Why militarized conservation may be counter-productive: illegal wildlife hunting as defiance

In response to the past decade's surge in the illegal hunting of rhinoceros for the global trade in their horn, conservation authorities in southern Africa have increased the securitization of protected area conservation. Political ecologists have warned that violent, militarized approaches are counter-productive, because they fail to address the root causes of "poaching" and may provide the moral justification for those already marginalized by conservation to hunt wildlife illegally. This is a powerful set of critiques, but scholars have yet to provide an empirically derived explanation of how and why militarized conservation might contribute to, rather than prevent, environmental harms. This lack of explanation emerges from a concerning scarcity of scholarly attention to the people who are actually living with militarized conservation. In response to the increased monitoring and arrests that accompanied the securitization of Mozambique's Limpopo National Park (LNP), Park residents expressed fear, anger, and resentment. Yet they continued to engage in the illegal subsistencebased practices of fishing and hunting in the face of mounting enforcements against these practices. In interviews and public demonstrations, residents consistently questioned the legitimacy of rangers' actions and authority, neutralizing the problem of their own potential involvement in illegal activities by focusing instead on the unjust and morally questionable behaviors of rangers. According to defiance theory, environmental harms will increase as the legitimacy of conservation policies, tactics, and authority decline. As defiance figures into and informs residents' decisions to continue to engage in subsistence-based practices, it may also motivate and justify people's involvement in the more lucrative and harmful activity of hunting threatened and endangered wildlife for the global trade. This article demonstrates why there is a need to fundamentally rethink the increased securitization of protected area conservation and addresses the need for improved understandings of human agency in the face of militarized conservation.

The need for improved environmental security has also justified land grabbing for conservation, economic marginalization, societal abandonment, and dispossession (Massé and Lunstrum 2016;Ramutsindela 2016;Witter and Satterfield 2019). Thus, critical scholars have argued, militarized conservation relies on and legitimizes the employment of illicit, discursive and structural violence in pursuit of improved environmental protections (Büscher and Ramitsundela 2015;Duffy 2014;Duffy et al. 2019;Lunstrum 2014;Witter and Satterfield 2019).
The emergence of violent environmental security approaches in southern Africa's premiere "Peace Park", the Great Limpopo Transfrontier Park (GLTP), is a painful contradiction (Büscher and Ramutsindela 2015). The GLTP connects the Kruger National Park (KNP) in South Africa with the Limpopo National Park (LNP) in Mozambique and Goharezhou National Park in Zimbabwe (See Figure 1). Transfrontier conservation areas aim to achieve cross-border protections for biodiversity, sustainable economic development, and international peace and cooperation. Over the past decade, the GLTP emerged instead as a hotbed for the socalled "war on poaching" (Duffy 2014;Lunstrum 2014). This occurred as the hunting-related death rates of black and white rhinoceroses (Diceros bicornis, Ceratotherium simum) in the KNP skyrocketed from an estimated low of 10 in 2007 to 36 in 2008, to 425 in 2012, reaching a height of 892 by 2015 (DEFF 2015;Ferreira et al. 2019;Lunstrum 2014: Poaching Facts 2015. To kill the rhinos, illegal hunters have employed violent and dangerous techniques; some have reportedly also carried weapons to use against rangers (Lunstrum 2014).
In response to significant threats both to wildlife and to rangers, and in light of the fact that a significant portion of those moving into Kruger to hunt have traveled over the border from Mozambique (Hübschle and Joost 2017), conservation and state authorities have taken a militarized approach, focused on the securitization of the Mozambican borderlands (Lunstrum 2014;Massé and Lunstrum 2016;Ramutsindela 2016). Antipoaching measures have resulted in a dramatic slow-down of poaching-related traffic coming over the border from Mozambique, and a steady drop in the numbers of rhino deaths since 2015 (Hübschle and Joost 2017;DEFF 2019). 2 Despite these indicators of progress, several scholars have warned that the increased securitization of conservation may be counter-productive (Duffy 2014;Duffy et al. 2015;Duffy et al. 2019;Massé and Lunstrum 2016;Hübschle 2017;Witter and Satterfield 2019). There are several logics operating here. First, there is a narrow understanding of the root causes of the "poaching crisis." 3 If one assumes, for example, that wealth rather poverty drives hunting, then heavy-handed anti-poaching approaches that target the poor and the desperate may be miss their target (Duffy et al. 2015;Hübschle 2017). In the case of the illegal global trade in rhinoceros horn, criminal networks stretch from the consumers in Asia to the procurers in Africa (Milliken 2014;Haas and Ferreira 2018). Yet, violent anti-poaching enforcements have been directed primarily at the socalled "foot soldiers" who incur the most risks and least financial benefit from entering protected areas to make the kills for the commercial trade (Milliken 2014;Haas and Ferreira 2018). 4 Second and related, violent approaches to anti-poaching may exacerbate the need and moral justification for those already disenfranchised by conservation to engage in or otherwise tacitly support illegal hunting (Witter and Satterfield 2019). Thus, anti-poaching efforts may "backfire" in the long run (Massé and Lunstrum 2016: 236) and contribute to advancing the very harms that conservation has been implemented to address. The prospect that anti-poaching measures may undermine, rather than support, long-term conservation goals deepens the paradox of violence in Peace Parks. 2 In 2013-2014 an estimated 75% of poachers were Mozambicans, including those who resided in or moved through the LNP (Hübschle and Joost 2017). By 2017 that number had decreased to 30% (ibid). 3 Herein I sometimes use the term "poaching", in reference to the illegal hunting of wildlife for the commercial trade, aware that the term is problematic and as a means to explicitly conjure the problematic nature of the terminology. I also use the term "illegal wildlife hunting" to signal the need for a more neutral and objective treatment of the topic. See Duffy et al. 2016 for fuller discussion of the merits of the latter terminology. 4 Shoot-to-kill-practices are not an official component of South Africa's anti-poaching programming (DEFF 2015;Hübschle and Joost 2017), but these have been employed in the KNP. By 2015, an estimated (and contested) 200-500 hundred Mozambican deaths had been associated with rhino poaching (Smith 2015;Hübschle 2017: 439). This powerful critique has major consequences for threatened and endangered species as well as for those living with and working in protected area conservation. Yet, not least in my own work (Witter and Satterfield 2019), the dots have yet to be fully connected. That is, there is not yet an empirically derived explanation of how and why the increased securitization of protected area conservation might backfire and contribute to, rather than prevent, environmental harms. As a result, the warnings of critical scholars can be lost as "mere" scholarly provocation instead of elucidating the need to fundamentally rethink militarized conservation. This lack of explanation emerges from a worrying scarcity of scholarly attention to the agency of those living with militarized conservation.
