Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Growing political : how forest-related violence shapes community-based forest management in Cambodia Grant, Victoria Hollie 2017

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2018_february_grant_victoriahollie.pdf [ 18.78MB ]
Metadata
JSON: 24-1.0362111.json
JSON-LD: 24-1.0362111-ld.json
RDF/XML (Pretty): 24-1.0362111-rdf.xml
RDF/JSON: 24-1.0362111-rdf.json
Turtle: 24-1.0362111-turtle.txt
N-Triples: 24-1.0362111-rdf-ntriples.txt
Original Record: 24-1.0362111-source.json
Full Text
24-1.0362111-fulltext.txt
Citation
24-1.0362111.ris

Full Text

GROWING POLITICAL: HOW FOREST-RELATED VIOLENCE SHAPES COMMUNITY-BASED FOREST MANAGEMENT PRACTICES IN CAMBODIA by  Victoria Hollie Grant  B.Sc. (Hons.), Durham University, 2008 M.Phil., University of Oxford, 2012  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Forestry)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  December 2017  © Victoria Hollie Grant, 2017 ii  Abstract  Community-based forest management (CBFM) projects support communities to take a central role in managing their local forests. Communities are asked to enforce the exclusion of illegitimate forest users and prevent illicit extraction of forest resources. However, the practices that CBFM aims to reduce (deforestation and illicit extraction of forest resources) are often facilitated by the use of direct violence. In seeking to reduce these practices, some participants in CBFM experience violent conflicts over forest use (hereafter, forest-related violence). Such violence presents a threat to the lives and human rights of CBFM participants and has the potential to undermine the effective implementation of forest conservation activities. Yet, the extent of forest-related violence in community-managed forests, exactly how it manifests, who is involved, or the outcomes of such violence for CBFM participants and their CBFM practices are not well known.  This dissertation explores these issues in the case study country of Cambodia. I draw on data collected between May and December 2015 through a national survey of eighteen NGOs involved in CBFM in Cambodia, semi-structured interviews with one hundred and fifty participants in forty CBFM sites, and participant observation in forest patrols and CBFM training sessions.  I demonstrate that forest-related violence is widespread across Cambodia affecting seventy-five per cent of CBFM groups interviewed and seventy-two per cent of NGOs surveyed. I argue that these ‘incidents’ of violence are not discrete but, rather, are the manifestation of a succession of iii  violent processes in which Cambodia’s neopatrimonial socio-political system is central. Furthermore, neopatrimonialism facilitates processes of structural and symbolic violence that preclude effective responses to direct forest-related violence.   As a consequence, forest-related violence acts as a disciplinary mechanism inciting fear and undermining the effectiveness of CBFM practices. Yet, it also acts as a catalyst for the re-politicization of CBFM practices and re-conceptualization of CBFM participants’ relationships with the environment, other forest users, and the government. Thus, this dissertation directs attention to the lived experience of forest-related violence and exposes the ‘tomorrow of violence’ – the enduring legacies of violence that configure the way people conceptualize themselves, their government, and the international development community.        iv  Lay Summary  Community-based forest management (CBFM) projects support communities to take a central role in managing their local forests. Communities are asked to enforce the exclusion of illegitimate forest users and prevent illicit extraction of forest resources. However, the practices that CBFM aims to reduce (deforestation and illicit extraction of forest resources) are often facilitated by the use of physical violence. CBFM participants are often targeted by this violence, which presents a threat to their lives and human rights and also undermines the effective implementation of forest conservation activities.  This dissertation explores the contexts in which forest-related violence occurs in Cambodia and how it affects the implementation of CBFM activities. It examines responses to violence and how these can be improved. Fieldwork was conducted between May and December 2015 and used a national survey of NGOs involved in CBFM in Cambodia and interviews with one hundred and fifty CBFM participants.      v  Preface  This dissertation is an original, unpublished, independent, intellectual product of the author, V. H. Grant. The fieldwork reported in Chapters 4-6 was covered by UBC Ethics Certificate number H15-00578. vi  Table of Contents  Abstract .......................................................................................................................................... ii	Lay Summary ............................................................................................................................... iv	Preface .............................................................................................................................................v	Table of Contents ......................................................................................................................... vi	List of Tables ................................................................................................................................ xi	List of Figures .............................................................................................................................. xii	List of Abbreviations ................................................................................................................. xiv	Acknowledgements .................................................................................................................... xvi	Dedication ................................................................................................................................. xviii	Chapter 1: Introduction: Conceptualizing forest-related violence ...........................................1	1.1	 Introduction ........................................................................................................................ 1	1.2	 A theory of violence ........................................................................................................... 6	1.2.1	 Violence as structured and structuring ........................................................................ 9	1.3	 ‘Causes’ and ‘consequences’ of environmental violence ................................................ 11	1.3.1	 Factors influencing occurrence and forms of environmental violence ..................... 13	1.3.2	 The structuring effects of environmental violence ................................................... 19	1.4	 Conceptual framework ..................................................................................................... 22	1.5	 Discussion ........................................................................................................................ 24	1.5.1	 Research objectives and introduction to the case study of Cambodia ...................... 29	1.6	 Structure of the dissertation ............................................................................................. 31	vii  Chapter 2: The environmental, socio-political, and institutional context of community-based forest management in Cambodia .....................................................................................33	2.1	 The socio-political context ............................................................................................... 34	2.2	 The ecological context: contested forests ........................................................................ 41	2.2.1	 The political importance of timber for Cambodia’s ruling elites, 1989-2017 .......... 42	2.2.2	 Forest-related violence in Cambodia since 2002: a review of the literature ............. 45	2.3	 CBFM in Cambodia ......................................................................................................... 51	2.4	 Conclusion ....................................................................................................................... 59	Chapter 3: A methodology for recognizing, researching, and representing direct violence in Cambodia’s community-managed forests .................................................................................61	3.1	 Methodological approach ................................................................................................. 62	3.2	 Methods............................................................................................................................ 66	3.2.1	 Selection of case studies ........................................................................................... 66	3.2.2	 Data collection methods ............................................................................................ 72	3.2.3	 Data analysis method ................................................................................................ 75	3.3	 Practical and theoretical challenges and limitations to the research ................................ 76	Chapter 4: Patterns of direct violence in Cambodia’s community-managed forests ............84	4.1	 Introduction ...................................................................................................................... 84	4.2	 Methods............................................................................................................................ 87	4.3	 Forest-related violence by different actors in Cambodia’s CBFM sites .......................... 87	4.3.1	 A typology of ‘neck kab prey cheur’ (people who cut the forest) ............................ 88	4.3.2	 The use of forest-related violence by different actors .............................................. 92	4.3.2.1	 Organized logging groups and forest-related violence ...................................... 97	viii  4.3.2.2	 The military and forest-related violence ............................................................ 98	4.3.2.3	 Local loggers and forest-related violence ........................................................ 102	4.3.2.4	 Migrants, deforestation, and forest-related violence ........................................ 107	4.3.2.5	 Economic land concessions and forest-related violence .................................. 111	4.4	 Case Study: Overlapping and interlocking moments of violence in Oddar Meanchey . 114	4.5	 Discussion ...................................................................................................................... 125	Chapter 5: Unrooted responses: how CBFM participants and development organizations (fail to) address forest-related violence in Cambodia .............................................................130	5.1	 Introduction .................................................................................................................... 130	5.2	 Literature review and conceptual framework: recommended responses to violence against environmental defenders ............................................................................................. 132	5.2.1	 Addressing the ‘root causes’ ................................................................................... 133	5.2.2	 Improving the rule of law ....................................................................................... 134	5.2.3	 Promoting human rights .......................................................................................... 137	5.2.4	 Recommended responses to reduce violence .......................................................... 139	5.2.5	 Upholding the rule of law in Cambodia: a review of relevant institutions ............. 143	5.3	 Methods.......................................................................................................................... 152	5.4	 Responses to forest-related violence in CBFM projects in Cambodia .......................... 153	5.4.1	 Addressing a ‘root cause’ by reducing forest conflicts ........................................... 154	5.4.2	 Using the justice system to implement the rule of law ........................................... 157	5.4.3	 Alternative ways of implementing the rule of law .................................................. 165	5.4.4	 Other ad hoc responses ........................................................................................... 170	ix  5.5	 Discussion: the role of development organizations in responding to forest-related violence and paths forward ..................................................................................................... 174	Chapter 6: Growing political: violence, community forestry, and the shaping of ‘environmental’ subjectivities ...................................................................................................181	6.1	 Introduction .................................................................................................................... 181	6.2	 Conceptual underpinnings: Violence, subjectivities, and performance ......................... 184	6.2.1	 Revisiting concepts of violence .............................................................................. 184	6.2.2	 Subjectivities and performance ............................................................................... 187	6.2.3	 Subjectivities and performance in community-based forest management ............. 189	6.2.4	 Using subjectivities to understand the impacts of and responses to violence ......... 193	6.3	 Violence and environmental subjectivities in Cambodian Community Forestry .......... 195	6.3.1	 Constructing environmental subjectivities through NGO-led CBFM .................... 196	6.3.1.1	 Moral conservationist ....................................................................................... 199	6.3.1.2	 Rational forest manager ................................................................................... 201	6.3.1.3	 Empowered Cambodian citizen ....................................................................... 203	6.3.1.4	 Educated global citizen .................................................................................... 205	6.3.1.5	 Summary .......................................................................................................... 206	6.3.2	 How forest-related violence re-shapes environmental subjectivities ...................... 207	6.3.2.1	 Re-affirmation of environmental subjectivities ............................................... 207	6.3.2.2	 Re-politicization of environmental subjectivities ............................................ 210	6.3.2.3	 Performance and performativity of environmental subjectivities .................... 216	6.4	 Conclusion ..................................................................................................................... 222	Chapter 7: Conclusion ...............................................................................................................225	x  7.1	 The implications of forest-related violence for CBFM practices .................................. 225	7.2	 Research limitations and future directions ..................................................................... 231	References ...................................................................................................................................235	Appendices ..................................................................................................................................264	Appendix A ............................................................................................................................. 264	A.1	 Survey of NGOs ........................................................................................................ 264	A.2	 Interview guide: NGO and donor representatives ..................................................... 266	A.3	 Interview guide: CBFM participants ......................................................................... 267	A.4	 Interview guide: additional questions for management committee leaders .............. 268	Appendix B ............................................................................................................................. 269	B.1	 Violations of the Constitution of the Kingdom of Cambodia ................................... 269	B.2	 Violations of the Cambodian Criminal Code (2009) ................................................ 269	 xi  List of Tables  Table 4.1 Differences and similarities between actors engaged in deforestation and logging in Cambodian CBFM sites ................................................................................................................ 91	Table 4.2 CBFM participants encountering and experiencing violence by each type of non-CBFM member ............................................................................................................................. 93	Table 4.3 NGOs encountering and experiencing violence by each type of non-CBFM member 94	Table 4.4 Forms of and factors contributing to forest-related violence in CBFM sites ............... 95	Table 4.5 CBFM sites targeted by migrants ............................................................................... 108	Table 4.6 Violence related to ELCs used against CBFM participants ....................................... 112	Table 6.1 Summary of CBFM participants’ interpellations and the effect of forest-related violence on environmental subjectivities .................................................................................... 198	 xii  List of Figures  Figure 1.1 Galtung’s (1990) ‘violence triangle’ ............................................................................. 8	Figure 1.2 Conceptual framework ................................................................................................ 24	Figure 3.1 Cambodia in Southeast Asia and case study provinces. Map adapted from OpenDevelopment Cambodia (2017). .......................................................................................... 71	Figure 4.1 CFs in Oddar Meanchey showing those with military bases and/or controlled by the military. ....................................................................................................................................... 100	Figure 4.2 Roots of a valuable ‘beng’ tree (Afzelia xylocarpa) outside a house in a village establishing a CLT, Ratanakiri ................................................................................................... 104	Figure 5.1 Conceptual framework: interlocking strategies to respond to and reduce forest-related violence ....................................................................................................................................... 133	Figure 5.2 Responses to forest-related violence in Cambodia .................................................... 154	Figure 5.3 Training Community Forestry Management Committee members on non-violent conflict resolution, Oddar Meanchey, 17th June 2015. ............................................................... 157	Figure 5.4 A soldier joins a patrol in the Monks’ Community Forest, Oddar Meanchey, 12th December 2015. .......................................................................................................................... 167	Figure 5.5 Freshly cut trees blocking a patrol group’s path through a community forest .......... 172	Figure 6.1 Image used in posters displayed in CF villages promoting sustainable use of the forest. Copyright: Cambodian Forestry Administration and the International Tropical Timber Organization (2009). ................................................................................................................... 200	Figure 6.2 A tree in a CF ordained with Buddhist monks’ robes symbolizing its value as a sacred living being, Oddar Meanchey. ................................................................................................... 201	xiii  Figure 6.3 Sign at the entrance of a CF, Kampong Thom proclaiming, “wildlife is our property.”..................................................................................................................................................... 204	xiv  List of Abbreviations  ADHOC Cambodian Human Rights and Development Association1 CBFM  Community-based forest management CBNRM Community-based natural resource management CCHR  Cambodia Center for Human Rights CF  Community forest/forestry CFMC  Community forest management committee CLT  Communal land title CPA  Community protected area ELC  Economic land concession FA  Forestry Administration FAO  Food and Agriculture Organization of the United Nations LANGO Law on associations and non-governmental organizations LICADHO Cambodian League for the Promotion and Defense of Human Rights2 MAFF  Ministry of Agriculture, Forestry, and Fisheries MoE  Ministry of Environment  NFP  National Forestry Programme 2010-2029 NGO  Non-governmental organization  NRM  Natural resource management                                                 1 Known by its French acronym 2 Known by its French acronym 3 For a discussion of the concept of violence, see section 1.2 below. 4 Thus, while not all environmental defenders are CBFM participants, all CBFM participants can be considered 2 Known by its French acr ym xv  NTFP  Non-timber forest product OECD  Organization for Economic Co-operation and Development RCAF  Royal Cambodian Armed Forces REDD+ Reducing emissions from deforestation and forest degradation RGC  Royal Government of Cambodia UNDP  United Nations Development Programme UNTAC United Nations Transitional Authority in Cambodia    xvi  Acknowledgements  I would like to express my sincere gratitude to my supervisors, Philippe Le Billon and Rob Kozak, for their unwavering intellectual and emotional support throughout my doctoral studies. I am grateful for their patient guidance and words of encouragement on my academic journey. I would also like to offer my sincere thanks to my committee member, Hannah Wittman, for her incisive critique and comments, which helped me to clarify my arguments.  I am deeply indebted to the many people who made it possible for me to conduct my research. First and foremost, I am grateful to all the CBFM groups and interviewees who not only patiently answered my never-ending questions and willingly discussed sensitive topics with me but also extended the utmost hospitality from teaching me how to sling my hammock so it wouldn’t fall down in the night during patrols to frying up special (pregnant) frogs to stop us going hungry when interviews turned into day-long conversations and forest-walks. To Narin and Pay, my dedicated Research Assistants, without whose extraordinary hard work, sense of adventure, natural ability to put interviewees at ease, and commitment to the project, we would never have been able to execute such an ambitious research project. I am also grateful to Tem Ven, Sylvie Kee, Va Moeurn, Venerable Bun Saluth, Phally, Sovannara, Alex Diment, and Amanda Bradley for logistical support and conversations that helped me understand the problem I wanted to explore and the data I had collected. Finally, thank you to Channa, my Khmer teacher, who taught me the meanings of language rather than a list of vocabulary.  xvii  I am grateful for the financial support of my PhD studies and fieldwork from the Canadian Rhodes Scholars’ Foundation, the National Geographic Society (grant number GEFNE156-15), Forest Trends, the University of British Columbia, the Faculty of Forestry, and the Namkoong Family Fellowship.  I also extend my thanks to my friends and family who have supported me from near and far – whether by providing a floor, sofa, or bed during fieldwork and visa runs, listening to me hash out ideas, or simply reminding me to get away from my desk once in a while, you have helped make the last five years go by more smoothly. In particular, my sincere gratitude goes out to the Grants and Gormleys for their moral and financial support of my research, and especially to Joe, whose friendship has helped me navigate the squalls of PhD life since picking me up at the airport when I first arrived in Canada. Special thanks to the yogis at Acorn and the Down Under School of Yoga whose friendship and wisdom has helped me find moments of peace when the research project felt overwhelming, and who encouraged me to remember the value of the process as well as the end result. Finally, I offer my enduring gratitude to my parents, who have supported my exploration of the world through both travel and education, and to Tyler, without whose unwavering love, encouragement, understanding, and cooking, I may never have completed this project.   xviii  Dedication  For Samnang, and all those risking their lives to protect Cambodia's forests   1 Chapter 1: Introduction: Conceptualizing forest-related violence  1.1 Introduction Community-based forest management (CBFM) projects support communities to take a central role in managing their local forests with the aim of improving sustainable forest management practices for conservation of locally and globally valued ecosystems, to support forest-dependent livelihoods for their economic and cultural value, and to promote political development through participation in democratic processes (Menzies, 2007). Although organizational structure varies between different CBFM projects, the approach implies the decentralization or delegation of forest management responsibilities and, to some extent, governance (Berkes, 2007). It is implied, and sometimes a formalized legal requirement, that with increased rights over local resources, communities will exclude illegitimate forest users and prevent illicit extraction of forest resources thereby assuming the responsibility of law enforcement at the local level. In this framework, CBFM participants are not passive recipient beneficiaries of integrated development and conservation projects but are tutored to become “responsibilized” (Cruikshank, 1999; Li, 2007a, p. 234) members of an environmental community of care, enacting the behaviours they have been taught to value (Agrawal, 2005; Sundar, 2001). However, the practices that CBFM aims to reduce (deforestation and illicit extraction of forest resources) are frequently associated with physical violence and threats thereof as illicit forest users attempt to take forest resources by force and scare environmental defenders into silence (Forst, 2016; Franck & Hansen, 2014; Global Witness, 2014, 2015a, 2016a, 2017; Gonclaves, Panjer, Greenberg & Magrath, 2012). Thus, in seeking to reduce these practices, some participants in community-based forest   2 management experience physical violence, threats, and intimidation during conflicts over forest use (hereafter, ‘forest-related violence’).3   The global extent of forest-related violence, exactly how it manifests, and in what circumstances are not well known; since forests are often far from major urban centers and the local communities who want to protect them are also geographically isolated and socially, politically, and economically marginalized, the forms and processes of forest-related violence are usually hidden from public view. Furthermore, like many other forms of physical violence, it is likely under-reported and complaints poorly recorded (Amnesty International, 2017a; Global Witness, 2015a). However, the numbers of reported murders of environmental activists suggest that forest-related violence is prevalent in many countries where CBFM is implemented. Global Witness (2017, p. 6) estimates that 2016 was the “deadliest year on record” for “environmental defenders,” defined as “people who take peaceful action, either voluntarily or professionally, to protect the environment or land rights. They are often ordinary people who may well not define themselves as ‘defenders’” (The Guardian & Global Witness, 2017a, para. 2).4 During 2016, at least 200 environmental defenders were murdered around the world,5 of which approximately 12% of cases were directly related to the logging sector (Global Witness, 2017). This figure likely obscures the true scale of forest-related violence as some of those killed in relation to their protests against the mining and oil sectors, agribusiness, poaching, and dams appear to have been trying to protect forested land (Global Witness, 2017, see also Global Witness, 2015a, 2016).                                                  3 For a discussion of the concept of violence, see section 1.2 below. 4 Thus, while not all environmental defenders are CBFM participants, all CBFM participants can be considered environmental defenders. 5 Not only are the absolute numbers of murders increasing but they are also occurring in more countries – twenty-four countries in 2016 compared to sixteen in 2015 (Global Witness, 2017).   3  Non-lethal forest-related violence is also thought to be prevalent although it is less easily documented or quantified than murders, which are usually reported to authorities (The Guardian & Global Witness, 2017a). Families of assassinated environmental defenders regularly report a history of threats and non-lethal violence preceding assassinations (Gibler, 2017; Lakhani, 2016; Tremblay, 2017) and advocacy organizations and news outlets sporadically report forest-related violence manifesting as physical attacks, intimidation, and verbal aggression against those who try to prevent forest exploitation (Hall, Hirsch & Li, 2012; Yasmi, Kelley, Murdiyarso & Patel, 2012). Perpetrators are known to include illegal loggers, government-backed concessionaires, and, significantly, representatives of the state, such as military or police (Global Witness, 2014; Peluso & Lund, 2011; Wolford, Borras, Hall, Scoones & White, 2013).   Although violence specifically against CBFM participants has not been extensively explored, broader discussions of CBFM projects reference violence against participants (Honey-Rosés, 2009; Goncalves et al., 2012; Klooster, 2000a) and broader discussions of environmental and natural-resource related violence reference CBFM projects (Forst, 2016). This indicates that violence is often related to the activities CBFM participants choose to oppose or conservation organizations ask them to oppose in order to maintain forest ecosystems. Thus, forest-related violence not only has the potential to undermine the effective implementation of forest conservation activities but, more importantly, presents a threat to the lives and human rights of CBFM participants. Therefore, the study of such violence, as opposed to more generalized ‘conflict,’ addresses theoretical issues of the biopolitics of CBFM projects – the politics of the   4 intrinsic value of life and power over life (Foucault, 1978; see also Arendt, 1970).6 In exploring who is involved in forest-related violence, in what capacity, and in what contexts, this dissertation explores how life is divided into that which is or is not included in and protected by national and international political institutions and development organizations, which Fraser (2007, p. 23) identifies as “the political dimension of justice.” Given that a government has the duty to uphold human rights and protect the lives of its citizens (Nickel, 1993) and other actors, such as development organizations, assume this responsibility voluntarily, violence can provide insight on the willingness of a state and development organizations to “make live or let die” (Li, 2010, p. 66) in pursuit of their goals, such as conservation or exploitation of forests. This theoretical inquiry has practical application as it brings into question how and to what extent CBFM projects and the development organizations and government partners that implement them are able to uphold the principle of ‘do no harm,’ widely held as a central ethical tenet of development interventions (Organization for Economic Co-operation and Development [OECD], 2010).   This is an important issue upon which to reflect given that the international development community continues to promote CBFM for its potential environmental, economic, and political benefits and encourage its expansion around the world (Almeida, 2017; Oxfam, International Land Coalition, & Rights and Resources Initiative, 2016; Shackleton, Campbell, Wollenberg &                                                 6 Definitions of forest-related conflict are usually vague (Hellström, 2001; see for example, Dhiaulaq, de Bruyn & Gritten, 2015; Eckerberg & Sandström, 2013; Kane et al., 2017; Koning et al., 2008; Mola-Yudego & Gritten, 2010). One of the more concrete definitions is “disagreements and disputes over access to, and control and use of, natural resources” (United Nations Food and Agriculture Organization [FAO], 2000, p. 1). Relatedly, studies of forest-related conflict rarely attend to issues of biopolitics, the right to freedom from violence, or the moral obligation for forest management projects to ‘do no harm.’   5 Edmonds, 2002), including in Cambodia (RECOFTC, 2015; Royal Government of Cambodia [RGC], 2010). Thus, this dissertation argues, it is important to understand not only how and why conflicts occur over forests but also how and why forest-related violence occurs, with what consequences for CBFM participants both in terms of their project implementation and their lives beyond the forest, and how CBFM participants and supporting development organizations are responding to violence. This research project addresses this topic in the context of Cambodia from the perspective of CBFM participants and employees of supporting NGOs. Findings contribute to the fields of political ecology and community-based forest management studies and specifically respond to Agrawal’s (2007, p. 128) call for increased theorization of how “macro-level processes structure what happens at the local level” of CBFM projects.   This introductory chapter provides the theoretical basis of this study. I first review the concept of violence to establish the focus of this research. I explore the definitions of three widely recognized forms of violence – direct, structural, and symbolic – and explore how these are linked in theory and practice. I then explore how violence can be both a structured, observable event and a structuring, invisible process with wide-reaching and long-lasting impacts. This establishes the basis for my approach to the study of forest-related violence and sets the scope for the literature review of environmental and natural resource related violence in which I explore the factors that contribute to its occurrence and the consequences of such violence. The literature review finds that existing research is weighted towards understanding the factors that structure environmental violence rather than how it can also be a structuring mechanism with the potential to transform people’s lives and engagement in natural resource management. These discussions   6 inform the conceptual framework presented in section 1.4 and the research objectives presented in section 1.5.   1.2 A theory of violence Violence is a slippery concept in that what counts as violence and the meanings of its manifestations can vary considerably within a given society, let alone between cultures and times (Taussig, 1989). Springer and Le Billon (2016, p. 1) concede that violence is “extremely difficult to make sense of” and “one of the most complex concepts” in the social sciences. Contemporary theoretical analyses of violence typically distinguish between direct, structural, and symbolic (also referred to as ‘cultural’) violence while also emphasizing their interconnectedness (Bourdieu, 1977; Galtung, 1969, 1990).7 These concepts have been particularly useful to social scientists seeking to move away from evolutionary theories explaining why individuals are violent and moving towards an understanding that violence occurs in all societies and everyday lives in various forms, each of which reveals much about social norms and values and their historical construction and political contestation (Accomazzo, 2012; Springer & Le Billon, 2016). These interconnected categories have also encouraged a broader definition of violence as not merely a physical act but any action that affects the “conditions of another, thereby reducing one’s potentiality” (Tyner & Inwood, 2014, p. 9).   Examining these categories individually, direct violence is understood as “practices (brutal acts) that cause direct harm to humans” (Peluso & Watts, 2001, p. 26). Those harms may be physical,                                                 7 See Accomazzo (2012) for a review of historical and current theories of violence proposed by biological and socio-cultural anthropologists.   7 economic, cultural, or emotional. There is an identifiable person “who acts” against another and can be observed doing so (Galtung, 1969, p. 170). Structural violence is that which is “built into the [social] structure and shows up as unequal power and consequently unequal life chances” (Galtung, 1969, p. 171). Farmer, Nizeye, Stulac, and Keshavjee (2006, p. 1686) propose that these processes result from social, political, economic, and institutional arrangements of our world that “put individuals and populations in harm’s way,” for example, through poverty, racism, sexism, colonialism, and governmental policies. Although its effects are apparent, no immediate or individual perpetrator can be identified. Nevertheless, some individuals may purposefully work to keep these structures in place to perpetuate inequalities so structural violence should not be assumed to be a ‘natural’ phenomenon. Structural violence is exerted systemically and indirectly, and, Galtung (1990) theorizes, can cause and perpetuate direct violence.   In turn, direct and structural violence are often legitimized and obscured by symbolic violence (Bourdieu, 1977, 1998), in which dominant ideologies create insidious inequalities, hierarchies, and social norms that are reproduced throughout society, even by those who are harmed by them (Accomazzo, 2012). By manipulating ‘knowledge’ and ‘truth,’ powerful actors make “invisible” and craft the “misrecognition” of violent acts, processes, and systems (Burawoy, 2008, p. 20; Scheper-Hughes & Bourgois, 2004, p. 21).8 Galtung (1990) proposes that symbolic violence (or, as he refers to it, ‘cultural violence’) is “an invariant, a ‘permanence’” (p. 294) that “makes                                                 8 Although this concept has similarities to Gramsci’s concept of cultural hegemony, Burawoy (2008) argues that while hegemonic ideas are recognized and consensually reproduced by the dominated classes, symbolic violence is so “deep and unconscious” (p. 20) that “submission is not a matter of consciousness but of habitus” (p. 21, emphasis in original).   8 direct and structural violence look, even feel, right – or at least not wrong” (p. 291). It becomes “part of the expectation of the living” (Lawrence & Karim, 2007, p. 12). Bourdieu goes further to argue that symbolic violence creates an “extraordinary inertia” that prevents people from naming or opposing it even if they do recognize its harmful effects (Bourdieu, 1997, p. 712, cited in Burawoy, 2008, p. 1). In doing so, symbolic violence facilitates the use of structural and direct violence by dominant actors. This indicates that in addition to considering the physicality of direct violence, analyses should also consider how such harms are (mis)recognized and (mis)represented by observers as well as those immediately involved.  Recognizing the significance of each of these three forms of violence – direct, structural, and symbolic – and their interrelatedness, Galtung (1990, p. 294) envisions them as three corners of a “violence triangle” (Figure 1.1). This conceptual framework suggests that wherever one form of violence is apparent, it will be supported and enabled by others. I draw on this conceptual framework throughout this dissertation.    Figure 1.1 Galtung’s (1990) ‘violence triangle’  Studying these heuristic categories of direct, structural, and symbolic violence, and the relationships between them have broadened understandings of how violence manifests and   9 causes suffering in various ways, often unequally distributed across society or unseen or misrecognized completely (Bourdieu, 1999; Nixon, 2011; Schlosberg, 2007). Importantly, it has shown that although direct violence may manifest as a site-specific phenomenon affecting distinct groups of people, it is rooted in networked social relations, local and global histories, and power dynamics (Peluso & Watts, 2001; Springer, 2011a). Accordingly, analyses of what appear to be ‘events’ of direct violence can provide insights into related processes of structural and symbolic violence (Galtung, 1990; Tyner & Inwood, 2014). Similarly, Springer (2009, p. 368) contends that violent ‘events’ are better thought of as “moments,” or analytical windows, in on-going processes of violence involving a coalescence of people, institutions, cultures, environments and “things” (Harvey, 1996, p. 50) that span time and space. These processes and ‘things’ are not static but are “relatively enduring crystallization[s] of internal relations into an analytically isolable thing or system” (Demeritt, 1998, p. 267). This approach supersedes Galtung’s (1990, p. 294) assertion that direct violence is an “event” while structural violence is a “process” to build a conceptual framework that encourages the exploration of all forms of violence as being moments that link people, places, and institutions across space and time.  1.2.1 Violence as structured and structuring  Conceptualizing direct violence as a ‘symptom’ of structural and symbolic violence or a ‘moment’ in a dynamic socio-political assemblage is useful in terms of directing attention to the processes that generate violence and asking why violence occurred and how it is sustained. It can be used as a “theoretical vantage point” for a more comprehensive and sustained analysis of social and spatial relations (Tyner & Inwood, 2014, p. 6). This is central to the recognition of violence, an important step in preventing its automatic acceptance (i.e. symbolic violence), and   10 thereby disrupting its cycle of reproduction (Bourdieu, 1999; Schlosberg, 2007). However, to focus only on identifying violence and the factors that structure it is to risk conceptualizing violence as a ‘product’ or the terminus of power – sporadic singular or serial episodes resulting from other processes, or assuming analysis is complete once a cyclical and self-perpetuating process of direct, structural, and symbolic violence has been recognized. This, feminist geographers argue, can deny the agency of those who experience violence, portraying them only as ‘victims.’ It can obscure the ways in which violence is embodied as a deeply personal experience with the potential to transform the way we conceptualise ourselves, relations to other people and institutions, and the places we create (Fluri, 2009, 2012; Hyndman, 2004; Koopman, 2011). Robben and Nordstrom (1995, p. 5), for example, argue that the social relations of violence cast “ripples that reconfigure lives in the most dramatic of ways, affecting identity in the present, the hopes and potentialities of the future, and even renditions of the past” (see also Baird & Le Billon, 2012; Das, Kleinman, Ramphele, & Reynolds, 2000). Violence is cumulative and boundless, it “spills over” into other aspects of life, structures beliefs, actions, and relationships, and “creates and recreates new norms of collective self-understanding” (Lawrence & Karim, 2007, p. 12; Scheper-Huges & Bourgois, 2004; Tyner & Inwood, 2014). Consequently, feminist scholars (Butler, 2004; Das, 2008; Fluri, 2012; Schröder and Schmidt, 2001) frequently draw on Foucault (1996) to suggest that rather than conceptualizing violence as an isolated process, violence can be understood as a social relation imbued with power.  Conceptualized in such a way, violence is seen as simultaneously both “structured and structuring” (Dumont, 1992, p. 139; Bourdieu, 1971). As a social relation, it is ‘structured’ and created by specific socio-political, economic, and historical contexts, yet it is also ‘structuring’   11 and generative of thought and action (Grenfell, 2008). This suggests that violence cannot be understood in terms of its physicality alone. Rather, it is the social and cultural dimensions of violence that give violence meaning when it is internalized (Scheper-Hughes and Bourgois, 2004; Tyner and Inwood, 2014). This suggests that violence only becomes a means of exerting power when it changes the way people think and act, for example, towards each other, themselves, or the environment. Conceptualizing violence in this way opens up the possibility for inquiry into the way it (re-)shapes the materiality and meanings of our worlds. It draws attention to the way experiences of violence – direct, structural, or symbolic, observable or transparent – structure people’s situated knowledge of themselves and others, and, correspondingly, their interactions with the world in the context of violence.   This brief discussion of theorizations of violence provides two main insights upon which I draw throughout this research. Firstly, studies of violence should consider its generative socio-political, economic, historical, and potentially ecological context and processes. Secondly, violence has structuring effects on people, systems, and processes. As will be explored in the following section, progress has been made in understanding the factors and processes that structure environmental and natural resource related violence. However, the structuring effects of this violence on societies and socio-natural landscapes, such as community-managed forests, are underexplored.  1.3 ‘Causes’ and ‘consequences’ of environmental violence By the term ‘environmental violence’ I refer to physical and brutal acts perpetrated by humans, against humans, in which the environment or natural resources are “objects” of violent struggles   12 and conflicts (Watts & Peluso, 2014 p. 184) or violence is used “for environmental motives” (Le Billon & Carter, 2012, p. 171).9 Although these violent conflicts are also at times referred to as ‘natural resource related violence,’ this privileges a capitalist orthodoxy and economic logic that implies nature is only contested due to its financial or instrumental value, obscures the ways that natural resources are constructed as such (Li, 2014), and encourages a vision of humans as separate from other parts of the natural world (Harvey, 1996). The term ‘environmental violence’ remains more neutral (although not entirely uncontested [Harvey, 1996]) and thus I use it throughout this discussion. Similarly, since the term ‘forest’ can be ambiguous in its definition (Peluso & Vandergeest, 2011), the term ‘forest-related violence’ does not imply that violence relates only to entitlements over forest resources such as timber but rather allows for the varied meanings of the term ‘forest’ that mark it as a spiritually, culturally, and politically important and constructed place.   Although scholars working in a range of paradigms have sought the ‘cause’ of environmental violence in various contexts, an interdisciplinary review of the literature suggests that there is never a single cause. Rather, various context-specific factors coalesce to create the potential for violence and then to enable or encourage conflicts to “escalate” into violence (Castro & Nielsen, 2003, p. 2). Four of the most commonly implicated factors are environmental dimensions, NRM policies and institutions, economic and socio-political systems, and issues relating to the agency                                                 9 The emphasis on the environment as a ‘thing’ that is contested differentiates this conception of environmental violence from the way they term is used by scholars seeking to draw attention to violence exerted upon the non-human environment (Salcedo, 2015; Trigo & Rodarte, 2015) or when that leads to subsequent harms against human bodies (Auyero & Switun, 2009; Nixon, 2011; Stoett, 2012). Although these forms of environmental violence can motivate violence between humans for environmental motives (Le Billon & Carter, 2012), these approaches are beyond the scope of this dissertation.   13 of those involved such as identity, subjectivity, and relationships with the environment. In this section I discuss these dominant themes and factors that contribute to the occurrence of environmental violence.   