The Ebola Crisis in Sierra Leone: Mediating Containment and Engagement in Humanitarian Emergencies

This article explores how framings of the 2014–16 outbreak of Ebola as a crisis, its causes, nature and consequences gave rise to two seemingly contradictory types of interventions within affected communities in Sierra Leone: a militarized state of emergency on the one hand, and efforts to foster local engagement and ownership on the other. Teasing out explicitly the underlying logic of these two modes of response, we are able to discern the convergence between containment and engagement approaches that are at the heart of contemporary humanitarianism. Rather than being opposed or contradictory, the article shows how they were mutually constitutive, through negotiations between different ways of knowing and responding to the Ebola crisis. The resulting divisive practices, juxtaposing ‘Ebola heroes’ and ‘dangerous bodies’, re-ordered the landscapes that individuals had to navigate in order to manage uncertainty. Tracing these logics through to the ‘subjects’ of intervention, the article tells the story of one traditional healer’s ‘epistemic navigations’ in his efforts to survive both the epidemic and its response. Bringing these dynamics and their consequences to the fore in the Sierra Leonean case invites broader reﬂections on a humanitarian assemblage increasingly reliant on the mutual constitution of containment and engagement, security and resilience, in its approach to managing ‘at risk’ populations. This


INTRODUCTION
In March 2014, the first official cases of Ebola were recorded in Guinea. By August of that year, the World Health Organization (WHO) had announced a Public Health Emergency of International Concern (PHEIC) as the epidemic reached hundreds of cases a week in Liberia, Guinea and Sierra Leone. By the end of the epidemic in March 2016, over 28,000 people had been affected by the disease: 11,310 died and countless others suffered as a result of the significant socio-economic setbacks associated with the outbreak (WHO, 2016). The tortuous two-year battle that seemingly caught the world off guard gave rise to innumerable analyses (e.g. Abramowitz, 2017;DuBois et al., 2015;GHRF Commission, 2016;Kamradt-Scott, 2016; Panel of Independent Experts, 2015; Moon et al., 2015;Ross et al., 2017). These have in turn generated recommendations for the reform of national health systems and global response mechanisms, pointing to failures at all levels during the West African outbreak: from systemic underfunding of the global health apparatus, slow resource mobilization and inadequate leadership on the part of WHO, to disastrously ill-prepared national health systems and ineffective engagement of affected communities. Yet the road from the first infection in Guinea's forest region in December 2013 to the publication of compelling explanations of what went wrong after thousands lost their lives was not straightforward. Retracing that road entails telling the story of how scientific and humanitarian communities, as well as national and international publics, came to understand the kind of problem emerging in West Africa, including disputes and shifts in Ebola's meaning and adequate avenues for response.
This article focuses on a particular aspect of this process by looking at the relationship between the humanitarian response apparatus and affected communities in Sierra Leone. In particular, it aims to explore how different ways of understanding Ebola as a crisis, its nature, causes and consequences, collided to give rise to seemingly contradictory types of interventions within communities: on the one hand, a militarized state of emergency, and on the other, efforts to foster local engagement and ownership. Outlining the logic of these two modes of response in the context of Ebola allows us to elaborate on a dynamic that increasingly characterizes contemporary humanitarianism, namely the convergence between a logic of containment and one of engagement, between militarized interventionary imaginations and a commitment to inclusion and the fostering of local 'resilience'. The epistemic politics around the Ebola outbreak in Sierra Leone offers a case study for how containment and engagement become entwined through negotiations amongst different kinds of actors as emergencies unfold, and highlights the material consequences of this entanglement.
The story of how Ebola was imagined as crisis and its implications is the subject of a growing body of literature. Moments of crisis provide a particularly useful standpoint from which to study knowledge production and the role of expertise in processes of problematization. 1 The notion of crisis, or what Caduff (2015) terms a 'culture of danger', is ubiquitous in contemporary public discourse, from the threat of terrorism to the menace of financial collapse and environmental devastation. The triumphant tones of Fukuyama's (1992) The End of History have given way to gloomier perceptions of reality as persistently unstable and uncertain. Taking a step back, as Roitman (2013) invites us to do in her book Anti-crisis, highlights how epistemological claims such as 'this is crisis' engender certain types of action whilst making others impossible. In the field of international development and humanitarianism, for instance, Duffield (2007) has pointed to an increasing trend towards securitization, whereby the dangers of underdevelopment in an interconnected world have underpinned justifications for particular kinds of aid.
Global health is a good example of this trend as concerns with 'biosecurity' and international contagion have similarly driven action. In this context, scholars have explored how scientific expertise engages with uncertainty including discursive and material technologies for making the future knowable. Caduff's (2015) exploration of scientific prophesy in The Pandemic Perhaps, for example, traces the politics of influenza preparedness in the USA and the way that certain claims about impending catastrophe become more compelling than others, regardless of their predictive accuracy. The perpetual production of scientific knowledge about influenza and narratives about a looming pandemic, 'at once affirm and deprive the world of confidence' so that 'truth is forever deferred in a circuit of infinite testing' (ibid.: 31). Similar prophetic claims contributed to the escalation of fear surrounding a possible spillover of Ebola across borders, not least when the Centre for Disease Control (CDC) predicted that without intervention we could foresee over a million infections by January 2015 (Meltzer et al., 2014). Delays in declaring Ebola an emergency raise important questions about when such claims become compelling: when does a crisis become a crisis? Lakoff (2017: 151) points to the role of 'technocratic classifications' as techniques for making the unknown manageable. For Lakoff, failure to respond adequately and in a timely manner to Ebola can be seen 'at least in part as a failure in administrative imagination', as the outbreak was initially misclassified based on historical precedent of epidemics that were controlled through humanitarian medicine (ibid.: 141). The declaration of a PHEIC then was a significant 'epistemic shift' as it ushered in a new 'imagery' of crisis, in which 'evidentiary charisma' was mobilized to argue for moral obligations, such as those to carry out research into vaccines and therapeutics (Kelly, 2018: 135, 137).