In this article, I address this important dearth in scholarship, through an examination of the ways people residing in Mozambique's Limpopo National Park (LNP) experience and respond to increased securitization. Drawing from Erica von Essen and colleagues (2014), I argue that 'defiance theory' provides a key analytic for understanding why militarized conservation may be counter-productive and for understanding, more broadly, human agency in the face of structural constraints and injustices. Ethnographic research conducted in the context of increased securitization in the LNP reveals that anti-poaching measures -initiated in response to rhino poaching -have also been directed against subsistence-based practices. In interviews and public demonstrations against the Park, residents expressed anger, fear, and resentment in response to increased monitoring, intimidation, arrests, and imprisonment. As these enforcements ensued, residents became increasingly united in their defiance against conservation: a majority of residents continued to fish, some continued to hunt for the pot, and some engaged in strike action against the parks. Residents consistently explained and justified these actions by questioning the legitimacy of rangers' actions and authority. Defiance theory holds that "sanctions perceived as unfair by way of harsh and disrespectful treatment from the sanctioning agent or by a lack of procedural fairness will result in a delegitimization of authorities and furtherance of crime" (von Essen et al. 2014: 643;Sherman 1993). The idea, in short, is that environmental harms, including in the forms of illegal wildlife hunting, will increase as the legitimacy of conservation policies, tactics, and authority decline. As defiance figured into residents' decisions to continue to engage in illegal subsistence-based activities, it may also inform their decisions to take up or tacitly support other forms of illegal activity, including the hunting of threatened and endangered wildlife for the lucrative commercial trade.
Before I proceed, it is important to distinguish between the two predominant forms of illegal wildlife hunting in conservation decision-making. Commercial hunting or "hunting for business" often involves the use of firearms to kill highly valued megafauna that are protected from species-level threats of endangerment or extinction. Subsistence hunting or "hunting for the pot" generally involves the killing of smaller, less valued, and less at-risk animals. The hunting is undertaken by local resource users -often former residents of areas now designated as conservation reserves -using traps or snares (Duffy 2014). Both typologies of hunting are illegal, yet the public scrutiny attached to commercial poaching is far more intense (e.g., Büscher and Ramutsindela 2016;Lunstrum 2016). Where hunting for the pot is sometimes tolerable in conservation contexts, commercial poaching is untenable (Neumann 2004;Ramutsindela 2016). Yet, the distinctions between these typologies can become blurred; for example, when hunters' motivations are multiple (i.e., for income and for food) or when hunters begin hunting for food then opportunistically turn to illegal wildlife trafficking (Duffy 2014: 830;Duffy et al. 2016: 15). In the context of increased securitization, there is spillover, moreover, in the enforcements against the practices. Thus, while fishing, subsistence-based hunting, and commercial poaching are, in some respects, fundamentally different forms of illegal resource use, they are entangled in a broader context of increased securitization wherein residents' ideas about and experiences with one act overlaps with and informs their ideas about and justifications for the other acts.

Poacher Economomicus
Despite relatively widespread recognition that there are multiple possible drivers for illegal wildlife hunting (Forsyth and Marckese 1993;Kahler and Gore 2012;Muth and Bowe 1998) 5 , an extensive review of scholarship found "an overwhelming reliance in the literature on instrumental economic theories" (von Essen et al. 2014: 635). As von Essen et al. (2014: 635) pointed out, the predominance of economic explanations is, to some extent, "intuitive" given that the illegal wildlife trade is among the most lucrative in the word (Ayling 2013;Challender and MacMillian 2014). Economic explanations of poaching motivations in the GLTP context are justified, moreover, by the high levels of poverty and marginalization among those living near and within these protected areas (Duffy et al. 2016) and by the monetary value of rhino horn, which reportedly increased from approximately $4,700 per kg in 1993 to $65,000 by 2012 (Biggs et al. 2013(Biggs et al. : 1038. However, an overreliance of economic explanations in the literature presents a number of shortcomings, including narrow conceptions of how and why humans make decisions.
Despite relatively widespread recognition that there are multiple possible drivers for illegal wildlife hunting (Forsyth and Marckese 1993;Kahler and Gore 2012;Muth and Bowe 1998) 6 , an extensive review of scholarship found "an overwhelming reliance in the literature on instrumental economic theories" (von Essen et al. 2014: 635). As von Essen et al. (2014: 635) pointed out, the predominance of economic explanations is, to some extent, "intuitive" given that the illegal wildlife trade is among the most lucrative in the word (Ayling 2013;Challender and MacMillian 2014). Economic explanations of poaching motivations in the GLTP context are justified, moreover, by the high levels of poverty and marginalization among those living near and within these protected areas (Duffy et al. 2016) and by the monetary value of rhino horn, which reportedly increased from approximately $4,700 per kg in 1993 to $65,000 by 2012 (Biggs et al. 2013(Biggs et al. : 1038. However, an overreliance of economic explanations in the literature presents a number of shortcomings, including narrow conceptions of how and why humans make decisions. The instrumental perspective draws from the conventional economic logics of rational choice, utility maximization, and opportunity costs to suggest that "individuals are driven by self-interest and respond to immediate incentives and penalties associated with the crime" (Peterson et al. 2017;von Essen et al. 2014: 634). In the GLTP context, Haas and Ferreia (2018) created a model that predicts poaching behaviors based on the assumption that poaching decisions are a function of maximizing utility in the pursuit of two overarching goals: making money and avoiding prosecution. Messer (2010) assigned numerical value to a number of factors (e.g., time spent poaching or not poaching, the price for the poached animal, the probability of getting caught, and risk of death) then developed equations that model poaching choices according to these predicted constraints and opportunities.