Direct environmental violence can manifest as immediately-apparent ‘spectacular’ events, or occur as serial incidents of low-level violence that have cumulative damaging effects.10 Determining how different factors structure environmental violence has been a key endeavour of political ecologists, unsurprisingly since, as Peluso and Watts (2001, p. 23) note, “understanding the processes by which some factors become causes is surely central to the social science enterprise.” The following section reviews how different factors have structured environmental violence in different contexts with an emphasis on literature discussing violence related to forests. Although one factor may initially appear most salient, it can often only structure violence because of the co-existence of other factors.11 After a discussion of these causal factors, I then explore literature discussing the ways that environmental violence has been influential in structuring socio-political and socio-natural relations and the materiality of the environment.   1.3.1 Factors influencing occurrence and forms of environmental violence  One of the most prominent although contested theories linking the environment and violence is that increased scarcity of natural resources, intensified by increased demand and economic                                                 10 I follow Nixon’s (2011, p. 2) definition of ‘spectacular’ violence as that which “erupt[s] into sensational visibility.” It can be compared to his conceptualization of ‘slow violence’ that “occurs gradually and out of sight, a violence of delayed destruction that is dispersed across time and space, an attritional violence that is not typically viewed as violence at all” (Nixon, 2011, p. 2). 11 The academic paradigm within which each researcher is working can influence which factors appear more or less important in causing violence.    14 inequality, will lead to civil violence (Homer-Dixon, 1994; Kaplan, 1994). Others (Busse & Gröning 2013; Collier & Hoeffler, 2004, 2005) have argued that it is not only resource scarcity but also resource abundance and/or dependence that drives environmental violence. Rent-related violence occurs between powerful groups and political factions that seek to control territories on which rents are generated and the shadow economies of resource exploitation, such as in the cases of illegal logging or oil exploitation (Tacconi, 2012).12 Similarly, when governing institutions are “grabber friendly… due to weak rule of law, malfunctioning bureaucracy, and corruption” (Mehlum, Moene & Torvik, 2006, p. 3), violent conflict can occur as a result of poor national economic development and increasing inequality. However, such “quasi-environmental determinism” (Le Billon, 2001, p. 564) has been criticized for assuming overly simplistic linkages between natural resources and violence (Gleditsch, 1998; Hartmann, 2001; Watts & Peluso, 2014). In response, scholars working in range of paradigms have nuanced and deepened these theories.   Those working within the field of New Institutionalism have replied that ineffective political institutions, including government ministries, laws, and NRM policies, exacerbate the potential violence associated with resource abundance by failing to manage economic ‘boom and bust’ cycles (Robinson et al., 2006) while political geographers have pointed out that corruption often works “through governance institutions – whether national or local – rather than despite or                                                 12 Illegal logging is defined as “the felling and extraction of logs from forests that is not in conformity with an approved forest management plan or an official license issued by a forest authority in accordance with operations authorized by prevailing forestry laws” (Mir & Fraser, 2003, p. 278).   15 against them” (Remy, 2017, p. 142-3, emphasis in original; Robbins, 2000).13 Although anthropologists and political ecologists have argued that the concept of ‘corruption’ is often culturally relative (Petrova, 2014; Ali & Nyborg, 2010), the role of corruption among political institutions and those enforcing NRM policies remains prominent in explanations of resource conflict and violence (Global Witness, 2017; Klooster, 2000a, 2000b). For example, Kaimowitz (2007, p. 114) attributes government officials’ use of violence against forest defenders to their need to “protect their interests or those of illegal loggers and poachers,” although he does not support this claim with any ethnographic data, which reflects the difficulty and dangers of conducting research into corruption among governmental NRM authorities (Robbins, 2000). In such cases, environmental violence cannot be addressed effectively by the criminal justice system because high-level law enforcers are often complicit in grand corruption or state capture while local officials are often involved in petty corruption or collusion with offenders (Dauvergne, 1997). This indicates that although the environmental properties of a natural resource can contribute to violence, this is exacerbated by local economic and socio-political factors and the (selective) application of NRM policies.  Environmental violence can also be structured by the properties of contested resources including their materiality and geographic distribution. These affect their strategic or economic value and subsequently the “micropolitics” of their production, protection, contestation, and associated violence (Peluso & Watts, 2001, p.34; Le Billon, 2012). These intersect with NRM institutions as well as place-specific socio-political and economic factors. For example, Neumann (2001, p.                                                 13 This is facilitated by institutional arrangements including the confidentiality of most NRM contracts and the discretionary power of government officials exercised over a country’s resources (Williams & Le Billon, 2017).   16 307) argues that across Africa, violence in national parks and wildlife reserves is “made ‘invisible,’” and therefore more possible, by virtue of geographic isolation. The wildlife most prized by poachers are in rural areas where there is sparse population, few public services, little infrastructure, high rates of poverty, and thus where people are “most tempted to challenge the enforcement of wildlife conservation laws” (Neumann, 2001, p. 308). Furthermore, conservation laws reflect colonial ideologies and do not match local perceptions of justice or economic needs.  There are two major implications from accepting that the materiality of natural resources “play a role in the dynamics of violence and struggle” (Peluso & Watts, 2001, p. 26). First, in attributing agency to nature, ‘the environment’ becomes more than a ‘thing’ that is shaped by human actions but also an actant that can structure and transform human agency and actions. Second, the materiality of different types of resources engenders the formation of different social relations around them, including different forms of violence. In questioning how context-specific socio-historical conditions shape the way natural resources are valued, extracted, and come to be contested, this approach exposes weaknesses in Homer-Dixon’s (1994) universalizing model of environmental scarcity-induced violence in which social and historical factors are external explanatory factors introduced only when observed outcomes do not match the model’s predictions of conflict.  In other contexts, NRM policies and supporting political institutions appear to be the most salient factors that structure environmental violence, or at least put people in harm’s way, although closer inspection reveals that they are usually only one of an assemblage of factors contributing to environmental violence. For example, the militarization of National Park Rangers and   17 community-based wildlife defenders in Sub-Saharan Africa’s National Parks has become part of an arms race with poachers and contributes to violence against a backdrop of economic, socio-political, and historically-produced tensions that encourage poaching and frame its criminalization (Büscher & Ramutsindela, 2016; Duffy, 2014, 2016; Lunstrum, 2014; Shaw & Raemayer, 2016). A lack of understanding of legal issues and rights granted by a NRM policy can also create the potential for conflict and violence to occur (Kane et al., 2017). NRM policies also intersect with economic and political ideologies to facilitate violent ‘land grabs’ (Schoenberger, Hall & Vandergeest, 2017). The ‘global land grab’ has been interpreted as a contemporary form of primitive accumulation, or “accumulation by dispossession” (Harvey, 2003, p. 144), in which violence is assumed to be an inherent trait, manifesting as direct violence (such as violent evictions), structural violence (such as the privileging of purchased ‘property’ over customary possession and land rights), and ‘slow’ violence (such as the long-term violence of exclusion of local communities from natural resources central to their livelihoods and cultures) (Hall, 2013; Li, 2014; Peters, 2013; Sargeson, 2013). Furthermore, the normalization of these forms of violence as a ‘necessary’ price of development has been conceptualized as a form of symbolic violence (Springer, 2011b). A range of actors are implicated in such violence: national governments construct the necessary legal frameworks, military and police are often enrolled in conducting evictions, and international investors provide the capital and perpetuate discourses that attempt to justify such violence in the ‘name of development’ (Bulkan, 2014; Daniel & Mittal, 2010; Wolford et al., 2013).   The factors discussed thus far have been external to the (human) actors involved in environmental violence – they can be considered part of the ‘social structure’ – but ethnographic   18 work has challenged deterministic explanations for environmental violence by exploring people’s agency. Scholars have shown that individual and collective identities and situated subjectivities (sense of self in relation to others) influence people’s engagement in environmental violence (Peluso & Watts, 2001). Accordingly, environmental violence may be an attempt to reclaim or forge new identities in relation to the environment itself (Peluso & Harwell, 2001; Spronk & Webber, 2007), other resource users (Laungaramsri, 2000), or the state as a manager of natural resources (Le Billon & Carter, 2012; Valdivia, 2008).14   Drawing on Bourdieu’s (1977) conceptualization of agency, Peluso and Watts (2001, p. 6) suggest that violence can become “a sort of habitus,” that is, an embodied and deeply engrained way of thinking and acting in the world that is “structured” by personal histories, local culture, and socio-political structures. Thus, when environmental conflicts arise, a habitus of violence contributes to the escalation of conflicts. This reflects not only the current conflict but also memories of previous violence, encounters, and frustrations. For example, Peluso and Harwell (2001) propose that the violent reclamation of Dayaks’ historically occupied spaces and resources in West Kalimantan was both a defense of natural resource entitlements in the face of territorializing government policies and a reiteration of environmentally-based ethnic identities. The specific forms of violence used against Madurese people were fundamentally shaped by memories of historical ethnic conflicts performed in earlier periods under quite different conditions. Similarly, Bobrow-Strain (2001, p. 33) explores how past violence is “repaid” by violence in the present, thus implicating natural resources as “part and parcel of social relations.”                                                 14 Violence against state representatives or corporations is often labeled ‘eco-terrorism,’ illustrating how some forms of violence are considered less legitimate and more deviant than others depending on the identity of perpetrator and ‘victim’ (Liddick, 2006; Vanderheiden, 2005).   19 This reveals how environmental violence may be embedded in and reflective of other grievances and is not necessarily only about controlling resource entitlements.  In summary, environmental violence reflects not only the immediate desires and conflicting perspectives of those with an interest in a specific environment or natural resource but also bring into focus a range of historically-constructed NRM policies and institutions, socio-political and economic factors, and identities and subjectivities that intersect to contribute to environmental violence.  1.3.2 The structuring effects of environmental violence  The literature reviewed thus far seeks to understand how environmental violence is ‘structured’ by social, political, institutional, and ecological conditions. Yet, as discussed in section 1.2.1, violence can be understood as a complex social phenomenon that is both “structured and structuring” and can reshape the social relations from which it arose (Dumont, 1992, p. 139; Feldman, 1991; Peluso & Watts, 2001). This section reviews how environmental violence has been found to shape the behaviour of natural resource managers, the way they interact with other resource users and institutions, and the materialities of contested environments. Since this structuring potential of environmental violence has been much less studied than the factors that structure it, I also draw on studies that show how political violence can similarly shape NRM practices.15                                                 15 The number of studies investigating specifically how the social relation of violence shapes NRM is less than might be expected as many studies address the parallel question of how the loss of natural resources (appropriated through violent means) shapes communities. For example, Dhiaulhaq et al. (2014a) ask how violently established economic land concessions impacts CBFM projects in Cambodia. However, their focus is not on the social relation   20  Environmental violence can directly shape NRM systems within communities by limiting access to resources. For example, Bohle and Fünfgeld (2007, p. 665) find that violent struggles over environmental entitlements in Sri Lanka created alternative systems of control over natural resources at the community level. Furthermore, by excluding certain groups, these initiated “new processes of disentitlement and social vulnerability” indicating that a cascade of social and environmental changes can result from environmental violence (see also Sundar, 2001).   In other cases, natural resource claimants’ use of violence against the state in defense of their resource claims has prompted governments to impose new NRM policies, which may either restrict or improve access to natural resources (McCarthy, 2001; Valdivia, 2008). For example, Watts (2001) has argued that the Ogoni’s use of violence against the Nigerian state in conflicts over oil production resulted in the claimants themselves being discursively erased from the landscape and in NRM planning documents. This indicates that direct violence can reshape the socio-political position and resource rights of both those who use violence and those who are targeted by it and supports the theory that violence is never a discrete event.   Rather than operating directly on NRM systems, Peluso and Watts (2001, p. 33) suggest that “community” can be “created or reinforced” by environmental violence. This has implications for natural resource management since the motivation to participate in CBFM projects can be                                                                                                                                                        of violence per se but rather on the loss of agricultural, forest, and residential land. Thus, their conclusions are somewhat unsurprising – livelihoods and wellbeing are less secure because they do not have access to the resources on which they previously depended. The more nuanced question of how surrounding communities continue to manage forests in the context of conflict does not appear to have been systematically investigated.    21 influenced by the feeling of belonging to an ‘imagined community,’ which is discursively constructed and maintained to bound who is included and excluded (Agrawal, 2005; Anderson, 2006 [1983]; Kubo, 2009). Indeed, Sundar (2001, p. 333) argues that community-based natural resource management inherently requires participants to “buy into” a constructed identity and sense of place in order for them to enact the environmental behaviours they have been taught to value. Environmental violence can also have political outcomes at a wider scale by enabling some groups to appropriate valuable natural resources and prolong political conflict. For example, timber has been used to fund national and regional conflict and terrorist organisations in countries such as Liberia, the Democratic Republic of the Congo, and Cambodia, among others (Le Billon, 2012).  Violence that is not motivated by environmental conflict can also shape natural resource use and management. At the community level, Nightingale (2014) demonstrates that political violence in Nepal directly impaired community forestry practices as Maoist insurgents limited the mobility of rural people and made them scared to enter the forest. At the national level, Richards (2001) shows how political violence contributed to the formation of alliances between resource claimants and anti-government groups in Sierra Leone both changing the way that diamond resources were contested and managed, and prolonging existing conflicts. Although these forms of violence were not originally political in nature, it emphasizes the way that community-based natural resource management practices and institutions are shaped by violence occurring throughout the society.    22 Although the literature that deals directly with the structuring effect of environmental violence is limited, that reviewed above shows that it may initiate a cascade of effects throughout society. This includes disciplinary or exclusionary effects that prevent people from physically accessing resources or political effects that then limit or increase resource rights. Thus, environmental violence can contribute to state-making through territorialization and signalling who is included in or excluded from political life (Peluso & Vandergeest, 2011; Stonich & Vandergeest, 2001). Furthermore, environmental violence can be “transformative” of identities and subjectivities (Peluso & Watts, 2001, p. 27) and shape the way places and environments are perceived and ‘made’ (Massey, 1994, 2005). Therefore, to understand the effects of environmental violence, scholars must look beyond the disciplinary potential of its physicality and consider the social and cultural dimensions of violence that give it meaning (Sheper-Hughes & Bourgois, 2004; Tyner & Inwood, 2014).  1.4 Conceptual framework Theoretical discussions of violence show that it can usefully be thought of as a social relation that is structured by an assemblage of context-specific factors. In the case of environmental violence, this includes the materiality of the contested environment, economic and socio-political factors, institutions and NRM policies, and the situated subjectivities and identities of those involved as victims, perpetrators, and facilitators of violence. These factors create arrangements of power that result in structural violence, spread symbolic violence, and facilitate direct violence. This concept forms the first half of the theoretical framework for this dissertation and structured research into what forms of forest-related violence occur, in what contexts, and why.     23 Theories of violence also suggest that it is a structuring mechanism that has the potential to shape the way that people perceive, interpret, and interact with the world through, for example, expressions of subjectivity, the socio-political systems they create, institutions they build, maintain, or operate within, and the way they use and manage natural resources. In short, violence has the potential to shape CBFM practices. Thus, the second half of my theoretical framework focuses on understanding the structuring effects of violence. It avoids presenting violence as a static ‘product’ of an assemblage of people, processes, and ‘things’ but rather conceptualizes it as part of an ever-shifting ‘Nature-Society dialectic’ in which both are continually constructed, cited, challenged, and re-made (Harvey, 1996). Since the literature review showed that the agency of natural resource users has often been obscured or ignored in studies of environmental violence, this aspect was prioritized in this research project with the intent to understand how violence shapes people’s participation in and the implementation of CBFM activities. In doing so, it responds to the call of feminist social scientists for a more nuanced analysis of embodied experiences of violence and its after-effects (de Leeuw, 2016; Jewitt, 2008; Rocheleau, Thomas-Slater & Wangari, 2013). This conceptual framework also offers a way to use direct violence as a “vantage point” (Tyner & Inwood, 2014, p. 6) from which to analyze broader systems of natural resource governance and social relations, such as those among competing resource users, and with state authorities, as well as how different the rights and lives of resource users are protected or made vulnerable. Figure 1.2 represents this conceptual framework below.     24  Figure 1.2 Conceptual framework  1.5 Discussion Studies of environmental conflicts and violence have contributed significantly to the broader fields of political ecology, natural resource management, and rights-based development studies (Le Billon, 2015; Zaitch, van Solinge & Müller, 2014). However, a number of knowledge gaps   25 remain. First, there is no clear consensus on the extent of environmental violence within any one country, let alone on a regional or global scale. This is likely due to the difficulties measuring and recording its incidence. Using a quantifiable measure, such as number of deaths (e.g. Global Witness, 2015a, 2016a, 2017; The Guardian and Global Witness, 2017b), provides some insights although it is highly likely these figures are underreported (Global Witness, 2017). For example, when deaths are not accurately attributed to environmental conflict (potentially due to the authorities’ involvement in the conflict), when the victim is not a high-profile public figure, and/or assassinations occur in geographically, socially, politically, and economically marginalized communities (also where many contested natural resources are located), assassinations may not be documented.   Similarly, non-lethal, ‘everyday’ violence is not consistently documented and is likely also vastly underreported (see Chapter 4 of this dissertation). This embeds the perception of violence as ‘events’ that punctuate environmental struggles and are picked up by media outlets rather than promoting an understanding of violence as a process that permeates all aspects of people’s lives and shapes the way they move through the world (c.f. Robben & Nordstrom, 1995). More importantly, a lack of attention to everyday environmental violence prevents identification of patterns over time and space, which is not only of academic interest but could provide useful information as to whether and at what point interventions could be or should have been made to prevent further ‘spectacular’ violence occurring.   Case-based, qualitative studies are better able to capture and represent the different forms of violence occurring. However, focusing on violent relations around a specific resource in specific   26 context encourages the tendency to discuss each in isolation as a separate problem; for example, evictions by land concession companies are often treated as unrelated to violence by illegal loggers despite the fact that these environmental conflicts may occur in the same areas at the same time (Eckerberg & Sandström, 2013). While this conceptualization and fragmentation of space (Lefebvre, 1991) is a useful simplification in terms of parsing out contesting claims to resources and prescribing solutions to a specific problem, it risks obscuring the connections between different resource conflicts, underlying structural and symbolic violence, and the assemblage of discourses and technologies through which they are perpetuated. It also fails to reflect the complexities of natural resource management in the context of overlapping forms of violence.  Further specificity is also required in some reports to avoid homogenizing ‘victims’ of violence under the all-inclusive term “communities” and ‘perpetrators’ of violence as “outsiders” (The Oakland Institute, 2011, p. 33; Yasmi, Kelley, Murdiyarso, & Patel, 2012, p. 103). Such generic terms may be necessary to respect research participants’ anonymity and are useful heuristics for identifying general groups but can obscure important relationships and tensions between members of communities that may contribute to or ameliorate violence. This risks romanticizing community members as ‘inherently benign’ despite their potential implication in perpetrating violence (Brosius, Tsing & Zerner, 1998; Tsing, 2005) and dehumanizes those who use violence, obscuring the reasons for which they may be compelled or motivated to do so. Similarly, the homogenization of ‘victims’ obscures differentiated vulnerability to violence that may result from their ethnicity, gender, socio-economic status, or different reliance on natural resources (c.f. Agarwal, 2013).    27  A further weakness in the literature is that there has been much greater attention on the causes rather than the consequences of environmental violence. Studies that address the consequences of violence tend to address more easily quantifiable outcomes such as loss of natural resources (Dhiaulhaq et al., 2014a, 2014b; Kaimowitz, 2007; Reboredo, 2013). Again, this trend reflects the methodological challenges of investigating such a sensitive subject. Analysis of structural factors that contribute to violence (political systems, policies and institutions, resource properties) can be done from a distance but understanding how this is internalized, interpreted, and responded to by those experiencing violence requires qualitative research in the communities involved, methodological issues that are discussed further in Chapter 2. Secondly, the bias towards studying causes rather than consequences may also reflect the assumption that violence only operates as a repressive force with straightforward, repressive consequences – suppression of protesters, loss of resources, dominance of traditionally ‘powerful’ actors over the ‘weak’ – reflecting a failure to consistently theorize violence within political ecology and related disciplines (c.f. Springer & Le Billon, 2016; Tyner & Inwood, 2014).16   This lack of attention to the consequences of environmental violence has two important implications: first, it risks disembodying violence from those affected, obscuring its effects on people’s emotional wellbeing, sense of self, and relations with those implicated in violence; second, on a more practical level, it fails to illuminate how people continue to use and manage natural resources when living with the threat of violence. When insufficient attention is paid to                                                 16 Springer and Le Billon (2016, p. 1) review the “emerging” conceptual field of “geographies of violence” and argue that it “deserves greater attention an a more rigorous examination.”   28 these consequences of violence, studies tend to become exercises in identifying the underlying, structural, ‘meta-problems’ and recommending grandiose responses intended to achieve utopian ideals of justice (see Chapter 5). Instead, exploring the consequences of violence, whilst remaining attentive to underlying structural causes, has the potential to identify how harms can be reduced through more pragmatic responses by or with the communities involved even if violence cannot be avoided completely.   Turning specifically to forest-related violence, the lack of attention to forest-related violence as a contextual variable in the operation of CBFM is a significant lacuna since CBFM projects specifically aim to reduce forest degradation, including illegal logging, deforestation, and that caused by land acquisitions, processes which are often embroiled in forest-related violence (Franck & Hansen, 2014; Tacconi et al., 2016). Furthermore, CBFM is considered a key approach of conservation organizations and is promoted around the world and seen as a path forward to ensure conservation and development go hand in hand (Almeida, 2017; Oxfam, International Land Coalition, & Rights and Resources Initiative, 2016; Schleicher et al., 2017). Although CBFM literature claims to value participants’ personal interpretations of how and why projects succeed or fail and local responses to specific interventions (Agarwal, 2001; Shackleton et al., 2002), this has not been widely explored in relation to forest-related violence, nor are the effects on institutions of CBFM well understood. Although some reports note that those who experience forest-related violence are CBFM participants (Honey-Rosés, 2009; Goncalves et al., 2012; Klooster, 2000a), they are often only conceptualized as local forest users rather than being seen as implementers of a formal conservation strategy that is intended to provide global benefits such as avoiding carbon emissions from deforestation and protecting habitats for endangered   29 species that are thought to have intrinsic value at the global level. This distinction is important because CBFM not only relies on communities’ ‘inherent’ desire to protect the forest but also purposefully cultivates a sense of duty to act as environmental defenders (Agrawal, 2005; see also Chapter 6). Thus, to portray CBFM participants as self-motivated environmental activists, accidently caught up in violence, is to obscure the role that development organizations may have, for better or worse, in shaping the way that forest-related violence unfolds.   1.5.1 Research objectives and introduction to the case study of Cambodia In order to address some of the gaps and weaknesses in the literature identified above the aim of this research was to understand how and to what extent forest-related violence shapes CBFM practices. To achieve this aim, the project focused on a single-country case study of Cambodia with embedded case studies of four provinces. The rationale for a using a single-country case study was that this would reduce the variability within some factors that could contribute to violence. Specifically, the NRM policy context and national political context theoretically would be consistent and variability in local political, social, economic, cultural, and environmental factors would likely be less than comparing between countries. This would allow the research to focus on more subtle differences between and within communities and explore interpretations and lived experiences of violence more acutely than undertaking a large-scale, multi-country study.   The choice of Cambodia was guided by the highly contested character of its forests despite significant on-going investments by international donors and the Cambodian government in CBFM (Poffenberger, 2013; UN-REDD, 2011) and news reports of forest-related violence   30 occurring. This included the assassination of two prominent forest defenders in the two years before this research was planned and at least thirteen additional deaths from forest-related violence since 2005 (Global Witness, 2014; Mech, 2014; Ouch, 2014; Phak, 2014). Furthermore, the author had two years of prior experience working with NGOs implementing CBFM projects in Cambodia that provided knowledge of the cultural context and language, which somewhat reduced the barriers to entry in arranging and conducting fieldwork. Further background on the context of CBFM in Cambodia is provided in Chapter 2 and the researcher’s positionality and methodological implications are discussed in Chapter 3.  Building on the broader literature and the context of Cambodia as a case study country, the specific research objectives were to:   1) Outline to what extent, in what forms, and in what social, political, and ecological contexts forest-related violence occurs in Cambodia’s CBFM sites; 2) Assess how CBFM participants and development organization partners are currently responding to forest-related violence; 3) Understand how CBFM participants internalize and interpret forest-related violence and its underlying causal factors and how this shapes their participation in CBFM; 4) Understand why and how CBFM groups implement CBFM activities despite the threat of forest-related violence and what can be done to better support them in this.  Achieving these research objectives will guide the identification of ways in which communities can be supported to protect their forest resources while reducing the harms caused by forest-  31 related violence. This has practical application in Cambodia, where CBFM is heavily promoted by development organizations yet forests are highly contested, and elsewhere in the world where forest-related violence affects CBFM participants. Findings also contribute to theoretical discussions in the fields of political ecology and community-based forest management studies and, as mentioned above, specifically respond to Agrawal’s (2007, p. 128) call for increased theorization of how “macro-level processes structure what happens at the local level” of CBFM projects.   1.6 Structure of the dissertation Chapter 2 provides context on the political and ecological features of Cambodia that make it an appropriate case study country for this research. It characterizes key factors theorized to contribute to environmental violence – the properties of forest resources, the institutions used to govern them (CBFM groups, state actors, and the Forestry Law), and the socio-political system in which CBFM is implemented.  Chapter 3 presents the methodology with specific attention given to the researcher’s positionality and the theoretical, practical, and ethical considerations entailed in the study of violence and conducting research in the context of Cambodia.  Chapter 4 exposes the forms of forest-related violence occurring in CBFM sites and the actors involved. Drawing on the theoretical framework described above, it explores how socio-political, environmental, and institutional factors structure forest-related violence in CBFM sites. Data   32 from interviews with CBFM participants provide some insights to the extent to which subjectivities structure violence within CBFM participant villages.  Chapter 5 considers how CBFM participants and development organizations can and should respond to forest-related violence. Although the literature suggests that responses should take a three-pronged approach of addressing the underlying causes of violence, enforcing the rule of law, and promoting human rights and political empowerment, I find that in practice in Cambodia, responses are obstructed and neutralized by the socio-political system. I consider the effects this has on CBFM practices, thereby connecting the two halves of the theoretical framework.   Chapter 6 explores the structuring effects of violence and specifically focuses on subjectivities and, to a lesser extent, on socio-political dynamics, outcomes of forest-related violence that the literature review suggested have not been well explored in comparison to visible material outcomes such as environmental destruction or economic losses.   Chapter 7 concludes the dissertation with a reflection on its contribution to the literature and areas for further research.    33 Chapter 2: The environmental, socio-political, and institutional context of community-based forest management in Cambodia  This chapter discusses the environmental and socio-political context in which CBFM is implemented and introduces the institutional forms of CBFM in Cambodia and their purposes. This familiarizes the reader with the Cambodian context – a necessary endeavour given the theorized importance of local environmental, socio-political, and institutional context when discussing environmental violence (Peluso & Watts, 2001) and the implementation of CBFM (Agrawal, 2002, 2007; Armitage et al., 2009; Leach, Mearns & Scoones, 1999). I begin with a discussion of Cambodia’s socio-political situation with specific attention given to the system of neopatrimonialism, a concept to which I refer back throughout this dissertation as it has been central to political ecologists’ analyses of Cambodia’s forestry sector and, as I show in this chapter, influences the organization of social and political life. I then discuss the ecological context in which CBFM is implemented and how neopatrimonialism, ruling elites’ struggles for power, and international aid organizations have shaped forest governance and deforestation in Cambodia since the 1970s. This provides the context for a discussion of how forest-related violence in Cambodia has been covered in academic publications and news media. I then characterize the various forms of CBFM projects that are being implemented in this context of neopatrimonialism and violently contested forests. This discussion justifies the choice of Cambodia as a country case study of forest-related violence    34 2.1 The socio-political context  Cambodia has undergone rapid social and economic transformation since the signing of the Paris Peace Accords in 1991 that formally returned the country to peace after more than a decade of armed conflict, civil war, and foreign occupation. In the years immediately following the Accords, Cambodia received substantial international development aid, amounting to over half the national budget as recently as 2005 (Ek & Sok, 2010) – although more recently this has been replaced in part by loans from China and South Korea and in part by foreign investment in the extractive industries, agricultural land concessions (predominantly for sugar and rubber), and the garment manufacturing sector (Ear, 2013; Ek & Sok, 2010). The middle class has expanded markedly in the past decade and levels of education and literacy have increased while reliance on subsistence agriculture has declined (Hughes & Un, 2011).  Yet, despite these developments, Cambodian citizens remain some of the poorest in the Southeast Asia region with approximately 13.5% of people living in poverty and another 40% living precariously in conditions of ‘near poverty’ (World Bank, 2017).17 Freedom House (2017) classifies Cambodia as ‘not free’ with poor political rights and civil liberties. The state consistently fails to provide many core services and infrastructure to its citizens because it is “unable or unwilling to do so” (Hughes & Un, 2011, p. 10). These failings have widely been attributed to the nature of Cambodia’s political system, which is authoritarian and heavily based                                                 17 ‘Near poverty’ means that the loss of 1,200 riel a day (approximately US$0.30) would result in poverty (World Bank, 2017).   35 on neopatrimonialism (Morgenbesser, 2016; Pak et al., 2007; Un & So, 2009, 2011).18 The concept has been used to explain how Cambodian state institutions have become vehicles for plunder, predation, and kleptocracy to enable senior government officials and their cronies to maintain and accumulate personal power and wealth (Acemoglu, Robinson & Vertier, 2003; Pak et al., 2007; Springer, 2011).  Neopatrimonialism is a heuristic concept, derived from the works of Weber (1978 [1922]), which draws attention to the conjunction of two forms of political domination: patrimonial and legal-rational domination (Bratton & Van de Walle, 1997). Under neopatrimonialism, public norms appear formal and rational, often being codified in law, and ostensibly there is a distinction between private and the public powers. Yet, a hierarchical network of patrimonial relations extends through all levels of society from political and economic elites to the judiciary, civil servants, village-level authorities, and peasants thereby linking political elites to peripheral spaces and marginalized populations. These patrimonial relations mean that the social practice of legal-rational systems, such as laws, roles of public office, or institutions of land titling, is often informal and personal and secured by obligations to repay favours in kind or through the exchange of material goods or money (Cammack, 2007). It is not that two systems coexist but rather that patrimonialism penetrates the legal-rational system and “twists its logic, functions, and output, but does not take exclusive control over the legal-rational logic” thus giving neopatrimonialism a “hybrid” character (Erdmann & Engel, 2007, p. 105). Consequently,                                                 18 Even scholars who analyze Cambodia’s political economy and society using alternative conceptual frameworks, such as neoliberalism, kleptocracy, and illiberal democracy (Cock, 2010; McCarthy & Un, 2017; Springer, 2017), uniformly begin with the recognition that its legal-rational systems are “embedded in the patronage based political order” (McCarthy & Un, 2017, p. 100).   36 neopatrimonialism is not the result of individuals’ deviant behaviour or moral failings but a coherent, purposive, and coercive system of domination that pervades all aspects of society and relies on hierarchical and interpersonal intermediation between subject and the state.  Under neopatrimonialism, the legal-rational system is not applied equally to all people, at all times, or in all places. This can provide a certain degree of choice as to whether an actor uses legal-rational mechanisms or patrimonial relations to achieve goals and better one’s interests. Thus, Erdmann and Engel (2007, p. 105) argue that neopatrimonialism is “characterized by insecurity” about the behaviour and role of state institutions and agents and the degree to which they will support citizens to claim and enact their rights as promised by legal-rational systems. Insecurity is reproduced in all parts of the socio-political system as actors continually strive to connect themselves to both the patrimonial network and legal-rational system (either directly or through brokers) in the hope that one will provide protection if the other does not (Erdmann and Engel, 2007).   Patronage and hierarchy pervade all socio-political relations in Cambodia with patrimonial networks reaching from the Prime Minster “down through every level of government to the villages” (Springer, 2017, p. 237). Due to the prominence of contemporary politicians in orchestrating neopatrimonial arrangements, neopatrimonialism in Cambodia is commonly traced back to 1979 when the Vietnamese removed the Khmer Rouge from power and enabled two former mid-level Khmer Rouge officials, Hun Sen and Chea Sim, to forcibly install themselves as leaders (Heder & Ledgerwood, 1995). As such, they directed the establishment of post-conflict state apparatus so as to secure their positions as heads of competing political networks,   37 neither of which were intended to promote the best interests of Cambodian citizens but rather to secure their own political and economic power (Gottesman, 2003; St John, 1995). Tracing the trajectory of the current political elite’s emergence provides important insights into the nodes of power and personal networks through which domination is exerted in Cambodia today; many of those same individuals retain influential political positions and have strategically used the legal-rational system and outright corruption and kleptocracy to enrich themselves, their families, and associates (Cock, 2010; Global Witness, 2009). This includes Prime Minister Hun Sen and other leaders of the ruling Cambodian People’s Party (CPP). Yet, although it is well documented who sits atop Cambodia’s hierarchical neopatrimonial network (Global Witness 2007, 2016), the identity of those just below the “top rungs of patron power” and the precise details of legal, illegal, and extra-legal agreements between patrons and clients are unclear (Springer, 2017, p. 244). What is known is that the upper echelons of the “interlocking pyramids of patron-client networks” that were dominant in the early post-conflict years (Heder, 1995, p. 425) have expanded to incorporate the business elite as well as the military, police, judiciary and other public servants, and the political elites’ families (Global Witness, 2007; Springer, 2017; Un & So, 2011). Thus, it is fitting to talk of Cambodia’s ‘political and economic elite’ or ‘ruling class’ rather than restricting analysis to those in formal political positions (Cock, 2016). The party-driven patronage system has been formalized, legitimizing personalized power with selectively designed laws. For example, under a sub-decree promulgated in 1994, individuals who donate $US100,000 (or $500,000 as of March 2017) to the government for the purposes of building infrastructure are awarded the ‘National Construction Medal’ and the honorific title ‘Oknya,’ access to the Prime Minister as a ‘special advisor,’ and a social status to which the majority of   38 Cambodians show deference out of fear, respect, or a combination of the two (Kijewski & Touch, 2017).  Yet, these individuals have not seized and maintained power simply by forcing the public to acquiesce, nor is Cambodian neopatrimonialism simply a “quaint, exotic cultural curiosity” (Kent, 2006, p. 357). Rather, the reasons why this hybrid form of domination is so pervasive, resistant to change, and seemingly accepted, or even desired, by the majority of Cambodians (Coventry, 2016; Jacobsen & Stuart-Fox, 2013) can only be understood in the context of the historical, cultural, and political factors that nurtured its emergence including Buddhism, French colonialism, and the influence of the international aid industry and development community (Morgenbesser, 2016).   The philosophical foundations of Buddhism legitimize social hierarchies through the concept of merit. Those atop hierarchies are assumed ‘meritorious individuals’ who should not be challenged nor required to provide accountability for their actions but instead respected and revered (Hazra, 1981; Kent, 2006). Merit can be possessed at birth (most powerfully exemplified by the belief in the King’s divine qualities) or acquired by donations to the sangha (Buddhist monastic community) or animistic “magic” (Harris, 2005; Jacobsen & Stuart-Fox, 2013, p. 10). For many Cambodians, these Buddhist teachings (albeit, interwoven with animism and Brahmanism) form the canvas into which all other socio-political norms are woven and encourage acceptance, or at least tolerance, of the current neopatrimonial state and actions of those dominating the political hierarchy (Harris, 2005; Kent, 2006).    39 Cultural traditions of patrimonialism were more tightly knit with legal-rationalism during the French colonial rule of Cambodia (1863-1953). A series of ‘modernizing’ legal-rational concepts such as the civil service, political hierarchy, and private land ownership were introduced (Morgenbesser, 2016). These were intended to bring ‘order’ to an unfamiliar and ‘unruly’ countryside while also providing a paternalistic guiding hand to the peasantry who, it was assumed, lacked self-reliance and preferred to look to “a powerful saviour from above or outside rather than seeking a local solution” (Vickery 1984, p. 14).19 Following independence from France in 1953, the monarchy continued to fuse personal and political power. Prince Norodom Sihanouk abdicated from the throne in favour of his father in order to lead the People’s Socialist Party in the 1955 election, which the party won with 82.7% of the vote in an atmosphere of “open intimidation” (Morgenbesser, 2016, p. 53). In the three subsequent elections, he used the monarchy’s wealth to secure his position as patron to the poor by “swooping over Cambodian rice fields dropping gifts for the peasants out of his helicopter and personally overseeing the creation of schools and other infrastructure across the country” (Hughes, 2006, p. 473). Such gift giving even by contemporary politicians, Hughes (2006, p. 469) argues, continues to be portrayed as “traditional” yet is imbued with threats that make it difficult for the recipient to refuse and thus the ‘acceptance’ of gifts and inclusion in a patronage network is better understood as “surrender.”   