Taking responsibility: Testimonial practices in Rithy Panh’s S21: The Khmer Rouge Killing Machine

Testimony and genocide films

Film has long been recognised for its capacity to document and produce information and in so doing, transform audiences (Torchin, 2012: 6). In straddling questions of ‘what to show, how to show it, and how much is too much to show’ (Wilson and Crowder-Taraborrelli, 2012: 6), many filmmakers working on genocide films turn to individual testimony as a way to narrate and evidence the atrocity (p. 9). Amongst such testimony centred genocide films, Claude Lanzmann’s Shoah (1985) is outstanding. For Shoshana Felman, the film created together with Arendt’s (1963/2006) Eichmann in Jerusalem a ‘conceptual breakthrough in our apprehension of the Holocaust’ (Felman, 2002: 106, italics removed), which produced a new idiom – witness testimony. Shoah engages with three kinds of witness testimonies simultaneously: Its focus on eye-witnesses and their recollections tests the limits of testimony; it probes the role of the spectator as a witness and it ‘embodies the capacity of art . . . to take the witness’s stand’ (Felman, 2000: 105, italics in original). The film is built completely on witness testimony with no archival footage. Lanzmann did not want to make a film about Shoah, and it was important for him that the film was ‘not a product of the Holocaust, that it is not a historical film’. Because there are no images of that which is most important, the film is instead ‘in the present’ (Lanzmann, quoted in LaCapra, 1998: 104), ‘a sort of originary event’, a ‘film that is the Shoah’ (Lanzmann, quoted in Brody, 2012, emphasis added). While certainly celebrated, the film and Lanzmann’s approach have also been criticised. Cultural theorist Dominick LaCapra (1998: 95) ascribes the tendency in contemporary culture to ‘sacralise the Shoah itself and to surround it with a taboo’ with Lanzmann’s Shoah. According to him, this ‘displaced secular religiosity’ stems from the film’s ‘tendency to grant the highest, perhaps the sole legitimate, status to the witness who not only provides testimony but who self-rendingly relives the traumatic suffering of the past’ (p. 101). For LaCapra, this witness has in Felman’s work become the ‘shattered, traumatized victim’ (LaCapra, 1998: 111). Thus, Shoah, alongside its commentary, demonstrates the slippages between the witness and the victim in relation to atrocity.

This epistemological assumption is challenged through the recent wave of perpetrator centred documentaries which are prompting discussions on the ethics of engaging with perpetrators of atrocity. While some critics warn that a focus on the perpetrator risks undermine the spectators’ moral judgement, many perpetrator documentaries are premised on a conviction that in order to understand atrocities, to prevent new ones and to repair those that have occurred, it is necessary to also take into account the experiences of the persons who committed them (Canet, 2020: 159–160). Examining films on/from the atrocities in Darfur, Rwanda, Indonesia, former Yugoslavia, Uganda and the Holocaust, cinema scholar Raya Morag (2020) argues that it is possible to identify a specific genre of perpetrator cinema, characterised by, amongst other things, an exploration of the ongoing conundrum about how ordinary people (‘the ordinary man’) can commit atrocity.

Rithy Panh’s S21: The Khmer Rouge Killing Machine epitomises what Morag describes as perpetrator cinema. As noted above, the film brings together a handful of men who experienced, either as its staff or its victims, notorious Khmer Rouge security centre S-21. At the heart of the film stand two questions: ‘How do you survive absolute horror? How do you become a killer?’ (Press Release, 2003). Before filming, Panh spent years identifying and convincing the former staff to participate in the film. Chairman Duch had been located by a journalist in 1999 but was held in military detention and inaccessible for Panh to interview. ⁷ Instead, S21 focuses on the experiences of one of Duch’s deputies, Him Huy, and 10 other former staff with varying level of responsibilities. These are contrasted and compared to the experiences of two prisoner survivors, Vann Nath and Chum Mey. Throughout the film, Vann in particular appears as a kind of moral compass, again and again returning the former staff to the fate of their victims. Like Shoah, S21 is a film in the present. Most scenes were filmed on site either at S-21, now Tuol Sleng Genocide Museum, or at Choeung Ek Killing Fields. Although archival footage is not outright rejected, it is limited to the opening scenes that introduce the Khmer Rouge era.

