Making (In?)Visible: Selectivity, Visibility and Authenticity in Cambodia’s Sites of Atrocity

ABSTRACT This article examines victim representation in former atrocity sites in Cambodia. It concentrates on Tuol Sleng and Choeung Ek, two prominent sites for the detention and torture of suspected enemies of the Khmer Rouge (KR) regime and the subsequent creation of mass graves. These sites of trauma, memory, and transitional justice are inextricably linked to the experiences of victims. Yet, the voice of victims, their visibility, and authentic representations of the past are not always centred in Cambodia’s atrocity sites. Drawing on empirical fieldwork in Cambodia and an interdisciplinary literature, this article interrogates the intersections between selectivity, visibility, and authenticity in sites of atrocity associated with the Cambodian genocide. As this paper demonstrates, engaging with these themes of selectivity, visibility, and authenticity, and by extension issues of, for example, “who” is recognized as a victim, “how” responsibility for past horrors is represented, “which” sites become tourist facing attractions, and “why” certain narratives of the past are prioritized, is essential to recognizing the rights, humanity, and dignity of victims and survivors.


Introduction
This article examines victim representation in former atrocity sites in Cambodia. It concentrates on Tuol Sleng and Choeung Ek, two prominent sites for the detention, torture, and eventual execution of suspected enemies of the Khmer Rouge (KR) regime. A minimum of 12,000 men, women, and children were detained and tortured at Tuol Sleng. 1 At the nearby Choeung Ek, a former Chinese graveyard, an estimated 14,000-17,000 of the KR's victims are buried in 129 mass graves. 2 Tuol Sleng and Choeung Ek have subsequently become crime sites at the Extraordinary Chambers in the Courts of Cambodia (ECCC), places of commemoration and memorialization, sites of education, objects of study, and the focus of considerable attention from international tourists. These sites of trauma, memory, and transitional justice are inextricably linked to the experiences of reference, combining structure with flexibility. A purposeful sampling methodology was employed with interviewees approached on the basis of their relevance to the research questions. Participants included victims/survivors of the genocide, former KR cadres, ECCC staff, site managers, local non-governmental organizations, historians, film makers, psychologists, and others. In total, thirty-one interviews were completed. 10 All interviewees were provided with a guarantee of anonymity. Accordingly, in the text that follows, we have used basic descriptors to denote interviewees. 11 Interviewees were asked a range of questions relating to the project themes, including questions related to the representation of victims and survivors at the sites, the scope for victim voice and agency in relation to the sites, ownership of the sites and their design and curation, the relationship between the sites and memorialization, and the ways in which the past is constructed and understood. While all interviewees were asked questions relating to these themes, the research instrument was adapted for different categories of interviewee (e.g. victims/survivors; site staff; NGOs and other elite actors etc.). Interviews were transcribed and transcripts translated by our research partners, DC-Cam. Extracts from these translated transcripts have been cited verbatim below.
The research team undertook site visits to the Tuol Sleng Museum of Genocide and the Choeung Ek Genocidal Center (also known as the "Killing Fields"), two popular sites of tourism and memorialization. 12 Where appropriate, we have sought to illustrate our points with the inclusion of photographs taken in Tuol Sleng and Choeung Ek by members of the research team. The team visited Tuol Sleng and Choeung Ek in January 2018 and again in November -December 2018. Photographs were taken of the permanent and temporary displays at the sites, in addition to general photos showing site structures, layout, and visitor interaction with the sites. On the second visit, the contents of the sites' audio guides were recorded in detail. Immediately after the site visits, the team recorded detailed field notes of their observations and reflections.
The article proceeds as follows. We first briefly introduce the case study of Cambodia and the sites which form the focus of this article. We then turn to an exploration of how victims' experiences are represented in the sites, focusing on the three key themes of selectivity, visibility, and authenticity. Turning first to selectivity, we critically examine which sites of atrocity have become tourist facing attractions and which crimes of the KR are visible and represented there. As we argue below, such selectivity has meant that in Cambodia, many landscapes and legacies of violence are "hidden in plain sight," 10 Ethical approval was granted by the authors institutional research ethics committee in late 2017. The ethical approval process involved the submission of a detailed seven-page memorandum covering managing the risks of trauma-traumatization; consent; confidentiality and anonymity; data protection; data retention and data sharing; and safety of the researcher. All participants were provided with a participant information sheet in advance of the interview, translated into Khmer and read aloud where necessary. The information sheet explained the nature of the research; why that individual had been asked to take part; the conduct of the interview (including location); how to withdraw from the project; confidentiality and anonymity; and data storage and retention. Participants who agreed to take part in the research were then asked to sign a consent form. 11 Please note that in some cases it is not possible to provide further descriptive information without rendering interviewees potentially identifiable. For instance, a relatively small number of staff work at these sites, so to indicate an interviewee's specific role would risk identification. 12 While not the primary focus of this article, we also visited Kampong Chhnang Airfield and Kraing Ta Chan Peace Centre, places of mass atrocity under the KR which also draw visitors, but which differ in terms of their visibility and recognition of victimhood. While the experience of violence extends across Cambodia, as a matter of practicality, fieldwork was restricted to the vicinity of the Phnom Penh area. The purpose of our visit was to provide a counterpoint to the focus on Tuol Sleng and Choeung Ek. For reasons of conceptual clarity, we do not focus on either site in this article. The contribution of these sites to our research will be discussed in our other publications.