International aid organizations and the United Nations Transitional Authority in Cambodia (UNTAC) further enabled neopatrimonialism to flourish in the post-conflict transition years                                                 19 Coventry (2016, p. 54) suggests that this belief persists today within development organizations but, although there may be some truth to it, warns, “care must be taken in assessing cultural generalizations.”   40 (1979-mid 1990s) (Cock, 2010). Newly implemented legal-rational mechanisms, such as elections and later the establishment of Commune Councils, were intended to root out patrimonial tendencies and hasten the growth of democracy, a much-criticized yet normative goal of development (Heder & Ledgerwood, 1995; Przeworski et al., 2000; Sørensen, 2008). However, in practice, these enabled ruling elites to extend their patrimonial power and surveillance over the countryside while appearing to appease international donors’ demands (Morgenbesser, 2017; Slocomb, 2010).   Many domestic and international development organizations remain embedded in and perpetuate neopatrimonialism (Henke, 2011). Globally, Cambodia has one of the highest numbers of NGOs per capita (Coventry, 2016), many of which are engaged in service provision (as opposed to political empowerment) (Norman, 2014). Citizens’ dependence on NGOs has enabled the state to avoid the responsibility of providing basic services such as sufficient education, healthcare, or infrastructure without significant decent or protest from the public (Ear, 2013). Simultaneously, by requiring foreign donors to partner with government departments, political elites have been able to siphon off incoming finances for their personal gain while using the partnerships to maintain an illusion of supporting community development initiatives (Ear, 2013). At the government level, this has meant that economic and political reforms, such as formalized land titling programmes, are largely performances for the donor community intended to maintain lucrative sources of finance rather than indicative of a real impetus for change and are only implemented to the extent that they do not challenge existing ruling elites’ interests (Cock, 2010; Springer, 2015). At the community level, a result has been an increased tendency to seek out   41 “protection when the state has failed” from alternative sources including from NGOs and individuals in the ruling elite (Pak et al., 2007, p. 54; see also Jacobsen & Stuart-Fox, 2013).   Consequently, the logic of neopatrimonialism does not only operate within governmental institutions but is pervasive in the organization of Cambodian social lives, economic transactions, and normative expectations of agreements between citizens and development organizations (Ear, 2013; Jacobsen & Stuart-Fox, 2013). This suggests that submission to the norms of neopatrimonialism has become a matter of ‘habitus’ (Bourdieu, 1977) for Cambodian people, that is, behaviour that is regulated, systematic, and predisposed without being the product of obedience to rules (Maton, 2008). It is in this socio-political context that Cambodian people and NGOs implement and engage in struggles over CBFM.  2.2 The ecological context: contested forests Cambodia is the second most densely forested country in Southeast Asia (Lambrick, Brown, Lawrence & Bebber, 2014) and is one of the last remaining countries in the region where endangered wildlife species such as clouded leopards (Neofelis nebulosa) and pangolins (Manis javanica) can be found (WWF, 2017). Rare and endangered tree species also grow in its forests including Dalbergia cochinchinensis (Siamese rosewood, locally ‘kronoung’), D. bariensis (Burmese rosewood, locally ‘neang noun’), Afzelia xylocarpa (locally ‘beng’), Pterocarpus pedatus (paduak, locally ‘thnoung’), and Hopea ferrea (locally ‘korki’). The Law on Forestry (2002) made it illegal to log rare species at all except for when building infrastructure (Article   42 29) or to log other species without a permit (Article 24).20 Yet, Cambodia also has one of the highest rates of deforestation in the world at 1.2% per year (2005-2010) (Lambrick et al., 2014). Forest cover declined by approximately 26.9% between 1990 and 2015 (UNDP, 2016a), in part because of the value of the timber from these rare trees21 and in part because of the value (or expected future value) of the land on which they grow. This is driven by ruling elites who use forest resources as a means to secure their personal wealth and thereby maintain their political power. However, the exploitation of Cambodia’s forests for personal enrichment and political gain is not a recent aberration but, as explored in the following section, follows decades of similar behaviour by ruling elites and political insurgents in which “public transcripts” have obscured the operations of the “shadow state” (Le Billon, 2002, p. 564).   2.2.1 The political importance of timber for Cambodia’s ruling elites, 1989-2017 During the 1960s, the civil war (1970-1975), the Khmer Rouge genocidal regime (1975-1979), and Vietnamese occupation (1979-1989), timber was instrumental for political elites and guerrilla insurgents seeking to secure their financial status and patronage networks (Cock, 2016). Notably, this includes the current Prime Minister, Hun Sen, who was a Khmer Rouge commander and installed by the Vietnamese as Prime Minister in 1985 (Human Rights Watch, 2015).                                                  20 The Law on Forestry was written to satisfy donors, specifically the International Monetary Committee and World Bank, that the Cambodian government was serious about managing its natural resources sustainably and for the economic benefit of the country, again reflecting the government’s acquiescence to external drivers of reform rather internally-motivated reform (Cock, 2016). This is discussed further below in section 2.3. 21 Thnoung timber (P. pedatus) in sheets of 50 x 10 x 120cm can be sold for $150 at the village level, $350 at the Lao or Vietnamese border, and $1,000 in China. A 70 x 10 x 400cm sheet can be sold for $10,000 at the provincial level (NGO Forum on Cambodia, 2015).   43 After the signing of a Peace Agreement in 1991, the United Nations Transitional Authority in Cambodia (UNTAC) assumed administration of the independent state and began organizing for a general election. Mounting political insecurity associated with the prospect of political change reinforced a pattern of logging and the sale of massive forest concessions inside the country. Decades of conflict meant that Cambodia’s forests were relatively more intact than those of neighbouring Thailand or Vietnam and with the growing accessibility of the border areas could be logged with “unusually low extraction cost” compared to the rest of the region (Cock, 2016, p. 83). Political elites attempted to bolster their finances and patronage networks before the 1993 elections by awarding logging concessions to various ‘strongmen’ who had seized power in the years following the civil war and international logging companies based in East and Southeast Asia (Le Billon, 2002). Meanwhile, the Khmer Rouge withdrew from the peace process and resumed fighting, financing themselves by trading timber to Thailand. Unwilling to use force against the Khmer Rouge, UNTAC banned the export of timber in an attempt to restrict their finances (Le Billon, 2000).  The 1993 elections resulted in an uneasy coalition between two co-Prime Ministers, Hun Sen and Prince Ranariddh, both of whom continued to capture forestry rent to build their personal patrimonial networks throughout the 1990s. By 1997, they had granted concessions to foreign companies that covered all Cambodia’s productive forests and entered into deals with remaining Khmer Rouge units, with whom the government was still officially at war, and temporarily permitted them to export “already felled” timber to Thailand (Le Billon, 2002, p. 571). Both the Khmer Rouge and co-Prime Ministers captured tens of millions of dollars of revenue from this political accommodation and exploitation of the forests. Khmer Rouge leaders also enriched   44 themselves by hiring out units to guard logging concessions, which they did violently (Le Billon, 2000).  By the late 1990s, forestry rents were declining as the stocks of high value timber had been significantly degraded (Barney, 2007). Meanwhile, Hun Sen and the Cambodian People’s Party had strengthened their grip on power at the expense of their coalition partners, FUCINPEC, and the Khmer Rouge was weakening with large-scale defections within the movement.22 This enabled the government to respond to pressure from international donors such as the World Bank and International Monetary Fund (IMF) to better regulate the forestry sector beginning with a series of logging bans. However, these were primarily used to stop small-scale loggers (often impoverished peasants and de-mobilized soldiers) and “unruly political clients” rather than well-connected companies directed by cronies of the political elite (Le Billon & Springer, 2007, p. 29). At the end of 2001, logging concessions were cancelled and the Law on Forestry (2002) was drafted, which continues to govern forestry practices today. This law established the legal framework for CBFM, the establishment of which is discussed below in section 2.3.   As logging concessions declined in economic and political importance, they were replaced by ‘economic land concessions’ (ELCs) for agro-industrial plantations (primarily sugar, palm oil, and rubber) and mining. Between 2001, when the Land Law was legislated, and 2013, foreign and Cambodian companies, usually with personal connections to the ruling elite (Global Witness, 2007, 2013; MacInnes, 2015), have been granted control of 2.6 million hectares, 14%                                                 22 A few Khmer Rouge commanders defected to the Cambodian People’s Party (CPP), which was under the leadership of Prime Minister Hun Sen, thereby formalizing their positions in the ruling elite. This included the former Foreign Minister of the Khmer Rouge/Communist Party of Kampuchea, Ieng Sary (Ouch & Wright, 2016).   45 of the country.23 Over 80% of these were on forested land, including many protected areas and national parks (Forest Trends, 2015), and they have dramatically hastened the pace of deforestation across the country (Davis et al., 2015). Meanwhile, illegal logging of valuable rare and endangered species for export to Vietnam and China has continued, often using the excuse that it is not newly logged timber but that which was ‘already felled’ legally by ELCs (Environmental Investigation Agency, 2017). In doing so, they not only continue the forest exploitation of previous years but also the ‘public transcript’ used to legitimize it. Small scale clearing for subsistence agriculture and cash crops also contributes to deforestation (Hought et al., 2012; Work & Beban, 2016).  This deforestation, whether legal or illegal,24 does not go uncontested and, a review of news media, NGO reports, and academic studies of Cambodian land and forest politics suggests, these conflicts can and do involve violence. Literature relating to forest-related violence in Cambodia is reviewed below.  2.2.2 Forest-related violence in Cambodia since 2002: a review of the literature Between 2002 and 2015, at least 15 deaths from forest-related violence have been recorded in Cambodia, which includes the death of journalists, NGO-based environmental activists, forest rangers employed by the Ministry of Environment and Forestry Administration officers, and community-based activists and spectators of protests (Global Witness, 2014, 2016a).                                                  23 This number dropped to 12% by 2015 due to the cancellation of a several large concessions, although the companies had partly cleared much of the forested land (LICADHO, n.d.). 24 Milne (2015, p. 202) notes that in Cambodia’s neopatrimonial and kleptocratic political regime, “the distinction between legal and illegal logging fast loses importance.”   46  Physical violence related to illegal logging is regularly reported in the Cambodian news media (Mech, 2014; Phak, 2013; Sek, 2015; Van & Peter, 2017; Peter & Aun, 2016). An online search of the most consistently published, independent, and locally respected English language Cambodian newspaper, the Cambodia Daily, returned thirty-eight articles mentioning ‘illegal logging’ and ‘violence’ or ‘violent’ between September 2007 and September 2017.25 Direct violence related to illegal logging in Cambodia has also been acknowledged in academic papers (e.g. Beban and Work, 2014; Hall et al., 2012; Rudi, Azadi, Witlox & Lebailly, 2014). Reports describe threats of violence against community forestry members and other citizens, including death threats, and shootings, and armed robbery of Forestry Administration offices.  News media, NGO reports, and academic articles also indicate that physical violence occurs in relation to the establishment of large-scale commercial land acquisitions of forested, agricultural, and residential land, known in Cambodia as economic land concessions (ELCs) (Global Witness, 2007, 2013; LICADHO, 2009; Schneider, 2011; Yasmi et al., 2010). ELCs are intended to intensify agricultural production, increase employment and improve local living standards, and generate state revenues (Royal Government of Cambodia, 2005, Articles 3 and 5). An online search of the Cambodia Daily returned forty-five reports mentioning ‘economic land concession,’ and ‘violence’ or ‘violent’ between September 2007 and September 2017. This includes violent eviction from ELC land and beatings and shootings of protesters at the ELC site in question or in the provincial or national capitals.                                                  25 The Cambodia Daily was forced to close on 4th September 2017 amid charges of failure to pay taxes. These allegations are widely interpreted as a means of controlling media critical of the government in the run up to the 2018 general elections (Amnesty International, 2017; Robertson, 2017).   47  These reports clearly indicate that forest-related violence occurs across Cambodia. However, they are unreliable sources on which to base theorizations of forest-related violence in Cambodia as news reports have only selectively engaged with moments of direct violence while political ecologists’ analyses have primarily focused on causal structures rather than the effects of direct violence or responses to and interpretations of it, as explored below.   First, news reports cannot be relied upon to reveal the geographic or temporal scale of forest-related violence across Cambodia as the nature of journalism often directs attention to the most extreme examples of conflict and ignores less dramatic or less physically harmful events (Gritten et al., 2011). Even if a report could be considered representative of an on-going problem, this cannot be known as each report is typically written by a different journalist and treated as an independent incident with no contextualization with the nationwide situation. Furthermore, in Cambodia, some journalists are part of the circles of resource-related corruption and are implicated in bribing illegal loggers and land grabbing elites implicated in forest-related conflicts and violence, remaining silent for a fee and exposing illegal activities if they are not paid (Mech & Kossov, 2016; Ouch, 2014). Others self-censor for fear of violence or law suits if they expose ruling elites’ engagement in forest-related conflict and violence (Cambodian Center for Independent Media, 2016). This undermines the reliability of Cambodian media as a source from which to understand the extent of forest-related violence.   Human rights NGO reports are an equally unreliable source on which to base theorizations of forest-related violence in Cambodia as many NGOs only work in a limited range of provinces.   48 For example, LICADHO (2009) acknowledges that it only records violence in the thirteen provinces where it has offices. NGOs’ adherence to invisible but meaningful provincial boundaries reflects their history of engagement and current, limited finances (both of which speak to the lumpy disbursement of donor funds) rather than reflecting actually occurring resource conflicts (Biddulph, 2011a). Similarly, donor priorities and therefore funding availability can direct their work towards certain fetishized ‘categories’ of people (e.g. indigenous people) or places (e.g. protected areas with internationally recognized but politically-constructed designations such as national parks) (Brockington & Duffy, 2011; Sheil, 2001). It can also focus attention on certain actors in resource conflict and therefore their representation in press briefings and internal reports. In Cambodia, forest conflicts that involve economic land concessions have received significant academic attention and donor funding, which does not only reflect the actual scale of the problem but also international attention to the theme of ‘the Global Land Grab’ (Schoenberger, Hall & Vandergeest, 2017). Other forms of “intimate exclusion,” which are informal and consequently less easily mappable or verifiable than land concessions, slip under the radar (Hall et al., 2012, p. 21).   Media and NGO reports also fail to follow up on the long-term impacts of violence either in terms of the effect on the lives, sense of well-being, or emotions of those involved or in terms of their (dis)continued work defending forest resources. Although a number of academic studies have explored the long-term impacts of loss of natural resources, agricultural land, and homes to ELCs (Baird & Fox, 2015; Dhiaulhaq et al., 2014; Jiao et al., 2015; LICADHO, 2015), the more nuanced question of how surrounding communities continue to manage remaining forests on the periphery of ELCs, potentially in the context of conflict and violence, has not been addressed.   49 This is likely because eviction and exclusion are visible to researchers long after the incident occurs, whilst violence in forests is hidden, has few witnesses, and may not ever become public knowledge. At the practical level, this means the harms perpetrated are quickly forgotten while it also hinders theorizations of violence by perpetuating the notion that an observed moment of violence is a discrete event. This issue mirrors weaknesses in global reports of environmental violence, as discussed in Chapter 1.  Academic analyses by political ecologists have, to some extent, attempted to redirect attention away from moments of direct violence by centering their analysis on the ‘root causes’ of forest-related conflicts in Cambodia, otherwise conceptualized as the structural violence that contributes to direct violence. Neopatrimonialism and corruption among political elites have been prominent themes throughout these analyses. For example, MacInnes (2015) traces how the CPP-controlled state apparatus has enabled five political and economic elites to control over 20% of Cambodia’s ELCs amounting to over half a million hectares (see also Baird, 2014; Neef & Touch, 2012, 2013; Springer, 2011; Un & So, 2011). Through processes of “grand corruption” and using legal mechanisms, they have dispossessed indigenous and ethnic minority communities of forests and agricultural land and violently evicted them from residential land (MacInness, 2015, p. 98).26 MacInnes equates these processes to Galtung’s (1969) definition of structural violence since they entail the use of “power to maintain unequal distribution of (and control over) resources” (MacInnes, 2015, p. 99).                                                  26 Examples of corruption range from government officials accepting bribes to ignore or pervert laws to senior government officials, politicians, and their families directly owning companies that are granted concessions despite this being prohibited by law. These connections are often “kept deliberately secret” (MacInnes, 2015, p. 101).   50 Similarly, examining illegal logging, Milne (2015, p. 200) exposes how the contemporary political regime has a “codependent relationship” with illegal logging in which political elites use and abuse legal mechanisms associated with conservation and foreign investment projects (such as ELCs) to exploit forests for their personal financial and political gain (see also Cock, 2016; Global Witness, 2007; Work & Thuon, 2017). Such behaviour is a continuation of the ruling elite’s kleptocratic practices that have enabled them to finance their political campaigns and gain social and economic power since the 1970s (Le Billon, 2000, 2002; Le Billon & Springer, 2007).  This political analysis goes further towards exposing the processes that manifest as forest-related violence and, as this dissertation argues, accurately identifies Cambodia’s political system as a significant factor in forest-related conflicts around which violence occurs. Yet, in these analyses, direct violence is often skimmed over, given only a passing mention in descriptions of landscape-scale forest conflicts. An unintended consequence is its portrayal as an inevitable outcome of patronage, corruption, and elites operating with impunity. Thus, while such studies expose the structural problems that contribute to direct violence against Cambodian people and the environment, it does not deeply engage with the experience of direct violence. Instead, there is a tendency to render it invisible thereby obscuring actors’ agency to respond and precluding exploration of its structuring effects including those that may be other-than-disciplinary (c.f. Dumont, 1992; Foucault, 1996; Peluso & Watts, 2001). While this ‘big picture’ analysis is essential in understanding where political change is necessary and how it may be engendered in the long-term, it does not provide any indication of how environmental defenders can be better supported or protected from threats to their lives and physical safety in the near term. It is in this   51 context of violent contestation of forests that CBFM is being implemented in the four forms discussed below.  2.3 CBFM in Cambodia Community based forest management (CBFM) belongs to the community based natural resource management (CBNRM) family of approaches (Biddulph, 2015) in that it draws on theories of common pool resources and common property regimes. As such, the approach assumes that if a resource cannot be easily sub-divided, a community of resource managers can effectively exclude outsiders, and the resource offers significant benefits to the community in the long term, communal management may be preferable to state management or individual ownership (Biddulph, 2015; Ostrom, 1999).   In Cambodia, there are four main forms of community-managed forests: community forests (CFs), community-protected areas (CPAs), communal land titles (CLTs), and Reduced Emissions from Deforestation and forest Degradation (REDD+) projects. This section explores the different institutional forms of CBFM projects in Cambodia before discussing reasons for which the government acquiesced to pressure from international development organizations to establish the legal frameworks for CBFM.  Community forestry (CF) is the foremost form of CBFM in Cambodia. The Law on Forestry (2002) first granted communities the right to manage forests and the Sub-decree on Community Forestry (2003) established CF as a legitimate means of forest management. However, guidelines   52 on how to legally establish a CF were not provided until 2006 when the Ministry of Agriculture, Forestry and Fisheries (MAFF) issued a ‘prakas,’ or declaration, documenting the process for communities to identify, legalize and manage the forests (RECOFTC, 2014).   As of 2014, 338 CF sites were under development covering approximately 332,000 hectares, approximately 3.8% of Cambodia’s forested land (OpenDevelopment Cambodia, 2014).27 The Royal Government of Cambodia’s National Forestry Plan 2010-2029 sets a target of increasing this area to two million hectares by 2030, around 20% of Cambodia’s forested land (RGC, 2010). Community forests range in size from just 7 hectares to over 18,000 hectares.28, 29 Forests remain ‘state public property’ meaning the land has “public interest value” and cannot be granted as Economic Land Concessions or for other private enterprises (OpenDevelopment Cambodia, 2015a, para. 1). Communities30 can apply to the Forestry Administration (FA) for management rights over forested land, which may be granted for fifteen years after which they may be renewed. However, under the Sub-decree on Community Forestry (2003), the Forestry Administration can terminate a CF Agreement with six months notice if there is “an understanding of the Royal Government of Cambodia that there is another purpose which provides a higher social and public benefit to the Kingdom of Cambodia” (Article 28). CF                                                 27 The data on total forested land includes rubber and palm plantations, which many forest observers argue “are not forests” (World Rainforest Movement, 2017, para. 1).   28 Nationally, the average size of a CF is 984 hectares while the median is 466 hectares due to the establishment of many CFs less than 100 hectares in provinces where the forest is badly degraded such as Kampot and Kep. The average size of CFs in the provinces included in this study is 2626 hectares and the median is 1723 hectares. Site selection is discussed in more detail in Chapter 3.  29 Raw data from OpenDevelopment Cambodia (2014). 30 Communities are conceptualized by the Sub-Decree on Community Forestry Management (RGC, 2003) as geographical units. They can be individual, legally registered villages or multiple nearby villages cooperating together.   53 establishment involves an eight-step process in which the proposed CF site and Management Plan are approved by various levels of government up to the national level of the Ministry of Forestry and Fisheries. This is an arduous and bureaucratic process, often lasting at least four years.31 Financial, technical, and legal support from NGOs is usually necessary and no CFs are known to have been established without NGO support.   People who live in the participant villages may choose to be enrolled as ‘CF members.’ CF members elect a Community Forestry Management Committee (CFMC), which is responsible for administrative tasks such as drafting the CF management plan for approval by CF members, as well as day-to-day management of CF practices such as organizing members to patrol the forest, recording forest degradation, demarcating boundaries, accounting, reporting forest crimes (such as illegal logging for deforestation) to the FA and law enforcement authorities, and educating community members about the CF bylaws and sustainable forest management practices.   Unlike community forests in many other countries, Cambodia’s CFs do not yet have any community-managed commercial logging enterprises (RGC, 2010).32 Rather, they are managed for subsistence use and conservation of biodiversity and wildlife habitats (referred to by the Forestry Administration as ‘Protection Forests’). To improve the financial motivation for villagers to participate in CF protection, communities are encouraged to pursue alternative ways                                                 31 Many communities have been in the process of establishing CF sites for over eight years (OpenDevelopment Cambodia, 2016b) 32 The government aims to explore the commercial potential of CF through a pilot of Community-Based Production Forestry by 2029 (RGC, 2010).   54 to generate income from the forest such as establishing eco-tourism projects or non-timber forest product (NTFP) enterprises (honey, rattan, and resin processing are common pursuits) (Sunderlin, 2006). The government also plans to generate income for communities and the FA through three forest carbon credit projects (RGC, 2010). These are based on the CF model – multiple CFs in a landscape are ‘bundled’ and counted as one REDD+ project even if the forest is not contiguous (Bradley, 2011). The day-to-day management remains largely the same as other CFs but there is additional technical assistance for measuring carbon emissions and avoided emissions. Although Cambodia was one of the first countries in SEA to initiate a REDD+ project and the national framework was promoted as a promising model for sustainable forest management elsewhere in Southeast Asia (Broadhead & Izquierdo, 2010; Poffenberger, De Gryze, & Durschinger, 2009), its pilot project in Oddar Meanchey has been abandoned by REDD+ donors (Lang, 2014) and has reverted back to regular CF operations due to the FA’s inability to agree terms with a carbon credit buyer in 2014 (Khoun, 2014).  A second form of CBFM in Cambodia is Community Protection Areas (CPAs). In terms of organizational structure at the community level, purpose, and daily management, CPAs mirror CFs. However, CPAs are established in forests that already are already designated as Protected Areas, such as a National Park, Wildlife Sanctuary, or Biosphere Reserve (OpenDevelopment Cambodia, 2016a). These national conservation zones are under the jurisdiction of the Ministry of Environment (MoE) rather than the Forestry Administration, a department in Ministry of Fisheries and Forestry (MAFF), although this has little impact at the community level – both are equally bureaucratic and inefficient at processing applications despite NGO attempts to move the process forward (San, 2006). Due to the similarity of CPAs to CFs and because they are much   55 fewer in number (approximately twenty eight have been granted [San, 2006]), there is little literature specifically on CPAs as distinct from CFs or the larger Protected Areas in which they are located. Even implementing NGOs tend to conversationally refer to CPAs under the same umbrella term of ‘CF.’   Communal Land Titles (CLTs) are a form of CBFM available only to registered indigenous communities.33 The Land Law (2001) and Sub-Decree 83 (2009) provide the legal framework for indigenous communities to obtain indefinite collective tenure over customarily-used residential, agricultural, and forested land. This can include up to 7 hectares of ‘spirit forest’ and 7 hectares of ‘cemetery forest’ and unlimited ‘reserve land,’ which may be forested but must be intended for agricultural conversion for future generations. Thus, CLTs provide much less forest protection than CFs or CPAs. It does not protect forest-based livelihoods or cultures or promote sustainable forest management but, rather, encourages sedentary agricultural livelihoods (Fox, Vogler & Poffenberger, 2009) and places value on the land on which trees stand rather than on forest places themselves.34 As in CFs, communities are required to produce a Management Plan and establish a set of rules regulating forest use but CLTs do not incentivize communities to engage in specific conservation activities such as forest patrols or apprehending illegal loggers. Only eleven CLTs have been granted since NGOs began supporting the initiative in 2005 although at least one hundred and sixty communities have begun the process (CCHR, 2016). The CLTs discussed in this dissertation are under development, although some have submitted the final documentation to the National Task Force that reviews CLT applications.                                                  33 See Baird (2011) on the complex and politicized history of ‘indigeneity’ in Cambodia.  34 See Massey (1994) on the socially constructed and imagined nature of places.   56  The development of CBFM in Cambodia runs parallel to many of the processes of deforestation and politicization of the forests described above. It similarly reflects the ruling elite’s use and distortion of forest policy to sustain their political positions and personal wealth while appearing to acquiesce to the demands of international donors. From 1992 and throughout the 1990s, international conservationists and development organizations began to promote community forestry as a rational response to the deforestation of locally and globally valued ecosystems and the economic need of local communities whose livelihoods depend on them as well as a means of improving democratic governance of natural resources at the local level (Poffenberger, 2009, 2013). However, interventions in and restructuring of forest governance were not immediately accepted by ruling elites since decentralizing forest management to provincial-level FA authorities and communities had the potential to “undermine the basis of the ruling elite’s power” (Cock, 2016, p. 71). On the other hand, cooperation with international aid organizations offered the elite fungible sources of foreign assistance that could be used to enhance their patrimonial rule and free them from domestic pressures to rule for the benefit of the Cambodian public (Cock, 2016). Furthermore, improved management of forest resources provided more direct and organized means through which senior state officials could “capture a greater proportion of the streams of revenue to which the state has access” (Cock, 2010, p. 258). Thus, a decade after community forestry had first been piloted in Cambodia, the government acquiesced to demands by the IMF and World Bank for forest sector reform and drafted the Law on Forestry (2002) and the Sub-Decree on Community Forestry (2003) providing the legal framework for CBFM but maintained conditions that enabled provincial and national level officials (with close personal ties to the top ruling elites) to veto proposed CBFM sites or delay their implementation   57 with administrative demands (c.f. Wittman & Geisler, 2005). In the late 1990s, the government even publicized its compliance to international demands by televising the burning of trucks carrying illegally cut timber, a spectacle designed for international observers since few Cambodians had electricity let alone televisions (Cock, 2016).   The establishment of the legal framework for Communal Land Titles in the Land Law (2001) similarly reflected international pressure and donor demands to recognize indigenous rights but was further motivated by the opportunity to sedentarize indigenous communities, many of whom engaged in shifting cultivation in the remote highlands, and incorporate them into formal villages and communes headed by Village Chiefs and Commune Councillors approved by political elites (CCHR, 2016; Cock, 2016).35 This also enabled the government to neutralize the power of locally-respected traditional elders (Cambodian Center for Human Rights, 2016).  This political genealogy of CBFM helps to explain why, contrary to projections by development and conservation proponents (Poffenberger, 2006; RECOFTC, 2015), CBFM projects have not been as successful as projected in terms of preventing deforestation and illegal logging within their boundaries with some sites being completely deforested by external actors or the CBFM community themselves (Biddulph, 2015; Lang, 2016a; Paley, 2015; Work, 2015) (although Blomley et al. [2010] and Lambrick et al. [2014] find CFs are better protected than state-managed forests). Neither have CBFM projects improved participants’ livelihoods as projected                                                 35 It appears the established trend of donors’ disbursement of funds being contingent on the Cambodian government passing desired legislation is continuing: USAID (2016, p. 6) indicates that the Forestry Administration requested budgetary support of $50,000, which will be provided once the government approves “the subdecree protecting the core forest in Prey Lang Landscape.”   58 (Milne & Adams, 2012; Pasgaard, 2015; Pasgaard & Chea, 2013). The government’s apparent lack of commitment to CBFM also points to why projects are still entirely led by NGOs that are dependent on foreign funding (Bañez-Ockelford, 2011) despite international donors having spent “countless millions” of dollars (Milne, 2013, p. 324) on forestry reform and building the capacity of relevant government departments such as MAFF and MoE that could work directly with communities to govern forests. The power of provincial and national officials to veto proposed CBFM sites or direct them to alternative land also largely explains why many CBFM sites are on already-degraded forest (Biddulph, 2011a; Fichtenau et al., 2002).  Recognizing these weaknesses, some observers argue that CBFM is still in its infancy in Cambodia and problems may be resolved in the future as the government has “taken legal and policy actions that reflect a policy commitment to move forward” with the “devolution” of forest rights to rural villages (Poffenberger, 2013, p. 187). Others see such statements of intent by the government as a continuation of past trends and thus warn that CBFM is “unlikely to work” in the long term due to the lack of political support (Nathan & Boon, 2012, p. 396; see also, DeShazo, Pandey & Smith, 2016; Frewer, 2015; Lang, 2015, 2016a, 2016b). Other scholars remain pessimistic about the future of sustainable forest management in Cambodia but stop short of outright condemning CBFM in the absence of alternative options to support communities trying to manage local forests (Biddulph, 2011b, 2015; Milne & Adams, 2012; Un & So, 2009; Work, 2015).36 In recent years, responses from donors have also been mixed: some donors have quietly redirected their funding elsewhere, such as the European Commission that did not renew                                                 36 Academics have expressed similar concerns throughout the development of CF in Cambodia (e.g. McKenney, Chea, Tola & Evans, 2004; Sunderlin, 2006).   59 funding for CF in nine provinces in 2015 (Radio Free Asia, 2015). Other donors, such as the International Tropical Timber Organisation, USAID, Winrock International, and RECOFTC, have continued or increased their funding (RECOFTC, 2015, 2017; Rogers, 2017) and suggest that further investment, especially in developing legislation and building the capacity of government actors and forest law enforcement officials, will help reduce deforestation and improve the sustainability of CBFM (USAID, 2016). The Cambodian government appears willing to accept such financial aid and continues a discourse of promoting CBFM (RGC, 2010).  2.4 Conclusion This chapter has reviewed the environmental and socio-political context in which the institutions of CBFM are implemented and in which forest-related violence occurs. In doing so, I have shown that Cambodia is a suitable case-study country in which to research forest-related violence since forests are violently contested by those looking to exploit forest resources and local communities seeking to reclaim control over them. Yet, neither this violence nor its implications for affected communities and the future of forest management have been rigorously investigated. Thus, this dissertation seeks to understand how and to what extent forest-related violence shapes CBFM practices.   Several of the themes discussed in this chapter are woven through this dissertation, namely, the ruling elite’s cooption of forest resources, the extent to which neopatrimonialism shapes political systems and social norms, and the impotency of NGOs to bring about meaningful political reform. I take as a starting point the reality of oppression and unequal power relations as historically constructed injustices that are maintained by the current socio-political system and,   60 rather than focusing on the genealogy of these injustices, seek a pragmatic approach to addressing forest-related violence by identifying exactly how it is experienced, what influence it is having on current CBFM practices, how people are responding, and what might be done to reduce its harmful effects.      61 Chapter 3: A methodology for recognizing, researching, and representing direct violence in Cambodia’s community-managed forests   This research project explores experiences of direct violence against CBFM participants and NGO staff. It specifically examines the contexts in which violence occurs, who is involved, and how forest-related violence shapes the implementation of CBFM activities. Central to this is an exploration of the lived experiences and emotional consequences of forest-related violence. This study used an embedded case study design in which forty CBFM projects were studied across four provinces within one case study country. The purpose of the design was to gain a broad understanding of forest-related violence across the country to contextualize data collected in the embedded case studies of individual CBFM projects and their participants.  I begin this chapter with a reflexive narrative describing my positionality in this research project, specifically, how my professional work and personal interactions with CBFM communities in Cambodia since 2008 has fundamentally shaped the design of this research including the subject of inquiry, methodological approach, and the research methods I used. I then discuss the case study selection and methods used for data collection and analysis before examining how violence as a subject of inquiry and the country context of Cambodia pose a particular set of practical, theoretical, and ethical challenges for a researcher and how I resolved or adapted to these challenges.     62 3.1 Methodological approach From 2008-2010, I worked in Cambodia with two environmental NGOs on Community Forestry (CF) and REDD+ projects. One of these was a local NGO with which I spent much of my time working in CF participant villages and community-managed forests. The other was an international NGO for which my work was primarily writing reports for foreign donors who had little familiarity with the context of Cambodia. During that time, I was often frustrated by the chasm I saw between the orderly and rational institutions described by academics and development organizations (a discourse to which I contributed in reports and funding proposals) and the messy reality of the slow, ad hoc activities implemented on the ground. Yet, despite the discursive obfuscation of the complexity of Cambodian lives, the drawn-out bureaucratic dances of trying to get government approval for CF plans, and the disheartening rapidity with which one bulldozer could wipe out years of work establishing a CF, I remained adamant that CBFM was a practice worth supporting. After all, when CF participants, some of whom I regarded as friends, were investing their time and energy in an institution they had been told would protect the forest they valued, I felt a moral obligation to support them even if that meant accepting or turning a blind eye to the imperfections of CF as an institution. Although Economic Land Concessions, roads, and casinos have destroyed many of the CF projects on which I worked in 2008-2010, I maintain that sense of moral obligation. My personal connections to those who value community-managed forests and the CBFM groups of which they are a part make it difficult for me to condemn CBFM in Cambodia as a complete failure or suggest participants’ efforts are for nought as external observers have done based on the ‘objective’ environmental indicators that show deforestation is continuing (see Chapter 2). For me, disavowing my professional experiences and personal connections to CBFM participants would be to actively choose to be   63 complicit in the oppression of already-marginalized communities and in the exploitation of the forests they value (c.f. Hays-Mitchell, 2001; Scheper-Hughes, 1992; Taussig, 1987).  Recognition of my emotional bias despite the ‘objective evidence’ that CBFM has not been a ‘successful’ mechanism of conservation in Cambodia prompted an early iteration of one of my research questions, a modified version of which is explored in Chapter 5 – why do people continue to participate in CBFM despite limited successes? Yet, it was not until friends and former colleagues who had worked in Cambodia repeatedly warned against my conducting research into Cambodia’s forestry sector out of fear for my safety that I developed an interest in forest-related violence. How could my interlocutors be so concerned for my safety while asking rural Cambodians to confront the same people they warned me about – illegal loggers and the political elite implicated in deforestation? Although concern for my safety was valid and did shape my research design in ways described below (section 3.3), this double standard, combined with my aforementioned emotional connection to Cambodian environmental defenders, significantly shaped my approach to the research. It drew my attention to, first, the lack of clarity in academic and grey literature regarding the extent and forms of violence against environmental defenders anywhere in the world,37 and second, it highlighted the lack of representation of the voices of environmental defenders who experience violence, especially relating to their lives after the ‘spectacular’ incident of violence that had been reported. Thus, for reasons of my positionality and to address gaps in the academic literature, my research design is informed by feminist methodologies, which promote the exploration and validation of embodied experiences,                                                 37 Global Witness has since led the charge in filling that knowledge gap (Global Witness, 2014, 2015a, 2016a, 2017), although, as I discussed in Chapter 1, their approach is primarily quantitative rather than qualitative.   64 emotions, and interpretations of those typically marginalized within society (McDowell, 1992; Smith, 1990).   My positionality in relation to the research participants also exacerbates an ethical dilemma central to the study of violence – whether the researcher should be a “scientific spectator,” an “engaged witness,” or a “political activist” (Scheper-Hughes & Bourgois, 2004, p. 26; Whitehead, 2004). Robben and Nordstrom (1995) argue that ethnographers of violence should take a neutral stance since anything else risks judging the situation against normative, ethnocentric values and worldviews thereby representing violence in a way that does more harm than good for the people involved, for example, by reproducing narratives that contribute to the ‘Othering’ of perpetrators of violence as barbaric savages. Drawing on postcolonial and decolonizing methodologies (Spivak, 1988; Young, 1988), they argue that speaking ‘for’ other people is always ethically contentious (see also Alcoff, 1991; Smith, 1999; Staeheli & Lawson, 1994). This is especially so in the case of violence, they argue, as no representation is ever enough to convey the pain and suffering of those depicted and instead cause it to lose its “absurdity and incomprehensibility,” which, paradoxically, are the very qualities representations intend to convey (Robben & Nordstrom, 1995, p. 12).   Scheper-Hughes and Bourgois (2004) also identify the problem of representations of violence succumbing to sensationalism, which, they suggest, is likely when academics position themselves as ‘political activist.’ The possibility to avoid such slippage has been demonstrated by scholars who have successfully used ‘academic activism’ to effect positive changes in the communities with which they work (Maxey, 1999; Speed, 2006). Yet, the viability of this   65 position for my own research was undermined by the practicality that such a position usually requires alignment and coordination with an “organized group of people in struggle” to “accompany them on the contradictory and partly compromised path toward their political goals” (Hale, 2006, p. 96). My research questions were not suited to this position as during the research planning stage I did not yet know what the extent of forest-related violence was nor were my connections to active CBFM communities strong enough to establish participatory research.  In contrast to Robben and Nordstrom (1995), Scheper-Hughes and Bourgois (2004, p. 26) argue that researchers should “eschew the privilege of neutrality” and position themselves as ‘engaged witnesses’ of violence. They suggest that a conscious adoption of this position prevents the researcher from “objectifying” distant suffering yet does not “turn the sufferer into an object of pity, contempt, or public spectacle” (p. 26). Rather, the researcher can “resensitize” their audiences to violence and suffering that may otherwise become invisible (Scheper-Hughes & Bourgois, 2004, p. 26; Schröder & Schmidt, 2001).  I aim for this balance in my own research into and representation of forest-related violence and have thus conducted the research and written this dissertation as an ‘engaged witness.’ The only exception is that Chapter 4, the first empirical chapter in which I explore the different actors involved in forest-related violence and how it manifests, is purposefully written from the perspective of the ‘scientific spectator.’ I took this position in order to avoid demonizing those who use direct violence since their behaviour is likely motivated and structured by factors external and internal to the actor as theorized in Chapter 1. Since I was unable to gather the necessary data to fully explain reasons and justifications for using violence or engaging in illicit   66 forest use, a limitation of the research discussed further below, Chapter 4 is written from a more neutral position of ‘scientific observer.’   3.2 Methods This research project aimed to explore the extent to which forest-related violence influences CBFM practices. To do so required both an understanding of the broader extent of forest-related violence and the ways in which it manifests across the forest landscape as well as the lived-experience of violence and CBFM participants’ interpretations thereof. To achieve this aim, a multiple, embedded case study approach was taken in which data was collected through surveys, interviews, and participant observation and analyzed through conventional qualitative content analysis. Fieldwork was conducted from May-August and November-December 2015. Two Research Assistants were hired who conducted the surveys and translated interviews as needed. While I speak enough Khmer to converse about more practical matters, discussing sensitive subjects requires precision of language, both in terms of being culturally and emotionally sensitive when asking questions and understanding the nuances and implications of answers. My Research Assistants were also invaluable in noting when something appeared to have gone unsaid or when euphemisms were used.  3.2.1 Selection of case studies Cases are “empirical units, theoretical constructs, and subject to evaluation because scientific and practical interests are tied to them” (Scholz & Tietje, 2002, p. 1). Embedded case studies involve more than one unit of analysis and data derived from one unit can be linked or compared to that derived from the larger unit in which it is embedded (Scholz & Tietje, 2002). In this research   67 project, the largest unit of analysis is the single-country case study of Cambodia (the choice of which was justified in Chapter 1, section 1.5). Embedded within this are four case study provinces, within which are embedded forty CBFM project site case studies. Studying different provinces enabled me to explore to what extent forest-related violence differs as the properties of forest resources, actors involved in forest conflicts, and the networks of political power vary across the country. Although provincial boundaries are a somewhat arbitrary way to delimitate the study of forested landscapes that extend across the country, they have meaningful consequences for CBFM project design and implementation in Cambodia as NGOs and donors often target specific provinces and provincial-level political leaders may be more or less willing to cooperate with administrative processes. Boundaries, and especially those associated with international borders, are also meaningful for CBFM practices as they involve specific actors, rules of conduct, and socio-natural practices (Peluso, 1995; Sturgeon, 2004).  Case studies, including those embedded within larger units of analysis, are usually selected because they are thought to be either representative of a broader population or “deviant” or extreme examples of a given phenomenon (Seawright & Gerring, 2008, p. 295). However, a challenge for all case-based studies is that at least a general understanding of the broader phenomenon of study is needed before representative or anomalous cases can be identified (Seawright & Gerring, 2008). As discussed in the previous chapter, neither media reports nor NGO activities can be relied upon to accurately represent the spatial spread of natural resource related conflicts in Cambodia so it was not clear which locations might be either representative or anomalous. Thus, the case study provinces were selected to show variation in two factors   68 identified as important in the literature review, namely environmental properties of the forest and institutional arrangements of the CBFM project.  Ratanakiri province was chosen in part because it is one of the most densely forested provinces of Cambodia (Fox, Vogler & Poffenberger, 2009). It was also chosen because it is one of only two provinces where Communal Land Titles (CLTs) are being established due to the geographical concentration of indigenous people (the other being Mondulkiri). Both of those two provinces have comparable natural resources and share an international border with Vietnam but Ratanakiri was the preferred choice for the practical reason that it is the home province of one of my Research Assistants. Although he did not know anyone in the study sites, his knowledge of the landscape and acquaintances all over the province enabled us to travel far from the provincial capital to CLT sites in every district without worrying about finding accommodation. He also speaks enough of two ethnic minority languages to help establish a rapport with research participants even though interviews were conducted in Khmer.   Kampong Thom province was chosen in part because of the presence of the CBFM group, Prey Lang Community Network (PLCN), which was co-led by Chut Wutty, an environmental activist who was assassinated in 2013 in Koh Kong province while investigating illegal logging and the illegal processing of ‘yellow vine’ (Coscinium fenestratum), a non-timber forest product used in the production of narcotics (May & Ponniah, 2013). This indicated that other CBFM members in the province would have considered the possibility of forest-related violence occurring during their work even if they had not experienced it. Kampong Thom also offered a diverse range of environmental factors to study including different CF sizes ranging from less than 200 hectares   69 to more than 7000 hectares, and forests that have been degraded to different extents and thus have different forest resources to protect.   Oddar Meanchey province was selected as a study site because the thirteen CFs are, on paper, some of the most well-established in Cambodia: they were the first, and remain some of the only, CFs to receive approval at the national level and they formed the basis of Cambodia’s first REDD+ project and so received more technical training on reducing illicit forest use and forest-related conflicts than many other CBFM sites (Bradley, 2009; Poffenberger, 2009). This indicated that CBFM practices would be well established and could be a useful comparison to other less-well established CBFM groups elsewhere in the country. Furthermore, forest-related conflicts were known to have occurred but besides the violent evictions of three villages by ELCs, reports of forest-related violence were almost entirely absent in English-language newspapers. This raised the question of whether these CF groups had been able to avoid violence or whether it was simply not reported.  Preah Vihear is bordered by Oddar Meanchey to the west and Kampong Thom to the south. Correspondingly, the environmental properties of the forests in the west and north are comparable to those of Oddar Meanchey while those in the south are more comparable to those of Kampong Thom and are also partially covered by the Prey Lang Protected Area. Like Oddar Meanchey, Preah Vihear has an international border with Thailand, which has the potential to influence legal and illegal logging (Singh, 2013) and has also been the subject of an armed conflict with Thailand 2008-2011 that brought the military into both provinces (International Crisis Group, 2011). Since the literature review and initial fieldwork suggested that the military   70 were involved in illegal logging and forest-related violence in Oddar Meanchey, including Preah Vihear in the study provided the opportunity to contextualize the findings from Oddar Meanchey.  To select CBFM sites within each province, lists of all the potential sites were obtained from NGOs supporting CBFM. With the exception of Oddar Meanchey, where all CBFM sites were included in the research, two CBFM sites in each possible district of each province were chosen at random in order to include CBFM sites in a broad range of forest types. CBFM sites that had been destroyed by ELCs or abandoned for other reasons were omitted as it was assumed they would no longer have any CBFM practices to study. The case study provinces are shown in map 3.1. These CBFM project sites within each province are not shown to maintain the anonymity of participants.  71  Figure 3.1 Cambodia in Southeast Asia and case study provinces. Map adapted from OpenDevelopment Cambodia (2017) 72  3.2.2 Data collection methods Empirical data were derived from a survey of NGOs, semi-structured and key informant interviews, and participant observation. A survey of community members in CBFM villages was also conducted although the data collected are not used to a great extent throughout this dissertation due to concerns about their validity and reliability, as discussed along with other limitations in section 3.3 below.  To explore the extent of forest-related violence across Cambodia and, therefore, better contextualize the data from my case study sites, I conducted a survey of national and international NGOs that were currently or had in the past five years (2010-2015) implemented a form of CBFM including Community Forestry (CF), Community Protected Areas (CPA), or Communal Land Titling (CLT). The survey explored the extent to which NGO staff had experienced forest-related violence and the extent to which they believed participants in their projects were affected (see Appendix A.1). Participants represented eighteen NGOs that have supported CBFM projects in fifteen of the twenty-two provinces in which CBFM projects are implemented.38 The respondent for each NGO was selected by the organization after being asked who had the most experience of working with CBFM communities. Surveys were conducted over the phone or in person with follow up questions where appropriate. This survey was also used to identify eight NGO staff that had experienced forest-related violence and were willing to be interviewed further about this (see Appendix A.2).                                                  38 List of CBFM projects obtained from OpenDevelopment Cambodia (2014). 73  Semi-structured interviews were conducted with one hundred and fifty participants in forty CBFM projects across four provinces. This included one hundred and eighteen members of CBFM management committees and thirty-two members not in leadership positions. Multiple members of the same CBFM management committee were often interviewed as a group, which was their preference. Interviews explored experiences, perceptions, and interpretations of forest-related violence, how CBFM groups are responding, and to what extent forest-related violence shapes forest protection activities and their day-to-day use of the forest, that is, CBFM practices (see Appendices A.3 and A.4). I drew on the ‘oral history’ method (Perks and Thomson, 2015), inviting interviewees to consider how their experiences with CBFM and forest-related violence have changed over time. This approach reduced the emphasis on narrating the most extreme examples of violence and encouraged discussion of less explosive conflicts and more subtle forms of violence.   Semi-structured interviews were also conducted with four CBFM members who have engaged in illegal logging in Cambodia and/or Thailand. These interviewees were not sought out but were encountered during the scheduled interviews with CBFM participants, either because they were CBFM participants themselves or were related to someone who was. These interviewees were not asked if they had ever used violence as this could have been interpreted as an accusation but they were asked if they had ever experienced violence from community-based or government-employed forest defenders.  Further in-depth semi-structured interviews were conducted with thirteen key informants, including CBFM Management Committee members, the Head Monk of Oddar Meanchey 74  province, the Head of the Oddar Meanchey CF Network, the Head of the Ratanakiri NGO Network, leaders of the Prey Lang Community Network (Kampong Thom), and various experienced members of NGO staff. These key informants were selected based on their lengthy participation in CBFM, knowledge of forestry and land-related conflicts in and beyond their own village, and CBFM leadership experience, especially if working as an intermediary between communities and NGOs.   Participant observation was used to gain a better understanding of the types of forest resources available, how forests are used, and locations in which violence is likely to occur. This included joining two over-night and four daytime forest patrols, collection of NTFPs, and meetings and training sessions about CF activities. This provided opportunities to ask questions related to the activities around which violence occurs and, since conversations were in a more relaxed atmosphere than semi-structured interviews, prompted participants to reflect more deeply on their experiences and interpretations of forest-related violence. Participant observation also provided an opportunity to converse with or observe other people involved in or affected by CBFM such as Ministry of Environment rangers, FA officers, soldiers recruited to join CF patrols, and migrants clearing forests.  A survey was also conducted of three hundred and twenty households in thirty-two villages that participate in twenty-five CF projects in two provinces (Kampong Thom and Oddar Meanchey). This was conducted by the two Research Assistants and I observed and took notes. Every fifth house on the right hand side of the road was sampled, moving on to the next one if no one was available to answer questions. This assessed the extent of different households’ use of the forest, 75  membership and participation in CBFM activities, and perception of challenges to CBFM. Open questions were asked about challenges to conservation to assess whether concern over forest-related violence was raised. Due to the sensitive nature of the topic and the uncertainty as to the extent the respondent supported CBFM or was engaged in illicit forest activities, this cautious approach was preferable to asking directly about violence.   3.2.3 Data analysis method Wolcott (1994, p. 3-4) asserts, “everything has the potential to be data, but nothing becomes data without the intervention of a researcher who takes note – and often makes note – of some things to the exclusion of others.” This statement accurately represents the data analysis process for this dissertation. Faced with an extensive amount of interview data, survey responses, and ethnographic fieldnotes, not all of which were pertinent to the research questions, I employed a conventional qualitative content analysis (Hsieh & Shannon, 2005). This method has been usefully employed by social scientists faced with an abundance of data who are seeking to focus attention on “selected aspects of meaning” that relate to overall research questions (Schreier, 2014, p. 170). It also emphasizes the value of an iterative approach and is sufficiently flexible to accommodate new codes and themes that emerge from the data rather than being generated by the literature review (Schreier, 2014).  Data was initially coded under the broad themes of ‘cause,’ ‘forest properties,’ ‘actor’ (Chapter 4), ‘response’ (Chapter 5), ‘emotion,’ ‘participation’ (Chapter 6), and ‘outcome’ (Chapters 4-6). These themes were guided by the research questions. Sub-themes were derived from the data although guided by the literature review, which had indicated that environmental, socio-political, 76  and institutional context would likely be important factors in understanding the cause of violence and that identities, relationships with others, and emotions could be shaped by experiences of violence.  Initial interpretations of the data collected from May to August were discussed with three key informants during my second data collection trip in November to December. Initial interpretations were also discussed with my research assistants during the fieldwork and in several Skype conversations afterwards, extra work on their behalf for which I am grateful. Although all interpretations are my own, I feel confident that I have represented the research participants fairly and accurately. However, returning to Wolcott’s (1994, p. 3) observation that “transforming” always excludes others, this dissertation inevitably leaves out many personal stories that interviewees specifically requested be told and I have left unexplored several theoretical fields in order to focus on the key research questions. Both these forms of data exclusion were necessary since this dissertation is a foray into a relatively unexplored topic, especially in the context of Cambodia, and thus it aims to provide a baseline of understanding about the extent and impacts of forest-related violence rather than being an exhaustive account.   3.3 Practical and theoretical challenges and limitations to the research Violence as a subject of inquiry posed a particular set of theoretical and practical challenges for this research project, as did the country context of Cambodia. This section explores those challenges, how I resolved or adapted to them, and how this shapes the ‘knowledge’ produced.   77  The major theoretical challenge encountered during data collection and analysis was how to bound and recognize ‘forest-related violence.’ Delimiting what counts as ‘violence’ is a challenge to all researchers studying the topic and Feldman (1991, p. 4) warns “the danger lies in making definitions of violence appear too polished and finished – for the reality never will be.” The literature review had indicated that direct violence should be the main focus and used as a way to illuminate processes of structural and symbolic violence but my conceptions of what direct violence is were challenged in the field. Specifically, many interviewees described threats of physical violence, threats of imprisonment, verbal intimidation, and harassment such as being followed. I have included these in the analysis because, like Galtung’s (1969) conception of direct violence, they appear to indicate an intentional and targeted attempt to discipline behaviour and were, to some extent, successful in doing so. Furthermore, my methodological approach emphasized the importance of recognizing people’s interpretations of violence and, despite the lack of physical contact, these experiences appear to elicit similar feelings of fear and anxiety, which were also found to structure behaviour. By maintaining a flexible approach and open mind to what violence would ‘look like’ and being sensitive to what research participants labeled as violence, my definitions and boundaries emerged through the research.   A related challenge was what violence counted as ‘forest-related’ and what emerged from other sources of conflict. As the literature review discussed (Chapter 1), various socio-political and historical factors can contribute to the emergence of violence. To accommodate these complexities, I asked research participants specifically about violence they experienced during or related to their CBFM work but also explored the contributing factors during conversations that followed in order to understand the other contributing factors. The blurring of forest boundaries 78  was especially problematic in cases of intra-community violence as those involved may have had histories of social tension within a village. Again, interviewees’ own narratives shaped my definitions and I have illustrated local complexities in a case study in section 4.4. One form of violence that could be labeled ‘forest-related’ was excluded: violence during protests over loss of natural resources. Over the last decade, public protests have become a more regular occurrence in Phnom Penh, many of which include participants affected by forest land grabs, agricultural land grabs, and urban eviction, but can also morph into anti-government protests (Human Rights Watch, 2014; Robertson, 2017b; Schoenberger, 2017). I assessed that these protests are beyond the scope of regular CBFM activities and, not wanting to expand the scope of this research project too far, I excluded them from this research. By extension and using the same logic, I applied the same logic to protests at ELC sites.  The theoretical challenge of identifying forest-related violence also contributed to a practical challenge. I initially attempted to quantify the number of incidents of violence but soon found they were too common to count, even by the people involved. This was especially so when the interviewee was a particularly active patrol member. It was also difficult to verify whether incidents were being double-counted during the interview process as several people involved in one incident may each recount the same story. Pinpointing exact dates of incidents was also challenging because there is little documentation of forest-related violence and calendar dates, including specific years, are not always memorable. Thus, dates are only mentioned in this dissertation when the interviewee could be certain they were correct.   79  Violence is a politically and emotionally sensitive subject and thus its research entails specific practical challenges that limit the data the can be collected and the kinds of ‘knowledge’ that can be produced. Practically, talking about violence is difficult. It is a loaded term. For some, it implies discussing illegal behaviour and may cause the concern that the interviewee will be asked to implicate someone in a crime and potentially face repercussions or retributive violence. It also has implications of trauma or other strong emotions, which can be difficult for victims and perpetrators to talk about (Gillen, 2014; Hays-Mitchell, 2001). Talking to an ‘outsider’ such as an interviewer can be freeing or anxiety-inducing depending on the individual interviewee (Federman, Hanna & Rodriguez, 2003; Nuno & St. John, 2015). In Cambodia specifically, the need to ‘save face’ and hide vulnerabilities is an important issue to navigate, especially when considering masculinity and the acceptability of discussing ‘failure’ or ‘weakness’ (Brickell & Garrett, 2015). Finally, the interviewer may not initially know which ‘side’ of a conflict the research participant is on. Do they or do they not support CBFM or are they involved in the illicit activities that contribute to violence? Have they been a perpetrator or victim of violence, or both?  Considering these issues, I structured surveys with members of CBFM communities and initial interviews with CBFM participants around general experiences of participation and barriers to project success, and gently probed the issue of conflict to see if people brought up incidents of violence, a strategy suggested by Federman et al. (2003). In many cases they did and once they had raised the topic and I had re-affirmed consent to ask more questions, it was often very easy to open up the conversation and delve deeply into their experiences. This was especially so with key informants, with whom I was more explicit about the topic of my research and was able to ask more probing questions after they had the time to reflect on what they were or were not 80  comfortable to talk about. Although this strategy did enable me to collect a large amount of data about forest-related violence in Cambodia, the question remains of how much data did I miss by not being able to discuss violence freely with all interviewees? Thus it is likely that the extent of forest-related violence is under-represented rather than over-represented in this study.  An important practical limitation to understanding the use of forest-related violence was who could and could not be interviewed without jeopardizing the safety of other research participants or the research team due to political sensitivity of the research topic and Cambodia-specific factors. Specifically, I did not attempt to interview anyone who was thought to have used direct forest-related violence such as illegal loggers, soldiers, or ELC employees, among others. Interviewing those who feel able to use such violence would have risked the personal safety of research team and could have been motivation for retributive violence against local CBFM groups if they were suspected of making accusations. Furthermore, it would have required interviewees to implicate themselves in illegal activities and discuss behaviour that is often considered morally reprehensible – factors that can make it difficult to build sufficient rapport with the research participant to collect data that can be considered reliable during a relatively short fieldwork period (McDowell, 2010). Although this remains a gap in the research, some insights have been gained into why people engage in illicit forest user or forest-related violence from conversations with survey participants who oppose CBFM and with CBFM members who admitted they have previously engaged in illegal logging in Cambodia or Thailand. Furthermore, CBFM participants have often had long conversations with illicit forest users and those who use forest-related violence so were able to provide some speculative insights into their motivations. While this remains an area for further research, the practicalities of doing so are challenging. 81   Similarly, I was unable to interview Forestry Administration (FA) officers who, like CBFM participants, are tasked with disrupting illicit forest use yet who are also part of an institutionalized neopatrimonial system and therefore directly or indirectly personally connected to those whose activities they are responsible for regulating (Milne, 2015). While interviews with a FA officer would have allowed the government a chance to respond to questions about how CBFM participants could best be supported and forest-related violence reduced, the FA as an organization is notoriously wary of researchers (Jacobsen & Stuart-Fox, 2013) and could have prevented me from conducting any research.39 Therefore, the decision was taken to not alert them to my presence until the majority of the interviews with CBFM participants had been collected. At that time, and as expected, it was impossible to secure an interview with the FA, who demanded I first obtain a ‘research visa,’ which does not exist in Cambodia. However, I was able to talk with a FA officer during participant observation of a forest patrol although the precise research topic was left vague as it was unclear to what extent any probing questions might undermine his support for the CF group.  The data collected through the survey of villagers in CBFM communities are not used extensively throughout this dissertation because I have concerns about their reliability. While some participants were forthright in stating that they did not know much about it and others had strong opinions either for or against CBFM, many respondents gave contradictory answers                                                 39 The police have previously detained foreign and Cambodian researchers investigating land rights issues in Oddar Meanchey (Sherchan, 2015) and deported environmental activists (Tat, May, Pye & Tapper, 2015). Investigative journalists are frequently detained (e.g. Wright & Van, 2017). The distinction between academic researchers and journalists is not well understood in Cambodia (Schoenberger & Beban, 2017). 82  throughout the survey which, combined with my observations of their body language and demeanour, I interpret as them trying to give the answers they thought a foreigner would want to hear. For example, they answered that they participate in CBFM activities and the forest is very important yet in later questions could not give examples of CBFM activities that they had joined, forest products that they use (e.g. firewood or food) or other reasons for which they value the forest, or suggest whether the forest faced any challenges with illegal logging or deforestation. Furthermore, few survey participants were willing to talk about which actors might be involved in forest-related conflict and even fewer were willing to talk about forest-related violence. In hindsight, this is likely due to fear of village-level surveillance (Milne, 2015), expectations that white foreigners would have a particular political agenda which might not be aligned with their views (Schoenberger & Beban, 2017) and what Bourdieu (1977) would likely label the ‘habitus’ of not criticizing the politically powerful (Jacobsen & Stuart-Fox, 2013). Nevertheless, the method provided some useful background and contextual knowledge of the forests in each location, even if many participants were not reliable informants or knowledgeable about the issues at hand in the systematic fashion that a reliable survey would require. It also provides the empirical insight that active participation in CBFM is potentially lower than the number of enrolled members would suggest and the role of the forest in people’s livelihoods is less than the oft-stated but poorly qualified “important” repeated by proponents of CBFM (e.g. Kim, Sasaki & Koike, 2008, p. 345; Mulcahy & Boissière, 2014, p. 1). Epistemologically, it is also a reminder that knowledge production is always political and as much as the researcher can represent findings according to their own interpretation, the ‘subjects’ of research can purposefully shape what is ‘known’ about them (McDowell, 2010).  83  In summary, there are significant challenges in recognizing, researching, and representing experiences of violence, the emotions it generates, and the murky social and political networks with which it is often intertwined. The Cambodian context adds its own challenges for researchers. These limited the knowledge that could be produced in this research project and constrained the conclusions that can be drawn. However, in some cases, these limitations reveal as much as they hide about the processes of forest-related violence and the Cambodian socio-political context, such as the FA’s unwillingness to be interviewed and the fact that forest management has become such a political issue that cannot be spoken about freely unless the interlocutor’s position is at least partially known. Thus, this methodological discussion and explanation of my positionality in the research should inform the reader’s interpretation of the following chapters.   84  Chapter 4: Patterns of direct violence in Cambodia’s community-managed forests  4.1 Introduction This chapter presents the main patterns of direct violence affecting community-based forest management (CBFM) participants and the NGOs that support them. It documents and examines what forms of direct violence occur, who is involved, and what environmental, socio-political, and institutional factors create the potential for, motivate, and facilitate violence.  In doing so, this chapter presents findings from, what I believe to be, the most comprehensive review of forest-related violence in Cambodia that has yet been conducted. Since media reports of forest-related violence in Cambodia usually only highlight the most dramatic and attention-grabbing violent ‘events,’ this chapter exposes the ‘everyday violence’ experienced by CBFM participants as much as it describes more spectacular stories. Yet, each ‘event’ described is part of a larger picture and contributes to the identification of broader trends in the spaces in which violence occurs, who is involved, how violence manifests, and what factors and processes contribute to these patterns. It provides the necessary context for the discussions in the following chapters of how CBFM participants and NGOs respond to and interpret forest-related violence and how it shapes CBFM practices.  This chapter presents three main findings. First, that the potential for violence to occur differs between CBFM sites since forest-related conflicts occur unevenly across Cambodia’s CBFM sites. The spatial distribution of forest-related conflicts, which changes over time, is influenced by environmental factors (distribution and perceived value of forest resources), institutions 85  (natural resource management policies, national laws, and local bylaws), and socio-political and economic factors that motivate or facilitate illegal logging and deforestation. The case study presented in section 4.4 shows that prior experiences of direct violence, perceptions of (in)justice, and emotions of betrayal, frustration, and anger also contribute to and motivate the emergence of violence between members of the same village. This finding corresponds with the theory explored in the literature review regarding the range of factors that structure the potential for environmental violence (section 1.3.1).   Second, I find that different actors perpetrate forest-related violence in different CBFM sites. These actors can usefully be conceptualized as belonging to five heuristic categories – ‘organized illegal logging groups,’ ‘military loggers’ (who can also be considered a type of organized logging group), ‘local illegal loggers’ from CBFM participant villages or neighbouring villages, ‘land-seeking migrants’, who engage in deforestation rather than selective logging, and Economic Land Concessions (ELCs), which contribute to both illegal logging and deforestation.40   Third, I find that the identity of the actor involved in a forest-related conflict strongly influences the extent to which violence is used. Specifically, the motivation and ability of different actors to use violence in a conflict largely dependent on their patrimonial connections within Cambodia’s                                                 40 All references to ‘logging groups’ in this dissertation refer to illegal logging groups. The only logging permitted or considered legal in CBFM sites is that in which timber is for personal use, such as house building. This is conducted on an ad hoc basis by members and overseen by Management Committees, if no confiscated, illegally-cut timber is available for use. Legal logging is not a source of conflict or violence in the CBFM sites surveyed. 86  socio-political system that change the balance of risk and reward and enable and embolden them to use violence with impunity.41  The chapter proceeds by explaining the heuristic categories I use to describe different ‘types’ of actor discussed in this dissertation and providing an overview of the trends of forest-related violence occurring in Cambodia’s CBFM sites. I then explore the environmental, institutional, economic, and socio-political factors that influence each actor’s involvement in forest-related conflicts, and the extent to which, how, and why they use direct violence. For each, I highlight the factors that contribute to some of these conflicts becoming physically violent. I pay specific attention to why violence does not occur in all CBFM sites where these actors are present as the absence of violence can contribute to understanding its presence in other situations.42 This is followed by a detailed case study of two Community Forests (CFs) in Oddar Meanchey province that illustrates the overlapping forms of violence, how neopatrimonialism distorts the implementation of CBFM, and how forest-related violence is used as a means to discipline local forest users and CBFM participants, disrupting the relationships that tie them together in belief and action.                                                   41 This argument is strengthened in Chapter 5 in which I explore how different actors are able to distort Cambodia’s justice system and avoid punishment for the use of violence. 42 This research does not include violence against land-rights activists at protests and rallies. Although such incidents are common in Cambodia and do affect people who are protesting the loss of their forests (Neef & Touch, 2012; Springer, 2009; Lamb et al., 2017), this research focused specifically on violence occurring in the course of routine community-based forest management activities, rather than at confrontational political events. This focus is intended to draw attention to the ‘everyday’ violence occurring beyond the ‘spectacular’ events that make the news headlines, a focus justified in the introductory chapter to this dissertation. 87  4.2 Methods  The data used in this chapter were collected through a survey of eighteen local NGOs that support CBFM, semi-structured interviews with one hundred and fifty CBFM participants in forty CBFM sites and with thirteen key informants, and participant observation of CBFM activities including forest patrols and NGO training sessions between May and December 2015. I draw on qualitative data to identify broad trends in who uses violence, against whom, and where, and to provide detailed examples of lived experiences of violence and, when necessary, highlight where and why exceptions occur. I explore the context-specific factors contributing to the occurrence of violence, especially relationships between different actors. The reader is reminded that although data may appear anecdotal due to the lack of formal police reports of violent incidents, interviewees’ inability to pinpoint exact dates of incidents, and my decision to anonymise villages and CBFM sites, the intention of this research is to explore CBFM participants’ lived experiences and find broad trends – such as commonalities between experiences and patterns in the ways violence manifests – rather than creating a database of incidents of violence or a mathematical model to predict its occurrence.  4.3 Forest-related violence by different actors in Cambodia’s CBFM sites Five heuristic ‘types’ of actors are implicated in the use of forest-related violence against CBFM participants and NGO staff in Cambodia. These are organized logging groups, military loggers, local loggers, land-seeking migrants, and employees of economic land concession (ELC) companies. The data presented in this chapter indicate that these actors use direct violence in different ways. However, ‘illegal loggers’ are often discussed as a singular threat to the environment and/or environmental defenders regardless of the social and political positions of 88  the actors involved because the immediate environmental consequences of their actions are similar – trees of valuable species are lost. This homogenization of multiple, related actors involved in logging and deforestation means there is little specificity in the literature regarding the manifestations of associated violence, its spatial distribution, or ways in which environmental defenders interpret such violence or respond to it. Yet, my research illustrates that different actors use violence in different ways. With this in mind, I begin this section with a typology of the different ‘neck kab prey cheur’ – people who cut the forest – to whom I refer throughout this dissertation. I then analyze the different ways in which these actors use direct violence.  4.3.1 A typology of ‘neck kab prey cheur’ (people who cut the forest) In Cambodia, ‘illegal loggers’ can be conceptualized as ‘organized logging groups,’ ‘the military,’ and ‘local loggers’ from CBFM participant villages or neighbouring villages.43 ‘Migrants’ and ‘economic land concessions’ (ELCs) are also involved in deforestation through clear-cutting and, to some extent, illegal logging. The operations of these groups are discussed in more detail throughout the chapter but four factors that distinguish these groups are summarized below and in Table 4.1.   First, ‘illegal loggers’ are primarily engaged in selective logging of valuable tree species while migrants and ELCs tend to responsible for clear-cut deforestation. Organized logging groups target forests anywhere in the country with relatively abundant and mature ‘luxury’ hardwood                                                 43 All references to ‘logging groups’ in this dissertation refer to illegal logging groups. The only logging permitted or considered legal in CBFM sites is that in which timber is for personal use, such as house building. This is conducted on an ad hoc basis by members and overseen by Management Committees, providing no confiscated, illegally-cut timber is available. 89  species.44 Military loggers are most active near military bases. Of the four provinces where research was conducted, military loggers are a problem across the whole province in Oddar Meanchey (environmental and political reasons for which are discussed further in section 4.3.2.2). Local loggers are less mobile, usually only logging in forests with which they are familiar, and, besides targeting large logs of ‘luxury’ timber, will also dig out stumps and roots of previously logged valuable trees and less financially-valuable construction-quality timber if there are accessible trees of a suitable size (approx. diameter >70cm), which are not sufficiently valuable to be of interest to most organized logging groups.45 Migrants searching for land tend to target land where they can clear the forest without being observed or disturbed, which directs them towards larger CBFM sites of at least a few hundred hectares.   Second, the scale of operations varies among illegal loggers. Organized logging groups can work on a much larger scale than local loggers. Groups may consist of dozens of loggers and frequently extract several trucks of timber from a site in a day.46 Military loggers do not operate on a large scale, with military groups generally limited to five to ten soldiers at a time, and with different soldiers rotating in on different days. However, they have greater access to resources such as chainsaws and trucks than local loggers and can thus clear forest much quicker. Migrants operate in various group sizes – some are single families while others work in larger groups of multiple families and may burn large areas of forest to clear it rapidly. ELCs operate on the                                                 44 Selective logging primarily targets valuable hardwoods such as Dalbergia cochinchinensis (Siamese rosewood, locally ‘kronoung’), D. bariensis (Burmese rosewood, locally ‘neang noun’), Afzelia xylocarpa (locally ‘beng’), Pterocarpus pedatus (paduak, locally ‘thnoung’), and Hopea ferrea (locally, ‘korki’). 45 Exceptions always exist. Two neighbouring CF groups in Preah Vihear complained that organized logging groups connected to an ELC growing and processing sugar cane had illegally logged their resin trees. 46 Observations in Tumring district, Kampong Thom province, June and July 2015. See also NGO Forum on Cambodia (2015). 90  largest scale with a single concession often nearing 10,000 hectares. Heavy machinery enables them to clear a forest very quickly. Since few ELCs are granted with free, prior, and informed consent of local people, CBFM groups often have little warning their forest will be destroyed.  Third, the management structure of these groups varies. Organized logging groups have formalized management structures enabling them to not only control employee loggers’ work but also their bodies and time. Employees eat, sleep, and work as a group and managers closely monitor their movements beyond the logging camp, sometimes prohibiting them from leaving freely.47 Managers also sell amphetamines to employees, which they use as a stimulant for the gruelling work. Employees often become indebted due to drug use, which effectively traps them in the logging group.48 Military loggers and local loggers appear to be self-organized although in some cases operate on commission for organized logging groups.   Fourth, organized groups and military loggers have overt patronage from political and economic elites, referred to locally as ‘Big Men’ (neck thom) or ‘oknya’ (often translated as ‘tycoon’). These elites may have some form of legal logging rights, such as economic land concessions, or claim to do so such as around military bases, but employ people to log far beyond those boundaries (NGO Forum on Cambodia, 2015). In contrast, local loggers and migrants work independently but may have a ‘string’ (k’sai) connecting them to organized logging groups, such as brokers who request timber of specific species and sizes thereby encouraging logging and concomitant violence within communities, as discussed further below.                                                  47 Interview with former employee of logging group, Ratanakiri, November 2015. 48 Interview with NGO-based key informant, Phnom Penh, 20th July 2015. 91    Forest resources targeted Scale of operations Management Patronage Organized logging groups Selective logging of ‘luxury’ timber anywhere in the country where stands are relatively abundant. May log construction-quality timber once luxury species are gone. Accessible by existing roads or cut new road. Dozens of loggers extracting multiple trucks of timber per day over multiple days, weeks, or months. Use chainsaws and heavy machinery. Establish sawmills. Own trucks for exporting.  Formalized management. Employees’ time and behaviour often tightly controlled by managers. Overt patronage from political and economic elites who provide legal impunity. Often claim logs were felled inside ELC despite evidence to the contrary. Military loggers ‘Luxury’ and construction-quality timber near military bases. Also responsible for clear-cut deforestation for construction of bases. Do not appear to overlap with other organized logging groups. Groups of approx. five to ten soldiers at any one time although loggers rotate in and out. Use chainsaws, and use trucks for exporting logs. Sawmills in army bases give appearance of ‘legitimacy.’  Formal management near military bases and when officers claim timber is for military purposes. Informal organization when logging is purely for financial gain. Overt patronage from military and political elite who provide legal impunity. Often claim logs were felled for official military purposes despite evidence to the contrary. Local loggers Local forests. Selective logging of individual ‘luxury’ or construction quality trees, or stumps and roots of ‘luxury’ trees that have already been felled. Accessible by motorbike or ‘tractor’ (koh yun). Individuals or small groups of five to ten (occasionally larger). Use handsaws or sometimes chainsaws. Use motorbike or ‘tractor’ (koh yun) for transport. No formal management. Usually independent but may have been commissioned to find wood of specific species and size. Usually no patronage. Sometimes loose connections or ‘string’ (k’sai) to political/economic elite at local level but no guarantee of legal impunity. No claims to legality. Migrants Target land rather than timber and typically clear-cut and/or burn areas of forest. May fell valuable timber if it is near targeted land. Variable: small family groups of two to six people or multiple families working together. Often clear forest by hand using chainsaws but may burn many hectares or use bulldozers/diggers provided by patrons. Self-motivated groups have no formal management. Those employed by a ‘big man’ (neck thom) from the ruling elite may have loose oversight of operations. Often receive patronage from ruling elite, who directs them to land, funds transport and initial living expenses, and provides machinery to clear forest and sometimes guns to scare away forest patrols. No claims to legality. Economic land concessions (ELCs) Target land suitable for agro-industry (sugarcane, rubber, palm oil) or that contains valuable minerals or metals. Very large: each concession can be up to 10,000 hectares but subsidiaries of the ELC company may have multiple adjacent concessions. Formalized management although managers may claim ignorance of employees’ illegal logging beyond concession boundaries.  ELC companies often owned by members of the ruling elite or their foreign associates. Receive patronage from high-level political elite.  