The politics of expertise and the ways in which knowledge about crisis is produced are significant in their own right, but they also have consequences.
1. Here I follow Foucault's understanding of problematization as the 'ensemble of discursive and non-discursive practices that make something enter into the play of true and false and constitute it as an object of thought' (cited in Rabinow and Rose, 2003: xix).
The implications of epistemic claims include the 'different modes by which . . . human beings are made subject' (Foucault, 1982: 777). Amongst these are the 'dividing practices' embodied in 'modes of manipulation that combine the mediation of science and the practice of exclusion', such as the creation of subjects through scientific classification and the 'technologies of the self' through which individuals discipline themselves according to dominant discourses (Foucault, 1988;Rabinow, 1991: 8). 'Docile [bodies] that may be subjected, used, transformed and improved' are thus a product of these intersections between knowledge and power (Foucault, 1975: 136). Anthropological explorations of international development programming have taken these concerns seriously, analysing how the definition of problems and delineation of solutions configure social relations and subjects. Li's work on conservation projects in Indonesia, for example, reflects how interventions that intend to '[optimize] the lives of others' require identifying 'deficiencies that need to be rectified' (Li, 2007: 5, 7). It is by rendering these deficiencies intelligible, through data and theories of change as well as selective omissions, that it becomes possible to posit technical interventions. Such technical solutions, through the mobilization of expert knowledge, structure power relations not least by creating a separation between the experts who diagnose problems and the beneficiaries in need of improvement (Ferguson, 1994). Rather than considering development as the product of hidden agendas of powerful interests, these studies suggest that we take the normative aims of these projects at their face value. They ask us to redirect our focus onto what these framings produce, what identities and relations they make possible and which they foreclose. Emancipatory visions of development as a participatory process, for example, rely on the figure of the 'active citizen' (Robins et al., 2008). Interventions create claims-based identities, like the 'biological citizenship' manifest by victims of Chernobyl in Petryna's (2004) work. These configurations also often tell stories about individual and collective responsibilities, such as the currently popular emphasis on 'resilience' to disasters (see Chandler, 2016). Imageries of crisis have specifically generative effects. In a state of exception, different forms of life come into existence through appeals to extraordinary circumstances (Agamben, 2005). The crisis imagination creates both subjects of intervention and agents of instability: 'surplus populations' to be contained through projects aimed at improving, correcting or aligning these groups to normative modes of existence (Duffield, 2007).
Drawing on this literature, this article focuses on the identities, relations and practices conjured by particular framings of the epidemic in Sierra Leone. The aim of the article is modest: it retells a now familiar story about the Ebola outbreak, but drawing out explicitly the logic underpinning the interventions and its implications in practice. Outlining negotiations around ways to understand and tackle the drivers of continued contagion in the final months of the epidemic, it shifts the focus away from the politics of expertise to its ramifications for affected communities. In particular, the article points to the divisive practices that emerged from the convergence of containment and engagement: between 'good' and 'bad' citizens, Ebola heroes and dangerous bodies. Efforts to domesticate the epidemic, rendering it knowable and manageable, ordered new uncertain landscapes. These spaces, where these interventionary logics are defined through individual and collective efforts to navigate crisis, make visible the everyday realities of a contemporary humanitarian project that increasingly relies on this dialectic between containment and local ownership.
The article is based on two stints of ethnographic research, interviews and document reviews carried out during and after the epidemic between 2015 and 2017. I was based in Kambia District, Northern Sierra Leone for 14 months of ethnographic fieldwork (starting towards the end of the epidemic) as part of a large medical research project, where my role was to explore community experiences of the epidemic and perceptions of interventions. During this period I was also able to attend meetings of the National and the District Ebola Response Centres. This was followed by a further five months of research in Kambia and in Sierra Leone's capital, Freetown, in 2017 studying the effects of experience and memories of the Ebola response on ideas of citizenship and state-society relations. In both periods, I carried out interviews with individuals involved at different levels of the response including those working in community engagement and surveillance, military officers, community activists, health workers, traditional healers and ordinary citizens affected by the outbreak.