This and other work assumes not only that individuals make "rational choices", but also that an individual's decision to poach can be deconstructed into a set of independent functions then reconfigured and expressed as a formula that predicts behavior. Conservation practice and policy then proceeds as if the decisions and behaviors of would-be poachers can (and should) be modified through external restructuring of costs relative to benefits. Thus, addressing the poaching problem becomes a matter of weighing up policies that increase the benefits (rewards and incentives) for not poaching against those that increase costs (risks and distinctives) for poaching. Scholars and decision-makers recognize that increasing the costs for poachers via heavy-handed enforcements bears significant problems ethically and in terms of political feasibility (Haas and Ferreira 2018;Hübschle and Joost 2017;Messer 2010), but they have, by and large, failed to acknowledge other factors. These include diverse values and norms; concerns about well-being, safety and control; and the feelings of guilt, shame, belonging, and alienation that inform decision-making (Sen 1994;Ortiz 2012;Büscher et al 2012;St John et al. 2015;Witter in review). This brings me to a second important shortcoming of narrowly Journal of Political Ecology Vol. 28, 2021 180 conceived economic explanations of poaching motivations: the dearth of meaningful critical engagement with the socio-political contexts in which peoples' decisions and actions play out. There is, in particular, substantive need for theories of hunting motivations that emerge from empirically informed understandings of the relationship between structure (the broader social, historical, and political economic contexts wherein hunting unfolds) and agency (how people experience and respond to structural constraints and opportunities) (Duffy et al. 2016(Duffy et al. , 2019von Essen et al. 2014;Hübschle 2017;Peterson et al. 2017).

Poacher Politicus
Scholars across disciplines have elucidated a number of scenarios, sites, and tactics where people dispossessed of land, forests, rivers, and other "environmental resources" continued to harvest, harm, or destroy them (e.g. Thompson 1975;Carruthers 1989Carruthers , 1993Peluso 1993;Jacoby 2001;Neumann 2004;Benjaminsen and Svarstad 2010;Kelly 2011;Mavhunga 2014). This work reveals a legacy of concerns about protected area conservation being counter-productive. It also demonstrates why understandings of socio-political contexts matter for making sense of hunting motivations. Three points merit emphasis here.
First, illegal wildlife hunting often means very different things to those living and working in conservation. Poaching is broadly conceived as a legal matter. However, when conservation authorities interpret the continuation of banned activities inside protected areas as "encroachment, or a failure of regulation", they obscure the political and moral meanings of the act (Holmes 2007: 193;Witter in review). Thus, Clapperton Mavhunga (2014: 53) argued that "part of the problem" of deeming livelihood practices "poaching and anathema to biodiversity conservation… lies in the interpretation of poaching as a legal issue rather than one of environmental knowledge." In southern Zimbabwe, poaching occurred when former inhabitants (forcibly removed from their land to create national parks) re-entered forests "in search of firewood, mushrooms, and other wild vegetables, grasses for thatching roofs, fishing, edible mopani worms, and meat" (Mavhunga 2014: 53). Such practices involved the interpretation of landscapes according to emic meanings and the reaffirmation of people's spiritual and material "yearnings" for land. Thus, people sought to care for one another in culturally significant ways by re-establishing access to that which was prohibited.
Second and related, illegal resource use can emerge as an important site for negotiating identity and control over resources. Jane Carruthers (1989Carruthers ( , 1993 demonstrated how the early 20 th century development of the KNP unified English and Afrikaaner settlers in the interests of white nationalism at the expense of dispossessed Africans. Even though incidents of poaching among Africans were reportedly low, conservation authorities "abhorred" African hunters, "not so much because of the danger they presented to wildlife, but because they represented freedom of action on the part of Africans and therefore a corresponding lack of white supremacy" (1989: 201). Meanwhile, Africans hunted in the KNP for subsistence and to protest white claims. As conservation authorities sought to impose and enforce restrictions on hunting, rural Africans performed their resistance to that authority by continuing to hunt in the face of those restrictions.
Third, what is of interest in cases of illegal resource use is not only that people engage in an activity after or despite its prohibition, but also that people's environmental discourses and practices take on new or reconfigured moral meanings. Karl Jacoby (2014) showed how in late 19 th century North America, conservationists in the Adirondacks reconfigured "many long-standing local practices as crimes: hunting as "poaching", the cutting of trees as "timber theft", foraging as "trespassing." Tactics to undermine conservation included the continuation of livelihood practices, practices that now harmed what conservationists aimed to protect. Thus "some residents took revenge on the forest itself, which to their eyes became a symbol of their newly dispossessed status." (Jacoby 2001: 337). Hunters gained the materials benefits "from a dead animal" while "simultaneously and implicitly making a statement that they have a right to kill animals" (Holmes 2007: 193).
Annette Hübschle (2017) has emphasized the political content of illegal wildlife hunting in the GLTP context. Based on extensive research with people convicted of rhino poaching, she highlighted the need to recognize "how changes in the rules, the racist and classist biases regarding hunting rights, and violent enforcement strategies may lead to poaching decisions among those most adversely affected" (Hübschle 2017: 431). Thus, she argued, rhino poaching should be understood as "a form of social and political protest" undertaken in a context of racialized inequities and the "stigmatization and alienation of communities, who, under different circumstances and framing, might be agents of change and disruptors of illegal horn supplies" (Hübschle 2017: 429).