while the range of harms that we "see" as opposed to those we do not "see" is carefully calibrated. 13 Closely intersecting with the theme of selectivity is that of visibility. In this part of the paper, we examine how Tuol Sleng and Choeung Ek and their transformation into internationally visible sites and tourist facing attractions was premised on, for example, a pragmatic politics of reconciliation, a need to remove the links between the KR regime and the new Vietnamese-backed People's Republic of Cambodia (PRK) government from visibility, and an attempt to legitimize the new state internationally and domestically through the narrative concentration of responsibility onto a small group of specific perpetrators. 14 We argue that privileging the "self-evident power" 15 of Tuol Sleng's iconic photographic display of its victims has acted to further this aim, creating exhibits and photographs that are graphic and poignant, yet often devoid of history or geography. 16 One particular effect is that it is not immediately evident that the majority of those imprisoned at Tuol Sleng were KR cadres themselves. 17 Third, under the theme of authenticity and focusing on the display of artefacts associated with the KR regime and its victims, we argue that applying a more critical lens to the representation of the past in these sites allows us to see how authenticity can be contested and negotiated. In conclusion, we argue that a selective and politically expedient engagement with the past and Cambodia's sites of mass atrocity has shaped how victimhood is "produced, perceived and interpreted" in these locations, 18 and that more might be done to recognize the rights, humanity, and dignity of victims and survivors.

Sites of Atrocity in Cambodia
A complete historical account of the Khmer Rouge regime is well beyond the scope of this paper. 19 However, by way of brief introduction, the Communist Party of Kampuchea (CPK), also known as the KR held power in Cambodia from April 1975 -January 1979. Their ascent to power followed a period of violence characterized by the loss of over 500,000 civilians during the US's bombing campaign in the Vietnam war 20 and authoritarian repression following a right-wing military coup in March 1970. 21 While the Khmer Rouge's victory in 1975 was initially celebrated by some as heralding a return to peace, what commenced was a period of severe and widespread suffering. The regime emptied towns and cities, forcing the population to march into the countryside to undertake forced labour. 22 While experiences of the regime varied across the country and between different social and ethnic groups, some common experiences included communal living; the regulation of marriage and familial relationships; the prohibition of religion and cultural rituals; and limited access to food, rest and healthcare. 23 Particular groups, including members of the former regime, city residents, Buddhist monks, and minority groups such as the Cham, ethnic Vietnamese, and Khmer Krom were particularly targeted for violence as the regime progressed. 24 Indeed, in the quest to create a "New Cambodia," the CPK enacted a series of policies that directly or indirectly led to the deaths of approximately 1,700,000-2,000,000 citizensthe result of extreme exhaustion, disease, torture, murder, and execution. 25 One of the defining features of "Democratic Kampuchea" (DK) was the establishment of security centres across Cambodia for the detention and torture of suspected enemies of the regime, and the subsequent creation of mass graves. The best known of these sites and those that are firmly on the tourist trail are the Tuol Sleng Genocide Museum and the Choeung Ek Genocidal Center. Tuol Sleng, or S-21, is a former high school in Phnom Penh, and was converted by the KR in 1976 into an interrogation centre. Its purpose was to document and punish perceived criminal offences against the state. 26 A minimum of 12,000 men, women, and children were detained and tortured at Tuol Sleng. 27 Only twelve are known to have survived. 28 The nearby Choeung Ek, a former Chinese graveyard, was appropriated by the KR in 1977 to act as a site of execution and burial for the bodies of detainees from Tuol Sleng and those who lived in the surrounding areas. An estimated 14,000-17,000 are buried at Choeung Ek in 129 mass graves. 29 Beyond Tuol Sleng and Choeung Ek, Jarvis notes that physical remains of the CPK's victims remain in "every single province, district, subdistrict and almost every village" in Cambodia. 30 Final resting places include former wats and schools, converted by the KR into security centres and warehouses; nearly 400 mass graves; caves, rivers, and wells; and fields and forests. 31 In the following section, we explore the implications of the selective development of certain sites into tourist-facing destinations.