The film is based on a number of encounters: between the former prisoners and material archives; between the former prisoners and former staff and between the former staff and remnants of their past. Each encounter is a staged confrontation that raises its own set of questions concerning the ethics of the image, and the relation between mind, body and things in producing testimony. Each confrontation is a creation of testimony.

Techniques

Testimony

These are testimonial scenes, built on the testimonies by men who experienced S-21. Testimony is often understood as a verbal or written speech act, and at times explicitly contrasted with re-enactment or bodily expressions (see Margulies, 2019: 253). When contrasted, the difference tends to hinge on testimony’s alleged exclusive relation to rationality and reflection. There is no etymological foundation for such limitations. The testis is derived from the Latin word for third, terstis, and has little to do with the form of the testimony. Meanwhile, the English word ‘witness’ is a translation from the Greek word for ‘martyr’, and acts of martyrdom could (can) be in many forms. Neither of these confines the act to an oral act of reflection and rationality. Scholars that distinguish the re-enactment from the act of testimony as a question of rationality also point to the immediacy of re-enactment: it is hallucinatory (Rancière, 2009), providing access to the unconscious by ‘regurgitat[ing] the real in the form of repetitions that are unconscious, accidental, and compulsive’ (Margulies, 2019: 253). LaCapra (1998: 104) suggested the ‘acting-out’ in Shoah constituted a form of ‘reincarnat[ing] or reliv[ing of] the past in an unmediated transferential process that subjects one to possession by haunting objects and to compulsively repeated incursions of traumatic residues’. I do not depart from these readings, only that this would bar them from constituting testimony. Instead, I join scholars writing on bodies and testimony (Hirsch and Spitzer, 2009; Jeffrey, 2020; Medina and Henning, 2021) and suggest that in S21, the former staff testify with and through their bodies, by interacting with material object and in gesturers sharing past routines. This makes their bodies a ‘communicative medium’ in the face of a language destroyed by pain (Corcoran, 2020: 652, drawing on Scarry).

The fact that the men contributed, perpetrated or furthered the violence has no bearing on their ability to testify. In that sense, I disagree with Morag’s (2020) distinction between testimony and confession. According to Morag, our moral commitment to the victims of an atrocity means that a perpetrator can never ‘testify’ but only ever ‘confess’. Indeed, for her, the terms should be used to distinguish between victims and perpetrators, and the confession must be ‘based on an uncathartic, accurate, unconvoluted, cognitive acknowledgment of their crimes’ (p. 93). Yet very few (if any) of the perpetrators in the films she examines meet these criteria, and so at the heart of her analysis lies the failed confession. Her refusal of a perpetrator’s ‘testimony’ signals the broader moral and ethical concerns in the perpetrator encounter. Yet, an expectation of ‘full confessions’ misconstrues both concepts of confession and testimony, and, in extension, undermines attempts to understand how atrocities unfold and why and how people partake in them.

A ‘confession’ is a particular speech act with a long history in law and in religion. As Brooks (2000) explains, the confessional model so pertinent in the West developed over centuries of imposed religious confession by the Roman Catholic Church. This model ‘permeates our [Western] culture, including our educational practices and our law’ (Brooks, 2000: 2). Confessions are personal, self-reflective accounts that centre on one’s guilt or sin. As such, they have a different quality than testimonies in which a person describes what s/he saw (as third party testis) or experienced (as superstes). Indeed, a confession is never only about providing – or finding out – truth. There is always something more to it. Confessions are both constatives (‘I did this’) and performatives that stage ‘a scene of exposure, guilt, and retribution’, calling for absolution (pp. 20–21). Importantly, the performative of a confession positions the confessor in a hierarchical relation with the listener, where the latter is always in a position of relative power and authority, able to provide absolution or to punish for purposes of reintegration. Hence, confessions always sit uneasily in relation to questions of reliability and of voluntariness (p. 19). There are plenty of examples of forced confessions, and of false ones, which may have been given voluntarily but in no way reflect the truth. The forced confessions by the prisoners at S-21 are excellent examples of this. In recognition of these dangers, most legal regimes provide rules for when and how something can be accepted as a confession. While these rules are not safeguards against false or forced confessions, their existence is a recognition of the danger.