The body of work above provides important insights into what Matthew Minarcheck (2020) recently described as "anti-environmental subject making", while not necessarily theorized in these terms. Theories of subjectivity hold that individual choices are "not given but made" in relationships with others (Jackson 1998: 6) and that human actions are "performative" -these are not predetermined, but they are constrained by historical and political economic inequities (Butler 1988). Thus, the socio-economic conditions under which people come to harm (rather than to care for) the environment are of consequence for understanding the connections between peoples' beliefs and their actions (Agrawal 2005). Applied to the context of militarized conservation, the logic follows that in a context of state-sanctioned dispossession, the framing of conservation as a security issue and the corresponding (mis)treatment of those living with conservation but seen as threatening or as criminals has been self-fulfilling, at least in part (e.g., Neumann 2004;Ojeda 2012). Defiance theory helps to explain why. As elucidated by von Essen et al. (2014: 636): In the case of illegal hunting, interactions with others and their broader environment socially condition the individual with both a set of practical techniques (such as hunting, trapping, hiding game, evading capture) and those values and psychological defense mechanisms that eliminate guilt associated with committing the crime.
Such defense mechanisms include "condemning the condemners" -neutralizing one's own guilt by pointing instead to the harms and offenses of conservation authorities (Eliason 2003;Forsyth and Marckese 1993;von Essen et al. 2019). 7 Such neutralizations might serve as a way to defend one's "intentions in the motivational process for the crime" and as a way to justify crimes "afterward to relieve cognitive dissonance" (von Essen et al. 2014: 637). Importantly, then, the condemnation of others' behaviors may be used to justify a previous crime and figure into and inform the motivation for new criminal acts (von Essen et al. 2014).

Method and knowledge co-produced
Recent calls to decolonize research and publishing practices (e.g., West and Aini 2018) and for antiracism in academia (e.g., Vaughn et al. 2020) underscore the need for me to be accountable to the enormous contributions of my research participants. The contentious nature of my research also demands that I am careful to protect informants from suspicion and harm. I start by underscoring that my theorizations -and not only my findings -draw, first and foremost, from the knowledge (experiences, observations, theories, and ideas) of those who live in and/or work for the LNP; from the interpretations of my research assistant turned long-term collaborator, Divy Mavasa 8 , as well as from those I cite and acknowledge.
The findings I present draw from interviews and observations collected in the LNP in June and July 2018 and contextualized in reference to long-term ethnographic field research in this region (2003, 2006-7, 2011, 2016). The 2018 research took place, primarily, in Makandezulu B, located along the Shingwedzi River Watershed, approximately 10km east of the KNP. June interviews remained open-ended so that Mavasa and I could affirm research permissions and safety; engage with what our interviewees found to be exciting, concerning, and otherwise important; and so that we could strategize. We were unsure, for example, if we should broach the topic of hunting-related arrests since the potential stakes for discussing illegal wildlife hunting openly (if at all) seemed high. We found out in June, however, that the topic of arrests was relatively easy to raise and a highly productive one to address. People were concerned with, and talking about arrests with some incidents so fresh and contentious that they were the only topic we could address. In those cases where Mavasa and I did initiate the topic, we consistently posed questions in the past tense. Thus, those willing and in some cases eager to engage had already been arrested, and any incidents (while sometimes still sensitive) were public knowledge.
Mavasa and I then drew from what we learned (as well as from what alluded, confused, and evaded us) to conduct semi-structured interviews in July. We introduced key themes and, where possible, asked some common questions (e.g., regarding interactions between residents and rangers), but we also kept the interviews open and conversant enough to follow where the participants might lead (e.g., in the unanticipated direction of enforcements against fishing). Mavasa and I conducted 72 total semi-structured interviews -58 with LNP residents and 14 with conservation employees. The interviews took place -with Mavasa translating, interpreting, and conferring between English and Shangaan -in homesteads, Park offices, and ranger stations. They were over meals, while walking and fishing, while driving and while stopped at a road block.
Mavasa and I consistently interpreted and analyzed our interviews and observations, comparing our first impressions immediately following the interviews, and our second and third impressions in more formal debriefing sessions that took place after leaving the LNP. After I returned home to the US, I analyzed the data thematically using the software program Nvivo, sometimes while in digital communication (over WhatsApp or Facebook Messenger) with Mavasa. 9 The subsequent presentation of results includes conflicting accounts of contentious interactions between LNP residents and rangers. 10 In such cases, my analysis consistently relies more heavily on the accounts of residents than on the accounts of rangers. My intention is not to prove one side of a story over another. I have cause to take interviewees (residents and rangers) at their word. Instead, I seek to understand how people experience and respond to increased securitization. To that end, LNP residents form the primary target of increased enforcements, and are thus my primary focal point. Moreover, the objective of the subsequent interpretations is not to discover and reveal the "true facts" of these events (if in fact person x hunted illegally or not), but to understand what else about human subjectivity such stories uncover. In the context of increased securitization, these stories reveal the collective and widespread practice, strategy, and response of questioning and condemning the other.

Increased securitization and defiance in the LNP borderlands
Limits and controls on hunting in the area now comprised by the LNP are not new. In the twentieth century, the Portuguese colonial government managed this region as a hunting concession; hunting was restricted, but enforcements were relatively few and far between (Mavhunga and Spierenburg 2009;Matusse 2019). After Mozambique gained independence from Portugal in 1975, violence associated with a protracted civil war ensued (Roesch 1992;Lubkemann 2005). An estimated six million people, roughly half the nation's population, became refugees and wildlife populations throughout the nation declined dramatically (Hatton et al. 2001). Nearing the close of war in the early 1990s, expatriate hunters established Gaza Safaris, a private hunting concession along the Shingwedzi River, on land claimed by Makandezulu B residents (Hughes and Barbosa 1995). Relations between the hunting concessioners and villagers were uneven and exploitative, though sometimes also amicable (Hughes 1995). The concession owners provided residents with jobs, including as hunting guides and skinners, yet they also ramped up restrictions on hunting.