Selectivity
The concept of selectivity 32 can be found in different guises throughout the literature on genocide studies, transitional justice, victimology, tourism and heritage studies, and  31 Tyner, Landscape. 32 When we use the term "selectivity" we refer to the dynamics of choosing that which will be included in a particular process (or in this case, site) and the simultaneous exclusion of others. memory studies. This diverse literature illustrates that in the aftermath of genocide, selectivity emerges particularly in the context of the narratives of the past that are given space, the victims we are willing to acknowledge, and the harms that we are willing to respond to. McCormack for example has noted that selectivity can be found in relation to the limiting of acts identified as international crimes and in relation to "the particular alleged atrocities the international community is prepared to collectively prosecute." 33 Likewise, in truth commissions, choices are inevitably made regarding the use of "window cases" or thematic hearings or the elevation of politically expedient voices. In the context of heritage development, Lennon and Tiberghien note the importance of considering "which heritage is interpreted and developed and what histories are overlooked." 34 As explored in this section, such selectivity is evident in Cambodia in relation to which sites of atrocity have become tourist facing attractions and which crimes of the KR are represented there. The Closing Order in the ECCC's case against the regime's "senior leaders" (Case 002) acknowledges that by the end of the KR regime, approximately 200 security centres and countless execution sites had been established throughout Cambodia. 35 A mapping exercise carried out by DC-Cam and the Cambodian Genocide Program established the existence of 19,471 mass graves, estimated to contain the remains of more than 1,000,000 people. 36 Yet, the tourist gaze is almost entirely drawn to Tuol Sleng and Choeung Ek. As detailed above, in part, such selectivity is inevitable when constructing responses to atrocity. That Tuol Sleng and Choeung Ek were prioritized as a matter of practicality was recognized by a number of our interviewees. For example, one dominant explanation amongst interviewees was that the development of one site over another was down to geographical positioning. Simply, the proximity of Tuol Sleng and Choeung Ek to Phnom Penh made them obvious tourist attractions. 37 Distance from Phnom Penh was also highlighted as a barrier to the development of other sites around the country as "until the roads got built ten years ago, there really wasn't an option to develop other sites because there was no way to get tourists to them or people to them." 38 Furthermore, both Tuol Sleng and Choeung Ek had been the subject of preservation measures by the Vietnamese soon after the fall of the KR, which protected them against physical deterioration. As observed by an interviewee: A lot of sites have been destroyed because budget limitations, because of political reasons, because of other things … When the Khmer Rouge collapsed, we still had civil war. It's difficult for us to preserve the place when fighting with each other until 1998. 39 Yet, the impact as Tyner et al note is that "these highly visible and officially commemorated sites serve to obfuscate other, more mundane sites (and practices) of violence" and that everyday "landscapes and legacies of violence … are hidden in plain sight." 40 The consequences are two-fold. First, it has engendered the impression that all violence was concentrated in Tuol Slenga point on which one interviewee was particularly critical, arguing that in Cambodia, "we do not have the voices of the many, but we have the voice of Tuol Sleng." 41 For historian Henri Locard, there is a need to focus on the experience of the "ordinary Khmer people" and the experience of imprisonment and violence across the country -"It's a little bit if you study the Nazi Regime, you concentrate on Auschwitz, Auschwitz, Auschwitz, nothing but Auschwitz. Yes, but what about the other concentration camps all over Europe? This is a little bit how I feel here." 42 Locard also raised the concern that the focus on Tuol Sleng and Choeung Ek feeds into historical revisionism and governmental attempts to downplay the extent of violence and imprisonment across the country. 43 Second, it has meant that other sites of violence and the experience of victimization across Cambodia have been rendered much less visible, obscuring the lived experience of millions. 44 Speaking on his experience at a forced labour site, one interviewee argued, I want all stories to be included related to the Democratic Kampuchea regime … the killings of the people from all over the country took place in this province … There was not enough food, just a very limited ration. It means the transfer to work … is not the killing but the overwork leading to sickness that people died one by one. 45 Equally, and as the above quotation suggests, such selectivity determines the victimizations we "see" as opposed to those we don't "see." 46 It also renders certain harms unrecognized andin cases where this invisibility is reflected in responses to atrocitywithout redress. 47 Williams for example cautions against the use of one location to relay a history of atrocity. He notes that "when a historic site is made to stand for some form of historic atrocity, a focus on physical location might mean that we miss much else of what was lost." 48 Indeed, Tyner has noted that in the case of Cambodia, structures linked to the regime's infrastructure projects, principally sites of forced labour and violence, are visible only to those who experienced first-hand the horrific conditions occasioned by the Khmer Rouge. 49 A parallel critique can be made regarding the capacity of Tuol Sleng and Choeung Ek to represent the full range of harms that were experienced during the KR regime and the importance of hearing a plurality of experiences on the past. As one interviewee 40  reflected, there were "tens of thousands of prisoners detained there." 50 The importance of recognizing the multi-faceted experience of victimhood and the diversity of victims' voices has become a strong theme in transitional justice. 51 On visiting Tuol Sleng and Choeung Ek, the primary focus on violent death is immediately apparent in the exhibition descriptions, displays, and audio guides. With an estimated 1,000,000-2,000,000 lives lost, a focus on death is entirely expected. However, the focus on violent death masks the range of harms that resulted in victimization and death, such as labour, starvation, and sexual and gender-based violence. Given that ninety-nine percent of survivors have reported almost starving to death, and ninety-six percent said they had been forced into slave labour, this selectivity obscures significant elements of life under the regime. 52 Broader patterns of victimization are also largely missing from the historical narrative. For example, the years leading up to the KR regime saw civilians suffer the brunt of armed struggle between a repressive monarchy and community insurgents, 53 a four-year US bombing campaign, 54 and the authoritarian violence of a right-wing military government. 