The former S-21 staff are not confessing in the scenes discussed above, but they are testifying. They are sharing knowledge of what they saw and of what they experienced. The expectation that they would or should confess relates, I believe, to our desire for them to take responsibility for their actions.

Responsibility

Even as the men are not confessing but rather testifying in the scenes discussed above, the testimonies say something about responsibility. However, these expressions of responsibility are difficult to hear within the contemporary paradigm of international criminal justice and individual accountability. Here, events and suffering are translated into crimes, and those involved are represented as victims, perpetrators or witnesses, with little to no possibility of overlap in subjectivity. Testimony becomes evidence, or, if given by an accused, potentially confession, but experiences are only ever relevant if they contribute to the establishment of an individual’s criminal responsibility. The international criminal justice paradigm affects what and how we can hear. As noted briefly in the Introduction, much of the academic scholarship on S21 utilises legal lexicon in the analysis, comparing the film or particular scenes to trials (Rachlin; Margulies). Yet, the expressions of responsibility in S21 are of a kind that would not necessarily be registered in a criminal court.

The testimonies provided in S21 by the former staff provide insights into the way these former perpetrators make sense of their own role and responsibility. In Alexander Hinton’s extensive anthropological study into why people killed during Khmer Rouge regime, ‘responsibility’ only features as displacement. The men Hinton (2004) interviewed would resort to ‘euphemisms and deflect[] responsibility onto authoritative figures’ (p. 187). Their deflection frustrates. A man Hinton calls Lor claims he was only following orders, and Hinton comments that his statement ‘is precisely [the] type of response that victims find so unsatisfying, since it absolves the perpetrator of responsibility and the need to personally express remorse’ (pp. 217–218). This sense of frustration is also apparent in Morag’s (2020) assertion that the speech by former cadres constitutes ‘failed confessions’. Meanwhile, in his interviews with former low level Khmer Rouge cadres, Williams (2018) found that they distanced themselves from any sense of agency or responsibility by pointing to strict divisions of labour. While the persons interviewed by Hinton and Williams, and depicted in the films examined by Morag, do not conceive of their role in terms of individual responsibility, some of their testimonies nevertheless do incorporate notions of responsibility.

Consider again the staged confrontation between Vann Nath and the five former staff. The repetition by the former staff of Khmer Rouge propaganda frustrates a contemporary audience if the expectation is a reflected confession. Yet, their testimonies are revealing for how the men conceive of the responsibility for the violence they were part of. Not only do they confirm the violence, but through their repetition of Khmer Rouge propaganda they are testifying to a worldview where their responsibility lay with Angkar (the ‘organisation’) and more immediately with Chairman Duch. Their responsibility was not with their fellow human beings. This kind of responsibility is not based on individual accountability or free will, but on a world view where Angkar is all-knowing and individual lives had little meaning or value. The men’s ‘deflection’ to Angkar testifies to a world that had clear divisions of power.

Units at S-21 had little to do with one another (Chandler, 1999: 17; ECCC 001 Judgement, 2010: 144–160) and responsibilities were strictly hierarchical. These strict separations of power were facilitated by the youth of many workers. Apart from a few ‘older brothers’ in leadership roles, almost all of the staff were men in their late teens to early 20s (Chandler, 1999: 18, 32–35): ‘young and impressionable staff’ from poor backgrounds and with little to no education (ECCC 001 Judgement, 2010: 162), whom Duch trained to ‘obey orders, to be cruel, to detain, to interrogate, to torture and to kill’ (p. 165). Poev, discussed above, was one of these boys. Orphaned as a child, Poev was raised and trained in a children’s units. He spent a few months in 1976 at S-21 as a young teenager to guard first animals, then prisoners. Even as he left S-21 to work in a rice field, members of his group were repeatedly taken away and killed. Later, he explained that they were constantly afraid and therefore never asked any questions (ECCC Trial Transcript, 2009: 64–74).

