan (PMP: Plan d'Aménagement Participatif) for the forests.The PMP for a forest is a document that aims to incorporate the knowledge and needs of the local communities and forest administration in a sustainable forest management plan (PAMF 2006).It is elaborated by the stakeholders, such as the project team, local communities, nongovernmental organisations, socio-political authorities, forest experts, etc.The PMP describes the resources available in the forest and how they should be managed in a sustainable way for a period of time (10 years in the case of the Agoua Forest).The PAMF began the establishment of the Agoua Forest PMP with awareness raising campaigns among local communities on the importance of natural resource conservation, and the objectives and activities to be carried out by the project.At the same time, the zoning of the forest was undertaken by the PAMF staff members and discussed with farmers' representatives.For the zoning process, aerial photos were taken to sort out the different levels of occupation and degradation of the forest.In fact, during the repressive management regime (in force from the time the forest was put under state protection in 1953, until participatory management was introduced in 1993), despite the Agoua Forest being supervised by forests rangers, farmers settled there, and 13 villages and hamlets had been created in the forest (PAMF 2006).The Agoua Forest was eventually divided into four zones: service zones, protection zones, agroforestry zones, and production zones.
Service zones consisted of roads that enabled travellers to reach villages located inside the forest.Protection zones represented ecosystems that bordered the three main rivers in the forest: Zou, Ogou and Otio.They were integrally conserved and protected from any human intrusion.Agro-forestry zones concerned parts of the forest mainly occupied by farms, fallows, private plantations of more than 1 ha in size, and the villages.According to the PAMF ( 2006), the lands reserved for agro-forestry were three times larger than the lands initially occupied by the farmers in the forest, covering an area of 21,831.09ha, to take account of growth in the local populations and the required expansion of farms in the future.The agroforestry zones were dedicated to the cultivation of food crops.However, cashew plantations of more than 1 ha were to be preserved.Also, farmers installed on these lands should pay an annual fee for occupying the state's land.For land in the agro-forestry zone occupied before the zoning, the annual fee was to be 10,000 XOF per ha per year, and for new land occupation, 20,000 XOF per ha per year.Cashew plantation owners were to pay 20,000 XOF per ha every year to continue harvesting their plantations.The remainder of the forest was designated as a production zone, comprised of those parts of the forest intended for reforestation, and for exploitation by forest users under contracts with the forest management team during the implementation phase of the PMP.
Because forest protection and production zones had been subject to signifi cant human infl uences (logging, hunting, grazing, agriculture, etc.) their level of degradation was high and they needed to be restored.Their reforestation with fast growing forest trees was entrusted to local communities.Contracts were signed with local communities' committees for forest restoration that produced and planted trees in the degraded parts of the forest.They were to take care of these trees for four years after planting them, before leaving them to grow naturally.The local hunters were transformed into forest guards to assist forest rangers in preventing illegal logging, hunting, grazing, and fishing in the forest.The project also improved infrastructure (e.g., construction of roads, wells, pumps, and warehouses) in the villages adjacent to the forest, to help farmers store their agricultural produce.Income-generating activities (grasscutter breeding, rabbit breeding, beekeeping for honey, and butter production by women) were initiated and fi nanced by the project to reduce local communities' poverty level and thus their dependence and pressure on the forest.These project activities mobilised a signifi cant number of people from the local villages and provided them with substantial income.When referring to this period, local communities used the following kind of utterances: When the PAMF came we were happy as we heard that it would fi ght bush fi res and then our cashew plantations would not burn.We also believed that it would bring jobs to our region (Source: Focus group discussion with Bantè's farmers, March 2007).

And:
At the start, we were happy with the PAMF as it built infrastructure in our villages and enabled many villagers to earn money by working with the restoration and hunting committees.We thanked the government for choosing our region for the implementation of this project (Source: Focus group discussion with Bantè's farmers, March 2007).
In these interview excerpts, local communities presented the project and its activities positively during the elaboration of the PMP.The Mayor of the Bantè municipality described local communities' initial feeling about the project's activities as follows: Local communities accepted this project with enthusiasm, because of its vision and policy of forest conservation and restoration, while simultaneously improving livelihood and reducing the level of poverty (Source: Bantè's Mayor, November 2007).