Table 4.1 Differences and similarities between actors engaged in deforestation and logging in Cambodian CBFM sites 92  4.3.2 The use of forest-related violence by different actors The data I collected show that forest-related violence is widespread across CBFM sites in Cambodia – CBFM participants have experienced some form of direct violence in thirty of the forty research sites, or 75%.49 In CFs and CPAs, direct violence most commonly occurs during forest patrols when CBFM participants confront illicit forest users and disrupt their logging or forest-clearing, or shortly thereafter in their home villages, which participants interpret as retaliation for the patrol’s disruption of illicit activities.50 No incidents of physical violence against non-patrol members were reported, although people gathering non-timber forest products or passing through the forest were verbally intimidated and threatened by organized illegal logging groups (see section 4.3.2.1).51 This indicates that direct violence in CFs and CPAs is experienced as a result of people’s participation in these projects and attempts to disrupt illegal logging and deforestation activities. In these sites, more men than women appear to have experienced forest-related violence but this is largely explained by their more frequent participation in patrols.  Amputees are specifically vulnerable to forest-related violence because they cannot run away as easily and have been specifically targeted for attack because they were                                                 49 Of the ten sites that have not experienced forest-related violence, two do not have contested resources so there is no forest-related conflict to initiate violence; all their valuable ‘luxury’ tree species were removed by logging concession companies in the 1990s and now they are completely surrounded by ELCs, which makes them undesirable for conversion to agricultural land because if they were cleared they would likely be incorporated by the ELCs before local people could formalize a land claim. When the ELCs were initially established, the NGOs were able to negotiate the exclusion of the CF land from the concessions. Despite interviews with the NGO involved and the CFMC leaders, it is unclear why these negotiations were successful in protecting the CFs while so many others around the country have not succeeded. However, Biddulph (2015) shows that ELC managers can sometimes be persuaded to make deals with NGOs. 50 Interview with CFMC leader, Oddar Meanchey, 10th June 2015. 51 Other non-CBFM-participant forest users are also impacted in some instances, which are flagged in the descriptions. One interview with a Ministry of Environment Ranger (Preah Vihear, 4th December 2015) indicates that government-employed rangers also experience forest-related violence but analysis of their experiences would require additional interview data.  93  unable to escape or fight back.52 In CLT sites, in which members do not undertake formal patrols, economic land concessions are the main cause of violence and intimidation. This occurs when CLT members confront the employees of concession companies that are encroaching on community-managed forests.   Forest-related violence also affects employees of the NGOs that support CBFM. Employees at thirteen of the eighteen NGOs surveyed (72%) have experienced forest-related violence. This is predominantly related to their work against ELCs and, to a much lesser extent, against organized logging groups. This data is summarized in Tables 4.2 and 4.3.  Non-CBFM actors involved in forest degradation and forest-related violence  Organized logging groups Local loggers Military Migrants Economic Land Concessions CBFM sites degraded by non-CBFM member (n=40) 11 31 10 17 16 CBFM sites experiencing violence from non-CBFM member (n=40) 10 12 14a 12 7b  Table 4.2 CBFM participants encountering and experiencing violence by each type of non-CBFM member a. More CBFM groups experienced violence by the military than there are CBFM sites degraded by the military because soldiers are sometimes hired by other illicit forest users to offer protection (see section 4.3.2.2). b. This includes two CLTs that experienced physical violence and 5 that experienced verbal intimidation.                                                  52 Interviews with CF members, Oddar Meanchey, 22nd and 30th May 2015. Amputations are relatively common in rural Cambodia (mines and unexploded ordnance [UXO] kill or injure two people a week [MAG, 2017]). UXO remain from bombardments associated with the Vietnam War and from the civil war and the Khmer Rouge period. During fieldwork, de-mining efforts were observed in two CFs. 94  Non-CBFM actors involved in forest degradation and forest-related violence Organized logging groups Local loggers Military Migrants Economic Land Concessions NGOs working in CBFM sites degraded by non-CBFM member (n=18) 13 18 2 11 17 NGOs experiencing violence from non-CBFM member (n=18) 3 0c 10d 0 11  Table 4.3 NGOs encountering and experiencing violence by each type of non-CBFM member c. This does not include the five NGOs that experienced violence while joining or observing community forest patrols as these incidents were not related to them being employees of NGOs but rather were targeted at the whole patrol group. d. More NGOs experienced violence by the military than there are NGOs working in CBFM sites degraded by the military because soldiers are sometimes hired by ELCs and organized logging groups to offer protection (see section 4.3.2.2).  The following section discusses the forms of forest-related violence that each of these actors use and to what extent and how NRM policies and institutions, and environmental, economic, and socio-political factors create the potential for and facilitate the use of such violence. These findings are summarized in Table 4.4.    95 Actors Deforestation type Violence Factors contributing to occurrence of violence Institutional and policy factors Socio-political, and economic factors Environmental factors Organized logging groups Selective logging In forest: Beatings, shootings, and chased and threatened with guns, chainsaws, and machetes. In villages: Incite police and military to use and threaten violence on behalf of organized logging groups. ELC(s) granted creating legal front for illegal logging beyond ELC boundaries. Lack of law enforcement on illegal logging. Clientelistic relationships with ruling elites and/or impunity from law.  Availability and accessibility of valuable trees. Military loggers Selective logging In forest: Beatings and pistol-whipping, shootings, chased and threatened with guns and chainsaws, held hostage. ‘National security’ discourse legitimizes dispossession of forest resources. Forest tenure rights not upheld by law enforcement authorities. Lack of law enforcement on illegal logging. Clientelistic relationships with ruling elites and/or impunity from law. Hired by organized logging groups and ELCs to provide protection. Availability and accessibility of valuable trees. Deforestation for agriculture In forest: Chased and threatened with guns. Lack of law enforcement on illegal deforestation. Clientelistic relationships with ruling elites and/or impunity from law Proximity to military bases and contested national border. Local loggers Selective logging In forest: Fist fights, threatened with machetes, axes, chainsaws, and homemade guns. In villages: Death threats, beatings, arson. ELC(s) granted and lack of secure land tenure over agricultural land and/or forests. Lack of law enforcement on illegal logging. Loss of livelihood and economic security due to ELC. Extent of clientelistic relationships with ruling elites/organized logging groups and/or impunity from law. Availability and accessibility of valuable trees.  Table 4.4 Forms of and factors contributing to forest-related violence in CBFM sites  Note. Continued on next page.   96 Actors Deforestation type Violence Factors contributing to occurrence of violence Institutional and policy factors Socio-political, and economic factors Environmental factors Local loggers Deforestation for agriculture In forest: Fist fights, threatened with machetes, axes, chainsaws, and homemade guns. In villages: Death threats, beatings. ELC(s) granted and lack of secure land tenure over agricultural land and/or forests Lack of law enforcement on illegal deforestation Loss of livelihood and economic security due to ELC. Lack of available agricultural land. Migrants Deforestation for agriculture In forest: Fist fights, threatened with machetes, axes, chainsaws, and homemade guns, held hostage. ELC(s) granted and lack of secure land tenure over agricultural land and/or forests. Lack of law enforcement on illegal deforestation. Loss of land and livelihoods to ELCs locally or elsewhere in country. High price of land elsewhere so sell land and look for ‘free’ alternative. Clientelistic relationships with ruling elites and/or impunity from law. Larger size of CF and thus more land ‘available’ to clear and easier to avoid detection. Economic Land Concessions (ELCs) Deforestation for agriculture Against CBFM members, in villages: Intimidation and threats to take land/forests without compensation. One community shot at when tried to meet company to negotiate. Two communities violently evicted from homes and land. Lack of secure land tenure.  Lack of accountability and transparency enables dispossession of land without free, prior, and informed consent. Clientelistic relationships with ruling elites and/or impunity from law. Patron-client relationship with organized illegal logging groups. None found Against NGO staff, in villages and towns: Intimidation and death threats, threats to hurt family. Loopholes in NRM laws and policies have been exploited and patrimonial relationships have been used to secure concessions.  Clientelistic relationships with ruling elites and/or impunity from law. NGO begins to publish press releases or reports on harmful ELC activities. None found  Table 4.4 (continued) Forms of and factors contributing to forest-related violence in CBFM sites   97 4.3.2.1 Organized logging groups and forest-related violence  Across Cambodia, direct violence occurs wherever CBFM patrol groups encounter and try to prevent logging by organized illegal logging groups. Violence against CBFM participants during forest patrols has been used in all of the eleven CBFM research sites in which organized logging groups have or continue to operate with only one exception – a CLT site where no one has tried to confront them. Employees at three NGOs also experienced such violence while joining or observing forest patrols.53 Violence has included beatings, chasing patrol groups with machetes, and threatening or shooting at them with guns, which are supplied by the wealthy leaders of organized logging groups. Other forest users (e.g. NTFP collectors) who disturb organized logging groups are also verbally threatened54 but CBFM patrol members are much more likely to have such experiences as they specifically seek out illicit forest users, confront them and attempt to confiscate logging equipment. Violence is usually reactionary and intended to scare away patrol groups, although it is planned to some extent in that organized logging groups are armed and ready to use violence to protect their equipment and the timber.55   Organized logging groups also pay police and military officers to use violence and threats on their behalf against CBFM members, Village Chiefs, and NGOs outside the forest. Tactics include setting up check points in forests to prevent patrol groups from entering and intimidating                                                 53 Neither NGO employees nor community-based participants believe NGO employees’ presence on a patrol makes violence against community-based patrol members any more or less likely.  54 In two CF research sites, organized logging groups have verbally threatened local people while they traversed the forest in search of NTFPs or to reach a distant village. Similarly, a local tour guide reported having to alter trekking routes in Virachey National Park as the previous routes are now too dangerous due to the presence of an armed logging group (Interview with community-based tour guide, Ratanakiri, 16th November 2015). This suggests that forest-related violence is negatively impacting the potential for local communities to use forests to support their livelihoods. 55 Common assessment made in interviews by CBFM participants in all provinces, May-December 2015.  98 patrol members in their villages, sometimes verbally threatening them, other times loitering silently near CFMC members’ houses as a reminder they are always being watched.56   Similarly, three NGOs working in areas targeted by organized logging groups reported frequent low-level intimidation by police and military officers. For example, staff at one NGO experience regular intimidation due to their prominence in anti-deforestation campaigns in two highly-contested forested landscapes where organized logging groups operate and the logging tycoon, Try Pheap, has been granted a license to collect and process the controversial ‘yellow vine’ plant (Coscinium fenestratum), which can be used in the production of narcotics (May & Ponniah, 2013).57 Every time staff members visit partner villages to hold meetings or training sessions, the police follow them, verbally abuse them, and threaten them with violence. The police have broken up the NGO’s community meetings on multiple occasions in the past year (2014-2015) and now demand that the NGO obtain written permission for any further meetings of more than five people, threatening them with imprisonment and beatings if they do not comply, but these requests for permission are always denied.58 This illustrates that organized logging groups contribute to violence and fear beyond the edges of the forest and use “the violence of law” (Springer, 2012, p. 5) in parallel with direct violence.   4.3.2.2 The military and forest-related violence  Soldiers from the Royal Cambodian Armed Forces (RCAF) are a significant cause of forest-related violence in Oddar Meanchey province, where they engage in illegal logging and                                                 56 Interview with CBFM members, Kampong Thom, 7th July 2015. 57 NGO interview, Phnom Penh, 8th May, 2015. See May and Ponniah (2013) and Otis (2014) for links between organized logging groups (especially Try Pheap) and the drug trade. 58 NGO interview, Phnom Penh, 8th May, 2015.  99 embolden local villagers to do the same (see section 4.4.2.3), and to some extent in the rest of Cambodia, where soldiers from local military units are hired by organized logging groups to protect their illegal activities.   The military’s involvement in forest-related conflict in Oddar Meanchey is influenced by the historical and geographical context of the province.59 Between 2008 and 2011, military units were drafted into Oddar Meanchey and Preah Vihear provinces along the national border in response to a diplomatic spat with Thailand over ownership of the UNESCO site of the Preah Vihear temple complex (International Crisis Group, 2011). During this act of political showmanship by the Cambodian ‘strongman’ Prime Minister Hun Sen, RCAF established bases inside three of Oddar Meanchey’s CFs (see Figure 4.1), which remain occupied by soldiers and their families despite the easing of tensions with Thailand and withdrawal of some troops. Although the legality of these military bases is questionable,60 the military claims the Prime Minister personally mandated their construction in these locations to benefit ‘national security.’61 Besides clear-cutting up to ninety per cent of these three CFs, soldiers also engage in illegal                                                 59 The Cambodian-Thai border has been militarized since the 1970s and logging in the border region has been a major source of funds for various militarized groups, including the Khmer Rouge until the late 1990s (Le Billon, 2002). 60 The legality of the military bases in the CFs is questionable. The FA does have the authority to terminate a CF Agreement in accordance with the sub-decree on CF (2003) if there is “an understanding of the Royal Government of Cambodia that there is another purpose which provides a higher social and public benefit to the Kingdom of Cambodia” (Article 28). However, in such an event, the FA Cantonment “shall give the Community Forestry Management Committee a written notice” six months prior to termination, certifying the reasons of termination (Article 28). Within this period, the Forestry Administration Cantonment Chief “shall discuss and negotiate with the Community Forestry Management Committee to determine the fair policy for the Community’s loss” (Article 28). The FA has neither formally terminated the CFs nor provided compensation for the communities’ losses. Furthermore, it has not shown that there was no alternative or better place to build military bases such as degraded land or that which was not already under protection. By remaining silent on the issue of termination, the FA is shirking responsibility and accountability for failing to support sustainable forest management in the Oddar Meanchey REDD+ project.  100 selective logging in eight other CFs in the province.    Figure 4.1 CFs in Oddar Meanchey showing those with military bases and/or controlled by the military. Note. Image based on a map from OpenDevelopment Cambodia (2017).   In the three CFs with military bases, soldiers have shot at and threatened to arrest or kill local villagers who tried to farm the cleared CF land or collect firewood or NTFPs in the remaining forest. Such violence was frequent in the first two years of the military’s occupation as villagers tried to challenge the loss of the forest but since they have given up challenging the military’s right to occupy and clear the forests, conflict and related violence has decreased in the area immediately surrounding the military bases. However, the military continues to engage in illegal selective logging during which some of the most extreme violence against Oddar Meanchey’s Community Forestry participants has occurred. This includes pistol-whipping patrol members, holding patrol groups hostage for twenty-four hours, and shooting the Head of the Provincial CF                                                                                                                                                        61 Interview with Head of CF Network, Oddar Meanchey, 24th May 2015; interview with Head of CFMC, Oddar Meanchey, 12th June 2015.  101 Network in the back of the neck and shoulder as he walked away from a heated argument over logging, which he believes was intended to be fatal.62 Due to the prevalence of this violence, two CFs are practically inaccessible to community members, in addition to the three in which military bases have been established (see Figure 4.1).  Cambodia’s neopatrimonial political system directly structures forest-related violence by the military. The Civil Servants’ Law (1994) declares that state officials and members of the military cannot be prosecuted for any crime unless the court first gains the permission of their superior in the civil service or armed forces, which, Hughes (2003) notes, is almost never. This is compounded by the fact that their superiors also financially benefit from their illegal logging.63 Consequently, soldiers can log and use violence with impunity. CF members know that these neopatrimonial relations protect soldiers and they are unlikely to be punished if they abuse civilians and thus the guns that soldiers wield are even more feared and more keenly avoided than those of other illegal loggers, to the extent that they will abandon a CF.    Yet, the extent and limitations of the military’s power to log and use violence with impunity are exposed by comparing the situation of Oddar Meanchey to that of Preah Vihear. Given that the military is a significant cause of forest-related violence in Oddar Meanchey province, it was expected that similar patterns of violence would occur in the neighbouring province of Preah Vihear, which has similar forest resources and a similar political context (a militarized national border and conflict over the temple). However, interviews with CF groups and local NGOs                                                 62 Interviews with CF members, CFMCs, and Head of CF Network, Oddar Meanchey, 24th May, 14th June, 15th June, 10th December 2015.  63 Interview with CFMC, Oddar Meanchey, 10th December 2015. Military Officers are said to take a cut of the logging proceeds in exchange for ignoring absence from official duties.  102 indicate that significantly less military-related violence occurs in Preah Vihear’s forests than in Oddar Meanchey. In Preah Vihear, soldiers do participate in illegal logging and are occasionally implicated in forest-related violence in Preah Vihear.64 However, they often work with or for organized logging groups rather than monopolizing control of access to forests as in Oddar Meanchey. This nuances claims by provincial politicians that they cannot control illegal logging by soldiers (Peter & Aun, 2016). In Oddar Meanchey, it appears that the highest political elites, including Prime Minister Hun Sen, allow the military to control the forests and thus it is unlikely that provincial politicians can do much to alter this power dynamic. Yet in Preah Vihear, when another actor is in the political or economic position to monopolize access to timber, illegal logging by the military was curtailed. This indicates that the ability to engage in illegal logging and use forest-related violence with impunity is determined by local political and personal alliances and Cambodia’s “interlocking pyramids of power” (Heder & Ledgerwood, 1995, p. 5) rather than the military automatically topping the socio-political hierarchy. Power is largely linked to the individual and their neopatrimonial relationships, rather than to their official position.  4.3.2.3 Local loggers and forest-related violence  Financial gain motivates opportunistic, sporadic illegal logging by local people across Cambodia, a problem that CBFM projects such as CF are explicitly intended to address (Clarke, Reed & Shrestha, 1993; Sunderlin, 2006). In addition to chronic financial poverty, household-level economic stresses, such as unexpected medical bills or house repairs after floods, can                                                 64 Evidenced by a soldier shooting dead an FA officer and a policeman during a forest patrol in November 2015 (Cuddy & Phak, 2015)  103 motivate opportunistic logging if valuable trees are available.65 This can be exacerbated by NRM policies that undermine local livelihoods and economic security, such as the establishment of ELCs that dispossess local people of agricultural land.66 This is exacerbated when ELCs spawn organized logging groups that are willing buyers of illegally cut timber (see case study in section 4.4 and Global Witness, 2007).  Furthermore, the spatial and temporal patterns of logging by local people are strongly influenced by the presence of organized logging groups in nearby forests. Although they often violently enforce exclusionary monopolies over logging territories for a particular time period, they increase economic incentives for local people to log by providing a guaranteed market for timber and reducing the legal risk and financial cost of smuggling timber to market by sending buyers to villages.67 Moreover, that ‘outsiders’ are clearly profiting from local forests also creates a feeling of frustration and perception of injustice that CBFM rules prevents local villagers from such immediate financial benefits while not adequately protecting their long-term sustainability. This contributes to a sense of urgency to benefit from valuable forest resources before ‘outsiders’ do in a ‘tragedy of the commons’ (Hardin, 1968) scenario.68 Such illegal logging by local people often continues for some time after organized logging groups move on as local loggers grow accustomed to the increased cash flow and are often willing to spend the time and energy to extract harder-to-reach or lower-grade specimens of valuable species or digging up roots of                                                 65 Interview with CF members, some of whom admit to having previously illegally logged in other CBFM sites, Kampong Thom, 6th July 2015; Oddar Meanchey 10th December 2015. 66 Interview with CF members, Kampong Thom, 6th July 2015. 67 Interview with illegal logger, Oddar Meanchey, 29th May 2015. 68 Interview with illegal logger, Oddar Meanchey, 29th May 2015; interview with CFMC, Kampong Thom, 11th July 2015.  104 felled trees (Figure 4.2).69   Figure 4.2 Roots of a valuable ‘beng’ tree (Afzelia xylocarpa) outside a house in a village establishing a CLT, Ratanakiri  Although illegal logging by local people occurs to some extent in all CBFM sites surveyed where valuable trees remain, violence does not occur in all encounters between local loggers and CBFM groups.70 Table 4.1 shows that thirty-one CBFM sites are degraded by local loggers but they have only used violence against CBFM participants in twelve of those. In most incidents, violence resembles that perpetrated by ‘outsider’ organized logging gangs: if patrols meet local loggers in the forest, scuffles over timber and equipment may occur and machetes and axes are often used as weapons. In some cases, loggers carry homemade guns, which they use the scare                                                 69 Interview with CLT member and former CPA member, Ratanakiri, 28th November 2015. Observations of roots and small pieces of timber awaiting collection by ‘middle-men’, Ratanakiri, 18th, 25th, and 28th November 2015. 70 Interviews with leaders of 30 CFs, 2 CPAs, and 8 CLTs, May – December 2015.  105 patrol groups away so the can continue their activities. However, in six of those sites, local loggers have used and threatened direct violence against patrol members outside the forest in what appear to be planned, retributive incidents. This includes intimidation and threats (including towards children of CBFM participants), beatings while walking down the road through the center of the village, and at least three known incidents of arson, the latest in June 2015. This raises the question, why does violence occur in these sites and not in others?   In part, this is explained by the fact that the eight CLTs in Ratanakiri are not formally patrolled so even though local people do sometimes engage in illegal logging, there is less opportunity for violence to occur due to the way CBFM is designed and implemented. In the other eleven CF and CPA sites where intra-community violence has not occurred, CBFM participants suggest local loggers do not want to harm their neighbours or “have problems living” in the village community.71 They believe local loggers remain dependent on traditional social networks within the community (which, to some extent, resemble patron-client relationships [Chan & Chheang, 2008]). Accordingly, local loggers avoid facing patrol groups in the forest by, for example, postponing a logging trip if village gossip alerts them to a planned patrol or abandoning an illegal logging site before the patrol arrives.72 If caught unawares, they are often willing to quietly sign agreements not to log anymore and then leave. Although patrol groups know these agreements will rarely be upheld in practice, they consider this façade of compliance and their own acceptance of the promise to be understandable behaviour given the necessity of maintaining social harmony. It is the occurrence of intra-community violence in other CBFM sites that can be considered the ‘social deviance’ in need of explanation through analysis of the                                                 71 Interview with CFMC, Preah Vihear, 6th December 2015. 72 Forest patrol with CF group, Oddar Meanchey, 24-25th May 2015.  106 local socio-political context in which violence occurs and, the data indicates, specifically to the connections of local loggers to politically powerful patrons.   In all eleven sites where such violence has occurred, interviewees believe that local loggers have established strong patron-client relationships with organized logging groups or the military, which emboldens them to use violence. As one interviewee bluntly explained, local loggers are often “not scared of the CF group because they have a patron (anh-konaung).”73, 74 Specifically, local loggers use their knowledge of the forest to identify hard-to-find or hard-to-reach, valuable timber, which they sell directly to these patrons, who operate nearby. In return, patrons pay local loggers good prices and, in some cases, collect timber directly from their houses thereby reducing the cost of transport and risk of being fined or (more likely) bribed by the police. They also protect local loggers from formal punishment by police or local political leaders, either by ensnaring local political leaders and law enforcement authorities in their patronage network or threatening less compliant political leaders with violence.75 Thus, by participating in these patrimonial networks, local loggers gain economic security (at least in the short term), and, when their neighbours and local political leaders participate (willingly or unwillingly) in the same patronage networks, they strengthen their social and political security in the village. Local political leaders and law enforcement officers are unwilling or unable to enforce the law or ‘social sanctions’ (Bardhan & Dayton-Johnson, 2001) against illegal logging and local loggers                                                 73 Interview, Deputy Head of a CF, Oddar Meanchey, 30th May 2015. 74 The Khmer term anh-konaung is a composite of anh, meaning patron or someone who can provide protection, and konaung, the anatomical term for one’s back. Together, the words mean ‘someone who can protect your back.’ 75 In 2 CF sites where local loggers use violence, Village Chiefs have objected to local loggers’ practices but were, in turn, threatened with violence by outsiders, thought to be employees of organized logging groups. Interview with Village Chief/CFMC leader, Oddar Meanchey, 30th May 2015, and interview with Village Chief/CFMC leader, Preah Vihear, 6th December 2015.  107 are emboldened to use more violent measures to facilitate illegal logging because they do not expect to be punished for those crimes either and more motivated because they know timber will be easily sold at a good price. This changes the balance of risk and reward for illegal logging and the use of violence by which illegal logging is facilitated. This indicates that local socio-political systems are an influential factor in whether or not forest-related conflicts within communities will become violent. These patrimonial relationships are represented in Figure 4.3.   Figure 4.3 Representation of local loggers’ patrimonial relationships  4.3.2.4 Migrants, deforestation, and forest-related violence  Migration into community-managed forests is a growing cause of conflict around which violence occurs between patrol members and migrants. Seventeen of the forty CBFM sites surveyed  108 reported migrants attempting to clear and claim land within the last three years and participants in fourteen of these sites experienced direct violence when confronting them (Table 4.5).76   Province CBFM sites surveyed CBFM sites targeted by migrants Experienced violence by migrants Kampong Thom  12 (12 CFs) 7 CFs 6 CFs Oddar Meanchey 13 (13 CFs) 7 CFs (2 targeted by military migrants) 7 CFs Preah Vihear 7 (5 CFs, 2 CPAs) 2 CFs 1 CF Ratanakiri 8 (8 CLTs) 1 CLT 0 Total 40 17 14  Table 4.5 CBFM sites targeted by migrants  Migrants are often assumed to be landless as a result of population growth and migrating in search of economic opportunities (IFAD, 2014; Pasgaard & Chea, 2013). However, my research suggests that this assumption does not account for the range of factors driving forest-related conflicts and violence. Landlessness and poverty are exacerbated by NRM policies that are implemented elsewhere in the country, specifically, ELCs and hydropower dams that dispossess people of land and livelihoods (Baird & Fox, 2015; Jiao et al., 2015; Scheidel, Giampietro & Ramos-Martin, 2013). Furthermore, not all migrants are landless; many migrants sell their land in provinces where demand and prices are high in order to fund their migration to cheaper land or land that can be claimed for only the cost of their own labour, as in CFs (c.f. Beban & Gorman, 2017).77 This also reflects a widespread lack of understanding of Cambodia’s Land                                                 76 Migration is of particular concern to CF communities because migrants tend to clear-cut forests for agricultural land, which causes more damage to forest ecosystems than logging individual trees. All NTFPs in the area are destroyed leaving little of value for local people. The damage is exacerbated by the use of fire, which can quickly spread beyond the migrants’ target area. Two CFs reported that migrants have cleared over thirty percent of the forest. Four additional CFs have at least one settled household inside the CF, increasing the likelihood that others will follow or the household will expand despite CF by-laws. 77 Interviews with CFMC leaders who hear this directly from the migrants, Oddar Meanchey, 25th May 2015; Kampong Thom, 8th July 2015. See also Beban and Gorman (2015).  109 Law (2001) leading to the belief that “if you clear it, you can have it” (Work & Beban, 2016, p. 37).78   Political and economic elites also facilitate and encourage migration. They hire migrant workers to clear-cut forested land in order to profit from the illegally cut timber, which is sold to sawmills or exported, and then claim land for either immediate agricultural development (usually for fast growing cash crops such as cashews and cassava) or to be held speculatively. Although the Land Law prohibits this practice, political and economic elites can use their personal networks to ensure laws are only selectively applied, that is, not applied to them. The hired workers commonly retain a few hectares of the newly cleared land as payment on which they settle permanently or seasonally. On one observed patrol, the migrants encountered in the CF admitted to being paid and promised land by a wealthy businessman, but they claim to have never met him nor know his real name.79 In these instances, patrons remove economic and practical barriers to migration by paying for transport to target forests and providing equipment to clear it (sometimes including large machinery such as diggers and bulldozers). These patrons also identify suitable forests, usually those that are large enough for migrants to go undetected for a while, which helps to explain how migrants from hundreds of kilometers away know which forested land to target. This socio-political system of patronage makes migration more possible and increases forest-related conflicts.                                                   78 Migrants met during forest patrols in Oddar Meanchey, 24-25th May 2015, and Kampong Thom, 10th July 2015, argued that ‘if you clear it, you can have it’ was supported by Cambodian law. Whether they actually believed this is true or if they were hoping to undermine the patrol group’s confidence in their right to evict the migrants is unclear. In any case, it demonstrates that it is important for patrol groups to be sure what their rights are so they are not misled by common misconceptions. 79 Participant observation of patrol, Oddar Meanchey, 25th May 2015.  110 Patronage also shapes the violence that emerges from these conflicts. Incidents involving migrants included fist-fights, threats of armed violence, and one hostage incident. Migrants carry machetes, axes, and chainsaws, which can be used as weapons even if they are primarily intended as tools for forest clearance. All incidents occurred when patrol groups tried to arrest migrants or force them to leave the forest. Although the first one or two encounters between migrants and patrol groups are usually violence-free (in part because patrol groups are scared by previous experiences of violence so do not insist too vehemently), migrants often return to the forest with more weapons to prevent patrol groups evicting them in future encounters. In at least three instances, migrants returned armed with AK-47 rifles, a weapon that is too expensive for most Cambodians to own, especially ‘landless migrants,’ leading CBFM participants to speculate that it could only have been supplied by a patron from the economic or political elite.80 Migrants also often outnumber patrol groups and use this to their advantage in leveraging intimidation tactics such as threats of violence.   Patrol participants believe that migrants are willing to use violence because they expect patrons would help them avoid punishment from law enforcement authorities.81 Additionally, migrants are not restrained by the same social conventions that local loggers are – the social consequences of using violence are less since they have no prior relationship with or dependence on patrol members. Thus, the potential personal benefits of using violence are high while the risks are low.                                                  80 Interview with CF patrol group, Oddar Meanchey, 24th May 2015. 81 Interview with CF patrol group, Oddar Meanchey, 24th May 2015, and interview with CFMC, Kampong Thom, 10th July 2015.  111 4.3.2.5 Economic land concessions and forest-related violence  One of the most publicized forms of forest-related violence in Cambodia is that perpetrated by Economic Land Concessions (ELCs) (e.g. ADHOC, 2015; Forest Trends, 2015; Phak, 2014). Photographs of violent evictions, burning of homes, and beatings and shootings of protesters have been published around the world (LICADHO, 2010; Loy, 2012). These have been directly attributed to poor NRM policies that prioritize the interests of ruling elites over other Cambodian citizens and distort legal-rational mechanisms intended to protect natural resources (symptoms of neopatrimonialism) (Pak et al., 2007). However, the data from this research project suggest that the relationship between the presence of ELCs and forest-related violence is much more complex than has thus far been portrayed in academic and grey literature.   Direct violence does often occur during the establishment of ELCs and protests against the loss of residential, agricultural, and forested land (see Table 4.6), but interviewees emphasized that protests were primarily against the loss of agricultural land, which is more important for livelihoods than forests.82 This differentiates these incidents from other forms of forest-related violence described in this chapter that are more directly related to CBFM activities. Yet, such violence is not commonly used against CBFM participants in the course of regular CBFM activities, such as patrolling forests, even when these border ELCs.83 In one exceptional case in which the CF is surrounded by an ELC, interviewees proposed that the presence of the company                                                 82 Although the dispossession of forests entails the loss of natural resources that are important for cultural practices and supplementing livelihoods, interviewees uniformly emphasized the dispossession of residential and agricultural land as being the greater harm to their livelihoods. Even when forested land was lost in Ratanakiri, the forested land was considered “stand-by land” for future generations to clear and farm. It is these losses that protests were primarily intended to prevent. 83 Although this raises the question of whether or not protests are part of ‘regular’ CBFM activities given that they are intended to protect forests, I chose to exclude these from my analysis because I believe they are better compared to other forms of political protest occurring around Cambodia.  112 decreases the incidence of illicit forest use in CF thereby reducing the potential for associated violence to occur. This is because the only way to enter the CF is a single road through the concession plantation, which is constantly guarded by company employees, some of whom are CF members. One interviewee likened the constant vigilance of the CF to “pineapple eyes,” which both look out at the villagers but also in at the CF.84 This finding contradicted my expectations derived from media reports of violence by ELCs and demonstrates that the extent and forms of forest-related violence in Cambodia cannot be accurately judged from media reports.   Type of CBFM project CBFM site bordering or overlapping ELC CBFM groups reporting ELC-related violence Notes on violence Community Forest (n=30) 9 2 Participants in 1 CF were violently evicted from their homes and agricultural land and lost half the CF to the ELC. Later experienced beatings by military guards when trying to reclaim lost agricultural land (see case study in section 2.3). Participants in 1 CF have been beaten by ELC employees and police on numerous occasions when protesting loss of agricultural land to a sugar-growing ELC. The CF members stole two bulldozers from the ELC company and have threatened ELC employees who have tried to reclaim them. Community Protected Area (n=2) 1 0 N/A Communal Land Title (n=8) 7 5 Participants in 4 CLT projects were threatened with violence or jail if they did not agree to let ELC take their forest and agricultural land. Participants in 1 CLT were shot at by police when they went to negotiate with the ELC company. Total 17 7   Table 4.6 Violence related to ELCs used against CBFM participants                                                  84 Interview with head of a CFMC, Kampong Thom, 2nd July 2015.  113 In contrast, ELCs are commonly implicated in violence against the staff of NGOs investigating their activities or trying to use legal mechanisms to protect communities’ land and forests. Of the eighteen NGOs surveyed, seventeen work with CBFM sites affected by ELCs and employees at eleven of those have experienced violence or received threats of violence from employees or representatives of ELCs. In all cases, violence occurred after the NGO initiated investigations into or filed formal complaints against concession companies and was often experienced with increasing frequency and intensity as the NGOs’ investigations posed a greater threat to ELCs by exposing corrupt and harmful social and environmental practices (c.f. Global Witness, 2017). For example, one NGO involved in the campaign against Hoang Anh Gia Lai (better known as HAGL) and the Vietnam Rubber Group received almost daily phone calls threatening to kill staff when it began working with international NGOs to involve the Compliance Advisor Ombudsman of the International Finance Corporation. “Gangsters” on motorbikes also intimidated staff members by following them to their homes and when they went out to drink with friends.85 In another case, during the night following the release of a report into land grabbing, a machete was stuck up through the floor of an NGO employee’s house where he would likely be sleeping, only missing his body by a few centimeters.86 This suggests that ELC companies view formal complaints and legal proceedings as real threats to their operations and are willing to use violence to try to discipline NGOs’ behaviour. Furthermore, that violence and intimidation were targeted at NGO staff rather than community members suggests that the beneficiaries of ELCs recognize NGOs as central to the momentum of communities’ legal struggles against ELCs and see community members as a relatively harmless nuisance (at least until they collaborate with an                                                 85 NGO interview, Ratanakiri, 21st November, 2015. 86 NGO survey data, phone conversation, 29th May, 2015.  114 NGO). This can usefully be compared to organized logging groups who primarily target local villagers rather than NGOs.   ELCs also indirectly increase the use of violence against CBFM participants by facilitating the illicit activities of organized logging groups and local loggers. They provide a legal front for the operations of organized logging groups, which log in the forests surrounding ELCs and get permits to export the timber claiming it was legally cut inside ELC boundaries (NGO Forum on Cambodia, 2015; Peter, 2015). Furthermore, the economic and food insecurity resulting from dispossession by ELCs further motivates illegal logging by local people with the potential to increase intra-community violence.87   4.4 Case Study: Overlapping and interlocking moments of violence in Oddar Meanchey  The above sections have heuristically characterized various forms of physical violence and the contexts in which they occur, but the reality is much more complex. The following case study explores the overlapping and interlocking nature of moments of violence in and around two CFs in Oddar Meanchey province, Ratanak Ruka CF and the Monks’ CF. It traces how different actors used violence against and later within these two communities during the establishment and eventual cancellation of three Economic Land Concessions (ELCs). By presenting this case in all its complexity, I demonstrate that, as Springer (2009, p. 89) notes, “violence begets violence” and a violent act is never discreet (Robben & Nordstrom, 1995) but is structured by and has implications for environmental and socio-political contexts. This case study also provides insights into the agency, subjectivity, and beliefs of those using violence against their                                                 87 Interview with CFMC, Oddar Meanchey, 15th June 2015, and interview with illegal logger, Oddar Meanchey, 11th June 2015.  115 neighbours. While it can offer only a glimpse of these, this is invaluable in exploring the complex nature of violence and how external factors can ‘override’ the social capital upon which CBFM is often said to depend (e.g. Pretty and Ward, 2001). The relationships between the actors in this case study are summarized in Figure 4.4.   Figure 4.4 Social, political, and economic connections between actors in Ratanak Ruka CF  In 2008, the Ministry of Agriculture, Forestry, and Fisheries (MAFF) granted three ELCs to subsidiaries of one Thai company, Mitr Phol (EC/IDI, 2013). The ELCs reduced the size of  116 Ratanak Ruka CF by more than 50%, from 26,036 ha. to 12,872 ha. (Sherchan, 2015).88 The ELCs covered 7,941 ha. of forested land and 5,223 ha. of agricultural and residential land used by villagers who lived within the CF boundaries. When the concession was granted, local villagers and NGOs had been working on the establishment of the CF for eight years and the CF Management Plan had been submitted to the Provincial Governor for approval (step seven of the eight legal steps). The boundaries of the ELCs, Ratanak Ruka CF, and the neighbouring Monks’ CF (a.k.a. Sorng Rokavorn CF) are shown in Figure 4.5.                                                 88 Of the 19,736 ha. of concession land, 13,164 ha. was part of the proposed Ratanak Ruka CF, including 7,941 ha. of forested land and 5,223 ha. of agricultural and residential land used by villagers who lived within the CF boundaries (Sherchan, 2015). When CFs are established, households that already have agricultural land within proposed CF areas are usually allowed to continue farming it if they agree not to expand the borders. This is more practical than excluding small areas from the CF landscape (Bradley et al., 2013). Several participants in the focus group discussion (10th June, 2015) had farmed land inside the proposed CF since 1997 when they re-settled after being in refugee camps on the Thai border during the civil war and Khmer Rouge regime.  117   Figure 4.5 ELCs and community forests in central Oddar Meanchey province Note. Mitr Phol’s ELCs are Angkor Sugar (blue), Cane and Sugar Valley (purple), and Tonle Sugar (yellow). Ratanak Ruka CF is in the center, fragmented by the ELCs. The Monks’ CF adjoins it to the east. Copyright: Equitable Cambodia and Inclusive Development International (2013).  The establishment of the ELCs was marked by violent evictions. One village was completely destroyed and 214 families were forcibly evicted (Sherchan, 2015). Villagers’ homes were burnt to the ground and evicted families were beaten as they tried to save their belongings (Khoun & Tep, 2009; Tep & May, 2009).   