The article firstly shows how, in the particular context of the Ebola response in Sierra Leone, security approaches focusing on containment coexisted with community engagement narratives. Rather than presenting the discourse of the response as 'totalizing and seamless' (Li, 2007: 25), it describes instead the contestations between very different kinds of actors working in the hardest of circumstances and their efforts to control a constantly changing epidemic landscape. This is not incompatible with an assessment of how these efforts to interpret and manage crisis (re)ordered political relations and created subjectivities through particular dividing practices. Finally, the article turns to the 'subjects' of intervention, considering how the containment/engagement logic structured the space of possible action in affected communities. Here the article joins an emergent literature exploring everyday life and practices in relation to the Ebola response (Desclaux et al., 2017;Hoffman, 2016;Lipton, 2017;Shepler, 2017) but also draws on a broader literature on tactical agency in uncertain landscapes (de Certeau, 1984;Christiansen et al., 2006;Utas, 2005;Vigh, 2006). This is outlined through the experiences of a traditional healer, Pa Yamba, 2 who repositioned himself from being a potentially dangerous subject in need of containment to becoming a knowledge broker, embodying the containment/engagement logic in his role as a member of a burial team. Pa Yamba refashioned his 'traditional knowledge' to shift subjectivities from a 'dangerous body' to an 'Ebola hero' and, in turn, to enforce (but also remould) those dividing practices within his community. This allowed him to survive not only the epidemic itself but also the logic of its response, pointing to a form of 'epistemic navigation', a further tactic in Sierra Leoneans' long-standing repertoire of practices of accommodating crisis and its management in their daily lives.

OPERATION NORTHERN PUSH: WORLDS COLLIDE
In the summer of 2015 Sierra Leone was entering its second rainy season marked by the Ebola epidemic. Kambia, a border town where the last days of the 'Ebola fight' were being fought, lies on the Freetown-Conakry highway, which at the time was strewn with checkpoints where soldiers and health officers took travellers' temperatures under the torrential rain. I was posted there as part of the social science team in a medical research project that was being set up in the district. The striking clash of temporalities, between the urgency of the emergency and the slower process of setting up clinical research, meant we were often observers of the District Ebola Response Centre (DERC) meetings, trying not to stand in the way of the unfolding emergency operation. Those who had been on the frontlines for the greater part of the epidemic were visibly tired and every new case announced was met with audible frustration at the evening meetings in the former Resource Centre which had become a hub for operations in the district. Occasional disagreements amongst members of different pillars making up the district-level response offered glimpses into the complex dilemmas faced by responders. One evening, the discussion heated up around the issue of bylaws that envisioned fines and jail sentences for those found in contravention of regulations put in place to combat the outbreak, such as the recording of 'strangers' arriving in town or the reporting of illness and death and performance of safe burials. A group of people in the district headquarters had been caught carrying out a burial in secret. The group had washed the body of their loved one -a practice central to local funerary rites but also carrying the highest risk of infection.
Ebola spreads through direct contact with body fluids of a person who is sick or has died of the disease, with the viral load being at its highest in a dead body. Funerals were identified early on in the epidemic as 'superspreader events' given the region's burial practices involving the washing of the dead (Richards, 2016). Because of the risks associated with funerals, as Lipton (2017: 804) notes, 'the regulation and transformation of mortuary practices were not collateral challenges but principal aims of the international response'. The response was to put in place procedures for 'Safe and Dignified Burials' to be carried out for any death, and not only those confirmed to have been caused by the virus, by a burial team dressed in Personal Protective Equipment (PPE). These procedures and their deviation from customary burial practices were a major source of contention because of the social meaning of death and the central role of funerary rites to the reproduction of social order (Lipton, 2017;Richards, 2016;Richards et al., 2015). Despite recognition of the fundamental disruption and anguish that these regulations were generating within families and communities, and efforts to integrate aspects of customary rituals in the safe burial protocols, the bylaws being discussed on that summer evening in Kambia had been deemed necessary to curb the persistent problem of secret burials performed to bypass the regulations.
The discussion centred on the fact that those caught burying their relative in the middle of the night, hoping to go undetected, were facing the possibility of time in jail. Some in the room expressed their discomfort with this decision, proposing that DERC consider how these bylaws could be reconciled with the recognition of the rights of the individuals involved. Another response worker, however, retorted: 'People have been lying to us, there is a background of deceit, we can't be too forgiving'. This tension between the protection of individual rights and the need to get community buy-in for the response, on the one hand, and the need to respond urgently to the risks posed to the collective by individual behaviour, on the other, was at the heart of the challenge posed by Ebola. In Kambia, these questions came to a head that summer, as the district, together with neighbouring Port Loko, became the centre of Operation Northern Push. The Operation was established to intensify efforts to 'identify, contain and eradicate EVD [Ebola Virus Disease]' from the last two districts where cases continued to be reported (NERC, 2015a). This would require 'intensive community engagement' (DERC, 2015) as well as a 'significant security element' to support the implementation of new regulations. These included a 6 pm to 6 am curfew and strengthened checkpoints, as well as 'strong efforts to find, isolate and track people who abscond and an increase in community surveillance, enhanced by a stricter enforcement of the Safe and Dignified Burial bylaws' (NERC, 2015a).
The Sierra Leone government's decision to 'take political risks by becoming more muscular' with Operation Northern Push, as one senior official involved in the international response recalled, reflected the growing impatience of the final efforts to end the epidemic. 3 However, the establishment of a heavy-handed intervention bringing together 'engagement' and 'security' elements can also be seen as the culmination of the entanglement, or even mutual constitution, of two seemingly opposed logics: one premised on containment and the other on community participation and behaviour change. Illustrating the assumptions and operationalization of these two types of logic, tracing their justifications and in particular their mutual 3. Interview, former international response representative, London, 21 December 2016. dependence, invites broader reflections on the implications of the collision of these two worlds across a growing number of humanitarian and development interventions.