These utterances indicate that, at the start of the project, local communities agreed with its objectives and activities.In their discourse, they praised the presence of the project in their region.They used positive frames to characterise the project and its activities.No confl ict or clash was perceptible in the discourse of the local communities until the end of the planning stage of the PMP, prior to its proposed implementation in 2006.

Explosion of a Confl ict
In 2006, the PAMF announced the beginning of the implementation of the PMP and this implied the execution of the Zoning Plan.The project management staff then decided to repossess forest lands occupied by farmers in the protection and production zones.Instead of throwing all the farmers out of the forest, the PAMF asked them to move into the agroforestry zones dedicated to farming.The PAMF planned to destroy all the farms and plantations in the production and protection zones in order to replace them with forest trees.This was contested by the farmers.Farmers agreed to abandon their food-crop farms but not their cashew tree plantations because they considered the latter part and parcel of the forest.Farmers installed in agro-forestry zones refused to pay the stipulated fees for land occupation.According to the farmers, the PAMF wanted to chase them from their lands, and they were struggling to resist the PAMF's attempt to extort their farms from them.The following interview excerpts highlight their thoughts: We were living here in peace and working on our farms when the PAMF came and created the confl ict.If somebody has his possessions and another wants to extort him, it means creating a confl ict (Source: Farmer informant, November 2007).

Another farmer added:
We don't agree with the way the PAMF manages our forests because they want us to leave our farms (Source: Farmer informant, November 2007).
In these testimonies, farmers presented the PAMF people as the troublemakers.Taking a closer look allows us to fi nd identity and characterisation frames that contributed to the start of the problem.The phrase 'We were living here in peace and working on our farms…' suggests that the farmers considered themselves as quiet, hardworking and peaceful beings, whereas the PAMF staff members were characterised as distorting their peaceful lives ('…when the PAMF came and created the confl ict', and '…because they want us to leave our farms').With the phrase 'If somebody has his possessions and another wants to extort him, it means creating a confl ict', the problem is brought to a higher, generalised level, aimed at getting support from the researcher as well.
The confl ict broke out when the PAMF destroyed a cashew plantation to replace it with a plantation of forest trees.Farmers started blaming the PAMF, saying that it had changed the initial objectives and agreements: When the PAMF started, its staff members organised meetings in our villages.At these meetings, they said the project will be implemented in our region.We asked them what they really wanted to do and they replied that they were coming to restore the protected forests of our region.We then asked them whether or not we would be chased away later.They told us that they would not chase us away; rather that they were coming to work together with us.
That was what we agreed upon together (Source: Farmer informant, August 2008).

Another farmer said:
When the project came fi rst, they did not tell us what they are doing now, namely, chasing people away from the forest.They said that they would give farmers some tree seedlings to replace the trees farmers had destroyed on their farms.We would grow our crops while simultaneously planting forestry trees.When these forestry trees were big enough, we would leave these sites.Suddenly, they asked some people to destroy our plantations and crops (Source: Farmer informant, August 2008).
The PAMF was pictured by the farmers as untrustworthy as the utterances show ('They told us that they will not chase us away; rather they came to work together with us', and 'When the project came fi rst, they did not tell us what they are doing now, namely, chasing people away from the forest').The farmers justifi ed the fact that they had not protested when the zoning was being established, and that they were now refusing to leave their farm by characterising the PAMF as liars ('They said that they would give farmers some tree seedlings to replace the trees farmers had destroyed on their farms... Suddenly, they asked some people to destroy our plantations and crops').
Also, the government offi cers were blamed for not having informed the farmers about the status of the forest in the past and instead allowed them to settle in the forest and helped them to establish themselves there, whereas the same offi cers were now allowing the PAMF to chase them away, as becomes clear in the next two utterances: When we came back from Ghana in 1970s, we heard that some farmers had been chased from one side of the forest.Then we settled on the other side.We were not informed that this side was also part of the forest until the PAMF came.When we were settling, nobody bothered us and now after working for so many years we are warned to leave (Source: Farmer informant, August 2008).

And:
In the past when the Ministry of Agriculture used to take care of the forests, the extension workers helped us to plant cashew trees in our farms in the forest.We established villages in the forest and the government built schools, dug wells and even opened health centres in these villages (Source: Farmer informant, August 2008).