These moments of ‘spectacular violence’ exemplify how physical violence manifests from the structural violence of accumulation by dispossession (EC/IDI, 2013; Springer, 2012). They also illustrate the direct influence of Cambodia’s ruling class in violence against economically and  118 politically marginalized citizens. Of note in this case is the association between Mitr Phol and Ly Yong Phat, a Cambodian People’s Party Senator and Special Economic Advisor to the Prime Minister. Ly Yong Phat is President of the LYP Group, which shares offices in Thailand with Mitr Phol (LYP Group 2017; Sherchan, 2015). His economic interest in the ELCs was indicated during the evictions as ELC employees were wearing t-shirts printed with “Donated by H.E. Ly Yong Path” [sic.] (Sherchan, 2015, p. 22). Evictions were also conducted by RCAF troops from Battalion 42, which is sponsored by Ly Yong Phat’s casino as part of a formal “partnership” (EC/IDI, 2013; O’Toole & Cheang, 2013, para. 1; Sherchan, 2015). This case illustrates how patronage and formal institutional mechanisms enable the use of violence with impunity while exploiting Cambodia’s forests. Yet, this was only the beginning of a series of moments of violence.    Figure 4.6 Houses being burnt in Oddar Meanchey during eviction in 2008 by Angkor Sugar, a subsidiary of Mitr Phol. Photo copyright: Phnom Penh Post, 2008.   119 Following the establishment of the ELCs, the forest was cleared but only 219 ha. of sugar cane was planted. Villagers who had lost farmland (their livelihoods) attempted to replant corn, rice, and cassava but were violently beaten and chased away by soldiers hired by Mitr Phol to guard the land.89   In May 2015, the Royal Government of Cambodia (RGC) cancelled the ELCs as part of a national review of concessions. The land was reclassified as ‘state private land’90 and, rather than being returned to the community, soldiers claimed to have “the rights” to it and continued to aggressively exclude local villagers.91 A local Village Chief supported this claim although no documentation was shown to villagers and they suspect that his complicity was bought.92 Soldiers work with the logging tycoon Try Pheap to operate a sawmill inherited from Mitr Pohl.93 Local villagers fear Try Pheap’s employees and the soldiers due to their reputation for violence against people who try to impede illegal logging operations.94 Thus, villagers have begun to seek alternative land in the neighbouring Monks’ CF.                                                  89 Focus group discussion, Oddar Meanchey, 10th June 2015. 90 ‘State private land’ “does not have public interest value” (OpenDevelopment Cambodia, 2015b, para. 1), in contrast to CFs, which are ‘state public land.’ 91 Interview with former Ratanak Ruka CF member, Oddar Meanchey, 10th June 2015. 92 Interview with former Ratanak Ruka CF member, Oddar Meanchey, 10th June 2015. The Village Chief declined to be interviewed. 93 During fieldwork, no official confirmation of who owns the sawmill could be established. Neither the research assistants nor I felt able to approach the sawmill without risk to ourselves or increasing the risk that local villagers would be harassed. Few villagers venture near the sawmill because they are intimidated by the heavy military presence guarding it. In May, June, and December 2015, I observed wagons carrying logs and bearing the logo of Try Pheap arriving from the east (in the direction of Preah Vihear and Ratanakiri provinces) and leaving on the road north towards the Thai border. 94 These fears appear to be widespread and warranted. For example, in Ratanakiri, interviewees reported that Try Pheap’s employees had threatened to kill them if they reported his illegal sawmill to the provincial authorities (interview with CLT member, Ratanakiri, November 2015).  120 The Monks’ CF is only separated from the former Ratanak Ruka CF by a single-track road. Since the ELCs were established, incidents of illegal logging and deforestation for agriculture in the Monks’ CF have increased each year. Related violence by loggers and land-seekers against patrol members, including former Ratanak Ruka CF members, has also escalated in terms of frequency and severity, occurring both in the forest at the moment of confrontation and, increasingly, afterwards as premeditated and deliberate attacks.95 Three examples illustrate the violence they face: • In late 2014, approximately 60 villagers (who were previously members of Ratanak Ruka CF) went together to the Monks’ CF to clear land for agriculture. The patrol group had heard of their plan but as they approached many of the logging group rushed at them with axes and machetes. The patrol group immediately retreated on their motorbikes so avoided injuries.  • In June 2015, the house of a CFMC leader was burnt to the ground (Figure 4.7), which he believes was retribution for confiscating homemade guns and chainsaws from other villagers who were illegally hunting and logging in the CF.  • CFMC leaders are regularly verbally abused in villages by neighbours who are involved in illegal logging. They are called derogatory names and warned that they will be killed in their sleep. Their children are bullied at school by their classmates, who repeat what their parents say at home.                                                   95 Interview with Head Monk, Monks’ CF, Oddar Meanchey, 15th June 2015.  121  Figure 4.7 CFMC leader (formerly Ratanak Ruka CF, now Monks’ CF) in front of the smoldering remains of his house. Monks’ CF on the horizon. Copyright: Bruno Carette.  Such incidents are partly the result of economic desperation driving conflict and leading former Ratanak Ruka members to desperate measures. The demand for construction timber, luxury wood, and agricultural land, has significantly increased.96 There are now approximately double the number of villagers trying to use the Monks’ CF and demand for construction materials was exacerbated by the evictions so all stockpiles of previously confiscated illegally-cut timber have been depleted.97 Concurrently, the loss of agricultural livelihoods increased economic strains on households and motivates illegal logging of ‘luxury’ timber, such as Siamese rosewood, in the                                                 96 For NTFPs that are plentiful, available across the entire forest, and can be substituted if the desired species is unavailable, such as mushrooms, fruit, bamboo shoots, and medicinal plants, the increased use by Ratanak Ruka villagers is relatively unproblematic as there is neither much competition for these resources nor rules restricting their usage. Competition for these resources is further reduced by the distribution of villages around the CF; Ratanak Ruka villagers collect NTFPs closest to their homes on the west of the Monks’ CF while long-time users of the Monks’ CF usually access the forest via the north and east sides closest to their homes.  97 Interview with Head Monk, Monks’ CF, Oddar Meanchey, 15th June 2015.  122 Monks’ CF. The nearby sawmill in the ELC has proven an easily accessible and willing buyer with no questions asked about the origins of the wood.98   However, demand in itself does not necessarily cause forest-related violence and deterministic assumptions obscure and ignore the agency, motivations, and subjectivities of those who use violence. Besides the material and economic drivers of demand for land and timber, Ratanak Ruka villagers have lost belief in the concept of CF and sustainable forest management. Data from four former Ratanak Ruka CF members who agreed to be interviewed suggest that negative experiences with the ELC have made people less willing to forego potential benefits from the forest. They expressed their frustrations with the Head Monk, CFMC, CF system, and NGOs that “promised” to help them protect their forest.99 They accuse the CFMC of being “selfish” by limiting logging as timber is perceived to be “plentiful” in the CF. 100 A particularly strident anger is directed towards the Head Monk, who, interviewees claim, failed to use his position as a politically influential public figure to intervene and protect their natural resources and livelihoods but, instead, only sought to protect “his” forest. One interviewee remarked, “he loves the forest cows more than he loves the people,”101 indicating a belief that he has failed them as political patron – a position religious leaders are expected to hold in Cambodia (Davis, 2015).                                                  98 Interview with former member of Ratanak Ruka CF who now engages in illegal logging, Oddar Meanchey, June 2015. 99 Interview with former Ratanak Ruka CF member, Oddar Meanchey, 12th June 2015. 100 Interview with former Ratanak Ruka CF member, Oddar Meanchey, 12th June 2015. 101 Interview with former Ratanak Ruka CF member, Oddar Meanchey, 12th June 2015. The interviewee is referring to the herd of endangered banteng (Bos javanicus) in the Monks’ CF.  123 Interviewees expressed belief that the losses they suffered in their own CF, despite their “giving” time and effort to protect it, make them “entitled” to forest resources in the Monks’ CF.102 One interviewee argued that the Monks’ CF “belongs to us, the people of [this] village” because, he reasoned, he and his neighbours used that forest before the CF was established and only agreed to stop using it so that NGOs and the Head Monk would help them establish a CF with formalized management rights over the Ratanak Ruka forest.103 The Head Monk, he continued, was only “taking care of it for the people,” to whom it must now be returned. Consequently, his perceived failure to support them despite their prior conservation efforts has led to their loss of respect for him as an authority figure and their disregard for the restrictions he is seen as dogmatically imposing on forest use.104 These responses indicate that CF participants expect that the formal institutions of CF will accommodate Cambodian norms of patronage and feel betrayed when they do not. The Head Monk refutes these criticisms and argues that he tried to stop the ELCs being granted but was told by the Provincial Governor that he can only protect the Monks’ CF and not all the province. However, he admits to, and takes pride in, imposing stricter limitations on the use and extraction of forest resources than other CFs in the province.105   Local trust in the Community Forestry system and the rights it confers has also been diminished by the FA’s perceived complicity in granting the ELC on land that was in the process of being registered as a CF. Interviewees argued that the CF system is doomed to fail as powerful business people, politicians, and the military are expected to forcibly appropriate the remaining                                                 102 Interview with former Ratanak Ruka CF member, Oddar Meanchey, 12th June 2015. 103 Interview with former Ratanak Ruka CF member, Oddar Meanchey, 13th June 2015. 104 Interview with former Ratanak Ruka CF member, Oddar Meanchey, 13th June 2015. 105 Interview with Head Monk, Monks’ CF, Oddar Meanchey, 15th June 2015.  124 CF in the future. As one interviewee argued, “If we do not claim that land now, someone else will take it soon and we will be left with nothing. It is meant to be for the community. What right do [the CFMC] have to stop us using it?”106  These feelings of insecurity and lack of trust in state authorities are compounded by the on-going construction of a hydroelectric dam and mine on another ELC adjoining the south of the Mitr Phol ELCs as a partnership between the RGC and a private Chinese company, Sinohydro, financed by a loan from the Chinese Government (ActionAid, 2016). Like the previous ELCs, the Sinohydro concession was granted without free, prior, and informed consent of impacted communities. At the time of research, there was much confusion and concern among interviewees regarding how many families may be displaced, how much agricultural and forested land would be flooded, and when.107 Two community members were under arrest for ‘inciting protests’ and the Provincial Governor had warned them not to get NGOs involved or they would not get any compensation for flooded property. Interviewees, including those who support CF and oppose it, feel that they are being targeted from all directions literally and figuratively: the villages, which lie along one road, are surrounded by ELCs to the north, south, and west, and the Monks’ CF to the east, and their economic, political, and cultural lives are being destabilized and destroyed.108                                                   106  Interview with former Ratanak Ruka CF member, Oddar Meanchey, 12th June 2015. 107 It appears their concerns were warranted; the dam will displace approximately 158 households and flood 3000 ha. of forested land in the Monks’ CF (ActionAid, 2016; Phak, 2016). There is no published estimate of how much agricultural land or land formerly in Ratanak Ruka CF will be flooded but interviewees believe that land around their villages will be inundated. 108 Focus group discussion, Oddar Meanchey, 10th June 2015.  125 A lack of trust in the political system and the politicians, who are meant to defend local people’s interests, is reducing some community members’ willingness to contribute to community-based forest management. Although this situation and the emotions it causes are largely a product of the structural violence of the neopatrimonial socio-political regime that allowed the ELC to destroy forests, homes, and livelihoods, the CFMC is an obvious scapegoat for community members’ current problems and difficulties in restructuring and securing their livelihoods and well-being. The CFMC leaders are seen as politically weak and lacking authority to address the systemic causes of forest loss and the injustices experienced so they cannot help villagers. Yet, they have just enough authority to block villagers’ access to immediate financial rewards from logging, land conversion, and wildlife hunting. The resulting anger, desperation, and frustration are widespread among the dispossessed villagers and increasingly manifest as forest-related violence within the community, specifically targeted at the CFMC and patrol groups. The emergence of such attitudes was catalyzed by the environmental, economic, and social disruption caused by the ELCs and compounded by the military’s hostile takeover of the land, both of which are themselves symptoms of Cambodia’s systemic inequalities and injustices. Understanding these perspectives helps to explain why former CF members are less willing to abide by CF bylaws and use violence when denied these forest resources.   4.5 Discussion This chapter has demonstrated that forest-related violence against CBFM participants and NGO employees is widespread in Cambodia. It is much more than the murders recorded in national headlines and counted in human rights watchdogs’ reports. Rather, it manifests as pistol whipping, death threats, and arson, or, more mundanely, as fist fights in the forest, harassment on  126 village roads, and intimidation while at home. These ‘incidents’ occur too frequently to count in many of the CBFM sites surveyed but each alone is not sufficiently ‘spectacular’ to be recorded or even noticed by anyone outside the communities where they occur. Yet, when taken together, they reveal important trends in the occurrence of forest-related violence in Cambodia, which, in turn, can help to explain the challenges CBFM participants and NGO staff face and how these shape CBFM practices.   This chapter illustrates how forest-related conflicts are structured by the properties of natural resources, NRM policies and institutions, and economic and socio-political factors. The case study of Ratanak Ruka CF also shows that individuals’ agency and behaviour is shaped by their perceptions of (in)justice, expectations that socio-political norms will be upheld, and their own previous experiences of forest-related violence. These findings support the theorizations of environmental violence explored in Chapter 1 (e.g. Peluso & Watts, 2001). However, the tendency for these conflicts to escalate into violence is most strongly shaped by the identity of the actors and their position in Cambodia’s patrimonial networks. Military loggers and organized logging groups are quick to use extreme forms of violence as they know their affiliation with specific groups headed by powerful patrons from Cambodia’s ruling elite enables them to act with impunity. Meanwhile, ELCs recognize this impunity and thus often ‘outsource’ their violence to these same actors or the police. In contrast, local loggers only use violence if they have individually secured patrimonial support that that provides them with a greater source of socio-political and economic security than they would otherwise have in their communities. These patron-client relationships change the balance of risk and reward for an individual using violence.  127  The findings presented in this chapter also demonstrate that the identity of the actor also matters when discussing forest-related violence because different actors use violence in different ways and, to some extent, for different purposes. Loggers use physical violence and threats primarily against other forest users (CBFM participants and people collecting NTFPs) in order to claim territory and forest resources directly. In contrast, ELCs use (criminal) physical violence primarily against NGOs in order to disrupt investigations or the organization of local campaign activities and protect the concession contracts that then ‘legitimize’ the use of direct violence against local communities, such as evictions and breaking up protests.   Comparing against whom ELCs and organized logging groups use direct violence also provides some indications as to why it is used. Organized logging groups used direct violence against 3 of the 13 NGOs (23%) that have supported communities in opposing them. In comparison, ELCs used violence against 11 of 17 NGOs (65%) that were opposing them (see Table 4.2). This comparison indicates that organized logging groups see little need to stop NGOs’ investigations, likely because their leaders are so closely connected to political elites that they do not fear their operations will be disrupted by the legal system (Global Witness, 2007, 2016b; MacInnes, 2015). Their freedom to act with impunity, a consequence of their position in Cambodia’s neopatrimonial socio-political system, means there is little risk in them using violence against anyone who tries to disrupt their operations but much to be gained from doing so. In comparison, ELCs are under greater pressure to appear to comply with the law since their activities can be  128 disrupted by international legal proceedings (e.g. Peter, 2016; Phak & de Bourmont, 2017),109 which they are less able to influence through patronage or other extra-legal means.110 Thus, using violence against NGOs to prevent them from travelling to communities or to deter them from engaging in these issues is a means of disrupting legal challenges in their early stages before international regulatory bodies are involved. These findings illuminate the interconnectedness of the multiple forms of violence that Cambodian forest defenders and rural communities experience while also nuancing and countering the narratives that homogenize perpetrators of forest-related violence as “outsiders” (e.g. Yasmi et al., 2012, p. 99; The Oakland Institute, 2011, p. 33).  This chapter also highlighted the role of state actors in forest-related violence. Not only do soldiers use violence to facilitate their own illegal logging practices but they, along with police officers, use physical violence and intimidation to enable the ruling elites connected to organized logging groups and ELCs to remain nameless and disconnected from the outcomes of their crimes, or at least deny accountability if their connection is exposed. By coopting the legal-rational institutions that should offer citizens protection from harm, these ruling elites both benefit from Cambodia’s socio-political system of neopatrimonialism and perpetuate it by ensuring that legal-rational mechanisms are selectively applied. The inequalities and injustices this produces are a form of structural violence, defined as that which is “built into the [social]                                                 109 In an agreement mediated by the International Finance Corporation’s Compliance Advisor Ombudsman, a controversial Vietnamese rubber firm, Hoang Anh Gia Lai, which had been granted an economic land concession in Ratanakiri was forced to return land to eleven villages, restore streams filled or polluted by its activities, and repair roads and bridges (Phak & de Bourmont, 2017). The ruling is a result of the work by a coalition of national and international NGOs funding and conducting investigations and a legal campaign against the rubber company. 110 Interview with NGOs, Ratanakiri, 16th November and 1st December 2015.  129 structure and shows up as unequal power and consequently unequal life chances” (Galtung, 1969, p. 171).   These findings indicate that direct forest-related violence can be seen as both part of a process and an outcome; it is shaped by structural violence and a means through which other forms of violence are reproduced. Thus, the outcome of a particular act of violence (such as the injury or death of a CBFM participant) should not be seen only as an outcome of a singular act of violence but as the manifestation of a succession of violence processes. This raises an important question. Given that the factors that contribute to forest-related violence are so complex and deeply rooted in the Cambodian socio-political structure, which is in turn shaped by local histories and global political and economic forces, how can and should CBFM participants and the NGOs that have assumed the responsibility of supporting them respond to forest-related violence? This question is addressed in the following chapter.    130 Chapter 5: Unrooted responses: how CBFM participants and development organizations (fail to) address forest-related violence in Cambodia  5.1 Introduction When international development projects are proposed to donors, there is usually a requirement to consider risks involved with the project. This is not only intended to identify factors that could undermine project success but also to minimize the risk that participants will be harmed during implementation or as a result of the proposed project (Kwak & Dewan, 2001; World Bank, n.d.). At the very least, it is expected is that development organizations should aim to uphold the humanitarian principle of ‘do no harm,’ accepted as the bare minimum requirement for community-based conservation projects, or indeed any development intervention (Brockington & Schmidt-Soltau, 2004; Brown, Seymour & Peskett, 2008). Thus, from a normative, human rights perspective, forest-related violence is a risk that should be managed since it endangers the lives, personal safety, and emotional and psychological well-being of CBFM participants across Cambodia, many of whom participate in CBFM at the request or encouragement of NGOs and donors in order to generate environmental benefits for both local and global communities. From an economic perspective, forest-related violence is a risk that should be managed since it hinders the implementation of necessary CBFM activities and thus reduces the likelihood of achieving desired environmental outcomes. With this in mind, this chapter asks, how are CBFM participants and development organizations globally, and in Cambodia specifically, responding to and trying to reduce forest-related violence?   131 In this chapter, I first explore three main themes within common recommendations of how to respond to natural resource related violence – addressing root causes of resource conflict, improving the rule of law, and supporting human rights and political empowerment. Taken together, these three themes form a normative conceptual framework that can be used to coordinate responses and design a long-term strategy to reduce forest-related violence. Using this framework to analyse responses that have been recommended in the Cambodian context, I show that commentators have predominantly focused on improving the rule of law because systemic political corruption and patronage are seen as both the root cause of forest-related conflicts and factors that undermine sustainable forest management. I then explore the theoretical assumptions on which these recommendations are based and the practical challenges associated with implementing them and discuss how these weaknesses are shown in the few recommended responses in the Cambodian context. Since improving the rule of law is so central to recommended responses to forest-related violence in Cambodia and the responses development organizations are trying to implement, I then review the current state of the Cambodian law enforcement institutions to establish the legal and socio-political context in which responses are conducted in Cambodia. This contextualizes ethnographic data examining the lived experience of law enforcement in CBFM projects. I find that attempts to deter or punish forest-related violence using the rule of law are obstructed and neutralized by neopatrimonialism and associated corruption among law enforcement institutions. This is compounded by social norms that hinder access to the justice system if patrimonial networks have rendered local law enforcement officers ineffective. Furthermore, since CBFM participants have little trust in the justice system due to the way it is routinely distorted and disrupted by patronage and corruption, they are often hesitant to report forest-related violence. Alternative strategies to respond to forest-related  132 violence are not rooted in a comprehensive strategy to address violence and consequently are only effective in reducing violence because they remove CBFM participants from encounters with illicit forest users and allow forest degradation to go unchallenged, thereby negating the conservation objectives of CBFM. This leaves CBFM participants feeling that they have no option but to resign themselves to acceptance of illicit forest use, forest-related violence, and violation of their human, environmental, and political rights.   5.2 Literature review and conceptual framework: recommended responses to violence against environmental defenders As attention to violence against environmental defenders and activists has expanded over recent years, academic researchers and human rights watchdogs have provided recommendations of appropriate responses by those affected (environmental defenders), those responsible for protecting the rights of those affected (the state), and those who volunteer to advocate for the rights of those affected (development organizations). Three major interlocking and interdependent themes emerge. First, address the “root causes” of natural resource related conflict, which, as Global Witness (2017, p. 39) recognizes, is the “only effective prevention in the long term.” Second, improve the rule of law to clarify rights to resources and resolve conflicts before violence occurs and punish and deter the use of violence. Third, promote the rights of environmental defenders and empower them to pursue justice for environmental and physical violence experienced. These themes and their overlaps are demonstrated in Figure 5.1 and discussed in more depth below.    133  Figure 5.1 Conceptual framework: interlocking strategies to respond to and reduce forest-related violence  5.2.1 Addressing the ‘root causes’ Addressing the ‘root causes’ of forest-related violence means identifying why conflicts occur and resolving them before they “escalate” into violence (de Koning, Capistrano, Yasmi & Cerutti, 2008, p. 17). This approach is central to mainstream recommended responses to forest-related violence. Yet, the task is immense and challenging given the range of contexts in which conflicts occur and the complex assemblages of socio-political, economic, institutional, and environmental factors on which they are based (Peluso & Watts, 2001). Consequently, diagnoses of ‘root causes’ tend to be drastically oversimplified and ‘rendered technical’ – that is, they are subject to a set of practices concerned with representing “the domain to be governed as an intelligible field with specifiable limits and particular characteristics ... defining boundaries,  134 rendering that within them visible, assembling information about that which is included and devising techniques to mobilize the forces and entities thus revealed” (Rose, 1999, p. 33).  When the ‘root cause’ of violence has been attributed to poverty as a driver of illicit resource use, such as illegal logging or poaching (McElwee, 2004), recommendations often focus on developing alternative income-generating options and securing livelihoods – an approach that underpins Integrated Conservation and Development projects, in which Community Forestry is included, and Payments for Ecosystem Services schemes such as REDD+ (Cooney et al., 2011; Sunderlin et al., 2005; Tänzler & Ries, 2012). Such approaches are intended to support economic and cultural rights, thereby reducing conflict and preventing violence and contributing to other aspects of rights-based development.   5.2.2 Improving the rule of law In many cases, addressing the root cause of conflict and supporting human rights may necessitate improvements in the rule of law and related institutions. For example, when conflict is attributed to lack of secure tenure rights, land titling has been recommended as it can enable the title-holder(s) to protest competing claims and/or claim compensation for forced dispossession (Smalley & Corbera, 2012; Sunderlin et al., 2009; Yasmi et al., 2010). Similarly, improving the rule of law has been widely recommended when conflicts and related violence have been framed as the result of a specific natural resource management (NRM) policy such as the commercial exploitation of natural resources “in the name of development” through hydroelectric dams, mines, and agricultural land concessions (Carter, 2015, p. 1) or evicting people from National  135 Parks or REDD+ forests111 (Cernea & Schmidt-Soltau, 2006; Franco et al., 2017; Lyons et al., 2014; Vedeld et al., 2012). In such cases, recommendations are commonly policy and/or legal reforms at the national level to change the way projects are implemented since there is “no specific ‘hard law’ instrument in international law to deal with the challenges posed by land grabbing” (Subedi, 2015, p. 25). For example, in response to violent land grabs, accumulation by dispossession, and concern that profit is being prioritized over people (seen as root causes), the Food and Agriculture Organization of the United Nations (FAO) recommended governments follow principles of Responsible Agricultural Investment and voluntary guidelines on responsible governance of land tenure (FAO, 2014, 2016; Stephens, 2013) or the United Nations’ Guiding Principles on Business and Human Rights (UNHRC, 2012). Such legal and policy reforms have also been promoted when corruption among state actors motivates the introduction of policies that have violent outcomes (Carter, 2015; De Schutter, 2016). Such reforms support the promotion of human rights when a legal reform clarifies resource rights or formally recognizes the right to maintain customary resource usage practices for all citizens or those of specific groups such as indigenous people (Munang, 2015).112 Additionally, legal instruments “give advocates a language” to use in campaigns that is difficult for governments to dismiss (Subedi, 2015, p. 33).  In order to effectively implement legal and policy frameworks, enforce sanctions against those who use violence, deter future violence, and protect the rights of citizens, law enforcement institutions must be impartial and independent (UNDP, 2016b). In many places where natural                                                 111 Notably, REDD+ has been identified as both a cause of violence and a potential solution to reduce conflict. 112 The United Nations Declaration on the rights of Indigenous Peoples is an important but non-binding international normative foundation for the legal framework on the recognition and registration of indigenous people and their right to collective land title although it is non-binding (Subedi, 2015).  136 resource related violence is highest, law enforcement institutions, such as the judiciary, police, and government ministries, are weak if not outright corrupt (Global Witness, 2016a; Yasmi et al., 2012). This not only allows individual citizens to use violence without fear of punishment but enables state officials to use their position to orchestrate violence as a means to control natural resources and accumulate personal wealth and power, sometimes through policies such as Economic Land Concessions, mines, or infrastructure projects discussed above (Global Witness, 2015b, 2016b, 2017; Peters, 2013; USAID, 2011). In such cases, anti-corruption measures have been recommended along with institutional reforms, capacity-building and technical assistance for law enforcement authorities with the expectation this will encourage them to apply existing laws effectively (De Jong et al., 2006; Duran et al., 2011; Global Witness, 2015b; MacInnes, 2015).   Improving the rule of law is also recommended to ensure national justice systems appropriately punish those who use violence and deter others from doing so in the future. Legal reform can change the balance of risk and reward for the use of violence by ensuring that the use of direct violence and other human rights abuses are appropriately criminalized (Global Witness, 2017). Although many countries in which natural resource related violence occurs are signatories to the Universal Declaration of Human Rights, these are not always fully reflected in national laws. These can be strengthened by legal reform to protect vulnerable groups, such as indigenous people, from natural resource related violence (Amnesty International, 2017a; de Koning et al., 2008). Again, advocating for such legal reform necessitates promotion of human rights and entails citizen empowerment.   137 Legal reform is also a necessary response to the “increasing” criminalization of activism, including environmental activism (Amnesty International, 2017a, p. 17). This can effectively or actually legalize direct violence and enable governments and companies to intimidate activists into silence or distract them with costly legal battles through threats of imprisonment (Atiles-Osoria, 2014; Eddy, 2005; Human Rights Watch, 2016). This was exemplified in the case of the Sioux tribe protesting the Dakota Access Pipeline at Standing Rock in North Dakota, USA, when state legislators debated a bill that would allow motorists (such as pipeline construction workers) to run over and kill environmental protestors without facing jail (Agrawal, 2017; Global Witness, 2017).113 In such cases, legal reforms are often necessary to protect citizens’ rights (Front Line Defenders, 2016).  5.2.3 Promoting human rights Upholding the rule of law is not only the responsibility of state institutions but also a process that necessitates that empowered citizens demand that justice systems protect their rights. One method of achieving this is for journalists and researchers to report on violence against environmental activists (Global Witness, 2017). Researchers have an important role to play in exposing crimes, drawing attention to corruption and the socio-political networks through which it operates, and raising awareness of human rights abuses yet such violence is currently underreported, especially at the international level (Gritten et al., 2012; The Guardian & Global Witness, 2017b). National and international human rights advocates can use such research to pressure states to protect human rights through improved rule of law. In starting public discussions about violence and the root causes of it, human rights advocates can empower                                                 113 Although it was narrowly defeated, it inspired a similar bill in North Carolina that was passed in the State House of Representatives and is awaiting approval by the State Senate (Global Witness, 2017).  138 citizens to demand protection of their rights ensuring that change is not only driven by externally-imposed reforms but is also driven by citizens’ demands. Notably, in countries with limited civil and political rights, reporting may only be an effective response if supported by legal reforms to protect the right to freedom of speech.   This review suggests that the three themes of addressing the root causes, improving rule of law, and promoting human rights are complementary and are the basis for a normative conceptual framework that can be used to coordinate responses and design a long-term strategy to reduce forest-related violence. Yet, in the context of Cambodia, there have been few specific recommendations made of how to respond to forest-related violence. Those that have been made almost exclusively focus on enhancing the rule of law.114 For example, given the widely acknowledged role of (neo)patrimonialism and corruption as structural drivers of forest conflicts (Baker & Milne, 2015; Davis, 2005; Global Witness, 2007; MacInnes, 2015; Milne, 2015; NGO Forum on Cambodia, 2015), enhancing the rule of law is seen as a way to both punish and deter violence directly while also addressing the ‘root causes’ of violence. Specific recommendations include improving communities’ and authorities’ capacities to mediate conflicts by referring to legal guidelines before they turn violent (Yasmi et al., 2010, 2012). Others focus on eliminating corruption (Global Witness, 2016a, 2016b), including by improving the “capacity and willingness of government and investors to take a community’s interest into account” (Dhiaulhaq                                                 114 Given the scarcity of academic studies focusing specifically on forest-related violence in Cambodia, there are few recommended responses. Most reports of violence merely report what happened and do not consider appropriate responses. Others only recommend responses to conflict over forest resources in which it is implied that both sides are antagonistic (e.g. Backstrom et al., 2006) or a lack of legal clarity and formalization is contributing to conflict (Ratner & Parnell, 2011).  139 et al., 2014a, p. 214), echoing the generic, mainstream recommendations made in global overviews of violence (e.g. Global Witness, 2015a, 2016a, 2017).   5.2.4 Recommended responses to reduce violence Recommendations of how to reduce natural resource related violence are certainly well-meaning and identify necessary long-term changes but there are several potential impediments to their success in practice. Reforms of laws, policies, and institutions, and recognizing and prioritizing the protection of human rights require significant political will or even a change of political culture (de Schutter, 2011; Fitzpatrick, 2015). These are gradual processes even when not complicated by the governmental corruption that often motivates and facilitates rent-seeking NRM policies and complicity of companies and consumers in resource extraction processes that prioritize profit over citizens’ rights (Acemoglu & Robinson, 2005). Although development organizations can encourage reforms through aid conditionalities, this can have perverse unexpected outcomes, for example, externally-driven reforms that are not representative of political will to decentralize natural resource management rights may actually give ruling elites more power over resources or exacerbate existing inequalities in communities (Cock, 2016; Hall et al., 2012; Ribot et al., 2006). Furthermore, proposed reforms must appropriately reflect local perceptions of ecological and social justice (Martinez-Alier, 2002), what counts as corruption and (im)moral behaviour (Ali & Nyborg, 2010; Robbins, 2000), and development aspirations (Sen, 1999).   A further concern is that when the ‘root cause’ of violence is ‘rendered technical’ (Li, 2011), for example, labelling it as ‘poverty,’ ‘corruption,’ or ‘lack of secure tenure,’ it may mask other  140 processes, discourses, and social, cultural, and political relations driving violence. As political ecologists have repeatedly demonstrated, NRM policies, legal and illegal resource usage patterns and conflicts, and normative expectations of rights are shaped by historically-rooted ideologies and cultural values, complex political and economic arrangements often involving domestic and foreign actors, and an international industry of NRM ‘best practices’ often rooted in flawed Western scientific discourse excluding local knowledge and priorities, among other factors (Benjaminsen et al, 2013; Horowitz, 2009; Leach and Fairhead, 2000). Ignoring these factors may mean that ‘solutions’ to existing conflicts lead to more violence or displace it in time or space, or draw different actors into the conflict.   Relatedly, the ideological perspective from which violence is analyzed can bias which ‘root cause’ is identified and thus the recommendations made. For example, those used to working with neoliberal-inspired or “neo-institutionalist approach” (Sandström et al., 2013, p. 124) to natural resource management theory are prone to diagnosing institutional weaknesses at the macro-, meso-, or micro-societal level as the cause of conflict due to the underlying assumption that “the quality of institutions trumps everything else” (Rodrik et al., 2004, p. 131). As Springer (2009b, p. 306) notes, violence is portrayed as a “barbarian” practice that “only the ‘civilising’ logic of neoliberalism may conquer.” In contrast, human rights activists tend to recommend Rawlsian-inspired, utopian ideals of justice that do not account for unequal power relations as historically constructed injustices, normalized through symbolic violence. Consequently, they may overlook more pragmatic, incremental improvements to justice outcomes that could be obtained through existing, albeit flawed, structures and institutions (c.f. Sen, 2009).    141 This indicates that the potential for recommendations to reduce conflict and violence depends on correctly diagnosing the root cause(s) and determining at what point in its genealogy an intervention should be made. For example, addressing poverty as purely an economic phenomenon may not be sufficient (Duffy & Humphreys, 2016). Rather, interventions may need to address social and political marginalization and inequality and support aspirational development goals as a means of reducing conflict instead of solely maintaining existing natural resource dependent livelihoods that may no longer provide everything rural people want or need (Duffy & Humphreys, 2016; Shankland & Hasenclever, 2011). Yet, recommendations must find a balance between being relevant and specific to the manifestation of problems affecting resource users given the rights-protection mechanisms and state institutions available to them whilst maintaining awareness of the histories and ideologies that feed the structures that generate violence. While resolving the ‘big picture’ root causes is certainly desirable, it often requires external oversight for an extended period. In the meantime, those experiencing natural resource related violence may need recommendations that are immediately actionable at the community level.   Several of these problems are exemplified in Cambodia-focused, mainstream recommended responses to violence discussed briefly above. Since neopatrimonialism and corruption are so overtly central to the organization of social, political, and economic life in Cambodia, it appears to be difficult for commentators to look past this divergence from the Western socio-political norm when seeking ways to reduce forest-related violence in the country. Consequently, violence is ‘rendered technical’ and represented as an outcome that can only be governed through neoliberal mechanisms of improving the accountability, transparency, and democratization of  142 forest governance and the justice system (c.f. Norman, 2014). This oversimplifies reasons for which people participate in illicit forest use, seek patronage from those involved in illicit forest use (as opposed to other political, economic or religious patrons), and choose or feel compelled to use violence. It also obscures other geographically-distant drivers of forest-related conflict and violence such as demand for luxury wood (Challender & MacMillan, 2014; EIA, 2017), a global food system that encourages the expansion of Economic Land Concessions for sugar and palm oil production (Cotula, 2013), and the complicity of international development and conservation organizations in sustaining neopatrimonial structures, failing to question the government’s “façade of reform” (Milne, 2015, p. 207) and “physically, discursively, and economically” accommodating and facilitating extractive projects (Work & Thuon, 2017, p. 362). Thus, while it may be argued that commentators on forest-related violence in Cambodia are trying to provide recommendations that address the problem at the national level rather than tackling implicated global political and economic systems, these recommendations are still focused on change at too high of a political level, are too idealistically utopian, and fail to account for the realities of living within a society organized by neopatrimonialism.   To counter this tendency, I propose that recommendations need to be more focused on how those affected by forest-related violence (environmental defenders) and those supporting CBFM projects (development organizations) can act to reduce forest-related violence and mitigate its harms when it does occur. Instead of aiming to resolve forest-related conflict and violence completely, this actor-oriented approach will seek out potential points of intervention where CBFM participants can respond more effectively or support can be provided by external actors that is both context-specific and appropriate and that would also incrementally but importantly  143 improve the lives and safety of those participating in CBFM. This pragmatic approach brings attention back to the agency of those affected by violence rather than directing attention to the violent socio-political macro-structures with which actors interact. In doing so, it would seek ways to protect human rights and promote the political empowerment of citizens, an aspect missing from currently recommended responses in Cambodia but which, in other contexts, has been theorized as an important step in reducing forest-related violence.   In order to explore what these alternative recommendations might look like and their workability given the socio-political context, it is necessary to analyze how those affected by forest-related violence are currently responding both in terms of trying to enforce the rule of law and finding alternative measures. It is necessary to understand exactly how patronage and corruption undermine the rule of law in the lives of CBFM participants and why current alternative responses are not fully effective. This is the task of the remainder of this chapter but first I review the current state of Cambodian law enforcement institutions to give the reader a more thorough understanding of the legal and socio-political context in which CBFM participants and development organizations act.   5.2.5 Upholding the rule of law in Cambodia: a review of relevant institutions  For the rule of law to be implemented to reduce forest-related violence, three elements of the justice system must function effectively. First, relevant laws must be in place – those that criminalize violence and those that regulate forest use to reduce conflict over resources. Second, laws must be upheld by a justice system in which actors work with integrity, independence from outside influence, accountability, and transparency. Third, citizens must know their rights as  144 granted by law and feel empowered to call on the justice system to defend those rights. This section reviews these three elements of Cambodia’s justice system to establish the legal, socio-political, and cultural context in which CBFM participants respond to forest-related violence.  