55  audio-guide (introduced in 2015 57 ) refers to forced transfer and forced labour as forms of victimization and is accompanied by maps of transfers and photographs of forced labour in Building B. 58 The museum's director has previously indicated that such changes reflect a desire "to make the museum a more nuanced reflection of Cambodia's darkest time." 59 Likewise, at Choeung Ek, the audio guide (introduced in 2011) 60 notes that thousands died of starvation and disease, while others were "simply worked to death." It also contains the story of a survivor who was raped and draws attention to the gendered nature of torture and execution (especially the act of stripping women). 61 Tuol Sleng has also curated a range of temporary exhibitions, including a "Children of Angkar" exhibition (2017-2018) 62 which highlighted policies of forced labour, conscription, and education, and a "Sorrow and Struggle of Women" exhibition (2016) 63 which highlighted policies of forced marriage. Other organizations have also played a role in this diversification. For example, DC-Cam has contributed a range of photo exhibitions over the past twenty years, 64 and in 2009 collaborated with Sweden's Living History Forum to produce an exhibition reflecting on a Swedish delegation's encounter with the KR regime's propaganda. 65 While these are important contributions towards constructing a more nuanced version of the past, they continue to be overshadowed by the emphasis on violent death. As a result, experiences of harm which fall between "the cracks" can all too easily be collapsed or silenced. 66 Explored further below, Sontag's work on the visual can be extrapolated to the selective representation of the crimes of the KR in Tuol Sleng and Choeung Ek. Exploring the unique contribution of photography to the world, Sontag noted that photography introduced a new visual code which expands "our notions of what is worth looking at and what we have a right to observe." 67 Moreover, drawing on Sontag, Judith Butler argues that certain images are brought into focus "on the condition that some portion of the visual field is ruled out," thus governing "when and where a life counts as human" and therefore "when and where such a life is grievable." 68 The observations made by the research team spoke to the creation of such "grievable" and "ungrievable" lives with the promotion of some harms, such as violent death, and the relative invisibility of other harmsstarvation, exhaustion, and sexual violence, creating hierarchies of harm and hierarchies of victimhood. This has left some of the survivors we interviewed feeling that their history and their voices have not been adequately recognized in these sites. Being able to "see the whole picture of the person when they suffered" was considered essential to promoting dignity and an accurate memory of the past amongst interviewees. 69 As one interviewee argued, recognizing the full range of harms is also key to establishing "the full truth of what happened and why." 70 The processes of curation and selection that occur when representing the horrors of the genocide may therefore both orchestrate and foreclose ethical responsiveness to suffering, simultaneously engendering points of high and low visibility. 71

Visibility
Rafter argues that in terms of understandings of power, social control, rituals of death, and means of self-empowerment, much can be gained by considering the visual. 72 Accordingly, in the quest to take ownership of the past, forms of commemoration and remembrance, including the creation of memorials, the display of objects and the accompanying contextual narrative can become essential tools for new post-conflict or transitional governments. 73 Reflecting on this point, Lisle notes that "monuments, memorials, exhibits and installations about conflict always project backwards into the time they are commemorating, but their construction is determined largely by present-tense geopolitical concerns." 74 In Cambodia, the demands of political expediency and fostering the legitimacy of the new regime locally and later in the eyes of the international community demanded that the new PRK government delegitimate and distance itself from the atrocities perpetrated by the KR. 75 Tuol Sleng and Choeung Ek and their transformation into internationally visible and tourist facing sites became central to this cause. 76 In this section, we examine how the displays at Tuol Sleng in particular have created a visible and narrative concentration of responsibility onto five members of the KR while simultaneously masking who the dead are, where they died, and why they died. Our understanding of "visibility" thus relates to the "visuality" of the past at the sitesin particular what we are "made" or "allowed" to see, 77 and what is also hidden from view.
As detailed above, the objective of both Tuol Sleng and Choeung Ek was to expose the horrors of Pol Pot's regime to the wider world and legitimize the PRK. 78 Doing so relied on defining DK as a "fascist" regime and framing the regime's distorted socialist ideology as "genocidal," promoting comparisons with the Holocaust rather than the assassination of political enemies that at different times had marked the history of the Soviet Union, Communist China and Vietnam. 79 Similarly, the connections drawn between the KR and Nazi Germany served to distinguish the regime's atrocities from the "legitimate" communism practised in Vietnam. 80 Reflecting these objectives, a 1980 report from the Ministry of Culture, Information and Propaganda notes that the objective of the Tuol Sleng museum was "to show the international guests the cruel torture committed by the traitors of the Khmer people." 81 Likewise, the display of physical horrors associated with the regimephotographs of victims taken when they arrived at Tuol Sleng, 82 blood stained cells, and implements of torture further helped to justify the Vietnamese presence in Cambodia and its image as liberators from the "genocidal clique" of Pol Pot and others and to legitimize the new PRK government, itself dominated by former KR cadres. 83 Indeed, the audio tours which accompany visitors to Tuol Sleng and Choeung Ek identify Vietnam and defectors as the liberators of Cambodia, but contain little recognition of early Vietnamese alliances with the KR. As Sion suggests, the neatly arranged skulls at Choeung Ek may even be thought of as providing a "clean" "aestheticized" image of the past "neatly arranged in a window case," that furthers the demarcation between the current political regime and its predecessor. 84 What is also notable in both sites is the relative visibility and invisibility of certain forms of responsibility. Both sites contain exhibitions detailing the membership of the Central Committee, while references to the fanatic leadership and "absolute control" of Pol Pot appear at various junctures. 85 This focus on senior leadership is also reflective of broader political strategies. 86 Since developing a "win-win" policy to encourage the deflection of KR cadres through amnesties, reconciliation, and "burying the past" in the 1990s, 87 the current government has promoted a narrative that centralizes responsibility for KR crimes on a small and specific group of perpetrators. 88 This strategy has been particularly prominent in the debates surrounding the remit and personal jurisdiction of the ECCC; the Cambodian government has consistently resisted the prosecution of any accused beyond those indicted in Case 001 (Duch) and Case 002 (Nuon Chea, Khieu Samphan, Ieng Sary and Ieng Thirith). 89 As opined by one interviewee, this focus on senior leaders is popular in Cambodia: "the worst people of all were the ones at the top level, the ones who stayed on running the Khmer Rouge regime and they're the ones they wanted to see out on trial." 