Through these utterances, farmers presented themselves as the ultimate victims.Apparently, the Bantè's Mayor acknowledged the responsibility of the state for the actual situation: You may be better informed than me about the status of this forest as a protected forest, the fact that it belongs to the state and cannot be occupied by anyone.However, in the past the state made the mistake of abandoning the forest for more than forty years, and this facilitated and favoured the settlement of the local communities in this area (Source: Bantè's Mayor, August 2008).
The CeCPA Director in Bantè Municipality also argued: In one way or another, the state is the one responsible.In the 1950s, this forest was declared a protected area.Since then, people have been allowed to settle in it.The state built some infrastructure, such as schools, health centres, wells, etc., in the villages created in the forest.Some of these villages are nowadays offi cially recognised villages.Then suddenly, one day farmers are told to leave as they are in a public domain.It would have been better if the state had taken its responsibility from the start by fencing the forest.Then, this situation would never have occurred (Source: Bantè's CeCPA Director, August 2008).
However, contrary to these vocalised stakeholder perspectives, the PAMF rejected the analysis of both the farmers and the Bantè municipality.This is highlighted in the following testimony of the PAMF offi ce head in Bantè Municipality: For them [the farmers] the project will just establish the Zoning Plan and stop by the end of the last year (2006) as it is a fi ve-year project.So then, they will go back to their initial places in the forest.It was clear in their mind that 'we will help them make it and the project will fi nish before they implement it.They will leave and we will go back again to our places'.The forest will become what it was before the project.When they realised at the end of 2006 that the project started again with the implementation of the PMP, they said 'we will never leave' (Source: Bantè's PAMF offi ce head, September 2008).
In this utterance, the PAMF explained the situation by the fact that, when the zoning was agreed upon, farmers did not protest because they thought that the actual implementation of the Zoning Plan would never take place.As demonstrated by the PAMF staff members' utterances below, the PAMF characterised the farmers' protests as selfi sh, since those who were complaining were the ones who had to move, and thus had a direct stake: After the zoning of the forest, everybody agreed it was not acceptable for farmers to be scattered everywhere in the forest.So, they had to be concentrated in the agro-forestry zones.However, at that time we did not know who had to leave the forest and who could stay.When the details of the Zoning Plan were demonstrated in the fi eld, those who found their lands in production and protection zones of the forest, and who had to be displaced, started to complain that they didn't agree with the zoning, nor would they leave (Source: PAMF staff member in Bantè, September 2008).

They added:
The farmers to be relocated are those who settled deep in the forest and are still extending their farms.We told them that they had to leave.They said that they agreed with us but that they had to be compensated before they left.That's the whole problem.The farmers are just asking to be compensated for the plantations they have to leave (Source: PAMF staff member in Bantè, September 2008).
In both testimonies, the PAMF suggested that farmers in general agreed with the zoning, but only the few who had to be relocated insisted on the preservation of their cashew plantations or compensation.
The utterances so far have been parts of conversations of stakeholders, who share a set of perceptions, norms, and expectations, as constructed and expressed in their talks (we-groups) with the researcher (Gumperz 1965).What is striking is that in such we-contexts both farmers and the PAMF managers rejected any responsibility for the situation and overtly blamed and accused the other side.By doing so, they presented themselves as victims and the other party as causer of the confl ict.Farmers blamed the PAMF for not telling them the truth at the start of the project, and the government for encouraging their settlement in the forest for many years before asking them to leave.From the PAMF's perspective, the farmers ignored the Zoning Plan as they thought that it would not be implemented before the end of the project.
These interview excerpts also show a shift in farmers' characterisations of the PAMF from positive to negative, and the emergence of confl ict frames in their discourse from the moment the implementation of the Zoning Plan was announced.Accusations that the PAMF was untrustworthy and a liar, emerged when the PAMF started to talk about its decision to really implement the Zoning Plan.The PAMF rejected these accusations and constructed the problem as a fi ght started by those farmers who were concerned with their personal interests.Victim identity frames, and blaming and accusing characterisation frames emerged in the stakeholders' discourse, reinforcing the tension between the different stakeholder groups.Each party was tacitly attributing the causes of the situation to the others.Here, we recognise the situation as described by Ford et al. (2002), who argued that change and resistance to change are a function of the ongoing background conversations constructed in we-group conversations, and consequently form the context for both the change initiative and the responses to it.