STATE OF EMERGENCY: THE LOGIC OF CONTAINMENT
The road towards identifying Ebola as an emergency and then the even harder task of bringing it to an end are products of contemporary configurations of humanitarian practice. Chronologies of the outbreak begin with Patient Zero, a young boy who played with a dead bat in the Guinean village of Meliandou in December 2013 (Leach, 2015). Despite calls, in particular by Médecins Sans Frontières, to intervene, the WHO only declared an emergency on 8 August 2014 after 1,700 cases had been recorded across the three affected countries. In the 'epistemic shift' ushered in by the declaration of a PHEIC, a number of mechanisms for intervention became possible (Kelly, 2018). 4 The epidemic was declared a 'threat to peace and security' by the UN Security Council in September 2014, in keeping with the increasing securitization of the global health agenda (Abraham, 2011;Benton and Dionne, 2015;Bernard, 2013;Chigudu, 2016;Davies, 2008;Heymann, 2003). 5 In a well-practised discursive twist, the securitization of Ebola increased its urgency on the global agenda. This particular way of understanding Ebola as a security problem had important consequences for how it could be managed. The UN set up its first-ever peacekeeping-style health mission, UN Mission for Ebola Emergency Response (UNMEER). Alongside a large complex of national and international actors from all corners of the development, global health and humanitarian fields, the securitization of the epidemic also ushered in a military response, through international deployments and the leading role of local armies (Benton, 2017). Concerns with the implications of militarization were countered by delinking military apparatuses from military logics (ibid.): militaries were simply seen to be better equipped, from a logistical perspective, to deal with particularly challenging circumstances.
Militarization was deeply intertwined with the particular narrative of emergency that was taking shape. The security complex is not new to Sierra Leone. Since its 11-year civil war, the country has become emblematic of the increasing depiction of underdevelopment as a security concern, with structural fragilities used to explain cyclical crises (Enria, 2018 Securitization in Sierra Leone, as elsewhere, has relied on narratives of containment of unruly, potentially dangerous populations and the spectre of instability as the driving force for development interventions (Duffield, 2007;Enria, 2018). As the disease spread, appeals to security were transposed to dealing with a new crisis and containment was enacted through the 'extraordinary measures' established in the declaration of a State of Emergency on 30 July 2014, including quarantines and restrictions on freedom of movement. In some parts of the country, overwhelmed by the epidemic and in the absence of a concerted response in the early stages, communities established their own isolation measures and bylaws to control movement (Richards, 2016). In the fall of 2014, as the epidemic appeared to be spiralling out of control, responsibility for coordinating the response was moved away from the Ministry of Health into a newly established National Ebola Response Centre (NERC) under the leadership of then Minister of Defence Paolo Conteh. Following the successes of local initiatives, especially in southern districts, the response was increasingly decentralized as authorities in other districts were encouraged to implement restrictions and punitive measures against the violation of Ebola regulations. The logic of security, and its corollary of containment through military intervention and 'social distancing measures', were not solely the product of the 'martial global imagination for responses to epidemics' (Desclaux et al., 2017: 212) or a simple mirror image of previous interventionary landscapes. The particular configuration in the context of the Sierra Leonean response, and especially in its final expressions such as Operation Northern Push, also relied on specific understandings of why the epidemic had become so intractable -ways of making sense of the uncertain terrains faced by response workers.

'SENSITIZATION': THE LOGIC OF ENGAGEMENT
As the disease spread across the region, response workers pointed to chains of transmission to highlight the factors that stood in the way of defeating Ebola. Episodes of resistance to disease control measures and practices involved in taking care of the sick and the deceased were especially prominent explanations. These explanations focused on how a lack of understanding or unwillingness to comply with public health regulations were hindering efforts to end Ebola. Various critics pointed to the fallacies of these interpretations (Abramowitz et al., 2015a(Abramowitz et al., , 2015bBolten and Shepler, 2017;Chandler et al., 2015;Jones, 2011;Richards, 2016;Wilkinson and Leach, 2015;Wilkinson et al., 2017). Indeed, 'behavioural and culturalist' interpretations that individualize, depoliticize and cast blame are not new to Ebola, and have been defined as 'as ineffective as they are unjust' in widely different contexts (Fassin, 2007: xix). The effectiveness, accuracy or fairness of these narratives is, however, not the main subject of this article: the focus is rather on the internal logic of the behaviour change argument and how it came to play a role in the security approach to epidemic management.