90 This narrative of specific individual criminal responsibility is visible in Tuol Sleng and Choeung Ek, 91 as is a narrative that the ECCC is the appropriate response to the crimes of the regime. Indeed, Lesley-Rozen has argued that "sites that were originally created primarily to justify Vietnamese occupation … have now been repurposed to serve as explanatory devices for the work of the ECCC." 92 Certainly, both sites contain specific (if out of date) 93 references to the work of the ECCC in their displays, audio guides and, in Tuol Sleng's case, temporary exhibitions. Similarly, each centres accountability in a way which emphasizes a small number of individuals as intentional actors, rather than the structural context and situational conditions which facilitated the torture and murder of thousands of people. 94 This limited focus and association with the ECCC was reassuring to some interviewees who had visited the sites and who admitted to having played a role in perpetrating violence during the regime: "I had been named in the past, scared, afraid of the arrest … now, no longer scared." 95 However, in restricting the narrative of responsibility, these sites also preclude any discussion of differences in experience during the period or across different areas or groups. 96 As noted by Manning, dissonance can emerge between the memories and perspectives of some survivors and the "projection of blameas substantiated and exemplified by Tuol Sleng." 97 Such observations link to the discussion above in relation to the selectivity of site development, as highlighted by one interviewee: "it's not the complete history but we need other factors, we need other things in order to get a fuller picture of the past." 98 Pertinent to the discussion of restricted responsibility is the fact that there are few indicators that many of those killed had been KR cadres themselves. 99 Indeed, as noted above, though graphic and poignant, the exhibits and photographs displayed at Tuol Sleng are often devoid of history or geography. 100 There is little to indicate who the dead were, where they died, and why they died. 101 Most photographs in Tuol Sleng are displayed as collectives arranged by a common trait such as gender or age, making them "at first … indistinguishable," 102 an "aggregate of victims. This was a deliberate choice on the part of the curator, and a policy which has extended to banning visitors who found their loved one from writing their name on their photograph. 104 The anonymity of the subjects and relative absence of explanatory information appears to suggest that these photographs are "self-evidential" in naturethat the images alone tell us what we need to know. 105 Yet, in Regarding the Pain of Others, Sontag argues for greater specificityarguing that compassion can become abstract when violence is pictured as a universal problem. 106 For those impacted, what matters may be precisely who was killed and by whom. Yet, the anonymity of the photos means that few questions are answered about the people who were killed at Tuol Sleng and Choeung Ek and precludes discussion as to why the regime killed so many of its own. As Cowie argues, while "visual images can be a window into other people's lives, it is not necessarily a transparent window." 107 Likewise, some photographs are repeated multiple times within Tuol Sleng, with particular images given more prominent positions than others. This appears to be largely the case with photographs of young peoplefor example Bophana, 108 a particularly wellknown victim of the regime whose image appears at least four times throughout Tuol Sleng, or the photograph of a mother with her young child, both reproduced below.
Williams suggests that such framing relates to an attempt to create "visual pleasure" in the tourist experience. 109 The number of photographs of women in Tuol Sleng also inaccurately reflects the number of women who were actually held there, as evidence suggests men were the majority of the sites' victims. 110 As Elander argues, the identification of a few "prominent" victims frames them as "martyrs" while the complexity around the rest of Tuol Sleng's victimsas former cadresis obscured. For example, the Tuol Sleng audio guide notes that 150 victims had previously been Tuol Sleng staff and acknowledges that many of the original CPK were ultimately brought to Tuol Sleng. Similarly, the Choeung Ek audio guide notes that some of those who were killed were KR soldiers from the Eastern Zone of DK. 112 However, this complexity is rarely visible in the visual exhibitions themselves. This was explained by one interviewee connected to the sites as being the correct, pragmatic choice in order to encourage engagement with the sites: "when they hear this is mostly the Khmer Rouge, they might not want to come or recommend to others." 113 Yet, obscuring this reality creates a situation where the victims are represented as having "neither real choice nor agency," 114 limiting the possibilities for engagement with "moral uncertainty" and depriving the viewer of insights into the complexities that separate "good and evil, life and 112 Author's fieldnotes. 113 Interview, 16 January 2018. 114 Williams,Memorial Museums,134. death, guilt and innocence." 115 Instead, "there are only perpetrators and victims, no one in between and no one both." 116 Given that many international visitors will have limited background knowledge of Cambodian history and politics, there is little opportunity for tourists to critically engage with the narratives presented at Tuol Sleng or Choeung Ek or to understand them in respect to their larger context or current relevance 117especially if they do not listen to (all of) the audio guide. Rather, sites such as Tuol Sleng have come to act as a lens through which the interaction between political power, atrocity narratives, and international audiences is filtered in a very visible way. As detailed above, patterns of visibility and invisibility dominate understandings of responsibility and victimhood. It has been argued that Tuol Sleng in particular has tried to achieve this aim through "a strategy of offending the eye and scandalizing the spirit." 118 In the absence of historical, social, and political context, the PRK and subsequent government's use of Cambodia's atrocity sites has created the impression that the violence of the KR was exceptional, unimaginable, and other. As atrocities across the world attest, it is not. 119 Yet the visual experience at Tuol Sleng in particular has contributed to a framing of the past in which the horrors of the KR regime are separated from present day politics, are capable of being boxed off and packaged for (predominantly international) public consumption and in which ethical distancing between political responsibility, victims and survivors, and the tourist gaze has been created. Such distancing and framing of the horrors of the KR as something to be gazed upon and left within the confines of Tuol Sleng and Choeung Ek we argue has the capacity to further create "a built-in estrangement from the world of other people's suffering." 120

Authenticity