Evolution of the Confl ict
The destruction of a cashew plantation incited farmers to action.They wrote a letter to the President of Benin on October 12, 2006, with copies to the Ministry of Environment and Nature Conservation, and the Ministry of Agriculture, Animal Breeding and Fisheries.In this letter, they called upon the President of Benin for help to preserve their plantations and farms in the forest.To persuade the President of Benin, they used an argument that they thought the President of Benin would fi nd relevant: the PAMF was hindering their efforts to increase crop production in their region by destroying their farms.In informal conversations however, the farmers also claimed their right to protection from the state as citizens represented by the President of Benin, arguing: We agree that the land is owned by the state, but we as human beings also belong to the state (Source: Farmer informant, August 2008).
Implicitly, they accused the state of not taking care of them by refusing to take their side.Both the ministries contacted sent a separate team to investigate the situation.They brought back different pictures about the situation.Then the President of Benin asked for an inter-ministerial commission to assess the situation and to come up with propositions to solve the problem.
The inter-ministerial commission was the fi rst to call a meeting to hear the different stakeholders' opinions on the situation.The meeting was held in the conference room of Bantè Municipality on March 3, 2007.The representatives of the Ministry of Agriculture, Animal Breeding and Fisheries, and the Ministry of Environment and Nature Conservation (forming the inter-ministerial commission wished by the President of Benin), all the farmers' representatives of all the local villages, the PAMF staff members, the Mayor of the Bantè municipality, and the Director of CeCPA were present at the meeting.
At this meeting, the PAMF staff members reminded the participants how the zoning had been determined.According to the PAMF, fi ve meetings had been organised with local communities' representatives-the village councils for sustainable development (CVDD: Conseil Villageois de Développement Durable) of all the local villages-to talk about the Zoning Plan.On each CVDD, farmers had at least one representative.As far as the PAMF was concerned, the CVDDs were in charge of informing the farmers of the outcomes of the meetings.The different zones, their surface areas and their functions were shown to the farmers.The PAMF reassured the farmers that nobody would be chased out of the forest, but rather that those who had their farms in the production and protection zones would be relocated to the agro-forestry zones.
The farmers publicly expressed their opposition to leaving the forest.The farmers argued that they were not involved in the elaboration of the Zoning Plan; in particular, the farmers contested the five meetings held to discuss the zoning.According to the farmers, some meetings had been organised in the framework of the participatory management of the forest but not especially to discuss the zoning of the forest.The farmers did not even remember the PAMF showing them the agro-forestry zones before they were unveiled at the meeting: When they tell farmers to move to the agro-forestry zones, which are occupied by other farmers, where will the newcomers be settled?Are they going to dispute land with their brothers or is it the government who is going to redistribute these lands to the farmers?We don't even know where these places [agro-forestry zones] are (Source: Bantè's farmers at the meeting, March 2007).
Later, in an interview with him, the Director of the CeCPA supported this argument by the farmers and recognised that: The fundamental problem that exists is where the farmers will settle if they leave their plantations in the forest.The farmers estimated that about 10,300 households settled in the forest had to leave.Where will they settle?Bantè Municipality is bordered by the Monts Kouffè Forest in the north, in the north-east by another forest, the Wari-Maro Forest, and in the south by the Agbado River.The Agoua Forest occupies the whole western part of the municipality (Source: Bantè's CeCPA Director, August 2008).
For the farmers and the Director of the CeCPA, the lands the PAMF presented as representing the agro-forestry zones were occupied by other farmers who were forbidding their peers to settle there, arguing that these lands were their fallows.According to them, there was no land left for them in that area.Even if they got lands in these zones they would not be allowed to recreate their destroyed cashew plantations, which represented their main source of livelihood.Some farmers stated that they had been chased away from other countries, and would not accept being chased anyway from the lands of their ancestors.They warned whoever intended to destroy their farms to be ready for a fi ght.Either the farmers would kill him or they would be killed.At the end of the meeting, the farmers refused to sign the report of the meeting.The behaviours of the PAMF and the farmers' representatives in this we-versus-they interaction setting show that each party was fi ghting to gain credibility in the eyes of the inter-ministerial commission.The PAMF tried to convince the participants at the meeting that it had done its job properly.According to the PAMF, the situation was due to farmers' representatives who did not pass on to their constituency the zoning information communicated at the meetings.The PAMF also tried to calm the situation through utterances like: We will not chase you out of the forest, but relocate you.Only those who have their farms and plantations deeper in the forest will be moved into the agro-forestry zone (Source: PAMF staff members at the meeting, March 2007).