Through points of conflict and convergence, containment collided with an approach focused on 'sensitization' and the engagement of communities to create a particular, if fractured, understanding of the causes of crisis. Of the seven challenges outlined in a Ministry of Health and Sanitation EVD Response Plan in July 2014, four were: 'inadequate understanding'; 'denial, mistrust and rejection' arising from 'misinterpretation of the cause of the new disease'; 'close community ties and movement' that made contact tracing difficult; and 'customary burial practices' leading to 'panic and anxiety' resulting from community deaths. The rest were logistical challenges associated with the capacity of healthcare workers (including their own fears of the disease) and geographical spread (MoHS, 2014). Alongside containment (or 'social distancing') measures such as quarantines and lockdowns, then, the Ebola response developed measures for social inclusion. An organizational pillar was set up to deal with communication and mobilization in affected communities. The integration of 'sensitization' in the response can be crudely described as a continuum of participation: starting with interventions aimed at correcting misinformation on the one hand and bottom-up approaches engaging communities in the response on the other. Over time, the response apparatus moved much closer to the second aim. Initial efforts at public health messaging relied on the notion that communication strategies needed to change 'risky behaviour' related to 'traditional practices' and 'misinformation' (Chandler et al., 2015(Chandler et al., : 1275. From this point of view, people needed more biomedical knowledge so as to protect themselves from the disease. Billboards were put up insisting that 'Ebola is real' and encouraging people to report the sickness to a national helpline, not to wash dead bodies, and so on. In these framings, cultural beliefs and 'customary practices' either prevented a full understanding of transmission chains or actively worked against the need to adapt to the exigencies of crisis. People, as a representative of national NGOs in the NERC argued, need to 'understand beyond their cultural perception'. 6 As the response unfolded, the core model of behaviour change and sensitization remained intact, but the mode of delivery changed, as initial negative messaging such as 'Ebola kills' was found to be counterproductive. A redesign of the sensitization model to focus on 'community ownership' was in line with intimations by anthropologists and community advocates about the need to understand reasons for mistrust, explore the adaptive potential of cultural practices and learn from existing community-based responses (Laverack and Manoncourt, 2016). This had crystallized by the end of the epidemic, as reflected in July 2015 when Operation Northern Push was put in 6. Interview, national civil society representative, Freetown, 20 January 2017. place, and the NERC published a new 'Getting to a Resilient Zero' strategy. This located drivers of transmission in 'fear, inadequate trust and collaboration from communities' tendencies to seek healthcare through informal structures', and highlighted the need to 'understand community behaviour' and ensure that communities took leadership of the response (NERC, 2015b). In other words, the problematization was similar, but the solution was more sophisticated and culturally sensitive: change in practices had to come from inside affected communities.
One of the earliest examples of this perspective was the Social Mobilisation Action Consortium's innovative 'Community-led Ebola Action (CLEA) Field Manual' developed to train community mobilizers, with the stated aim of 'inspiring communities to understand the urgency and the steps they can take to protect themselves from Ebola' (SMAC, 2014: 6). These efforts, 'unlike previous mobilisation efforts, which have mainly used education and one-way communications . . . focus on the community as a whole and on the collective benefits of a cooperative and community-led approach' (ibid.: 6-7). Part of this process was to ensure that communities committed to temporarily putting aside knowledge and 'traditional' practices that were deemed to be risky. The theory of change of these interventions was therefore aimed at shifting risky behaviours, initially framed as misconceptions, later understood to require a deeper, more inclusive engagement with communities to address barriers to necessary behavioural changes. Like the security complex, this turn drew explicitly from the long-standing repertoire of development and health interventions in Sierra Leone and beyond, with its focus on participation and behaviour change, on resilience and individual responsibility (see, e.g., Klein, 2016).
However, during Ebola, the entanglement of security and development agendas came to the fore in ways that are rarely so visible. As is often the case, moments of crisis amplify and make evident underlying regimes of knowledge. In the logic of epidemic management, the engagement paradigm was not opposed to the logic of containment. The security approach relied on the same interpretation of the problem. An Office of National Security (ONS) official described the response as requiring a 'carrot and stick' approach: restrictions and punitive measures ('stick') had to complement community engagement ('carrot'). 7 Security personnel's justifications for containment measures similarly followed the premise of the engagement model, as security measures were reserved for those who failed to collaborate and change their behaviours. Speaking of his involvement in the response in Kambia, an officer of the Republic of Sierra Leone Armed Forces (RSLAF) emphasized his understanding of reasons that may drive affected individuals to defy quarantines, such as the need to tend to one's farm. At the same time, he argued that unwillingness to commit to behaviour change, regardless of the 7. Interview, ONS representative, Freetown, 12 January 2017. costs, and failure to comply with emergency regulations were threats that needed to be addressed: In some chiefdoms that I don't want to name, it was the lawlessness that made the sickness spread . . . . Why do you think they [invoked] this state of emergency? If they had just relaxed, the thing would have been worse, so they saw that the best thing they could do was to bring in security. It was not violence per se, but just for people to comply with the law and for them to be able to listen to the medical advice. 8 The sentiment was shared beyond military circles, including a civil society leader who argued: When the president announced the involvement of the military I thought: 'These people with guns, how could they fight Ebola? Something that they cannot see?'. But when after some time they came and I saw the rationale, because when we were also going into the communities some people when you would give them simple instructions to follow so as not to cause problems in the community they would not want to follow it. If we had not had the law enforcers like the military it would have been difficult! 9 Security and engagement, in other words, were two sides of the same coin, although this is not to say that there were not fundamental tensions between sets of actors engaged in quite different sides of the response. This was clear, for example, in the altercations at the DERC surrounding the fate of those caught performing unsafe burials. As in every large-scale intervention, and certainly in one as complex as that mounted to counter Ebola, the motivations, ethos and practices across different actors were extremely varied. Despite these tensions, the uncomfortable but mutually constitutive co-existence between these different sides of the response meant that specific crisis subjectivities emerged in often-unintended ways.