In the Authenticity Hoax, Andrew Potter observed that authenticity "is something people definitely want. That is, when something is described as 'authentic', what is invariably meant is that it is a Good Thing." 121 The quest to preserve or engender authenticity, often through location, the use of genuine artefacts, and interpretations of death and disaster has hence become a key concern for former sites of atrocity. 122 That Tuol Sleng and Choeung Ek represent an "authentic" experience and accurate representations of the Cambodian genocide has become something of a leitmotif in the literature on dark 115 Ibid., 134. 116  boarding (author's fieldnotes). The Choeung Ek audio guide also ends on the following note: "Cambodia was not the first genocide, but we never thought it would happen here. It can happen anywhere … Tragically, it will probably happen again. So, for your sake, remember us, remember our past as you look into your future" (author's fieldnotes). tourism and promotional materials surrounding both sites. 123 For example, writing on Tuol Sleng, Tyner notes that "Tuol Sleng, as a materialized, authorized discourse, capitalizes on the site's authenticityit is, after all, the actual building in which prisoners were detained, tortured, and often forced to confess prior to their execution." 124 The "untouched experience" is manifested in the physical preservation of the perimeter fence of corrugated iron and the freedom given to visitors to walk into cells stained with blood and containing implements of torture in the absence of guards or other site personnel. 125 Likewise, at Choeung Ek, the preservation of approximately 129 mass graves, the housing of skulls and other bones in the memorial stupa and across the site at large, as well as the ongoing process of collecting and storing victims' clothes in a glass display box are used to testify to the horrors of the past. 126 The desire to preserve the sites as evidence was also reflected in our interviews, with interviewees noting that the sites need preserved "as evidence for the world" 127 and "the best place to show the public about the crimes committed." 128 Indeed, the latter interviewee also noted the role the sites had played in the prosecutions at the ECCC, highlighting the value of the authentic preservation of atrocity sites to broader transitional justice measures. Yet, there may be "strain between authenticity and evidence, on the one hand, and the desire to create an emotive, dramatic visitor experience, on the other." 129 Having spent time exploring Tuol Sleng and Choeung Ek and watching visitors interact with the artefacts and through interviews with survivors, we concur with Williams' proposal. Our observations prompted us to reflect on three particular aspects of the sites that speak to this "strain" between authenticity and experiencethe display of skeletal remains; the strategic positioning of implements of torture; and the contested provenance of certain displays in Tuol Sleng.
The prominent display of bones has long been associated with Choeung Ek, both in relation to the mass graves that mark much of the physical landscape, but also those which "appear" in seemingly random places as a result of topographical changes and are collected for safe keeping by staff members. 130 The rationale for the display of remains is twofold. First, to physically evidence something of the sheer scale of loss and the absence of care or dignity with which bodily remains were dealt with. Where there is little media coverage of an atrocity, as was the case in Cambodia, "human remains may be rare evidence in situations where there was otherwise little media entrusted to account for what was occurring." 131 Second is the attempt to capitalize on the inherent power of the body. 132 Renshaw for example argues that "The dead body as witness holds a power that little else can match because of its authenticity; the materiality of dead bodies enables representations of the past to be made without apparent authorship or mediation." 133 Indeed, "corpses are often made integral to the understanding of an event." 134 To this end, the visible display of skeletal remains assumes prominence, for the "authority of bodies is one that no other document or artefact can claim: they were once people." 135 Yet, just as Verdery points to the use of the dead for political purposes in Eastern Europe -"Dead people come with a curriculum vitae or resumeseveral possible resumes, depending on which aspect of their life is being considered," human remains have been "made to work" in Cambodia. 136 Memorials or memorial sites, Dwyer and Alderman remind us, are important symbolic conduits not only for expressing a version of history but for casting legitimacy upon it as well. 137 Perhaps unsurprisingly given the context, in Choeung Ek, the bones of the deceased quickly became political artifacts, acting to authenticate the crimes of the KR. 138 Thus, as Guillou explains, human remains, such as those on display at Choeung Ek, were not made physically visible as an act of memorialization or to promote reconciliation, but "became the best proof of the genocide and therefore a symbol of the legitimacy of the newly installed Vietnamese-based government." 139 As Verdery puts it, "if one wants to revise the past … it is comforting to have actual bones to hand." 140 At Choeung Ek, forms of manipulation of human remains continue today. 141 This is both part of the attempt to consolidate the government's preferred narrative of the past and to continue to "sell" Choeung Ek as a site of dark tourism. One of the ways in which this has been done at Choeung Ek is through the "strategic" placing of human remains at key junctures and points of visibility on the tourist trail around the site. 142 It would be clear to almost any visitor that the human remains exhumed at the site have beento varying extentsdisarticulated, collated and relocated (the placing of skulls in a stupa (see below) is a particularly striking example of this). However, while one might (perhaps) expect some disruption of remains during the establishment of a site such as Choeung Ek, observations by the authors and others suggest that some disruption continues. Writing in 2018, Caroline Bennett noted that every day at around 2pm, "a group of cleaners and caretakers gather to sweep leaves and other debris from the mass grave pits, and to collect bones and clothing emerging from the soil," and how during one visit, "a fragmented femur lay fully exposed in a chained off area that the day before had been empty except for leaves and dirt." 143 During the fieldwork on which this paper is based, a member of the research team had a similar experience, observing on one visit a collection of human teeth on a small metal tray at the edge of a path. These teeth had not been there when the same researcher had visited the site three days previously. These observations suggest that the mediation and staging of human remains continues to be practiced at Choeung Ek as a form of "staged authenticity" and means through which to "improve" the tourist experience. 144 The visible display of bones and skulls inside Choeung Ek's memorial stupa is a further point of contention regarding the use of skeletal remains. 145 The stupa, pictured below, was erected in 1988 and contains more than 8,000 skulls and bones which are visible through glass windows on all four sides of the structure. At first glance, the veneer of Buddhist authenticity is compelling and the memorial is seen as a culturally sensitive place of remembrance. 146 In practice, it is "wholly antithetical to Khmer religious practice." 147 According to Khmer Buddhist beliefs and funerary practices, people who die suddenly should be buried or cremated on site, as quickly as possible after death. 