With the implementation of the national and transfrontier parks at the turn of the century, controls and enforcements on illegal hunting increased further. According to Mozambican Conservation Law 16/2014, hunting and fishing, along with forestry, agriculture, mining, and livestock management are strictly and "rigorously" prohibited in national parks. LNP conservation authorities have enforced the prohibition on hunting, in particular of large mammals, but they have not enforced prohibitions on farming for those residing in the Park, nor for the collection of wild plants, fruits, roots, and firewood or fishing -with the understanding 9 The presentation of results includes a number of Shangaan terms, all of which emerged in the field and have been translated, interpreted, and double checked by Mavasa. 10 For the most part, rangers and residents form separate groups. LNP rangers are often, though not exclusively, hired from locations outside the Park. While a handful of Makandezulu B residents do work for the Park (e.g., maintaining and guarding a nearby tourist camp and cooking and cleaning for the nearby ranger station), the only active LNP ranger from this village has been stationed elsewhere in the Park).

Journal of Political Ecology
Vol. 28, 2021 183 that authorities will ramp up the enforcement of these and other prohibitions following residents' long pending resettlement. In some respects, and as I demonstrate next, the lack of enforcements against subsistence-based activities like fishing appears to be changing. First, though, understanding how people residing in the LNP experience and respond to the increased securitization of conservation necessitates reckoning with the larger history of conservation-related, statesanctioned dispossession. The Mozambican government established the LNP in 2001, and by 2003, conservation authorities targeted nine villages within the Park for resettlement as means to ensure environmental protection (Witter 2013). Since then a series of pushbacks by these groups and delays linked to ongoing funding shortages, poor planning and communications, and failed infrastructure have ensued (DeMotts 2017;Lunstrum 2015;Milgroom 2015;Witter and Satterfield 2014) 11 . Thus, the majority of targeted residents have remained living in the Park under increasingly challenging conditions. Since the founding of the national park, residents have experienced increased food and water insecurity, intensified conflict with wildlife, and the criminalization of their livelihood practices (Witter 2013;Witter and Satterfield 2019). Dispossession comprises state and conservation authorities who failed to respond meaningfully to the concerns expressed and the harms experienced by LNP residents and the inability for residents to plan for the future (DeMotts 2017; Witter 2013; Witter and Satterfield 2019). In these and other ways, there has been a consistent lack of recognition of residents' customary rights and a lack of appreciation for how they understand and actualize these rights in the implementation and practice of conservation (Witter 2013;Satterfield 2014, 2019). More recently residents have faced the threats and actualization of violence that has accompanied increased securitization (Witter and Satterfield 2019; Witter in review).
More than a decade following the establishment of the LNP, international concern about rhino poaching stimulated the investments needed to ramp up the enforcement of controls on illegal wildlife hunting. In recent years, conservation authorities have improved capacity by partnering with, among others, the Dyck Advisory Group, which uses extensive military expertise to counter poaching through securitization strategies. The LNP has adopted a more mobile, field-oriented base for security operations enabled through the procurement of a helicopter, more 4x4 trucks, the deployment of a canine unit, the hiring of more guards and rangers (e.g., 29 new rangers hired in early 2018), and extensive ranger trainings. Also important are improved cross-border communications and coordination between LNP and KNP rangers, collaborations between LNP rangers and Mozambican border police -some are stationed together in the LNP -and changes to Mozambican law that mandate longer sentences for those convicted of illegal hunting (PPF 2019a).
Even though concerns about rhino poaching prompted the increases securitization of the protected area, LNP decision-makers have not directed those resources solely, or even primarily, at preventing rhino poaching. This is partly because, from the standpoint of LNP employees, rhino poaching is not the killing of most concern in recent years. Poachers (including rangers) reportedly killed the last of the rhino population in the LNP in 2013 (AWF 2013), and since then LNP employees are more concerned about the hunting of elephants and lions -to trade their animal parts and out of retaliation for destroying crops and cattle. They are also concerned about the more common mode of illegal hunting in the LNP, "hunting for the pot" as well as with day-to-day practices of food procurement, like fishing. In 2018 these latter activities provided for a hotbed of interactions between LNP rangers and residents. As rangers exerted their new found powers, residents lamented the increased instances of monitoring and arrests, questioned the legitimacy of ranger authority, and defied it.

Fishing, fear, and sauce
For generations, Makandezulu residents and their ancestors have fished along the Shingwedzi River. This involves about an eight kilometer walk, west towards the border with the KNP. Residents typically arrive at the river at a station shared by LNP rangers and border police (the former Gaza Safari hunting lodge), and either fish there or pass by it to reach other preferred spots. It is common, therefore, for residents to encounter Journal of Political Ecology Vol. 28, 2021 184 rangers while or en route to fishing. Park implementation and conservation authorities have typically, if sometimes grudgingly, allowed residents to fish. In July 2018 and in previous years, I observed that a number of homesteads fished several times per week. In contrast to hunting, which is rare and covert, residents fish in the open and in the light of day -often in full view of rangers. In recent years, however, conservation authorities have expressed concerns about overfishing, and rangers have attempted to introduce some restrictions. For example, I observed a ranger encourage a group of women fishers to limit their keeps to adult fish, and rangers reportedly discussed with village leadership a plan to implement a ban on fishing nets, including burning any nets they found. Most importantly here, rangers regularly monitored, surveilled, threatened, and intimidated fishers. Thus, in and beyond contexts of fishing, fear (ku chava) and danger or sense of being scared (chavisa) were common themes in 2018 interviews. As had been the case in the past, residents expressed fear of dangerous animals (nyarhi ya chavisi) (Witter 2013;Witter and Satterfield 2019). More so, they also expressed fears of rangers (mafiscal) 12 , and of travelling outside the village where they might encounter rangers.
Residents reported fear during encounters with rangers. One example was an older woman who described fishing with a small group of other women -residents tend to fish in small, gender-exclusive groups. Approximately four rangers, all of them male, approached the fishers, remaining at a distance, but staring at them for an extended period of time. Whereas male fishers may use a line or a net, female fishers typically use nets, which require them to enter the water and to work in pairs to manage the nets. In this case, the women felt uneasy and afraid and ran from the rangers who then confiscated their fishing equipment. Similarly, another interviewee reported being with a group of agemates, young men, on their way to the Shingwedzi River to fish. They encountered more than five rangers who threatened them, saying that they would take their fishing equipment if they found them fishing.