DANGEROUS BODIES AND EBOLA HEROES: CRISIS SUBJECTIVITIES
In August 2015, the Kambia DERC was preparing to celebrate its achievement of 42 days without an Ebola case, and thus to declare the district free of the virus. In a dark twist of fate, the evening meeting of the DERC opened instead with the announcement of a new case: a woman had died in a village within the district and a swab had confirmed her positive. The military officers in charge of the briefing told a sombre room that the district had been 'invaded' and 'attacked' once more. Despite resolute pleas by public health officials not to speculate about the circumstances of the case until they had been confirmed, the meeting centred on rumours that the victim had been infected through a secret affair with an Ebola survivor. In the midst of uncertainty surrounding the possibility for sexual transmission by survivors, visual representations of epidemiological chains depicted dotted 8. Interview, RSLAF representative, Kambia, 18 November 2015. 9. Interview, local civil society representative, Kambia, 29 June 2016. lines, expressing tentative but intimated connections between cases, telling individual stories of traditional medicine, unsafe burials and unprotected sex.
Efforts to render the epidemic manageable relied on particular problematizations of the crisis, operationalized through pre-existing interventionary paradigms. Security and engagement were brought together into a single solution necessitated by the urgency and uncertainty of the emergency as it unfolded. Change had to happen within communities, driven by individuals willing to commit to changing their practices and convincing their communities to do the same. Containment measures would enforce compliance and deal with deviance. The political dimensions of crisis were muted as responsibility was individualized and containment measures depicted as inevitable. Yet, these problematizations were deeply political in practice, not least in the way that they produced a dichotomy of subjects: the 'Ebola heroes', the active citizens that accepted biomedical expertise and took charge of sensitization drives; and those who were holding up progress, the dangerous bodies putting their society at risk. These subjectivities emanated from both the 'material and immaterial' dimensions of disease control strategies, through the language, spatial practices and visual artefacts of the response such as representations of epidemiological chains at response meetings (Hoffman, 2016: 247).
In their 'Standard Operating Procedure for Social Mobilisation and Community Engagement', the NERC (2015c: 1) tells prospective mobilizers that: 'To stop Ebola transmission, communities and individuals themselves must make changes to some of their social and cultural practices. Social mobilisation and community engagement aims to help communities and individuals to understand and take ownership of their situation'. In the CLEA Field Manual such ownership of behaviour change was to be led by 'community champions', who were 'critical to success, because they have the commitment and energy to follow up with their neighbours and to encourage changes in community norms and implementation of the agreed action plan' (SMAC, 2014: 29). Community champions were to carry their communities to what the CLEA manual calls an 'ignition moment', that is, the 'collective realisation that due to community practices (of good, caring people) community members are currently at serious risk of catching Ebola' (ibid.: 26). The notion that communities, led by active and concerned individuals, are central to the response was further underlined by the ubiquitous billboards asserting that 'You can help to Stop Ebola' and posters pasted across the country picturing doctors, contact tracers, survivors and police officers as 'Ebola Heroes'.
These narratives of community ownership and individual responsibility for the common good were undoubtedly reductive. 'Ebola heroes' such as nurses and volunteers faced challenges regarding remuneration and hazard pay, as well as suffering significant stigma and mental health repercussions (Kingori and McGowan, 2016). Similarly, community engagement language romanticized communities, evading the realities of power and hierarchies and the impact these would have on the implementation of engagement practices (Enria et al., 2016;Wilkinson et al., 2017). However, as has been recognized by scholars of development discourse, it is often precisely through such simplifications that it is possible to create powerful mobilizing narratives (Mosse, 2004). As such, politics and the challenges of mistrust and community tensions had to be put aside in order to elicit action.
The figure of the Ebola hero was premised on its opposite: the dangerous bodies of the sick, the dead and, most importantly, those at risk of contracting and transmitting the disease. These distinctions came into being through 'dividing practices' such as body bags, PPE outfits, quarantines and triage points. In their study of different containment and contact-tracing measures in Guinea and Senegal, Desclaux et al. (2017: 229), drawing on Mary Douglas's seminal work on purity and danger, point to the role of containment as an 'ideal order based on the visible distinction between the risky and the safe, a cultural process fundamental for the symbolic control of disorder'. Their cross-country comparisons show that although quarantines were negotiated differently, resulting in different models across contexts, containment practices could result in stigma as communities assumed that individuals involved had been contaminated. Amongst those in Kambia, a border town, who had experienced quarantine and other measures especially under Operation Northern Push, the challenge of the state of emergency was often recounted through comparison with neighbouring Guinea where procedures did not involve being 'locked' in, or held in a 'civilized prison' as a respondent described it during a parallel Operation in Freetown. 10 The language used to describe the response also contributed to the production of dividing practices, such as the war imagery and combat metaphors described above. The 'invisible enemy' was to be defeated by finding and containing people who were a risk to themselves and others. In September 2015, for example, a reward was announced for finding a missing contact related to the last transmission chain in Kambia. Response officials emphasized that her dangerous behaviour might be due to a lack of knowledge and understanding and so, whilst not intending to treat her like a criminal, she needed to be found: The National Ebola Response Centre (NERC) has announced a five million Leones reward for the arrest or information leading to the whereabouts of . . . believed to be an Ebola high-risk contact.