In resorting to the hierarchical structures they were in, the former cadres’ testimonies echo testimonies given by persons testifying from/on other institutional and bureaucratic contexts (see e.g. Cohen, 2001) including those beyond contexts of mass atrocity. Scholars of institutional responsibility have argued that the emphasis on individual criminal responsibility and subjective culpability in contemporary regimes means that legal institutions are poorly equipped to allocate responsibility. Officials called to testify in an Australian national inquiry into systematic child sexual abuse – a seemingly very different context – repeatedly referred to institutional divisions of power when trying to make sense of how the abuse had continued (Crofts, 2016: 2). For Penny Crofts, this displacement demonstrates how the emphasis on individual responsibility in criminal law and other legal regimes facilitates practices of irresponsibility rather than responsibility (also see Veitch, 2007). The way the former staff at S-21 make sense of their responsibility (or the limits thereof) is thus not unique to institutions where genocide or crimes against humanity are committed.

However, where Hinton and Morag read the institutional references as forms of displacement or failure, and Crofts points to the way the legal regime organises irresponsibility, Felman (2000) suggests that an act of testifying can in and of itself constitute an act of responsibility. This act of responsibility is, perhaps, not as ‘strong’ as an act of confession but it can be seen as a step towards (re-)establishing relations. This understanding of responsibility emphasises the way it is a fundamentally relational concept (Cane, 2002: 3), organising ‘key sets of relations – between self, others and the state – as relations of responsibility’ (Loughnan, 2020: 2). A breach of responsibility is thus a severing of a relation, a failure to uphold and maintain a relation between self and another. In that sense, attempts to re-establish relations by ‘commit[ting] oneself, and to commit the narrative, to others’ is a way ‘to take responsibility . . . for history or for the truth’ (Felman, 2000: 103–104). Here, the very act of testimony is an act towards (re)affirming relations: between the self and the audience, and between the past and the present. In filming S21, Rithy Panh told the former staff that he ‘would accept everything that they said’ on one condition: that he would ‘not find proof to the contrary of what they say’ (Panh in Oppenheimer, 2013: 246). The former staff, under the direction of Panh, commit to the truth. In testifying, they take some responsibility for what occurred.

Conclusion: Listening to testimony

In S21: The Khmer Rouge Killing Machine, men who experienced the institution provide testimonies. At times, this is in conjunction with material objects, like in the scenes where Vann and Mey engage with the archives, and where Poev takes on the ammunition bucket in his re-enactment. At other times, the men bear witness through their bodies, like in the scene with Poev. And at yet other times, testimony is produced through staged confrontations between the perpetrators and their victims, like those between the former guards and Vann. In the end, these are filmed, cinematic, testimonies. To conclude this article, I turn to how this has bearing on the way we, spectators of S21, engage with them.

There is a cinematic quality to the testimonies in S21. These are not confrontations and re-enactments within a judicial setting, nor literary testimonials, but rather staged and created as part of a documentary film. This matters particularly in relation to the bodily testimonies. The criticisms levied against the Act of Killing point to the risks in staging re-enactments with perpetrators of atrocity, both vis-à-vis the victims of the harm and the perpetrator. But as discussed above, S21 differs. The way testimonies are staged here provides for the spectator’s ethical engagement with the testimony.