At the same time, the farmers' representatives denied all the actions that the PAMF claimed to have undertaken to inform them about the zoning process.They even characterised the PAMF staff members as 'killers', like in this piece: For us, the PAMF wants to kill us alive.They want to kill farmers alive, which means killing us before our age to die, as some of the farmers are old and unable to create new plantations, which are their only one source of livelihood now and during their retirement period.Since God created Bantè our municipality, no factory has been built here.We only grow crops and the land is our wealth (Source: Farmer informants, March 2007).
With the aim of gaining the support of the commission (Shmueli et al. 2006;Marichal 2009), they again framed themselves as victims of the PAMF's actions.They expressed their distrust to the PAMF by refusing to sign the report of the meeting at the end.At that meeting, farmers expressed themselves freely as they constituted a majority compared to the members of the commission and the PAMF staff members.Emotion, anger and violence were noticeable in all their utterances.The PAMF staff members, in their turn, used positive identity frames to describe themselves.The difference between the farmers' negative characterisation frames of the PAMF staff members and the PAMF staff members' own identity frames increased the distance between these stakeholders and thus reinforced the confl ict.
Identity and characterisation frames were used by the different stakeholders in this we-versus-they interaction setting to persuade the members of the inter-ministerial commission to support their position.The commission heard the farmers' complaints and the PAMF's reactions.As discussions raised tension, the commission decided to stop the meeting and call another negotiation meeting with only the farmers' representatives, the PAMF and the other stakeholders involved in the confl ict.Before the commission members left, they asked the farmers to stay calm and reassured them that they would solve the problem.However, despite the commission's warning, on August 3, 2007, the farmers held a demonstration march, that the national television was invited by the farmers to cover and broadcast.This action by the farmers aimed at gaining governmental support: Because they did not follow what the President said, that is, to discuss with us and look for our agreement before taking any action and starting to destroy our farms, we decided to have a demonstration.That will enable the government to know what is happening here (Source: Farmer informant, September 2008).
When broadcast on national TV, this news indeed pushed the President of Benin to take the situation seriously, and he sent the Minister of Environment and Nature Conservation to solve the problem.The Minister of Environment and Nature Conservation organised a tour in the project area to see what the project was doing and had a discussion with the members of the project and the farmers.After listening to all the parties, the Minister of Environment and Nature Conservation appeased the farmers and promised to organise a meeting between the representatives of the parties and herself.

Escalation of the Confl ict
After two postponements, the last negotiation meeting was fi nally held on October 18 and 19, 2007, in Bohicon but without the presence of the Minister of Environment and Nature Conservation, who sent representatives.Representatives of the Ministry of Agriculture, Animal Breeding and Fisheries, the Ministry of Environment and Nature Conservation, the PAMF project, and those of the Municipalities of Savalou and Bantè, the farmers, the forest restoration committees, the parliament member of the Bantè Municipality, and the local radio reporter were present at the meeting.
The meeting opened with a résumé of the history of the confl ict as some participants had not been involved since the beginning.The process that resulted in the zoning was presented by the PAMF.The fl oor was then opened to the farmers' representatives.Farmers started by accepting the zoning of the forest and the principles involved.Next, they started negotiating by proposing a wish list.They wanted the occupation fees payable by the farmers in agro-forestry zones to be abolished or at least reduced.They also proposed that cashew trees and forest trees should be allowed to co-exist in the forest; they wanted all the plantations in the forest to be saved.If this was not feasible, they suggested assisting the farmers to create new cashew plantations outside the forest.In this case, they asked for a moratorium of about twenty or thirty years to continue harvesting their cashew plantations in the production and protection zones, in order to be able to create new cashew plantations outside the forest.They asked for fi nancial compensation to be paid by the government if they had to abandon their plantations.And they asked the PAMF to accept the creation of new cashew plantations in agro-forestry zones.