[She], who is 'not a criminal', according to the CEO, Palo Conteh, has been out of the radar of Contact Tracers for the past 20 days . . . . 'It is also possible that she does not understand how vital she is to the response', the CEO maintains. (Awoko, 2015) 10. Interview, community member, Freetown, 11 September 2017.
Problematizations of the crisis and the practical solutions they made possible thus relied on the production of subjectivities to be acted upon and the distinction between active citizens to be included and recalcitrant ones to be distanced. Willingness or ability to embrace the knowledge of the response, and adapting one's beliefs and practices accordingly, marked the boundaries between the two.

PA YAMBA: EPISTEMIC NAVIGATIONS
Regulations over local practices were a central arena for these contrasts between different forms of knowledge. For example, the practice of traditional medicine was banned under the state of emergency as the application of herbal remedies on patients' bodies was deemed to be a major conduit of disease. Moreover, consultation with herbalists rather than visiting health centres meant that potential cases of the disease were going undetected. Many healers felt aggrieved as they were put out of work; some continued in secret, risking fines and their lives as many were exposed to the virus. In Kambia, a fine of 500,000 Leones was imposed and several prominent healers in the district were quarantined after having come into contact with suspected Ebola patients. With the growing role of community engagement in the response, traditional healers were invited to join mobilization drives. Such integration was, however, accompanied with suspicion in Kambia, emblematic of long-standing ambivalent attitudes towards herbalists and their role in healthcare provision in the district.
Amongst those who had lobbied before the crisis for the inclusion of traditional medicine in the health system was Pa Yamba, a powerful herbalist from an area of the district renowned for producing formidable healers. His advanced age and wisdom made him the keeper of the trade's history, with prospective apprentices always welcome in his crowded house. As we sat on his veranda towards the end of the outbreak, Pa Yamba reflected on his own decision to take an active role in the response. In the early months of the state of emergency he volunteered for a job nobody wanted to do: he joined the burial team, tasked with burying the dead according to Infection Prevention and Control protocols as well as following the 'safe and dignified burials' guidelines. Burials, as noted above, were at the centre of the response and like traditional healing, they brought to the fore 'the core social conflict of the Ebola crisis: local beliefs and practices versus those of the international response' (Lipton, 2017: 804). The new regulations prohibiting the washing and dressing of dead bodies and the use of body bags by PPEclad burial teams caused significant tensions between the Ebola response agents and the communities where they intervened. Burial team members came to inhabit a challenging in-between space, as representatives of the response measures contributing to social disruption, but also taking the lead in efforts to humanize these measures and negotiating funerary protocols within communities (Lipton, 2017). Of course, the burial teams' task was also extremely risky and Pa Yamba remembered that when the job was advertised many refused. However, he argued that this was the only way he was able to continue feeding his family as he had stopped practising. In our conversations he emphasized his refusal to accept patients throughout the outbreak, theatrically demonstrating how he would turn people away and report them to the police in secret.
Two incidents he recounted stood out as symbolic of the ways in which people like Pa Yamba sought to navigate the crisis, mediating the local knowledge and the knowledge produced by the response in an effort to manage uncertainty. In Pa Yamba's narrative, it was this kind of epistemic navigation that allowed him to survive both the epidemic and the difficult terrains created by the response and its regulations. The first story began with Pa Yamba and his team being called out to a village in a different chiefdom. As they arrived in their response vehicle and approached the house, they met several women boiling pots of water, preparing to wash the body of the deceased. Pa Yamba gestured to his team to proceed slowly. Making no mention of the water being prepared, he asked one of the women to show him to the room where the deceased lay. He asked one of his colleagues to discreetly jot down the names of the caregivers as the woman enumerated them, to be used later for tracing any at-risk contacts. As he left the room and stepped into a backyard he met a huddle of men who, believing he did not speak their local language, made arrangements to attack him and his team. Pa Yamba remained calm and signalled to his colleagues that it was time to leave. As the men produced their machetes, the burial team jumped in the car and departed at great speed. On their way out, military officers on their motorbikes stopped the car, asking the team why they were driving so fast. After they recounted their story, the soldiers told them to turn around: they would escort them. When they arrived, Pa Yamba described how the soldiers forced the village men onto their knees, hitting them with the butts of their guns, whilst the burial team quickly changed into their PPE and buried the body. Before leaving, Pa Yamba made the villagers hold out the kasanke, the Muslim burial cloth, while he set it on fire -half procedure, half punishment.
The point of this story is not to report malpractice amongst Ebola response workers or to sensationalize the violence that was part of both disease containment and resistance to it. Rather, it is to highlight how someone like Pa Yamba, in telling this story, aimed to reposition himself from the danger he embodied as a representative of traditional knowledge to active citizen, enforcer of the response. As traditional practices were declared barriers to disease containment, he identified a threat to his livelihood alongside the risk that the virus posed to his life. His identity as traditional healer framed him as a potentially dangerous citizen to be disciplined. Through his role in the response, Pa Yamba was able to present himself as malleable and ready to change his behaviour, in contrast to the dangerous bodies in the village.