148 Violent deaths are considered to be particularly problematic and are believed to necessitate immediate cremation to allow "the spirit to move into the next karmic realm." 149 This point was stressed by one survivor, who explained that "Some people believe that if we don't cremate them, the victims will not rebirth." 150 From this perspective, Choeung Ek's stupa, complete with the comingled remains of multiple uncremated individuals who are neither separated by age or sex fails to offer its residents a dignified and traditional Buddhist rite of passage. Yet, as the interviewee continued, "If we cremate the remains, how could the world see it?" 151 These conflicted feelings were wechoed in other interviews: "if the bones are cremated, all evidence will be gone. If we keep, we disrespect the victims who died." 152 Another interviewee noted the importance of considering not only those who died but those who are left behind: "the cremation or the preservation has nothing to give the value to the victims … The important thing is our feeling. We pay the respect to them … we keep the evidence alive." 153 Tyner has suggested that the physical remains of those who died at Choeung Ek have become "commodities" in the quest for tourist revenue and "continue to represent the 'historical erasure' of Cambodia's victims." 154 Certainly, this point was raised by interviewees: "you sell the skull and the bones of the victim to show to the visitors rather than to respect." 155 As these quotations suggest, the stupa at Choeung Ek speaks to a host of complex and mixed emotions with regards to the balance between preservation, commemoration and the culturally appropriate treatment of the dead. 156 Human remains are not of course the sole vector of authentic or inauthentic representations of the past in Cambodia. The display of instruments of torture (shackles, chains, and iron rods for example) in Choeung Ek and Tuol Sleng is a case in point. Williams alludes to "the sticky question of authenticity" and the use of physical artefacts: if the museum conceives of objects as tools or even props supporting a historic story, how necessary is it really that all objects displayed have a provenance definitely known to be of the event, linked to a certain person and place? This idea … points to the importance of dramatic effect in making the history contained in any memorial museum sharp and compelling. In this, memorial museums are implicitly subject to the notion that being faithful to documentary evidence and serving the historical imagination may pull in different directions. 157 Choeung Ek's display of torture implements associated with the regime more broadly rather than the site itself exemplifies this tension in the historical imagination. 158 Their inclusion serves as a visceral means of driving home the brutality of the regime as experienced beyond the perimeters of the site itself. Yet, such a reminder is arguably unnecessary given the emotive and tragic nature of the surroundings in which they appear, and perhaps speaks more to the design and promotion of a site that capitalizes on the darkest aspects of our imaginations rather than the rights and dignity of the deceased.
Indeed, depictions of torture and torture implements in the sites were critiqued by some interviewees, with one arguing "it is not the purpose of education, it is just attractive for tourists." 159 The same interviewee even queried whether visceral displays might be counterproductive, by showing forms of torture that visitors could "learn and they can practice and they can try it." 160 However, it should be noted that others valued that these sites captured the worst extremes of the regime's violence and placed emphasis on evidencing what happened in broad terms rather than on the specific occurrences of the site itself. 161 Similar challenges exist in relation to Tuol Sleng. Tyner argues that "At one level, the museum seemingly provides an 'authentic' experience, one where visitors can enter into cells or interrogation rooms just as the rooms were when prisoners were actually detained and tortured." 162 Indeed, Tyner claims that "The power of Tuol Sleng, as a memorial site, lies in its purported authenticity." 163 We do not dispute these points. Rather, the capacity to see, hear and feel the interior of Tuol Sleng is central to its capacity to act as a site of truth telling. 164 However, as Chhabra explains, "authenticity is often staged and commodified to meet the needs of the tourist." 165 In Tuol Sleng further commodification is evident in the inclusion of facilities to improve visitor comfortthe changing of a prison block into bathroom facilities and the introduction of shops selling gifts and soft drinks. Views on this differed amongst our interviewees. Some accepted these changes as necessary to facilitate visitors, as one pragmatically stated: "that's okay for the selling of the water because the visitors need drinks." 166 Others considered changes an affront to the authenticity of the site and indicative of the attempt to use the legacy of mass victimhood for profit making purposes. One interviewee for example argued, "There are so many changes that it loses its originality," 167 while another opined, "they can build a nicer building but it's not meaningful, the memory is not about the expensive place and so on but it's about the memory of what happened." 168 Another worried that enabling food to be sold could show a lack of respect for victims: "it is not a place that they go to enjoy … eating and chatting … it shows disrespect." 169 Two closely interlinked displays in Tuol Sleng were problematized by our interviewees on the grounds of authenticity. The first was the inclusion of Vann Nath's paintings depicting torture. At least one of the paintings has been verified by the ECCC as accurate and accepted as a way to "acknowledge the suffering." 170 The authenticity of the paintings has however been queried by survivors who testified at the ECCC, with one witness describing the paintings as coming "from the heart" of the victims, who might be "too emotional," "so we don't know whether those paintings are really 100% accurate … 80% or so might be." 171 The second display that drew criticism was the gallows-type structure positioned in the courtyard outside Block C and occupying a prominent space in the site. Signage beside the structure and the site audio guide note that the gallows were originally used by the school's students during physical exercise but were appropriated by the KR as a mechanism by which to torture prisoners, principally by hanging them upside down and dipping their heads into one of three pots of rancid water underneath. 172 As illustrated in the photograph below, the display includes a human figure, placed next to the pots of water. On one level, this display may be thought of as having a performative dimension. Sather-Wagstaff and Knudsen and Waade assert that tourists not only gaze at physical artefacts but are bodies performing at specific sights. 173 According to Tiberghien and Lennon, by weaving a tactile body, movements, actions and emotions into the notion of performativity, visitors can authenticate places through their emotional connection to them. 174 Tuol Sleng's gallows may therefore be thought of as representing a prime site for such emotional authentication, both of the site itself and its role in the construction of the broader narrative surrounding the KR regime and political transition.