The PAMF representatives reacted by arguing that farmers would not be allowed either to create new cashew plantations in the agro-forestry zones or to continue harvesting the existing plantations in the production and protection zones.The PAMF representatives feared that otherwise the farmers would continue exercising their property right on these lands, which their children would consider as their heritage.The PAMF agreed to compensate farmers by assisting them to create new cashew plantations outside the forest.They would receive improved cashew varieties and technical assistance to grow them.However, it was out of the question for the PAMF to give fi nancial compensation to the farmers who had to leave the forest, because the land belonged to the state.The only compensation that the project could give the farmers was to help them start beekeeping and raising grasscutters.Farmers would be allowed to continue harvesting cashew nuts in plantations that were more than fi ve years' old in the production and protection zones for eight years only, to enable them to create new ones outside the forest.Plantations that were less than fi ve years' old would be destroyed.The PAMF agreed to reduce the annual fees payable by farmers in agroforestry zones, from 10,000 XOF to 7,500 XOF per year for those who were already settled in the forest for annual crop, from 20,000 XOF to 15,000 XOF for new land occupation in the agro-forestry zones, and from 20,000 XOF to 15,000 XOF for cashew plantations in the forest.Finally, the PAMF promised to note in the report to be sent to the Minister of Environment and Nature Conservation that farmers would like compensation for their investments (cashew plantations) in the forest.This proposition was to be presented to government to be studied in detail and to see whether or not the government would take social measures to help them.These PAMF propositions constituted the points of the meeting's report.All the participants present at the meeting signed the report; this was interpreted by the offi cials as agreement.
However, although all the participants signed the report, later interviews with them showed that they had different interpretations and views about the meeting and the outcomes.The representatives of the farmers reacted in this way: Researcher: What do you think about the outcomes of the meeting?Farmers' representative 1: All the solutions retained there were proposed by them.They did not accept any of our propositions.We asked them to compensate us and they said that they could not ask the President to give us compensation as many other villages are in the same situation.If they compensate us they will have to do the same for these other villages.We told them that we were not asking for compensation for the land we are leaving but for our plantations, as we would need a lot of funding to start other ones.Many of us are old and weak and do not have the strength necessary to create a new plantation without funding.
Researcher: But why did you sign the report of the meeting?
Farmers' representative 1: We just accepted keeping in mind that we will not leave unless the state pays us every year what we get from our cashew plantations.We know that it cannot do that, so we are sure that we will never leave.Even when we told this to the farmers, most of them answered that they don't want any kind of compensation but only to let them continue harvesting their cashew plantations.
Farmers' representative 2: First of all, they invited us to a city far from the place where the problem is taking place, and at the end of the meeting they threatened us by saying that those parties who will not respect the agreements will be taken to court (Source: Interview with farmers' representatives, November 2007).
In this we-group discussion, farmers vocalised their hidden agenda, which was to conserve their farms in the forest.Even though they signed the report, they neither agreed with the PAMF's propositions nor expressed this at the we-versus-they meeting.According to them, they started to negotiate, proposed solutions, and signed at this specifi c meeting because they were a minority (four farmers out of 29 participants).As they were less represented than the other stakeholders, they did not feel able to express and defend their arguments as they did in the former meeting.Instead, farmers postponed their contestations to later, when the plan was being implemented.
For the PAMF, the problem was solved at the end of this meeting as the farmers signed the fi nal report; this means that in eight years they will leave the forest without any concessions.
The Minister of Environment and Nature Conservation said that she does not understand how a forest could be restored without any damage.She also said that enough concessions have been made by the PAMF.At this stage, she said, it is time to go to the essential matters.The awareness raising time is over, it is time to act (Source: PAMF staff members, November 2007).
The PAMF used this testimony of the Minister of Environment and Nature Conservation to give more credibility to its own plan, which was to throw farmers out of the forest.Like the farmers, the PAMF vocalised this plan only in interviews with the researcher, not in the meeting where it probably should have been discussed.As a result, the agreement that was reached at the meeting was no more than a pseudo consensus, hiding the confl ict that was nevertheless experienced by all stakeholders involved.
No action was taken after this meeting in the framework of the management of the confl ict.The project ended in June 2008 in this impasse, without implementing the Zoning Plan.Although a lot of discussions took place, both within we-groups and between the different stakeholders, at the end of the project nothing had been implemented.
The research shows that the confl ict was constructed by means of talk.Failing negotiations in combination with a lot of stereotyping and stigmatising that mainly took place in we-groups reinforced the confl ict.The project ended in a formal agreement that hid a serious impasse.Over time, the issue of implementing the forest restoration evolved into a protracted or intractable confl ict (Gray 2003;Shmueli et al. 2006;Coleman et al. 2007).