Increased monitoring and intimidation also contribute to feelings of frustration and anger expressed by several interviewees who explained that they are no longer able to move freely across the landscape, especially outside the village. As one man complained, "…we are afraid to go to the bush. During the drought, we go to gather fruits. Now you can't even reach that tree because you are going to be questioned [by rangers]." A second interviewee reflected, "We are scared because of the people who want to resettle us. When they find us, they try to intimidate us… anytime you meet [the rangers], they ask you where you are going." A third lamented, "[There are] areas that we used to go that we can't go anymore… we used to get fish for xexebo (sauce, explained in more detail next). Now if we go there, we get questions." Food insecurity in this region is exacerbated by prolonged drought, loss of crops due to wildlife incursions, and enforced prohibitions on hunting. Fish provide people with an important protein source and with culturally meaningful xexebo. Xexebo typically refers to any meat or vegetable-based accompaniment to the daily staple of xima, a type of corn porridge. LNP residents commonly referenced xexebo to explain their day-to-day struggles to subsist; for example, the lament of having only wild plants to use as xexebo or, worse, having no xexebo and eating solely corn porridge. Interviewees' use of the term demonstrated, moreover, the entanglement of their subsistence strategies with the day-to-day politics of conservation surveillance. Thus, the frustration of one interviewee who told the story of three rangers who stopped his fishing group to ask where they were coming from. I said, 'fishing.' I was afraid. These people are scary... They try to intimidate you… They even searched my bag [for illegally hunted meat] only to find there was fish. They thought we were hunting. Then, the rangers asked for "xexebo." The men complied, begrudgingly, giving away three fish to each ranger.
This brings me to a more nefarious meaning of the term. Indicative of the spillover between illegal practices, both residents and rangers, sometimes also employed xexebo as code for meat (nyama), specifically The rangers stopped us and asked 'what we ate for xexebo?' ... We told them, 'We don't have xexebo.' When I asked her what she thought the rangers meant by asking for xexebo, she replied that they were looking for "nyama" (in this context, bushmeat).
In such cases, rangers reportedly used their encounters with fishers, foresters, and others (e.g, the children of a suspected hunter in the next section) as testing grounds for residents' knowledge of or involvement in illegal hunting, a far more disdained practice than fishing.
In June and July 2018, fishing provided LNP rangers with an almost daily opportunity to monitor, intimidate, and potentially apprehend LNP residents. Residents expressed fear, frustration, and anger in their accounts of their interactions with rangers, describing these as unfair and akin to harassment; they persistently questioned the legitimacy of that treatment. Nonetheless, residents continued to fish. Fishing in the face of shared fears, angers, and anxieties -i.e., fishing nonetheless -was a shared phenomenon, an act of collective conscience and a continued performance both of food procurement and, as fundamentally, a moral practice of defiance.

Hunting for the pot and the notion of unfair arrests
The tensions and emotions I have described were intensified in the context of illegal wildlife hunting, and in particular, discussions and debates regarding hunting-related arrests. Recall that "hunting for the pot" generally involves the killing of non-threatened, non-endangered animals (e.g., antelope species or small mammals) for meat that is prepared, shared, traded, and consumed locally. To capture the animals, hunters typically use snares -wires linked together and placed strategically in the bush so that the unsuspecting animals become entangled. Hunting for the pot is less common but more problematic than fishing. Relative to restrictions of fishing, which are relatively new, controls on hunting have been consistently enforced since Park implementation; though limited in the past by the combined lack of labor, technology, finances, and the political will to carry out enforcements. With the increased securitization that has accompanied the rhino poaching crisis, such constraints have been overcome, at least to some extent, and rangers have ramped up enforcements on all hunting. There are, moreover, heavier consequences if caught. Whereas in the past, local residents might be fined for hunting in the national park, now they are being arrested and imprisoned and, reportedly, in greater numbers. 13 One interviewee explained the changing landscape of hunting-related securitization as follows: "In the time of the Portuguese [i.e., under colonization] they had rangers… but not a lot…." [Nonetheless], "they wrote our names down" [meaning authorities cited and fined, but did not arrest men caught hunting]. Postcolonization (but before park implementation), residents caught hunting were taken to the district administrative headquarters and fined but "then came back" to the village. He continued, "the laws are stricter now that we are living in a park. The Park came [to us] with their laws and they [now] have a lot of workers." When I followed up with a question about his own experience with any hunting-related fines or arrests, he kept his response vague, yet saying "they wrote [down] my name too…." His final comment on the matter was "Our names are on their lists, even now." This indicated that the struggle over hunting and hunting enforcements endures as a collective one.
In our 2018 interviews, hunting (ku hlota) emerged as important theme, in part because of tensions surrounding a suite of hunting-related arrests (ku khomiwa hi ku hlota) that occurred those months. At issue were stricter enforcements and the legitimacy of these enforcements. Resident accounts of these arrests reveal a tenacious condemnation of rangers, and of conservation authorities more broadly, in ways that exceed the interactions I described in the context of fishing. Among other arrests in other villages, authorities arrested at Journal of Political Ecology Vol. 28, 2021 186 least two men from Makandezulu B in July: one while driving towards the KNP with a weapon (a silencer) and another for subsistence-based hunting. A resident witness of the latter arrest told the story as follows: Three rangers entered the accused man's homestead, uninvited, while he was resting indoors. With other members of the homestead out fishing and collecting water, the rangers first encountered and addressed the man's young children, scaring them with "Did your father bring xexebo [bring home any meat]?" The rangers then yelled to the man inside his hut, "We are coming for you!", while "forcing" the man to open his door. Angry, the accused agreed to go with the rangers to the bush, "Show me what I have done!" Instead, the rangers searched the man's home, then called in reinforcements. Additional rangers arrived in a truck and took the accused away. Following the arrest, the man was imprisoned. The minimal sentence for subsistence hunting is reportedly two years. 14 In the days and weeks following the arrest, other residents, who did not witness the arrest, expressed deep consternation about the way it unfolded, in particular that rangers arrested (and arguably found) this man at home -not in the bush, as was being claimed by the rangers. One interviewee, reflecting on this arrest, bemoaned, "[People's] rights (ntimphanelo) are being violated." He further explained that rangers and residents had recently come to an agreement not to track suspected hunters in the villages. "They can be arrested in the bush, but not at home" or elsewhere in the village. Worse, the rangers who conducted this arrest failed to inform the village leadership.