Yet the authenticity of Tuol Sleng's gallows structure was questioned by some of our interviewees, who claimed that this structure was not present in Tuol Sleng during the regime and suggested that instead a gallows type structure may have been present elsewhere in Cambodia. 175 It is important to note that the use of the structure was verified in the ECCC's Case 001: torture was found to include "plunging detainees' heads in a water jar and lifting by the hands tied in the back." 176 Furthermore, as noted above, the court accepted as evidence of this Van Nath's painting, and the signage at the structure reflects Vann Nath's testimony to the court. 177 Competing perspectives can also be found in the scholarly literature. For example, for Eggert, the wooden structure is a "historic witness" 178 to the violence perpetrated at Tuol Sleng, while Williams believes it to detract from "the site's faithfulness to history." 179 Whether the gallows were there or not, this dispute over the presence and use of the structure and the accuracy of paintings displayed in the museum speaks to the ways that memories of the past can be sites of "collision" or contest, 180 and the notion that authenticity "is a quality that … varies according to who is observing the object and in what context." 181

Conclusion
This paper has interrogated the intersection between victimhood and selectivity, visibility, and authenticity in two Cambodian sites of atrocity that have become popular tourist destinations. Throughout, we have sought to make an empirically grounded contribution to the existing literature on the use and management of atrocity sites, bringing the experiences and perspectives of victims and survivors into view. Drawing from our field research as well as diverse bodies of literature, we have aimed to contribute a victim-centred analysis which highlights the myriad complex relationships that exist in Cambodia between political narratives of the past, recognition of victimization and perpetration, and the preservation of "authentic" evidence for visitors and future generations.
As this paper has demonstrated, Tuol Sleng and Choeung Ek are physical and emotional repositories for many of the horrors of the past. The positioning of such sites as "must see" venues for international visitors raises critical questions regarding how victims and victimhood are represented and engaged with. Selectivity in representation can lead to the creation of "grievable lives" and "grievable harms" while simultaneously filtering out the broader context of violence and other forms of victimization. As a result, victims/survivors of violence may consider these sites to tell a valuable, but partial, story, while international visitors' understandings of past harm may be shaped by and limited to what they have learned at these places. This selectivity can be compounded by the visibility of only certain, politically expedient aspects of the past; this can mask complicated narratives of responsibility. Selective visibility in this context can preclude a more complete understanding of past atrocities and the choices that face those that are caught up in such violence and human rights abuses. Moreover, political considerations and efforts to render these sites "tourist friendly" can result in physical artefacts and human remains being used in ways that raise questions around the balance between authenticity, evidence preservation and respect for the dead. Taken together, the complexity of the past and the full range of harms and forms of victimization can be rendered invisible.
Such conclusions invite reflection around the relationships between these Cambodian sites of atrocity and the victims and survivors connected to those sites. We are not arguing that theirs are the only views to consider. As this paper has highlighted, sites of atrocity speak to many diverse purposes and audiences. If one such purpose is to provide a contextualized account of a past political regime, then other views may play an important role. Victims and survivors cannot be expected to account for more than their own experience, or to have insights into the objectives and intentions of violent regimes. Nor do we wish to suggest that victims and survivors will speak with one voice, easily incorporated into site management and design. As we have argued, part of the challenge lies in capturing the multitude of experiences endured throughout periods of atrocity. In the Cambodian context, additional complexity arises in the context of victim-perpetrators; if it is accepted that Khmer Rouge cadres were also victimized, then questions arise as to whether/how their views are also acknowledged.
While recognizing these challenges, we argue that the views expressed by our interviewees suggest that, where possible, more might be done to recognize the voice and agency of victims and survivors. This might be through the development of diverse sites to better recognize the breadth of harms or through constructive engagement with survivors and victims' families around the commercialization and alteration of sites. Our research suggests that Tuol Sleng and Choeung Ek were appreciated by participants for their role in preserving evidence and educating visitors about Cambodia's violent history. Yet, it also suggests that more might be done to recognize the rights, humanity, and dignity of victims and survivors.