DISCUSSION AND CONCLUSION
Although strict natural resource conservation in the past has succeeded in preserving such resources in some cases, in many others it has created confl icts between local communities and public institutions charged with conservation (Rishi 2007).Participatory management was initiated to avoid these confl icts and promote sustainable management of natural resources.However, in the Agoua participatory restoration project, despite its positive start and agreed-upon plan, confl ict emerged, and local communities opposed the project.
The dispute began when the PAMF started talking about the implementation of the PMP.Farmers framed this situation as a threat, because they felt that the PAMF wanted to throw them out of the forest, and this triggered frame shifting on their part.In retrospect, they started blaming the government for not forbidding them earlier from settling in the forest, and the PAMF for not informing them about the zoning of the forest at the beginning of the project.Until then, the PAMF had considered that the farmers had been informed about the Zoning Plan by their representatives, who attended all the meetings organised by the PAMF to discuss the zoning.The PAMF was surprised by the farmers' reactions as its staff members had not monitored the feedback process of the farmers' representatives to their constituency.The support of stakeholder constituencies is critical to participatory management and confl ict resolution.Hence, meetings and discussions should not be limited to stakeholders' representatives but involve the wider stakeholder population.
The conflict was constructed, interpreted, enacted and maintained in stakeholders' discourses through which they emphasised different realities and had different senses of themselves and their world (Ford et al. 2002).As in other conflict situations, frames uttered in different contexts consisted of blaming, stereotyping and stigmatising, resulting in divergence and distancing among the stakeholders (Aarts et al. 2011;Gray 2003).All stakeholders portrayed their opponents negatively and projected responsibility for the confl ict onto them.At the same time, they presented themselves as being the victims of the others' behaviours.Stakeholders constructed and expressed their frames in wegroup conversations, recognising and affi rming each other.In we-versus-they interactions, depending on their framing of power positions and interdependence, stakeholders used denial and disapproval to attack each other or to start positional bargaining.In public they referred to frames constructed in we-group conversations, to look for support from powerful outsiders.
By analysing and connecting different conversation contexts (both over time and between different people), this study makes clear why negotiations did not result in stakeholders' reframing the situation, but rather ended in distrust, accusations, and even threats.
The conflict management process did not succeed in creating a new reality that would enable the deconstruction of stakeholders' former frames, as no effort was made inside or outside the meeting arenas to exchange and critically discuss these we-group frames to trigger integrative reframing (Ford et al. 2002).This may be explained by the fact that all the parties had a stake in the confl ict.No neutral party was involved in the confl ict management process as was requested by the farmers, who called upon the President of Benin to mediate because they did not trust the PAMF.Although the request to the President of Benin led to the interest and involvement of relative outsiders, they did not play a mediating role.They merely attended a PAMF management meeting that led only to pseudo consensus.We agree with Gray (2003: 32) who argues, "Since reframing requires perspective taking, it is often diffi cult for parties to reframe without the help of a neutral third party or someone who does not have a direct stake in the confl ict".Thus, one way of solving the confl ict in relation to the restoration of the Agoua Forest may be to involve a neutral mediator in its management.The main task of the mediator would be to bring the stakeholders together to discuss their frames with regard to the problem, including the causes and solutions they have in mind, with the aim of creating mutual understanding, which may trigger common reframing (Shmueli et al. 2006).
It can be concluded that in participatory management there is no shortage of power struggles between stakeholders (Leeuwis 2000(Leeuwis , 2004)).Our analysis shows that confl ict was constructed by means of conversations, both within we-groups and in negotiations with all the stakeholders involved.Conscious and unconscious, explicit and implicit accusations destroyed the relationships between the stakeholders, reinforcing the distance and thereby reducing the space for changing situations.To understand the evolution of a confl ict, an interactional framing approach (such as we used) showing how confl icts emerge, evolve and end, or persist in talk, seems to be effective (Aarts et al. 2011;Dewulf et al. 2009).
From this case study, we learnt that, if potential problems are not properly discussed at the beginning, confl icts emerge during implementation.On the basis of this study, we stress that people who are involved in participatory protected area management should pay more attention to the divergences that often exist between upper level and local discourses (Bosak 2008).Not only should one take into account what happens during the participatory meetings, but also what happens in we-group conversations in which stereotypes and stigmas are constructed.Hence discourse can create confl ict, even when utterances are not supported by concrete action.In our case, the confl ict was a discursive affair as very little changed on the ground.