The right of rangers to enter or pursue residents in the villages was a hot topic in 2018. It emerged in discussions between residents, including as justification for a series of coordinated strikes that residents led against the Park, as well as in interviews about hunting-related arrests. 15 In one such strike, a group of residents from a different Shingwedzi Watershed village, Machamba, blocked the Park's main road. I observed the strike as I attempted, unsuccessfully, to travel with Mavasa and three resident elders to the administrative headquarters of the LNP. Machamba residents closed the road following the arrest of three men from their own village (for a range of illegal activities, including hunting). They maintained the road closure for several days, effectively cutting off travel and severely limiting communications between the northern and southern regions of the Park, thus undermining the ability of Park employees to work, all the while calling for the release of members of their families and villages who had been, from their perspective, unfairly arrested.
My encounter with the strike involved a group of demonstrators who stood outside my rented 4x4 truck and told us to turn around, explaining "the road was closed" (patu ri pfariwile). One woman expressed anger, yelling: "Go back… We don't want the Park!" She continued, gesturing with her torso and arms, "We don't have husbands; they've been arrested!" A few minutes later, another demonstrator leaned through an open back seat window to address one of my travel companions. The woman emoted, "Someone was arrested in their field!" My passenger responded with similar news. Making reference to the arrest I described above, she explained: "We've got the same situation! Someone was arrested in their homestead!" Here again, the location of these arrests provided some explanation for the frustration and anger many residents felt. As noted above, in recent months, Park residents and rangers came to an understanding that rangers would not pursue arrests in the villages without informing village chiefs. In this and other cases, from the perspective of LNP residents, rangers did not uphold their end of the deal.
For conservationists, the arrest I used to introduce this section was a success; the rangers did their duty. For residents, however, this and other arrests were unjust, and, in response, residents expressed their frustration 14 Note that rangers provided a significantly different account of the arrest as follows: Three rangers found the man in the bush; they also found snares nearby. Upon encountering the rangers, the accused ran home. The rangers followed him there, where they searched the man's home and bag, finding additional snares in each. Upon questioning, the accused initially lied, saying that he was trapping frankolins (a type of ground bird prevalent in the area, Pternistis spp.). The rangers suspected the man was hunting antelope species like impala, nyala, or kudu. Rangers did not find a trapped or dead animal, nor did they find any meat. Nonetheless, they arrested the man and confiscated the snares. 15 In 2018 resident-led strikes primarily took the form of roadblocks that compromised Park operations. Other types of strikes reportedly involved residents who threatened, chased away, and/or physically harmed park rangers. In one case, a large group of residents chased park rangers away from one their stations; the rangers fled, escaping on foot and eventually -hours later -took cover at another station. A later strike prompted the consequential decision to close a border gate that typically enables tourists to travel between the KNP and the LNP. and their defiance against the Park. The ongoing dispute over whether the rangers found the suspected hunter in the bush (with snares) or in his own home (unarmed) informed these divergent perspectives. More fundamental and formative to the upholding and the entrenchment of these contradictory perspectives and associated condemnations are different interpretations of what hunting means, and the intensification of these meanings in the context of increased securitization.

What hunting means
From the perspective of conservation authorities and rangers, subsistence hunting is a serious problem. It involves the destruction of legally protected animals in a globally valued conservation area. Hunting amounts to a transgression against the state and an act of global environmental harm by people who authorities have variously described as "poor", "desperate", "sneaky", and "skilled." With respect to the latter, authorities have noted, and some authorities revere, the knowledge and abilities of some local hunters. Nonetheless, against a background of serious concerns about species loss, increased hunting in the LNP signifies that conservation efforts are failing and that authorities need to establish greater control over the Park. Preventing hunting is their responsibility, and they believe they have the legal and the moral authority to do so.
From the perspective of residents, hunting relies on knowledge and skill accumulated through generations and the resurrected performance of a suite of cultural norms, relations, and responsibilities, intimately linked to the sharing of labor, of tools and equipment, of fuel for cooking, and of meat. Such norms emphasize the sharing of xexebo with neighbors (vamakhelwani), including those most vulnerable to food insecurity (e.g., widows, widowers, and orphans). Against a backdrop of shrinking opportunities to obtain meat and of increased consequences for doing so, hunting means that people can care for one another in ways that are culturally significant.
Illegal hunting and fishing also mean defiance. Recall that the interviewee who lamented that "people's rights are being violated" did so in reference to the matter of unfair arrests, but the statement also captures a persistent and collective sentiment of disdain for an overall approach to protected area conservation that excludes and criminalizes those who have long lived with wildlife. The refrain residents have variously returned to in interviews since 2003, that "the Park found us here [first] living with the animals" underscores residents' persistent dismay that conservation hinges on the erosion of their longstanding (however long they have been sidelined) customary claims to land and authority in this region. In contrast, Park employees and rangers find illegal hunting detestable and highly actionable. But from residents' point of view, it is they who hold the high moral ground. Hunting in the face of their dispossession is a reclamation of their rights and an affirmation of their self-determination.
Of course, not all types of illegal wildlife hunting are the same. Take the 2018 interviewee who drew a distinction between those who hunt in the LNP to eat and those who go to the KNP for "business." The distinction between "hunting for the pot" and "hunting for business" is an important one not lost on residents, some of whom described the latter as an enormously risky, destructive, and unethical transaction and a matter of deep consternation. Yet, with the increased securitization of the LNP borderlands, there has been some conflation both in the enforcements against illegal acts (described above) as well as in people's motivation to undertake them.