Cambodia’s Forestry Law (2002) and Land Law (2001) define who may hold what kind of forest resource rights and how tenure can be formally secured. The Forestry Law also criminalizes logging several species of tree targeted by organized logging groups. These laws form a clear legal groundwork for CBFM. Yet, their implementation has been problematic and has contributed to the occurrence of forest-related conflicts and violence. The process is long and bureaucratic so legal conflicts arise when proposed sites are not promptly approved by state authorities and overlapping resource claims are made in the meantime, such as by ELC companies. Loopholes in the laws also enable the ruling elites and their associates to find ways to appropriate resources without breaking the law. For example, one company may only hold 10,000 ha. of land but multiple subsidiaries of one company can each claim their own 10,000 ha. lot facilitating mass accumulation in the hands of a few elites (Neef & Touch, 2012). Consequently, these laws have been criticized for being made to favour the interests of ruling elites and their associates (Dwyer, 2015; Springer, 2012; Work & Thuon, 2017). Thus, forest-related conflicts are not the result of a lack of legal clarity but a consequence of the ineffective implementation of these laws and the weak justice system, plagued by corruption and patrimonialism, which prevents challenges from being pursued.   Similarly, the Constitution of Cambodia and the Criminal Code criminalize the use of direct violence, including the forms of violence described in the previous chapter. Specifically, death  145 threats, detainment, and physical abuse are punishable by imprisonment of up to five years and fines of up to ten million riel (approximately US$2,500). Relevant Articles are detailed in Annex B. Donors have invested in establishing the appropriate organizations and structures to investigate crimes, enforce laws, and try crimes in the courts (AusAid, 2012). They have also established an Anti-Corruption Law (2010) that has been praised for meeting international standards, although enforcement is “weak” (GAN Business Anti-Corruption Portal, 2015, para. 15).   Yet, despite the existence of these laws, attempts to use the justice system to punish violence are often blocked by the “twin scourges” (Freedom House, 2017, p. 4) of patrimonialism and corruption among law enforcement authorities. Cambodia’s justice system lacks integrity, independence, impartiality, transparency, or accountability (Bertelsmann Stiftung, 2016), qualities that are all considered “preconditions for fair and effective access to justice and for the protection of human rights” (UNDP, 2016b, p. 23). Accountability and transparency experts recognize Cambodia’s justice system as one of the least effective and most corrupt in the world. The World Justice Project’s Rule of Law Index 2016 ranked Cambodia 112 out of 113 countries (World Justice Project, 2016). It scored Cambodia’s justice system 0.05 and 0.10 out of 1 for ‘effective enforcement’ and ‘no corruption’ respectively.115 The World Justice Report’s (2016) justifications for these low scores are exemplified in the four institutions responsible for enforcing the rule of law in Cambodia’s forests – the police, the judiciary, local political leaders,                                                 115 Mean scores in East Asia and Pacific were 0.58 and 0.60 respectively, and mean scores for lower middle income countries were 0.41 and 0.38 respectively. Similarly, Bertelsmann Stiftung’s Transformation Index 2016 scores ‘the separation of powers between government and judiciary’ as 2 on a scale of 1 to 10 (Bertelsmann Stiftung, 2016) and Freedom House (2017) classifies Cambodia as ‘not free’. Human Rights Watch (2017a, para. 4) criticizes Cambodia for “corruption” that remains “a huge problem” and argues that the “impunity” of high-ranking officials plagues the justice system.  146 and the Forestry Administration (FA) – which are fundamentally structured by patronage and nepotism.   Although the constitution mandates the separation of powers between the legislature, executive, and judicial system, the police force does not operate independently but, rather, is closely tied to the ruling Cambodian People’s Party (CPP) and the Prime Minster, Hun Sen. The Commissioner General channels orders directly from the PM through a strict institutional hierarchy. Nepotism secures this chain of command – the Commissioner General, Neth Saveoun, is the husband of Prime Minister Hun Sen’s niece, although he lacks formal qualifications for the role (Voice of America, 2008). At all levels, police officers are expected to join the CPP and those who refuse are often isolated at work and not promoted (Johnson, 2014). Nepotism and patronage also dictate who can become a police officer in the first place. Jobs in the police are not maintained or earned by qualifications or performance but rather by one’s ability to pay the right person or be connected to them by blood or marriage. An “upward bribery chain” operates with police officers routinely paying a percentage of their salaries to higher-ranking officers and buying promotions (Transparency International, 2016, p. 6). This creates a relationship of dependence with an officer who will likely always be one promotion ahead, which makes it difficult to question superiors (Johnson, 2014). Likewise, challenging the political party line is seen as a challenge to one’s superiors and is often punished so police officers are pressured to engage in corrupt and unjust practices regardless of their personal beliefs (Johnson, 2014). Police officers’ engagement with citizens is also fraught with corruption. Public sector salaries are generally “inadequate” (Cock, 2016, p. 61) which encourages a “rent-seeking mentality” and extortion of those seeking justice (Bertelsmann Stiftung, 2016, p11). In Transparency International’s Global  147 Corruption Barometer 2013, 65% of respondents had paid a bribe to the police in the last 12 months (Transparency International, 2013). Consequently, the police’s role in enforcing the rule of law is distorted by patronage politics despite donors’ investments in legal reforms and police training. Johnson (2014, p. 22) concludes, “police officers are not only subservient to their direct superiors, but must also submit to the dominant will of the ruling party.”  Like the police force, Cambodia’s judiciary fails to operate with independence, accountability, or transparency. Judges are poorly trained (Un, 2009) and are subject to pressure from the CPP (Freedom House, 2016). The judiciary remains a “tool” of the ruling elite to prosecute critics of the government, protect the economic interests of the wealthy, and “perpetuate impunity” for ruling elites and their associates (LICADHO, 2017, para. 2; Un, 2009).116 Trials are not free or fair (Freedom House, 2017). Transparency International reports that Cambodians view the judiciary as the “most corrupt” institution in Cambodia with over 68% of those who used the courts having paid a bribe in an attempt to obtain justice (2016, p. 6; Transparency International, 2013). Trials are widely regarded as a sham with people who do not bribe the judge seen as “not being interested in a favourable judgement” (Bertelsmann Stiftung, 2016, p. 11). Judges’ motivations to extort plaintiffs are exacerbated by low salaries,117 pressure to share earnings with direct superiors, and debt for bribes paid to secure jobs as judges that can be up to six-figure sums in US dollars (Bertelsmann Stiftung, 2016). A lack of transparency and accountability enable such payments to be taken without fear of being sanctioned for abuse of office                                                 116 Bertelsmann Stiftung (2016, p. 32) assesses that the CPP has “promoted a de facto suspension of the judiciary” since, for the last several decades, the government’s top decision makers have orchestrated all trials with any political dimension thereby undermining the judiciary’s ability or perceived legitimacy to protect human rights and freedoms or provide a check on other branches of government. 117 In 2015, provincial judges’ salaries were US $125 per month (Bertelsmann Stiftung, 2016) although they have since been raised to US $750 in an attempt to reduce corruption (Mech & Turton, 2016).  148 (Bertelsmann Stiftung, 2016). Even in civil cases, the ruling elite pressures judges by way of threats and bribes to ensure their interests are protected (LICADHO, 2017).  Local political leaders (Village Chiefs, Commune Councils, and Commune Chiefs) are influential actors in the Cambodian justice system as they are mandated to “mediate local conflicts and maintain law and order” in their jurisdiction (Romeo & Spyckeralle, 2003, p. 5). They also assist villagers with administrative tasks such as reporting crimes to law enforcement authorities and therefore can be seen as ‘gatekeepers’ to the justice system for those who lack the literacy or confidence to seek justice themselves. Local political leaders are part of a formalized, vertical, administrative structure that was developed through donor-driven legal reforms with the intention of increasing accountability and transparency through ‘democratic decentralization’ (Öjendal & Sedara, 2006; Sedara & Öjendal, 2009). This legal-rational system is inspired by the neoliberal ideology of ‘rolling back’ central government’s powers by transferring some administrative, fiscal, and political responsibilities from the national level to commune level (Chan & Chheang, 2008). Yet, local political leaders are still strongly controlled by patrimonial networks headed by CPP officials and economic elites (Chan & Chheang, 2008; Thon et al., 2009). Village Chiefs are appointed by and answer to Commune Councillors, who, in turn, answer to Commune Chiefs, who answer directly to the Minister of the Interior (Sedara & Öjendal, 2009).118 Village members elect Commune Councillors but Freedom House (2017) assesses that elections are ‘not free.’ Not only do candidates often pay for votes with cash or goods but the District Governor must approve nominees for the Commune Council, which                                                 118 The Minster of the Interior is Sar Kheng, who has been among Cambodia's political elite since 1979 when the Khmer Rouge was removed from power. He has served as Deputy Prime Minster since 1992 and is brother-in-law to the late Chea Sim, who was president of the CPP from 1991 to 2015, although maintains a rivalrous relationship with the Prime Minister, Hun Sen (Hughes & Un, 2011).  149 enables political leaders to use the legal-rational system to neutralize strong opposition candidates and maintain power in the hands of a select and vetted few. This patron-client network extends to Village Chiefs, who are appointed by Commune Councillors, seen as a personal favour that secures future loyalty (Eng, 2014). Furthermore, local political leaders become clients of economic elites, who demand or attract loyalty by offering wealth and power in exchange for favouring business interests over the well-being of their constituency (Thon et al., 2009).   The Forestry Administration (FA), a department of the Ministry of Agriculture, Forestry, and Fisheries (MAFF), and the Ministry of Environment (MoE) also play roles in upholding the rule of law relating to forests. Specifically, the FA is responsible for regulating, investigating, and suppressing various forms of forest destruction (RGC, 2010). FA Officers and MoE Rangers can also issue fines for illicit forest use (OpenDevelopment Cambodia, 2016b). It is unclear whether the FA or MoE has the responsibility or authority to press charges over forest-related violence but in their role of supporting CBFM implementation they make reports of illicit forest activities to the police and courts on behalf of a community. There does not appear to have been systematic inquiry into FA or MoE operations as there has been for other law enforcement agencies.119 However, Clements et al. (2010, p. 1284) note that “enforcement of laws is rare” and patronage, corruption, and lack of capacity often limit their effectiveness (Milne, 2015; Vong,                                                 119 My own experiences and conversations with fellow researchers indicate that FA is often hostile to academic researchers who are not embedded within the organization (and therefore under their control) or conducting ecological research that directly benefits the FA (c.f. Shoenberger & Beban, 2017). Even then, hostilities remain and interactions with outside observers are strictly supervised. For example, the capacity-building development organization, Voluntary Service Overseas (VSO), suspended their partnership after finding the FA “secretive,” “suspicious,” “too difficult” to work with, and “resistant to criticism or change” (interview with former VSO Livelihoods Programme Director, 6th May 2015).  150 2015).120 Like the police force, jobs are often obtained through nepotistic or patrimonial connections and promotions are bought rather than earned.121 Bureaucracy, inefficiency, and lack of resources also hinder operational capacity and both encourage and facilitate systemic, petty corruption (e.g. bribing CBFM groups to approve forms). Collusion with illegal logging groups, the military, and ELCs occurs (e.g. turning a blind eye to logging, selling confiscated logging equipment and logs back to known loggers) (Lang, 2012) and Milne (2015) reports that the logging tycoon Try Pheap recently donated a new office for MoE rangers, although such practices appear to vary between offices and collusion waxes and wanes over time reflecting the evolving nature of patron-client relationships.122  Journalists also have a major role to play in the justice system by exposing the power networks and informing a public debate on the need for accountability. However, Cambodian media under-reports forest-related violence and when it does, related corruption and neopatrimonialism are rarely explored. While this is, in part, symptomatic of the normalization of direct and structural violence, it is largely an effect of the suppression of free speech in the media. Direct violence is used to threaten and punish journalists, political commentators, and human rights activists who expose illicit forest activities and ruling elites’ implication (CCHR, 2010, 2014; Human Rights Watch, 2017a). The ‘defamation law’ (Article 305 of the Criminal Code) is also used to criminalize criticism of the ruling party, 123 illustrating how the ruling elite uses legal                                                 120 Interviews with NGO, Kampong Thom, 1st July 2015. 121 Interview with Head of CF Network, Oddar Meanchey, 17th June 2015; informal conversations with Cambodian friends regarding applying for civil servant jobs. 122 Interviews with CFMC heads, Oddar Meanchey, 22nd May, 15th June, and 17th June 2015; Kampong Thom, 5th July 2015. 123 Most commonly, individual reporters and political activists are targeted (e.g. Chhorn, 2017; Committee to Protect Journalists, 2014), although in 2017 the government launched targeted attacks on two English-language newspapers, the Voice of America and the Cambodia Daily, attempting to shut them down on dubious charges of tax avoidance  151 rational mechanisms to perpetuate and mask the violent, patrimonial networks through which they dominate Cambodian citizens and the environment. The media is also compromised by journalists’ participation in the patrimonial networks of illegal loggers in which they accept or demand bribes to omit a story or distort the details.124   For a justice system to function effectively, citizens under its protection must also participate in law enforcement procedures, specifically, reporting crimes and giving evidence in court (UNDP, 2016b). Yet, the Cambodian Center for Human Rights finds that not all Cambodians are aware of their resource rights so do not know when a crime has been committed against them (CCHR, 2016). Furthermore, as a consequence of the (neo)patrimonialism and corruption that plague the institutions of the justice system, citizens also have a “general distrust” in the justice system (Transparency International, 2016, p 3; Chan & Chheang, 2008; Hasselskog, 2009). As I discuss below, these two factors negatively impact their propensity to seek law enforcement against those who use forest-related violence.  In summary, the literature indicates that despite the establishment of relevant legal frameworks and institutions, it is difficult to enforce the rule of law in Cambodia due to actors in the justice system being entangled in patrimonial networks that oblige them to act in the best interests of patrons rather than of Cambodian citizens. These patrimonial networks facilitate and encourage corruption further undermining the effective workings of the justice system. This has the effect                                                                                                                                                        and ‘destabilizing the political situation’ (Gallo, 2017; Wright, 2017a, 2017b). CCHR (2014) challenges the constitutionality of Articles criminalizing defamation and ‘insulting’ politicians. 124 For example, journalists have been accused of lying in wait along roads near international borders and extorting loggers transporting illegally cut timber (Mech & Kossov, 2016).  152 of undermining citizens’ confidence in the justice system and reducing the likelihood they will report crimes or try to achieve justice through legal-rational mechanisms.   It is in this legal and socio-political context that CBFM groups and NGOs are attempting to respond to and reduce forest-related violence. Thus, the rest of this chapter asks, how do those affected by forest-related violence respond and with what effects? What are the exact mechanisms and socio-political relations that prevent effective responses? How can development organizations better support CBFM participants in responding to and reducing direct violence?  5.3 Methods The aim of this chapter is to expose how CBFM participants and supporting development organizations are responding to forest-related violence in practice and question the workability of suggestions made in the literature. A particular focus is given to lived experiences of implementing the ‘rule of law’ since this is so commonly promoted as a necessary course of action. To explore these issues, I take an actor-oriented approach (Long, 2001) to understanding responses to forest-related violence in an effort to depict the ways in which people “manage the dilemmas of their everyday lives” (p. 10) and counter the structural and institutionalist nature of mainstream recommendations.   I draw on data from interviews with CBFM participants and conversations and observations made during participant observation of CBFM activities, especially forest patrols. I also use data from surveys and semi-structured interviews with eighteen NGOs and three donors regarding their knowledge and understanding of forest-related violence in the projects they support and  153 their responses to this issue. Details of the survey and interview protocols were described in Chapter 3, section 3.2, along with the limitations of researching corruption among state actors.  5.4 Responses to forest-related violence in CBFM projects in Cambodia This section explores NGOs’ and communities’ responses to forest-related violence. I show that these responses are primarily aimed at implementing the rule of law, which is also intended to address ‘root causes’ of illegal logging in terms of corruption. This aligns with the institutionalist ideology on which CBFM is based but is ineffective against those illicit forest users most likely to use violence – organized logging groups and the military – as those actors can leverage their patrimonial relationships with state actors to circumvent the legal-rational disciplinary mechanisms of the justice system. I then present the alternative ad hoc strategies used by communities and NGOs that do reduce the incidence of forest-related violence but also undermine the effectiveness of CBFM as a means of forest protection. These responses are summarized in Figure 5.2. I argue that these responses are not rooted in a comprehensive strategy to reduce direct violence or address the structural violence of the socio-political system that maintains ruling elites in positions of power and enables them to use direct violence with impunity. Consequently, they do little to promote human rights or political empowerment thus neglecting an important area of potential intervention. Yet, analysis of the weaknesses in these responses and the obstacles to their success illuminates ways in which CBFM participants could be better assisted if they experience direct violence and supported to reduce its future occurrence.   154  Figure 5.2 Responses to forest-related violence in Cambodia  5.4.1 Addressing a ‘root cause’ by reducing forest conflicts NGOs’ primary response to forest-related violence is to attempt to reduce the number of conflicts over forest resources. When this involves the military or ELCs, NGOs have supported  155 communities and community leaders in negotiations although this has infrequently resolved the conflict (c.f. Yasmi et al., 2013).125 They have also supported communities in reporting conflicts over forest resources to law enforcement authorities, the problems with which are discussed below. When conflicts involve local loggers, all eighteen NGOs surveyed encourage and financially support CBFM groups to hold community education meetings to educate local loggers on the importance of sustainable forest management through which they hope to reduce local logging. However, loggers are often familiar with this rhetoric and claim they need the income from timber now more than they need the forest in the future. Not only do loggers find this approach patronizing since they are well aware of the role of the forest in local livelihoods or were even involved in establishing CFs or CPAs126 but it also fails to address their motivations for engaging in illicit forest use.127 If they do agree to stop logging locally, they often displace their activities to another area creating problems of logging and forest-related violence in a different community.128   Addressing violence more specifically, six surveyed NGOs deliver training courses on non-violent conflict resolution for communities engaged in forest-related conflicts.129 Participants are taught negotiation skills to help them deescalate conflicts that may turn violent and taught to “educate” people who break the forest bylaws rather than trying to physically stop illicit                                                 125 NGO survey data 126 Interviews with local loggers, Oddar Meanchey, 29th May and 12th June 2015, and interview with former member of organized logging group, Ratanakiri, 16th November 2015. 127 Initial interviews with local loggers suggest that motivations include financial desperation and debt, opportunistic financial gain, frustration at outsiders logging, lack of belief in the future success of CBFM, and loss of land to ELCs. Further research is needed to explore these motivations.  128 This dynamic is referred to as ‘leakage’ or ‘displacement’ (Atmadja & Verchot, 2012) in literature on REDD+ but is less commonly discussed with regards to CF and CPAs.  129 I observed training on this topic in Oddar Meanchey, 17th June 2015. It was combined with training on human rights.  156 activities.130 Although CBFM participants claim to be unwilling to use violence unless in self-defence, recognizing that it usually incites more retaliatory violence,131 one NGO employee noted “sometimes patrol members get very angry when they see someone cutting the forest because they have worked hard to protect it. We teach them not to lose their tempers and take some time to think before confronting loggers.”132, 133 This training emphasizes the value of discussion, and, if that fails, encourages patrol groups to walk away and call the FA and/or police for assistance (although that strategy has little chance of success, as discussed below). While attempting to address a ‘root cause’ of violence – conflict over resources – this approach is inadequate for the nature of the conflicts in which illicit forest users are knowingly in violation of bylaws and who have nothing to gain from entering into negotiations and more to gain by using violence to enforce their will. Furthermore, data from interviews with CBFM participants and participant observation of patrols indicate that patrol members do not need encouragement to avoid violence – their fear of physical harm means they are already prepared to back away from a conflict if it feels like violence will occur.                                                  130 Participant observation at training for CF groups, Oddar Meanchey, 17th June, 2015. 131 Interview with Head of CFMC, Oddar Meanchey, 15th June 2015. 132 Interview with NGO, Oddar Meanchey, 10th June 2015. 133 In the training I observed, the NGO promoted the role of women in deescalating conflicts proposing that women are less likely to use violence first and are less likely to be physically attacked by illicit forest users. This view was shared by a number of participants with whom I spoke during the workshop breaks but more research is necessary to determine the effectiveness of this approach in practice.  157  Figure 5.3 Training Community Forestry Management Committee members on non-violent conflict resolution, Oddar Meanchey, 17th June 2015.  5.4.2 Using the justice system to implement the rule of law In an attempt to implement the rule of law, CBFM participants and NGOs report, or assist victims in reporting, incidents of forest-related violence, as well as illicit forest use, to law enforcement authorities, specifically, the police and the Forestry Administration (FA), with the hope that those authorities will escalate the report to the courts, who would punish violent behaviour. NGOs are well positioned to assist in this regard as staff have the literacy skills and confidence to navigate the bureaucratic procedures of law enforcement. Although all eighteen NGOs surveyed have reported illegal logging or deforestation to law enforcement authorities, only seven NGOs have supported communities in reporting violence. They explained that this was because they only make reports if the victim suffered physical harm requiring medical attention and if the victim wishes to make a formal report. Yet, this support appears to be inconsistent as at least three interviewees who have experienced violence and needed medical  158 attention have not been offered NGO help to report the incidents. Threats of violence are rarely reported unless an imminent danger to life is suspected. This hesitancy to report all incidents of forest-related violence is discussed further below.  Community-based participants are more likely to report violence to the Village Chief than directly to the police because they believe that Village Chiefs are more literate and able to navigate bureaucratic procedures and are responsible for looking after village members. Also, they expect that the police are less likely to try to bribe Village Chiefs than other villagers, and in some cases believe that the Village Chief has direct influence over those using violence, especially if that person belongs to the local community. In contrast, police are widely believed to be corrupt, engaging in bribery and colluding with loggers. However, as discussed below, there are several reasons why some CBFM participants would not make reports to the Village Chief either.   Attempts to enforce existing laws in order to punish and deter forest-related violence appear to be wholly unsuccessful. During fieldwork, no cases were found in which those responsible for forest-related violence against CBFM participants or NGO employees had been arrested, let alone convicted, even in cases where the identity of the perpetrator was known. This mirrors the national trend as even cases that have been prominent in Cambodian and international media have not been fully investigated or have been closed prematurely (Amnesty International, 2017c; Cox & Ok, 2012; Lambrick & de Smet, 2015).134                                                  134 For example, the military police officer convicted of assassinating environmental activist Chut Wutty in 2012 was given a suspended sentence of eighteen months and released soon after the verdict (Aun, 2017). Chut Wutty was assassinated in front of witnesses in 2012 by a military police officer who, along with others, was guarding an  159  Examples from interview data of how law enforcement authorities fail to act on reports of forest-related violence are dishearteningly common and repetitive despite the brevity of each one: “we reported them to the police but they did not investigate”; “I reported him to the Village Chief but he did not report to the police or give any punishment”; “the police and the Village Chief came to see the remains [of an arson attack on a house] but up to now, they have not made any arrests.”135 Explanations for this repeatedly point to the authority figure to which the report was made being engaged in illegal logging networks and associated violence or being the political client of those who are. The precise details of these networks remain murky – a default given the illegal nature of their activities and the illicit and extra-legal arrangements that are made (Springer, 2017). Yet, a consistent picture emerges from interviewees’ claims and observations of village-level social dynamics indicating that law enforcement authorities do not pursue reports of forest-related violence because of influence from political and institutional patrons.136                                                                                                                                                         ELC company engaged in illegal logging and suspected of processing ‘yellow vine’ (Coscinium fenestratum) for use in psychotropic drugs (Lambrick & de Smet, 2015; May, 2012). Relatedly, the investigation into the murder of political analyst Kem Ley, shot dead inside a gas station in Phnom Penh in July 2016, has been prematurely closed and judged “deficient in several important respects,” with the assassin, who identifies himself only as ‘Meet to Kill’, taking all the blame and obscuring suspected connections to ruling elites (Human Rights Watch, 2017b, para. 3). Likewise, Global Witness was unable to identify any instance in which a member of the Hun family has been prosecuted or held to account for their part in any alleged crimes relating to natural resource appropriation and associated violence. The watchdog suggests that that the ability to evade prosecution appears to be greater if illicit forest users have closer ties to the Prime Minister, Hun Sen (Global Witness, 2016b).  135 Interview with CFMC, 5th July 2015, Kampong Thom; interview with CF member, 8th July 2015; Kampong Thom; interview with CFMC member, 15th June 2015, Oddar Meanchey. 136 For example, CBFM participants’ belief that some local political leaders and many police officers are involved in the patrimonial networks surrounding illicit forest use was illustrated in one village where many local people sell illegally cut timber to timber tycoon Try Pheap when I was warned not to let the Village Chief know I was there or it would cause problems for the CBFM group. Others requested meetings be held away from their villages and communes so local police officers did not see me. One other told me, “If the police stop you, tell them you are researching something else. Do not say anything about the forests or you will have problems.” Head of CFMC, Kampong Thom, 8th July 2015.  160 The overarching reason that existing laws are not effective in punishing forest-related violence is that the illicit forest users most likely to use violence – organized logging groups, the military, and clientelistic local loggers and migrants – are those most able to use personalized systems of power to circumvent the legal-rational disciplinary mechanisms of the justice system. These failures by law enforcement authorities in turn structure the behaviour of Cambodian citizens and shape how they interact with the justice system to defend their rights and respond to direct violence. Specifically, it makes many CBFM participants less likely to ask law enforcement authorities to uphold their rights and contributes to an attitude of resignation to systemic injustice, as I explore below.  Neopatrimonialism within law enforcement institutions not only deters individual officers from upholding citizens’ rights but also prevents those officers who want to help from doing so. Similarly, those who are tasked with joining patrols (an uncomfortable chore most bureaucrats prefer to avoid137) are often so low ranking that they lack the authority to prosecute environmental crimes if a politically influential patron demands that the arrested are released and are not supported in making arrests by their colleagues and superiors.138  The rigidity of legal-rational hierarchies compounds the effects of patrimonialism, especially when it occurs at lower levels of the justice system. It is almost impossible for CBFM participants or NGOs to circumvent the hierarchical system in asking authorities to investigate a case. Interviewees repeatedly complained that the four central law enforcement authorities refuse                                                 137 Conversation with FA officer joining patrol, Oddar Meanchey, 24th May 2015. 138 Participant observation of forest patrol and conversation with FA officer joining patrol, Oddar Meanchey, 24-25th May 2015.  161 to investigate reports of forest-related violence unless local level officials have already assessed it. This meant that although some interviewees mentioned corruption among judges and court staff, none had managed to elevate reports of violence to the courts yet so conversations were primarily about the police and local political leaders. The rigidity of legal-rational hierarchies was exemplified in one CF where the CF participants accused the Village Chief of complicity with the military in a land-grab of the forest. After the military established a base in the CF and began clearing it, he is said to have bought CF land cheaply from the military and sold it on to villagers who saw the opportunity to extend their own farms and to migrants seeking to move to the area.139 Consequently, if lower level authorities are involved in networks of patrimonialism with illicit forest users, reports of violence (or illicit forest use) will not proceed past that administrative level demonstrating how ruling elites use and distort legal-rational systems to protect themselves.140   The detrimental effects of this administrative rule are compounded by the cultural norm of deference to social hierarchies, especially when forest-related violence involves people from the same community such as ‘local loggers.’141 CBFM participants do not feel able to report community problems directly to police or other outside authorities including NGOs (even though this would not contravene legal-rational hierarchies discussed above) because social convention dictates that such matters ‘should’ be first reported to the Village Chief, whose responsibility it is                                                 139 Interviews with CF participants, Oddar Meanchey, 18th June 2015.  140  See Schoenberg (2017) for an example of how engaging with a change of national policy enabled on community to move their complaints past the District level and attract the attention of the PM only to have blame and responsibility passed back down the hierarchy again. 141 As opposed to organized logging groups. For definition and justification of this term, see section 4.3.1.  162 to maintain order within their village.142 They fear that breaking this social convention would likely lead to their ostracism from the community.143 However, when local political leaders have patrimonial ties to illicit forest users, either directly or via political elites to whom they owe their position and relative power, those local political leaders have been unwilling to help communities seek justice for forest-related violence. Consequently, some interviewees expressed frustration at their inability to achieve justice or even legal recognition that a crime had been committed. For example, lamenting how this situation robs CF participants of their dignity, agency, and political rights one interviewee asked, “To live as a human being, what should we do? The powerful people treat us like animals.”144 The combination of rigid legal-rational hierarchies and social conventions create barriers for CBFM participants seeking justice for forest-related violence.  The belief that law enforcement officials will not investigate cases fully nor prosecute the accused deters CBFM members from reporting incidents of violence or illicit forest activities. Interviewees expect that logging tycoons and their associates in the political elite will bribe or threaten law enforcement officials into dropping or closing cases regardless of whether the law enforcement officers had the will and/or capacity to investigate or not.145 Past experience has taught them that law enforcement authorities are, at best, uncommunicative, uncooperative, and under-resourced to investigate and at worst, dismissive, abusive, and predatory – seeking to bribe                                                 142 Interview with CF members, Oddar Meanchey, 22nd and 29th May 2015; Kampong Thom, 29th June 2015. 143 Interview with CF member, Oddar Meanchey, 22nd May 2015. 144 Remark made during training by NGO for CF groups on non-violent conflict resolution, Oddar Meanchey, 17th June 2015. 145 Interview with CF members, Preah Vihear, 6th December 2015. Given this lack of trust in the FA and the belief they enable deforestation by allowing loggers to buy back confiscated logging equipment, multiple CBFM groups corrode the gas tanks of confiscated chainsaws by filling them with acid or salt before handing them over to the FA.  163 villagers for making reports while also taking benefits from and protecting patrons. CBFM participants and NGOs say they are unwilling to pay because they believe it leaves them vulnerable to further bribery in the future. Furthermore, neither communities nor NGOs usually have the money to be able to pay. NGOs are also prevented from doing so because such payments are “difficult to write in the budget for donors”146 since they fall outside of what are considered appropriate use of project funds.147   CBFM participants are also hesitant to report violent crimes because they fear retributive violence if the law enforcement officer has a “string” to a patron involved in illegal logging.148 This fear is not only generated by their understanding of the pervasive patrimonial system but is reaffirmed by witnessing its effects when police participate in violence and the judiciary cooperates in false charges against citizens. For example, at the time of research in one site in Oddar Meanchey, two villagers were being held in jail awaiting trial for ‘disrupting the peace’ because they asked the police to investigate logging in their CF by soldiers who were guarding a local ELC.149                                                   146 Interview with Executive Director of NGO, Phnom Penh, 16th December 2015. 147 Despite the hesitancy to use project funds for bribes to punish violence and the concern over how to report this to donors, in my previous experience of writing project proposals and budgets for Cambodian environmental NGOs in 2008-2010, I was asked to increase some budget lines in order to accommodate the necessary ‘cost of doing business’ with local political leaders, which sometimes means paying for dinner and drinks but sometimes means paying ‘unofficial taxes’ or bribes in order to, for example, get CF boundaries approved. This indicates that while NGOs feel it is acceptable to use money to gain project approval, they are less willing to take such risks to ensure project success once the project is established. 148 Interview with CFMC, Preah Vihear, 5th December 2015.  149 Another villager was also arrested in the same case but was released after two days “because he was old and blind,” a reason that while benefitting the defendant, indicates the arbitrary nature of arrests. Interview with CF participants, Oddar Meanchey, 11th June 2015.  164 Not only do interviewees purposefully refrain from reporting violence to law enforcement authorities but, interview data suggests, those who experience violence do not always recognize their right to do so or their right to have responsive law enforcement and an effective criminal justice system. When interviewees were asked if they had reported violent incidents to anyone, several interviewees showed confusion or began telling me that they had reported illicit forest use to the FA or Village Chief. Upon clarifying that I was interested in whether physical violence had been reported, these interviewees inevitably responded that they had not reported it to anyone. One interviewee responded by asking my Research Assistant, “Why would I do that? Does your kid report to your wife every time you beat it?”150 suggesting that not only are some forms of direct violence considered routine151 but also that it is not widely understood that it is illegal and could be formally punished. These interviews suggest that there is a conflation of crimes against the environment and crimes against the person and that reporting is inhibited by a lack of understanding of their rights to personal safety and freedom from violence.  Recognizing that CBFM participants are not always aware of their rights or laws criminalizing direct violence, five of the eighteen surveyed NGOs have addressed the issue of violence in training on human rights and laws.152 Three of these NGOs conduct the training themselves as part of their programmes on indigenous rights protection and invite all interested CBFM members. Two NGOs hire external human rights specialists to conduct the training since they do                                                 150 Interview with CFMC member, 16th June 2015. 151 See Brickell (2017) on widespread nature of domestic violence in Cambodia and the failure of the law to protect victims. 152 The other thirteen NGOs all conduct training on human rights but focus on environmental rights and laws regarding CBFM establishment and do not explicitly cover rights to personal safety or freedom from violence or relevant laws – an important distinction to make.  165 not have the internal expertise.153 Only Management Committee members attend this training and are expected to “disseminate learnings” to other non-executive CBFM members.154 They propose that training on human rights is essential in building CBFM participants’ knowledge of and confidence in their right to live peacefully and without threat of violence. NGOs also suggested that such training begins to challenge the Cambodian cultural norm of social hierarchies in which one’s social or ‘moral’ superiors cannot be challenged (c.f. Harris, 2005).155 This training is intended to empower communities to report violence to law enforcement authorities, without which crimes will not be investigated. As the only human rights based approach currently implemented, it is important as part of a full strategic framework to address violence. Although its effectiveness is limited if not supported by structural and operational reforms in the justice system, such training on human rights influences how CBFM participants think about violence and their rights, which I demonstrate and discuss further in the next chapter.  5.4.3 Alternative ways of implementing the rule of law Finding that the justice system is ineffective in preventing or responding to forest-related violence, communities and NGOs have implemented alternative responses. Nine CF groups and one CPA group attempt to reduce violence by paying soldiers, FA officers, or Forest Rangers to join patrols using money from NGO-supported project budgets, sale of seized timber, and voluntary donations by patrol members willing to pay to protect themselves or participating                                                 153 Survey of NGOs; participant observation of training, Oddar Meanchey, 17th June 2015. 154 Survey of NGOs; participant observation of training, Oddar Meanchey, 17th June 2015. 155 Interview with NGOs, Ratanakiri, 21st November 2015; Oddar Meanchey, 9th December, 2015.  166 family members.156 This reflects the UNDP’s (2016b) assertion that corruption in the justice system erodes public confidence in it and leads them to seek alternative ways of enforcing the law (c.f. Honey-Rosés, 2009). Military accompaniment is an attempt to force illicit forest users to respect the rule of law. Soldiers and armed Forest Rangers are primarily recruited in response to military loggers and migrants, both of whom commonly carry guns. One CF also employs soldiers to join patrols against local loggers who have begun arming themselves with homemade guns.157 The accompanying soldiers may have family ties to the community or may be outsiders who have shown an interest in forest protection.158 The CBFM groups report that accompanying soldiers deter the use of violence by creating a threat of retaliation – physical violence is a greater threat to illicit forest users than the justice system. Military accompaniment is seen as preferable to the patrol members carrying guns themselves because patrol members do not want to encourage the proliferation of guns in their villages (c.f. Roe, 2015) and, they argue, loggers would kill a civilian but are unlikely to shoot a soldier since they would be punished to the full extent of the law.159 This assessment by CFMC members reflects their acknowledgment that the justice system and the ruling elites who control it do not protect or value all lives equally.                                                   156 FA officers join CF patrols since the Forestry Administration has jurisdiction over CF sites while Forest Rangers join CPA patrols since the Ministry of Environment has jurisdiction over CPA sites. No CLT groups are known to pay state officials to support them since they do not often engage in formal forest patrols. 157 Interview with CFMC, Oddar Meanchey, 13th June 2015. 158 Conversation with Head of CF during military-accompanied patrol, Oddar Meanchey, 12th December 2015. 159 Conversation with Head of CF during military-accompanied patrol, Oddar Meanchey, 12th December 2015.  167  Figure 5.4 A soldier joins a patrol in the Monks’ Community Forest, Oddar Meanchey, 12th December 2015.  FA Officers are also occasionally paid per diems to accompany patrols to remove migrants or local loggers from CF sites. Although FA Officers are not armed, patrol members believe they increase the appearance of authority over forests. CF groups also suggested that their presence can reduce the occurrence of forest-related violence because they act as witnesses of violence to whom police are more likely to respond than reports made by “a normal person” due to their “good connections.”160 In this case, CBFM groups still entertain the idea of using the justice system but they have to try alternative ways to get it to work in their interests.  Although paying law enforcement officers is expensive, they are often used to support evictions of migrants, who are often violent and can clear large areas of forest in a relatively short period                                                 160 Interview with patrol participant, Oddar Meanchey, 25th May 2015.  168 of time since if one household is allowed to establish within a community-managed forest, others soon follow.161 The immediate financial cost is seen as an investment to protect the forest in the long-term. However, the extent to which this tactic deters or successfully punishes violence is questionable. Violence does still occur when FA Officers join patrols (see Chapter 4), none of which have been prosecuted. Yet, participant observation of one patrol suggested this strategy might have some value. An FA Officer had been paid to join an overnight patrol to confront a group of migrants. The encounter was relatively calm and the FA officer and a small group of patrol members went ahead of the main group to survey a separate area of forest. As the rest of us later made our way back towards the village, approximately six men from the group of migrants blocked the path, surrounded us, and began arguing much more aggressively than they had done in the previous encounter. Although no more than pushing and shoving occurred, the patrol members were visibly shaken and one of the leaders told me that the aggressive confrontation would not have occurred if the FA Officer had remained with us.162   Regardless of the short-term effectiveness of the solution, militarizing forest patrols is not a sustainable solution to reducing violence in the forests. To do so is treating the symptom of the problem (direct violence) rather than the causes (structural violence and elites’ impunity to the law) (Duffy et al., 2015; Lunstrum, 2014). It displaces forest-related violence in time or space rather than eliminating it. For example, in at least two similar incidents, once the accompanying soldiers had left patrol groups back in the village after seizing illegally cut timber, military                                                 161 This response contrasts with the response to selective logging, which is often unwillingly accepted and for which many CBFM members no longer risk their own lives or their family’s safety, as discussed below. 