In his characterization of the villagers as resistant to change and potentially violent, Pa Yamba justified their disciplining by the soldiers and the powerfully symbolic burning of the kasanke as a final lesson in the importance of collaboration. This tactical shape shifting, however, also entailed a reinterpretation of the role of traditional knowledge in the response in ways that elevated him to the role of irreplaceable cultural broker. Indeed, he argued that he was a uniquely positioned member of the burial team -not despite his role as healer and 'society man' (a member of a secret society), but because of it. He recalled, for example, a second anecdote, where Temne members of the burial team had been sent to deal with the burial of a member of the Limba Gbangbani society. 11 Amidst widespread anxieties about new burial regulations, society burials presented particular challenges as they are 'secret events -not "secret burials" in the sense implied by Ebola responders, but closed events restricted only to members of the sodality' (Richards, 2016: 101). Pa Yamba noted that, since the team members that were sent over were 'outsiders' (not initiated 'society men' and from a different ethnic group), they were bound to fail in gaining the trust of those tasked with the funeral. As his prediction proved accurate, Pa Yamba was called in and was able to negotiate, convincing the Gbangbani to 'play the society' in the forest as the burial team buried the man according to the procedures. He described how the burial team members became afraid during the funeral as they heard the society's performance in the distance, but he assured them his efforts had pacified them. As a society man, Pa Yamba was able to be at once enforcer of dividing practices and a broker for reframing the response according to his superior knowledge of the community. Whilst accepting the premises of the emergency measures, he moulded the response in ways that made him and people like him essential to success, whilst simultaneously differentiating himself from dangerous others by disciplining them.
During the outbreak there was a spectrum of resistance to the knowledge and practices of the response, from refusal to comply with regulations to social commentary, rumours and conspiracies or the violent resistance plotted by the villagers in Pa Yamba's story. Yet juxtaposing engagement to overt resistance or rejection of dominant ways of knowing is a reductive representation of agentive possibilities. Pa Yamba moved tactically, renegotiating his role in the crisis. In his individual effort to manage uncertain terrain, Pa Yamba re-enacted and reproduced the logic of the response. His repositioning of traditional knowledge through the idioms of the response, of containment and engagement, reflects a particularly creative way 11. A male initiation society found primarily amongst members of the Limba ethnic group in Sierra Leone. As Goguen and Bolten (2017: 435) argue, the Gbangbani is particularly respected and feared in Sierra Leone, and because of its role in fighting witchcraft it is central to perceptions of safety within communities so that, for example, 'the importance to the body politic of handling the death of [an] initiator properly cannot be overstated'.
to manage uncertainty. This kind of epistemic navigation, in other words, echoes the fluid nature of identity and social order inscribed in the region's history, as individual efforts to survive insecurity have long required rapid responses to shifting terrains (Enria, n.d.;Lipton, 2017;McGovern, 2012).

CONCLUSION
The story of the West African Ebola outbreak will shape how future epidemics will be contained, how national health systems are rebuilt and how both governments and citizens in the three most affected countries rebuild their societies and their lives. In this article, I have aimed to retell the familiar story of the epidemic, by focusing on the epistemic struggles between different kinds of actors through their efforts to make the crisis 'knowable' and therefore manageable. In retelling this story, the article aims to render visible its underpinning assumptions and logic as well as their implications. A conjuncture of narratives about crisis and its roots, driven by different framings, from epidemiological considerations to the politics of international intervention and national-level negotiations, configured a particular set of relations and subjectivities during a state of exception. The coming together of state of emergency regulations on the one hand, and a growing concern for community engagement and behaviour change on the other, created a dichotomy between compliant and dangerous citizens in need of discipline and containment. At a time of great uncertainty, in other words, different framings collided to create a depoliticized plan for action with deeply political consequences. Appeals to security facilitated militarized containment practices and expedient complements to community engagement efforts, and thus dividing narratives about individual responsibility and the public goods. This in turn shaped the possibilities for individuals to navigate their own way out of the emergency. Some resisted, whilst others like Pa Yamba appropriated and reshaped the idiom of the response, thus reproducing its logic through his efforts to survive. Whilst subjects are made through interventions they also actively reproduce and mould them through individual efforts to navigate uncertainty. Drawing out the logic of crisis narratives and the response they made possible, brings to the fore a particular feature of contemporary humanitarian interventions, that is the co-existence of two types of approaches that appear opposed and are instead co-extensive: containment and engagement. Critical approaches to humanitarian and development interventions that paint them as efforts to contain 'surplus populations', in this view, are not contrasted by participatory visions. Rather, they rely on each other: securitization and resilience as the two faces of the same interventionary imagination. We see this as military actors increasingly become first responders in humanitarian crises, or as redirection of aid policies towards a focus on security goes alongside localization agendas. Ebola offered an especially visual representation of the containment/engagement logic through its quarantines and social mobilization drives -the symbiosis between social distancing and closeness. Yet these have been powerful mobilizing metaphors across crises, from migration to unemployment. As the humanitarian gaze has shifted onto the next emergency, including a current Ebola epidemic in the Democratic Republic of the Congo, it remains important to trace the everyday implications of the modalities of contemporary approaches to managing uncertainty.