From conflation in enforcement to conflation in motivation
Conservation authorities in the GLTP are concerned about spillover between subsistence-based and commercial hunts. Take an April 2019 example. Snares, placed to catch food, killed a waterbuck and three endangered wild dogs (PPF 2019b). There is suspicion, more broadly, that LNP residents use subsistence-level resource harvesting like this as a cover for more destructive activities linked to the illegal global trafficking of wildlife and animal parts. Even if this is the case, the thousands of people still living in the LNP do not share a common experience of illegal wildlife hunting -and they should not be held accountable for it.
For example, a few Makandezulu B residents have been variously engaged or implicated in rhino poaching. Some have been variously engaged or implicated in hunting for the pot, and a majority have openly engaged in fishing. 16 Yet there has been spillover in recent years in the ways conservation authorities have responded to the "poaching crisis." The extended infrastructure of surveillance and securitization has enabled conservation authorities to interrogate and to arrest residents suspected of rhino poaching, and it has also increased the monitoring of residents engaged in day-to-day subsistence-related resource use. The result: again, residents do not share a common experience of illegal wildlife hunting, but they do share the experience of increased securitization, and they have responded in common.
With increased securitization of conservation, residents share the experiences of being questioned, intimidated, and harassed while undertaking day-to-day activities that were once permissible. These include fishing and illegal hunting, meaning they live under suspicion and surveillance. In response, they have expressed fear, frustration, and resentment. Resident-led strikes conferred a collective sense of defiance against the Park, but so too did residents' persistent lamentations against the actions of rangers. Recall that in interviews and in heated conversations, residents consistently and persistently questioned the legitimacy of rangers' activities and authority (e.g., when they confiscated nets, demanded fish, or when they made arrests in the village). Doing this enabled them to in effect, "flip the script" -that is, to evade and to neutralize concerns about their own potential involvement in illegal activities by raising instead questions and concerns about the immorality and injustices of ranger behavior. Defiance theory holds that peoples' negative experiences with legal authority, especially under conditions where there is mistrust, may inform and validate their decision to engage in illegal acts (Sherman 1993;von Essen et el. 2014). My findings demonstrate some of the ways in which residents have experienced and responded to the scaled-up harms, aggressions, and violence that have accompanied the securitization of conservation. They reveal the full scope of motivations and justifications for illegal fishing and hunting (i.e., food, care, cultural continuity, and defiance). These findings also lay the groundwork, I contend, for understanding how and why residents' experiences with heavy-handed enforcements against hunting shape and inform the meanings, motivations, and justifications at stake (i.e., power, authority, poverty, profit, legitimacy, and dignity) in their decisions to take up rhino poaching in the face of militarized conservation.

The need for agent-full understandings of militarized conservation
Understanding illegal wildlife hunting as an act of defiance clearly resonates with Thompson's (1975) ideas about hunting as resistance to the enclosure of the commons and Scott's (1990) inclusion of poaching as a form everyday resistance. In these, as in other studies cited above, illegal hunting is a means to deriving material or economic gain, for communicating opposition and outrage, and for exercising highly contested rights. My intention here is not to substitute defiance for resistance, but rather to move beyond entrenched scholarly debates that can themselves tend towards reductionism. Rather than writing from a position of proving or disproving resistance or even entering into a scholarly performance where the points (and the order of citations) are already mapped out for me, I seek to engage with and to explain human experience and response from an emic perspective. Roderick Neumann argued that "the key to understanding the political meaning and intention of specific actions, particularly criminal acts, revolves around… an explication of the social and historical context within which they are performed" (1998: 48). Scholars seeking to understand illegal wildlife hunting in the GLTP have underscored the influence of rising levels of wealth inequality, the lack of viable alternatives for rural people living in and near national parks, and the repercussions of a discursive history of dispossession that sees rural black Africans as security threats (Duffy et al. 2016;Haas and Ferreira 2018;Hübschle 2016Hübschle , 2017Lunstrum and Ybarra 2018;Lunstrum and Giva 2020). Yet, scholarship seeking to explain illegal hunting motivations in the GLTP has paid surprisingly little attention to the socio-political context of dispossession, not least increased securitization. Even less attention has been directed to examining how people experience and respond to such processes. Thus, a recent review of emergent work on militarized conservation pointed out that "the experiences of the people living in the areas concerned are overlooked in debates about the militarization of conservation" (Duffy et al. 2019: 68). The authors went onto explain that this neglect is consequential, because "understanding what militarization means for those people can shed light on how the drive to save species by more forceful means has counterproductive effects in the longer term" (ibid). As political ecologists take up this charge, it will be important to direct their attention to the effects or impacts of militarization, but also to people's agency in the face of militarization.
In this article, I have shown why militarized conservation may be counter-productive through an analysis of how people experience and respond to increased securitization. Understanding illegal wildlife hunting as a moral practice of defiance involves de-centering and disrupting the default positioning central to (if sometimes unspoken in) conservation decision-making. The latter has three elements. First, that human behavior rests, and therefore can and should be amended, based on economic incentives; second, that human values, morals, and power relations can be adequately described and comprehended in legal terms; and third, that if and when legal and economic incentives fall short or incentivize the wrong behavior, the preservation of endangered wildlife necessitates and justifies violence. My intervention has also been a more agent-full understanding of increased securitization than in previous studies. Recent scholarship on "militarized conservation", "green militarization", "green violence", "green security", and "green wars" (Büscher and Fletcher 2018;Büscher and Ramutsindela 2015;Duffy 2014;Duffy et al. 2019;Kelly and Ybarra 2016;Lunstrum 2014;Massé et al. 2017) provides compelling frameworks for understanding the links between accelerating power inequities and conservation-related violence. However, in our efforts to decolonize research and scholarship, we must contend with the fact that improved understandings of militarized conservation will not stem from and follow existent frameworks, however insightful they may be. Instead, theory on the intersections between power, loss, justice, and environment needs to be better informed and guided by the people living with militarization -and reconstituted according to how local populations make sense of these structures, processes, practices, and relations.