162 Participant observation of a forest patrol, Oddar Meanchey, 25th May 2015.  169 loggers ambushed the patrol groups and reclaimed logs at gunpoint.163 In another incident in Oddar Meanchey, armed soldiers stormed the FA compound to retake timber and a truck the FA had confiscated (Peter & Aun, 2016) demonstrating that not only are soldiers willing to use violence but also that they are confident they outrank a government department in terms of political power. This indicates that ‘buying authority’ is not enough to undermine the system of political patronage and personalized power that organizes Cambodian society. Furthermore, some CF members worry that soldiers accompanying patrols use the opportunity to locate valuable trees that they can return to cut later,164 a distrust that reflects the military’s historical and contemporary abuses of power experienced by most Cambodian civilians.  NGOs supporting CBFM also sporadically work with journalists to publicize incidents of ‘spectacular’ forest-related violence occurring in CBFM sites. However, this practice is not widespread as it relies on good relationships with journalists who are willing to cover such controversial and politically-sensitive issues. Additionally, incidents of forest-related violence are rarely sufficiently ‘spectacular’ to make the news and are regularly overshadowed by far more gruesome incidents of direct violence elsewhere in the country or by current political events that affect more people. The present model of only covering the most severe examples of direct violence portrays each as a discrete, spontaneous, and unusual event thereby failing to expose the on-going violence and everyday intimidation that CBFM participants experience and the underlying processes.                                                   163 Interview with CFMC, Oddar Meanchey, 1st June 2015. 164 Interview with CFMC, Oddar Meanchey, 10th June 2015.  170 Several Cambodian human rights NGOs have issued press briefings and reports on violence perpetrated by ELC companies (ADHOC, 2013; Asian Human Rights Defenders, 2015; LICADHO, 2009), although this usually omits reports of forest-related violence by other actors such as illegal logging groups. An interviewee at a human rights organization lamented that such violence falls beyond the scope of their donor-funded programmes on land politics so they cannot dedicate much time or resources to investigating the issue,165 a response that reveals the way discourses around natural resource problems are framed and constructed.  5.4.4 Other ad hoc responses Since attempts to improve the rule of law to punish and deter direct violence have not been successful, CBFM groups have improvised alternative responses to reduce the likelihood that they will experience forest-related violence.   Some CBFM groups have adopted the strategy of publicizing the dates of patrols to reduce the likelihood of encounters and violence with local loggers who are part of their communities. Although this does reduce the incidence of forest-related violence as patrol groups and local loggers encounter each other less frequently, it does little to reduce forest degradation as logging is only postponed or displaced to another forest rather than prevented. Thus, the decision to use this strategy is an attempt to find a compromise between the environmental damage done by local loggers and the desire to minimize conflict within the community. It also reflects local awareness that their neighbours’ reasons for illegal logging are complex and not necessarily indicative of outright greed and disregard for the environment (Duffy & Humphreys, 2016;                                                 165 Interview with NGO, Ratanakiri, 1st December 2015.  171 Duffy et al., 2016). Notably, local loggers who do not use violence against CBFM groups also appear to cooperate with this strategy as it enables them to benefit from illicit activities without increasing direct conflict with neighbours. I witnessed this when one patrol I joined began with our convoy of ten motorbikes driving through the village with me prominently displayed at the front. A couple of hours into the patrol, we found our path blocked by very freshly cut small trees and branches, which the patrol group explained indicated that the loggers had been warned about our imminent arrival by people in the village (see Figure 5.5).166 Although this decision to ‘turn a blind eye’ to illegal logging by neighbours does leave tensions unaddressed, it enables both CBFM members and local loggers (their neighbours) to ‘save face,’ an important concept in Cambodian society (Chan & Chheang, 2008).                                                  166 Participant observation of a CF patrol, Oddar Meanchey, 24th May 2015. A similar incident occurred before a patrol in Kampong Thom, 10th July 2015.  172  Figure 5.5 Freshly cut trees blocking a patrol group’s path through a community forest  Several CBFM groups have taken the decision not to confront some illicit forest users in the future, especially organized logging groups and military loggers. Although this means some tree species are lost from their forest (and some damage is caused to surrounding trees by felling and transportation), these species have little direct value for local livelihoods and many interviewees believe that it is not worth risking their lives in order to protect individual trees. This pragmatic decision does not negate the significant emotional and spiritual value these trees hold for CBFM members. Specifically, there is concern that ancestral spirits that inhabit trees will be angered, bringing illness and misfortune to the village (c.f. Beban & Work, 2014).167 Nevertheless,                                                 167 Interview with CLT members in Ratanakiri, 18th November 2015, interview with CFMC, Kampong Thom, 30th June 2015, and conversation during patrol with CF members, Kampong Thom, 2nd July 2015. See also Beban and Work (2014).  173 violence against their cultural artefacts and sacred places is accepted in order to protect themselves and their families from bodily harm.   In the CFs where military bases have been established, CF activities have completely ceased, even where forest remains, because it is considered too dangerous to oppose the military. Yet, members of those villages continue to work for forest protection by joining CF activities in other parts of the province. They argue that this is necessary because, first, conservation efforts should not be motivated only by personal gain from the forest but by the greater conservation need, and second, they want to oppose political and economic elites’ domination of Cambodia’s forests. The motivations behind this continuation of their work are explored in more depth in Chapter 6.  In terms of responding to the physical and material consequences of violence, options are very limited. Professional medical attention is often unaffordable for patrol members so first aid is administered by friends and family unless injuries are serious. In one instance, an international donor provided $5,000 to cover emergency medical treatment when the Head of the Provincial CF Network in Oddar Meanchey was shot in the neck and shoulder by a soldier while on a forest patrol.168 Such support is not routine and the recipient questions whether it would have been offered if he were not such a prominent figure in Cambodia’s CF movement. No other instances were found in which development organizations offered financial or medical support to those who experienced forest-related violence. When forest-related violence causes damage to possessions, such as motorbikes, tractors and trailers, or, in the three arson cases, houses, minimal financial support is available. The house owners sought support from the local pagoda                                                 168 Interview with Head of Oddar Meanchey CFMC, Oddar Meanchey, 24th May 2015.  174 and were given timber that had been seized by illegal loggers in the CF but this was insufficient to replace the houses. There is a clear lack of support in this area.  5.5 Discussion: the role of development organizations in responding to forest-related violence and paths forward At present, no CBFM group or supporting development organization appears to have a coordinated strategy to reduce forest-related violence in Cambodia or to mitigate harms and support CBFM participants when violence does occur. Rather, responses are reactive, inconsistently implemented, and ad hoc. Consequently, they are ineffective in terms of reducing forest-related violence or the physical, emotional, or material harms it causes.   At the community level, this lack of strategic planning appears to reflect a feeling of powerlessness to change the underlying socio-political structures that contribute to direct violence, resignation to these risks being part of the cost of protecting forests, and a perception that they need to maintain the support of development organizations. These dynamics are explored further in Chapter 6.   Among supporting NGOs, this inconsistency and lack of strategic planning also appears to reflect the influence of the authoritarian, neopatrimonial political system but also illuminates problems within the broader conservation-development aid industry. Conservation-focused NGOs are unwilling to work outside of or push the boundaries of their environmental remit by addressing the overtly political issues of patronage and corruption as a means to tackle some of the ‘root causes’ of forest-related violence. In part, they attribute this conservatism to restrictions  175 placed on them by donors (c.f. Brockington & Duffy, 2011; Coventry, 2016; Frewer, 2013), such as when a donor only supports projects that fit within a specific theme. Consequently, proposed activities must directly and clearly contribute to that theme so, for example, human rights training may not be considered a relevant activity within a conservation project. Relatedly, there is then little scope to redirect funds to pay for medical care for a participant who is injured during CBFM activities, much less pay for material losses such as damage to property or recovery after an arson attack. Yet, despite the complementarities between environmental and political activities, the needs of CBFM participants are not always well communicated to donors by NGOs and problems of forest-related violence are not made clear. As one NGO employee observed, “we cannot tell the donors all the problems that happen in the project because they will think we are failing and will not want to support our next proposal.”169, 170   NGOs also succumb to pressure from local political allies (such as cooperative District and Provincial Governors and FA staff) to stay away from political issues such as human rights or political reform and therefore do not propose budgets to donors that include political activities. They fear that disrupting the political status quo will make it harder to get approval for future project activities such as designating CF boundaries, thus both failing the communities they intend to help and, by failing to demonstrate quantifiable successes for donors, risking the future of the NGO.171 NGOs are also subject to pressure from the government at the national level, which can use legal-rational mechanisms to constrain NGOs’ actions and protect the interests of                                                 169 Interview with NGO, Kampong Thom, 1st July 2015. 170 The extent to which this is true was not ascertained in this research but this interviewee indicates that violence is a sensitive subject and not readily discussed among development organizations. 171 Interviews with NGOs, Kampong Thom, 1st July 2015; Phnom Penh, 14th July, 2015; Ratanakiri, 17th November 2015.  See also Biddulph (2011a, 2011b) on NGOs’ tendency to only promote CFs where they face least political resistance rather than where they would really be needed.  176 the ruling elite. The Law on Associations and NGOs (LANGO), which came into effect in August 2015 in the middle of this research project, can be and has been used to revoke the registration of NGOs that engage in political action or support citizens to do so (Amnesty International, 2017c; Mech & Ananth, 2017).172 LANGO is likely to deter NGOs from responding to violence connected to ruling elites and thus makes it easier for those elites to use direct and structural violence with impunity.173 At the personal level of NGO employees, fear for personal safety, exacerbated by low incomes, poor medical insurance, lack of financial savings, and sometimes low job security, mean they are not personally motivated to direct an environmental NGO’s activities towards more difficult, conflictual political issues.174 Ultimately, the neopatrimonial system shapes NGOs’ agendas, constrains their behaviour, and limits the possibility for effective responses through existing legal mechanisms.   This illustrates two important dynamics that prevent NGOs from more strategically planning how to respond to forest-related violence. First, NGOs are part of the neopatrimonial system so they are unable to oppose political elites whose cooperation they need in different circumstances. Second, they are dependent on maintaining an appearance of success in order to secure future funding from donors (who are also somewhat beholden to maintaining a good relationship with political elites).175 Consequently, NGOs are disincentivized from recognizing or responding to forest-related violence since it is easier to ignore forest-related violence, make it invisible, or portray it as inevitable rather than trying to respond and fail or to have only negligible, intangible                                                 172 The Interior Ministry used the Law on Associations and NGOs to dissolve the Cambodian environmental activist NGO ‘Mother Nature’ in August 2017. 173 Survey of NGOs. This law was a cause of significant concern for all NGOs surveyed. See the Worldwide Movement for Human Rights (2015) for details of the restrictive nature of this law. 174 Interview with NGO employees, Kampong Thom, 1st July 2015, and Phnom Penh, 14th December 2015. 175 Interview with donor representative, Phnom Penh, 23rd July 2015.  177 success. Thus, misrecognition of forest-related violence is not simply an oversight or effect of living in a society where violence of many kinds is pervasive but misrecognition is also functional.  Despite these explanations and justifications, this is an insufficient and irresponsible way to manage what can be considered a known risk to the lives, physical and emotional well-being, and human rights of project participants as well as to the success of projects’ conservation goals. I do not propose that development organizations can eliminate forest-related violence – the ‘root causes’ run far too deep and are far too complex to suggest that – but donors, implementing NGOs, and international governments that actively encourage ‘local communities’ to confront illicit forest users and that strategize how to increase participation in conservation projects in order to increase global benefits, not just local ones (e.g. Awono et al., 2014), should recognize all the risks that could and do cause harm to participants, including forest-related violence. If development organizations claim their interventions are intended to improve the well-being of local people, they must also recognize how these interventions can expose those same local people to harm and must plan to mitigate those risks accordingly. Failing to recognize these harms or accepting them as unavoidable and inevitable is an act of symbolic violence, as conceptualized by Bourdieu (1977) in that it naturalizes injustice and thereby makes it invisible.   This raises the important question of how can CBFM participants and NGOs respond to forest-related violence? Are there ways in which these processes of violence can be disrupted and incrementally less-harmful outcomes sought while working within the difficult context of  178 pervasive neopatrimonialism, corruption, and violence? The data presented above provides some insights into potential areas where positive changes can be made.   First, given that risks can be better discussed and addressed when they are known, monitored, and analyzed to identify areas of specific concern, CBFM groups and development organizations should improve the documentation of forest-related violence. This includes recording details about when and where violence occurred, in what context, who was involved, what happened, medical expenses incurred, and legal procedures initiated. This information should be shared between communities, implementing NGOs, and donors as a first step towards recognizing forest-related violence and planning for the design, consistent implementation, and financing of appropriate responses. Importantly, neither NGOs nor donors should construe this information as a sign of ‘failure’ but rather as qualitative information on the socio-political situation in which their project is implemented and a risk that should be addressed. Improved understanding of forest-related violence and its consequences can be used to stress the necessity of linking environmental rights to political rights.  Second, given that journalists and the media play important roles in holding governments, political leaders, and businesses accountable for the outcomes of their actions, NGOs and communities could more consistently seek coverage of the experiences of CBFM participants by issuing press releases and reports on the forms and processes of violence in Cambodia’s forests. To avoid over-saturation of the media in small reports, long-term trends could be drawn from data collected by communities. Such coverage would not only contribute to the rule of law but  179 also would promote human rights and challenge assumptions that such violence is normal, inevitable, or acceptable.  Third, development organizations could more consistently support training on human and political rights during the establishment and running of CBFM projects, including their right to freedom from violence and right to effective law enforcement, and how to advocate for those rights when dealing with law enforcement authorities. Improving participants’ understanding of these topics would not only complement CBFM’s stated support of rights-based development but could also help Cambodian citizens demand more accountability from their elected officials and engage marginalized communities in the electoral process. This may necessitate coordination with local human rights organizations.  Fourth, in order to reduce the harms of direct violence when it does occur, NGOs should budget for and donors should support cash reserves to pay for medical care for injuries and contribute towards material losses such as damage to motorbikes and trailers. Professional support for emotional trauma may be appropriate in some cases, although such services are of limited availability in Cambodia. Such financial and psychological support would have been invaluable for the CBFM members whose homes were destroyed in retribution for their involvement in CBFM.   The difficulty of eliminating forest-related violence should not prevent development organizations from taking steps to better recognize it and support those who experience it even if those steps only make incremental improvements. Rather, those planning and managing CBFM  180 projects should take as a starting point the reality of violence in all its forms as historically constructed injustices that are reproduced in the present day and seek a pragmatic approach to upholding human rights by identifying where failings occurred and what actions can be taken to remedy or mitigate the harms caused by specific expressions of violence. They should not only seek improved rights (e.g. environmental and economic rights) as an outcome of conservation and development projects, but should also seek to ensure that the implementation process of those projects – in itself – also upholds and promotes participants’ rights wherever possible.   181 Chapter 6: Growing political: violence, community forestry, and the shaping of ‘environmental’ subjectivities   “A violent structure leaves marks not only on the human body but also on the mind and spirit”  - Galtung, 1990, p. 294.  6.1 Introduction The previous chapters examined how direct, physical violence is used to prevent participants in Community-Based Forest Management (CBFM) projects from effectively protecting forests. I argued that direct forest-related violence is made possible as a means of control by the structural violence engendered by neopatrimonialism. Furthermore, symbolic violence normalizes both direct and structural violence and perpetuates their use, creating a socio-political system in which forest-related violence is accepted by development organizations as an inevitable, if unfortunate, outcome of CBFM in Cambodia. In those chapters, I emphasized the ways in which forest-related violence is used as a repressive power – a means to disperse, silence, and isolate citizens, disrupting the relationships that tie them together in belief and action. Such use of violence is an important factor in the outcomes of CBFM in Cambodia. Yet, to assume repression and discipline are its only effects risks characterizing those who experience violence as ‘victims’ and obscuring or denying their agency to analyze, respond to, or resist violence. Thus, in this chapter, I move from considering direct violence as an act against the external, physical body to, instead, exploring how such violence is internalized and interpreted by those who experience it. I explore why people participate in CBFM despite significant threats to their personal safety, even their lives, and minimal ‘success’ in achieving environmental protection or securing livelihood  182 resources, behaviour that appears to contradict traditional common-pool resource management theory, which predicts that economically-rational actors are only motivated to participate in sustainable natural resource management if personally-beneficial arrangements can be established via a suitable arrangement of institutions, people, and resources. I argue that people not only participate in CBFM despite direct violence, but also participate because of the structural and symbolic violence that is illuminated by experiences of direct violence. Such experiences cause people to re-assess how they feel and think about themselves in relation to others – that is, their subjectivities – and how they perform those subjectivities through CBFM. In short, experiences of direct violence can re-affirm, re-configure, and re-politicize subjectivities, giving new meaning to environmental behaviour such as participation in CBFM. This approach furthers the theorization of geographies of violence by exploring how violence constitutes, constructs, and acts on and through people and places.   The first section of this chapter brings into conversation three concepts informing my analytical framework: violence, subjectivity, and performance. Beginning with a review of the theory of violence discussed in Chapter 1, I draw on feminist scholars who conceptualize violence as a deeply personal experience of an expression of power that can transform how we think about and act in the world. I then explain the related concepts of subjectivity and performance, and review how they have previously been used to understand behaviour and engagement with natural resource management and conservation. Reviewing insights from these studies, I highlight how the concept of subjectivity can help trace the way violence shapes environmental behaviour through a focus on the internalization of experiences and their expression in action.    183 In the second half of the chapter, I explore how forest-related violence has contributed to the making and re-making of environmental subjectivities in Cambodia and how these are performed through CBFM activities. I begin by exploring how development organizations have sought to shape environmental subjectivities by hailing Cambodians into the role of ‘environmental defender’ through CBFM project discourse. I then draw on qualitative data to illustrate how experiences of forest-related violence have shaped these environmental subjectivities. I suggest that experiences of forest-related violence not only re-affirm environmental subjectivities but also re-politicize them, transforming the practices of CBFM from seemingly ineffective charades into performances of “rightful resistance” (O’Brien, 1996, p. 33). I conclude with a discussion of the value of this approach for enquiries into the geographies of violence, focusing on the transformational impacts of violent processes on emancipatory environmental subjectivities and a deepened sense of purpose.  Data used in this chapter were collected through qualitative interviews conducted between May 2015 and February 2016 across thirty-two Community Forestry and Community Protected Area projects in three Cambodian provinces – Kampong Thom, Oddar Meanchey, and Preah Vihear. This chapter does not discuss participation in Communal Land Title (CLT) projects because participation in CLTs is not entirely voluntary176 and, as described in Chapter 4, does not usually                                                 176 Despite the apparent 'opt-in' model of CLTs, participation is neither completely voluntary nor determined only by the legal-rational logic behind this form of natural resource management. Rather, participation is driven by the insecurity generated by neopatrimonialism and the need to maintain personal relationships through which rights and benefits can be claimed. This becomes evident when examining reasons for which individuals and villages pursue CLTs. If a household chooses to exclude the land they use from a CLT, they cannot register as indigenous or claim any other benefits arising from that identity such as scholarships for children or, in some communities, voting on village issues. If a household or village chooses not to pursue a CLT, they may apply for private land titles but this is a complicated and expensive process that few pursue without patronage from an NGO (Grimsditch & Schoenberger, 2015). If a village does not have either a CF or CLT, there is no other means of securing rights over forested land and they are vulnerable to forced dispossession, a form of insecurity generated by neopatrimonialism and  184 entail forest patrols and so does not carry the same risks of physical violence or financial costs. Detailed explanations of the methodology and interview methods can be found in Chapter 3.  6.2 Conceptual underpinnings: Violence, subjectivities, and performance This section reviews theories of violence and introduces the concepts of subjectivity and performance. These three concepts are brought together in a theoretical framework which can be used to understand how violence impacts subjectivities and with what consequences for the practice and meaning of CBFM activities.  6.2.1 Revisiting concepts of violence Theoretical analyses of violence typically distinguish between direct, structural, and symbolic (also referred to as ‘cultural’) violence (Bourdieu, 1977; Galtung, 1969, 1990). Direct violence is often thought of as an observable “event” with identifiable, responsible actors (Galtung, 1990, p. 294). In contrast, structural violence can be thought of as a “process” in which there may not be an identifiable actor (Galtung, 1990, p. 294). Farmer et al. (2006, p. 1686) propose that these processes result from social, political, and economic arrangements of our world that “put individuals and populations in harm’s way.” Nevertheless, some individuals may purposefully work to keep these structures in place to perpetuate inequalities and thus structural violence should not be assumed to be a ‘natural’ phenomenon. Rather, Galtung (1990, p. 294) proposes that cultural violence is “a permanence” that “makes direct and structural violence look, even feel, right – or at least not wrong” (p. 291). Similarly, Bourdieu (1999) describes how “symbolic                                                                                                                                                        kleptocracy (Milne, 2015; Un & So, 2011). Consequently, an individual's participation in a CLT is not only based on preferences over natural resource management but also on the need to maintain social status and relationships in a community in order to access certain benefits and claim rights that are only accessible by leveraging personal relationships.  185 violence” is used to legitimize direct and structural violence, or make reality opaque, leading to the misrecognition of violent acts, processes, and systems. Thus, structural and symbolic violence are key characteristics of neopatrimonialism, being both outcomes of that socio-political system and tools purposefully employed by powerful elites (Allen, 1999).   These heuristic categories of direct, structural, and symbolic violence have broadened understandings of how violence manifests and causes suffering in various ways, often unequally distributed across society or unseen or misrecognized completely (Bourdieu, 1999; Farmer 2004; Nixon, 2011). Exploring the relationship between direct, structural, and symbolic violence, Springer (2009a, p. 368) conceptualizes episodes of violence as “moments,” or analytical windows, through which to see the coalescence of and dynamic relationships between processes, “things” (Harvey, 1996, p. 50), and the bodies upon which violence is exerted. This approach suggests that ‘moments’ are linked across space and time and analysis of what appears to be direct, localized, personalized violence can provide insights into related structural and symbolic violence. This framework has been used to study the geographies of violence, including violence related to natural resources, showing that although direct violence may manifest as a site-specific phenomenon, it is rooted in networked social relations, local and global histories, and power dynamics (Peluso & Watts, 2001; Springer, 2011b).   Such approaches have provided critical insights into how social and spatial relations contribute to violence, which, as I argued in previous chapters, is central to the recognition of violence and ensuring that it is not perpetuated by its misrecognition (Bourdieu, 1999; Schlosberg, 2007). However, to focus on identifying violence and its causes is to risk conceptualizing violence as a  186 ‘product’ or the terminus of power – sporadic, singular, or serial episodes resulting from other processes, or assuming analysis is complete once a cyclical and self-perpetuating process of direct, structural, and symbolic violence has been recognized. This, feminist geographers argue, can deny the agency of those who experience violence, portraying them only as ‘victims’. It can obscure the ways in which violence is embodied as a deeply personal experience with the potential to transform the way we conceptualise ourselves, relations to others, and the places we create (Fluri, 2009; Koopman, 2011). Robben and Nordstrom (1995, p. 5), for example, argue that the social relations of violence cast “ripples that reconfigure lives in the most dramatic of ways, affecting identity in the present, the hopes and potentialities of the future, and even renditions of the past” (see also Baird & Le Billon, 2012; Das et al., 2000). Violence is cumulative and boundless, it “spills over” into other aspects of life, structures beliefs, actions, and relationships, and “creates and recreates new norms of collective self-understanding” (Lawrence & Karim, 2007, p. 12; Scheper-Huges & Bourgois, 2004; Tyner & Inwood, 2014). Consequently, feminist geographers (Das, 2008; Fluri, 2009) often draw on Foucault (1996) to suggest that rather than conceptualizing violence as isolated exemplars of a ‘thing’ or an “event” (Galtung, 1990, p. 294), violence itself can be understood as an expression of or a social relation imbued with power (Schröder & Schmidt, 2001).   Conceptualized in such a way, violence is seen as simultaneously both “structured and structuring” (Dumont, 1992, p. 139; Bourdieu 1971). As a social relation it is ‘structured’ and created by specific socio-political and historical contexts, yet it is also ‘structuring’ and generative of thought and action (Grenfell, 2008). This suggests that violence cannot be understood in terms of its physicality alone. Rather, it is the social and cultural dimensions of  187 violence that give violence meaning when it is internalized (Scheper-Hughes & Bourgois, 2004; Tyner & Inwood, 2014). Conceptualizing violence in this way opens up the possibility for inquiry into the way it (re-)shapes the materiality and meanings of our worlds. It draws attention to the way experiences of violence – direct, structural, or symbolic, observable or transparent – structure people’s situated knowledge of themselves and others, and, correspondingly, their interactions with the world in the context of violence. In other words, violence has the potential to shape subjectivities – a concept explored in the following section.   6.2.2 Subjectivities and performance Subjectivity refers to one’s sense of holding a subject position in a specific place, time, or set of relationships (Morales & Harris, 2014). Butler (1997) proposes that a subject position is an effect of power. Drawing on Althusser, she argues that subjects are “hailed” or “interpellated” into subject positions by actions, discourses, symbolism, histories, socio-spatial relations, and other expressions of power experienced in everyday life (Butler, 1997, p. 96). She builds on Althusser’s infamous example of interpellation through language: a policeman hails a person on the street, who turns and recognizes himself as the one who is hailed. In this exchange, the offering and acceptance of recognition discursively produces the social subject. Subjectivity is the internalization of this subjection, acknowledging the power relations that shaped that subject position (Butler, 1997; Nightingale & Ojha, 2013). Rose (1996, p. 37) conceptualizes this as the “infolding of exteriority,” in which the effects of power become another “fold” in the individual’s “interiority.” The metaphor of the fold emphasizes the continuity in the process of internalization and possibility for multiple “vectors” of power to act simultaneously on subjectivities (p. 37). Each experience is interpreted by and incorporated into an already-existing  188 subjectivity. As power relations change through time and space, experienced as embodied, discursive, material, and social processes, so subject positions and subjectivities are re-shaped (Longhurst, 2003). Thus, subjectivities are considered relational (Probyn, 2003).   Butler theorizes that not only are experiences of power relations translated into an individual’s relational constructions of themselves and their relationships with others, but the resulting subjectivities are performed through their actions. She proposes, power “makes the subject possible… and is taken up and reiterated in the subject’s ‘own’ acting” (1997, p. 14). Thus, although subjectivity may not be verbally articulated, or even articulable (Pile, 2008), performances – what individual subjects “do, say, or ‘act out’” (Gregson & Rose, 2000, p. 434) – can provide an insight into how that individual has interpreted and internalized an expression of power used towards them.   The concept of subjectivity can help to expose and trace how power operates throughout societies. Central to this concept is the notion that power has multiple potentialities, being at once disciplinary, repressive, and generative. A subject may be hailed with the intention of invoking the performance of specific actions and thus, following Foucault, Butler (1997, p. 2) shows that power can subordinate and form the subject through domination. Consequently, subjectivities can give the appearance of being stabilized due to hegemonic psychosocial and socio-spatial processes that continually re-articulate and re-affirm these subjectivities (Butler, 1997; Butler & Athanasiou, 2013; Massey, 1994; Probyn, 2003). This resonates with Bourdieu’s (1998) theory of symbolic violence in that hegemonic ideologies normalize injurious social orders and processes of subjection. Yet, Butler (1997, p. 2) argues, power can also be generative  189 as it simultaneously provides “the very condition of” the subject’s existence, “agency,” and potential for “resistance,” an effect she calls the “recoil” (p. 6) of power. For example, an interpellation may be “heard or interpreted as an insult” or an act of symbolic “violence” (Butler, 1997, p. 96), depending on the context. Expressions of power intended to dominate or discipline can be internalized then subverted, resisted through performances other than those that might be expected, or by giving specific meaning to performances. Consequently, there is some degree of “instability and unpredictability” (Butler, 1997, p. 96) in the appropriation of a subject position and subsequent performances with unanticipated effects when individuals reject an injurious subject position (Freire, 2000 [1970]; Mahoney & Yngvesson, 1992; Weeks, 2007). Butler (1997, p. 99) refers to these consequent actions as “performative” and “citational” practices, in that they reproduce and/or subvert discourses and enable (rather than discipline) subjects. Thus conceptualized, subjectivities are not the fixed, end products of power but sites of potential where power can be contested, transformed, and performed anew.  6.2.3 Subjectivities and performance in community-based forest management  The concepts of subjectivity, and, relatedly, emotion and affect, have proved useful to NRM researchers and practitioners frustrated that the rational choice, ‘homo economicus’ model of human behaviour does not fully account for NRM behaviour in all contexts (Heras & Tàbara, 2016; Morales & Harris, 2014). Traditional common-pool resource management theory predicts that economically-rational actors are only motivated to participate in sustainable natural resource management if personally-beneficial arrangements can be established (Baland & Plateau, 1996; Hardin, 1968). This model has been nuanced by accounting for the influence of the specific characteristics of resources, users, or institutions (Agrawal, 2001; Demsetz, 2002; Hajjar et al.,  190 2016; Ostrom, 1990). Within this paradigm, low participation or adherence to rules regarding resource use continue to be seen as a result of poor institutional structures that either do not provide or allow people to access sufficient economic benefits, do not fit local perceptions of justice, or are not well designed for local livelihoods or cultural preferences (Nurse & Malla, 2006; San 2006; Sunderlin et al., 2005). Accordingly, ‘solutions’ tend to promote improving economic incentives (e.g. by integrating income generation schemes into natural resource management projects through ecotourism, community-based forest enterprises using NTFPs or timber, or providing payments for ecosystem services) (Angelsen & Wunder, 2003; Molnar et al., 2007), or re-configuring institutions to better reflect local livelihood priorities and forest resources (Martin et al., 2013; Oldekop & Hajjar, 2017; Ribot et al., 2006).   Although this work has improved the design and outcomes of many CBNRM projects, such approaches give little consideration to the agency of natural resource users or non-economic reasons for (non)participation. The ‘self,’ central to the economic model of rationality and self-interest, remains largely unquestioned and is assumed to be immutable and predetermined (Singh, 2013). Thus, the economic rationality model risks projecting misleading, stereotyped, essentialist identities, beliefs, and preferences onto ‘categories’ of people and places, such as atavistic attachments to ‘nature’ or ‘ancestral places’ of ‘indigenous’ or ‘peasant’ people. How social relations (including violence) shape individuals’ decisions and behaviours is ignored. This approach not only delegitimizes the emotions and subjectivities reflected in (non)participants’ decisions but also depoliticizes the decision-making context and reaffirms the supposedly objective, scientific and economic rationality (and implied superiority) of NRM projects. The policies and project designs resulting from such depoliticized, technocratic discourse are unlikely  191 to meet the needs and preferences of participant communities (Mosse, 2004; Scott, 1998). Consequently, rational choice models cannot explain behaviour that appears to be ‘irrational’, either when potential benefits are foregone due to lack of action or when people do engage in environmental care despite it being costly in terms of money, time, or, as in the context of forest-related violence, personal safety.   Where these economic models of human behaviour have fallen short, studies of subjectivity, emotion, and affect have provided valuable insights into people’s NRM practices. Three key themes come out of this literature: the emergence of new subjectivities through NRM, the performance of existing and newly emerging subjectivities through NRM, and the re-appropriation of natural resources as symbols of subjectivities.   The emergence of new “environmental subjectivities” through place-based, organized labour has been observed in studies of CBNRM. Sundar (2001, p. 333) argues that CBNRM inherently requires participants to “buy into” new environmental subjectivities in order for them to enact the behaviours they have been taught to value. Likewise, Pratt (1998, p. 28) asserts that particular places, such as workplaces or protected landscapes, “not only enable but exact the performance” of particular subjectivities. Agrawal’s (2005, p. 9) ethnographic study of CBFM schemes in India traces how technologies of governance, narratives of historical environmental use, and socio-political relations are employed to hail environmental subjectivities that encourage environmentally beneficial behaviour. As people participate in CBFM, their conceptions of themselves as “environmental subjects” grow and they “come to think and act in new ways in relation to the environment” (Agrawal, 2005, p. xiv). However, there is a tendency to emphasize  192 the disciplinary potential of power in shaping subjectivities, and the technologies through which it is exerted, thus obscuring individuals’ multilayered, diverse, and complex biographies that also inform subjectivity as well as the potentially generative “effect of power in recoil” identified by Butler (1997). Consequently, an ‘environmental subjectivity’ does not only reflect a relational position to the environment but also reflects a relational position in historicized, gendered, classed, racialized, and politicized socio-political network.  In an effort to bring focus back to individual agency and the “process of ‘becoming’,” Singh (2013, p. 190) suggests the concept of subjectivity can be complemented with Spinoza’s theory of affect and Hardt and Negri’s (2000) concept of affective labour. Singh (2015, p. 53) finds that forest protection can be “joyful and life-affirming” work for participants, through which they develop affective relations with the forests and develop new subjectivities as environmental defenders. Other researchers have explored how participation in NRM provides opportunities for the performance of subjectivities, including subjectivities not based on environmental values. For example, Nightingale and Ojha (2013) find that by participating in NGO-supported forest management, Nepalese villagers express newly developing political subjectivities, which enables them to resist patronage and define new political systems. Similarly, existing subjectivities can be performed through NRM activities. Lau and Scales (2016) find that oyster harvesting in The Gambia is not only a livelihood but also a means to perform intersectional, gendered, classed, and historically-contingent subjectivities in order to assert a specific, higher social status in relation to alternative employment options. Natural resources can also be appropriated as symbols of specific subjectivities, with, for example, petroleum and citizenship in Ecuador (Valdivia, 2008). In these examples, the authors not only eschew an essentialist understanding of  193 subjectivity as static and predetermined, but also demonstrate that environmental subjectivities are continually shaped by concurrent and multiple vectors of power, of which environmental conflicts are just one.   These studies use the concept of subjectivity to trace how power is applied, internalized, and reproduced (being at once disciplinary and generative), in the context of an individual’s socially and historically situated life. In doing so, they echo Butler’s (1997) theory of subjectivity: the performance of subjectivities reflects established systems of knowledge and produces, re-affirms, and re-shapes subject positions. Thus, the concept of subjectivity illuminates reasons for participation in NRM that go beyond economic rationality.   6.2.4 Using subjectivities to understand the impacts of and responses to violence  The concept of subjectivity can be used to understand the impacts of and responses to the social relation of violence and, more specifically, why some Cambodians continue to participate in CBFM despite experiencing forest-related violence and minimal ‘success’ in securing environmental or economic benefits. A study of subjectivities focuses attention on three key issues that illuminate the generative and more-than-disciplinary effects of violence: the interpellation of specific subject positions through expressions of power, the internalization and interpretation of embodied expressions of power, and the subsequent performance and performativity of emergent subjectivities. By exploring the construction of subject positions after experiencing violence and the potential for those emergent subjects to act or act differently, the concept of subjectivity can be used to trace how the repressive and generative powers of  194 violence, among other forms of power, influence the behaviours of participants in CBFM projects.   Furthermore, subjectivity is a particularly useful tool for exploring the impacts of violence as, feminist scholars argue, the body is a key site for the production of knowledge, emotions, affect, and history, all of which fundamentally shape subjectivity (Probyn, 2003). The embodied experience is held as central to understanding the construction and performance of subjectivity and workings of power. Lived experiences, discourses, and the ideas they represent are “intimately interwoven” in subjectivities (Longhurst, 2003, p. 285; Pile, 2008). Consequently, the embodied experience of violence has the potential to shape subjectivity when it produces emotions, (de-)legitimizes knowledge claims, and (re-)produces relationships. For this reason, the concept of subjectivity is well positioned to explore the intimate, embodied experience of violence, be that direct or indirect.  I proceed to demonstrate this with reference to Community Forestry in Cambodia with the caveat that although this analysis focuses on how forest-related violence shapes environmental subjectivities and performances of thereof, forest-related violence is but one of many social relations that simultaneously hail the individual. Therefore, while this study illuminates emerging geographies of violence in Cambodian forests, it does not offer ultimate explanations of the behaviour of all community members or why all individuals chose to participate or not.177 Indeed, subjectivity theory suggests that such an achievement would be impossible given that each individual’s subjectivities are unique and thus experiences interpreted differently.                                                 177 Pasgaard and Chea (2013, p. 3