Consider again the scene with Poev. In guarding prisoners, Poev inflicted violence and his practice took part of a regime that caused pain, suffering and death. Yet, although Poev is re-enacting his violent routine, the scene is not a repetition of the violence, nor a celebration thereof. In her work on re-enactments of speeches by mass murderers, Knittel (2019) suggests that the question whether re-enacting violent texts repeat their violence ‘hinges on the ability of repetition to create a difference’ (p. 173). Repeating the text does not automatically recreate the violence, but it does not automatically negate it either. Instead, for a re-enactment to be ethical, it must actively engage with the difference. While Boyle (2009: 95) suggests Poev’s re-enactment seems to collapse time, his commentary and out-of-body performance also ‘forfeit[s] its indexical bond to the original event’ (Nichols, 2008: 74). His re-enactment simultaneously evidences the original event, ties the present to the past and highlights the temporal gap. The re-enactment comes with a difference, one that produces distance to the event – different in kind from the way oral narration distances the original event but similar in quality in that it also enables ethical engagement.

This difference is accentuated through the camera work. The re-enactment is filmed and mediated by Panh and cinematographer Prum Mesa, whose positionality in relation to Poev makes all the difference. Whereas scenes like those with Vann and Chum in the archive are intimate with the camera closely following the men’s movements and gazes, this scene with Poev’s re-enactment remains at a distance. When Poev enters the room ‘filled with prisoners’, Panh and Mesa remain outside. Later, Panh explained that walking inside would have made him a perpetrator, as it would have involved walking on ‘the people who are lying there, who ask for help’ (w Boyle, 2016: 41). If the re-enactment itself entails both a collapse of time and an emphasis on the split, the camera’s distance accentuates the difference. For the spectator, this facilitates an engagement with their own positionality in relation to this difference.

It also matters for the way the spectators engage with the film that the men are testifying rather than confessing. To demand that a person who has participated in or furthered an atrocity only ever ‘confesses’ when sharing their experiences situates the listener/spectator in a position of superiority. As discussed above, as a performative act, a confession always takes places in a hierarchical relation where the listener holds the power to punish or absolve. To expect a filmed confession would assume that we, the spectators, should or even can provide absolution or punish.

This is also why S21 is nothing like a trial. There are aspects of the techniques employed in the film that mimic those used in legal investigations: In inquisitorial judicial systems like those of France and Cambodia, the suspect, witnesses and victims may, at the end of an investigation but before the trial, be brought together at the scene of the crime for confrontation and re-enactment. Indeed, a few years after S21 was filmed, this kind of reconstitution was organised as part of the investigation into the case against Duch at the Extraordinary Chambers in the Courts of Cambodia. Yet, there is a fundamental difference: While the men in the film S21 are testifying and there is plenty of evidence presented, there is no judge and no judgement. Instead, the former staff of S-21 testify to their experiences of having partaken in the atrocities at S-21.

Atrocities are (too) often announced ‘incomprehensible’. Trauma shatters, but responsibility for dealing with the atrocities does not only lie with those involved. To ‘comprehend’ is for Hannah Arendt (1967) to face up to reality, to not shy away from the hard stories. Rithy Panh’s S21 offers us, the spectators, an insight into the bureaucracy of torture and killings. The experiences of those who made up the bureaucracy need to be heard. For Panh, the aim of inviting the former torturers and killers to participate in S21 was based on a ‘belief in humanity’. ‘[T]he challenge’, as he saw it, ‘is to bring the torturers back to humanity. And that is done by the action of testifying’ (Panh in Oppenheimer, 2013: 246). In testifying, a person revisits the past and articulates events in words or in body. In listening to acts of testimony, a person is challenged to reflect on their own positionality. Listening does not, as Baines (2009: 164) argues in relation to the violence committed by former child soldiers, ‘deny the atrocities . . . committed, but . . . make[s] them intelligible’. Comprehension does not need to mean forgiveness or acceptance. For Arendt, it is as much about ‘facing up to’ reality as it is ‘resisting of, reality, whatever that may be’. And so, we listen so that we can resist. In S21: The Khmer Rouge Killing Machine, Rithy Panh stages testimonies in a way that challenges the men but without ever losing sight of the common humanity.