NATION-BUILDING CONFESSIONS: CARCERAL MEMORY IN POSTGENOCIDE RWANDA

The postconflict Rwandan state has crafted a “we are all Rwandans” national identity narrativewithout ethnicity, in the interest ofmaintaining a delicate, postgenocide peace. The annual genocide commemoration period called Kwibuka—“to remember”—which takes place over the course of one hundred days every year, is an underresearched part of this narrative. During the commemoration period, génocidaires’ confessions increase dramatically; these confessions lead the government to previously undiscovered graves all over the country, just as confessions given during the grassroots justice system—gacaca— did in themore immediate aftermath of the genocide. According to a prominent government official known for his prison outreach, the Rwandan government no longer provides incentives for prisoners to confess. Instead,he stated in a 2017 interview, thosewho speak up over twenty years later are simply“moved by the spirit of Kwibuka.”When confessions are made, memories of past action (“bad behaviour”) are used by the state, seemingly toward an ultimate end of reinforcing the nationalmaster narrative, to subsume the individual memories of innocent survivors into the national collective memory. This paper explores the questions around the state’s evolving use of prisoner confessions, both how those confessions are obtained, and how they factor into commemoration practices now.


ABSTRACT:
The postconflict Rwandan state has crafted a "we are all Rwandans" national identity narrative without ethnicity, in the interest of maintaining a delicate, postgenocide peace. The annual genocide commemoration period called Kwibuka-"to remember"-which takes place over the course of one hundred days every year, is an underresearched part of this narrative. During the commemoration period, génocidaires' confessions increase dramatically; these confessions lead the government to previously undiscovered graves all over the country, just as confessions given during the grassroots justice system-gacacadid in the more immediate aftermath of the genocide. According to a prominent government official known for his prison outreach, the Rwandan government no longer provides incentives for prisoners to confess. Instead, he stated in a 2017 interview, those who speak up over twenty years later are simply "moved by the spirit of Kwibuka."When confessions are made, memories of past action ("bad behaviour") are used by the state, seemingly toward an ultimate end of reinforcing the national master narrative, to subsume the individual memories of innocent survivors into the national collective memory. This paper explores the questions around the state's evolving use of prisoner confessions, both how those confessions are obtained, and how they factor into commemoration practices now.

INTRODUCTION
Memory is intelligent. It's a knowledge seated neither in the senses, nor in the spirit, but in collective memory. It is communal, though deeply personal. Involved with the self, though autonomous. At war with death.
-Etel Adnan In conflict-affected environments, states can manipulate collective memory in the interest of maintaining a delicate peace. In such cases, the state often sets the terms for what is and is not allowed into the public memory; peace is defined by the parties in power in government, and that definition often obscures or denies entirely any bad behaviour done by the state during wartime. This reinforces postconflict binaries of victors and defeated (corresponding to good and bad actors, respectively).
The postgenocide Rwandan state has crafted such a narrative. In the public, national memory, the ruling political party, the Rwandan Patriotic Front (RPF) liberated the country from génocidaires in 1994, taking back the country and putting a swift stop to 100 days of brutal violence. The RPF is frequently touted as a saviour force that has transformed Rwanda developmentally into "the Singapore of Africa," but it has also faced critique for its refusal to acknowledge its own violent role in the conflict before, during, and after the Rwandan genocide and for building what many consider an authoritarian government in the aftermath. As Phil Clark writes, "simply labeling the Rwandan government as authoritarian shrouds the complexity of political and social divisions" in the country. The RPF as a political party has a "tendency to pursue innovative social policies during the good times but to lash out during periods of perceived uncertainty." 1 Every year, Rwanda recognizes 100 days of official genocide commemoration, known as Kwibuka. 2 This period is distinct from the rest of the year in Rwanda for many reasons, many of which are easily observable, though discussion of the deeper nuances of and trauma within commemoration is very sensitive, 3 creating a "period of perceived uncertainty," per Phil Clark. During my 2017 field work in Rwanda, there was one particularly remarkable feature of Kwibuka that came up in a number of interviews and commemoration discussions that very few people seemed to have information about: during every Kwibuka, imprisoned génocidaires' confessions are said to increase dramatically. These confessions lead the Rwandan government to previously undiscovered graves all over the country. When a new grave is found, bodies are returned to their families and buried and mourned as part of their community's Kwibuka event. Less frequently, these bodies are interred in national memorial spaces. 4 The government's process for extracting or recording these confessions is opaque, which may account for why little to no literature on the Kwibuka spike exists. The confessions take place despite the fact that incentives (such as reduced sentences or even release from prison) for genocide perpetrators to confess ended over a decade ago. 5 According to a parliamentarian active in prison outreach, those prisoners who speak up now, over twenty years later, are simply "moved by the spirit of Kwibuka." 6 With this practice, the Hutu (perpetrator) contribution to national memory is one that reinforces its ostensible opposite: survivor memory. Memories of past action ("bad behaviour") are extracted by the state, seemingly toward an ultimate end of reinforcing the Rwandan nationalist project, which supports, follows, and elevates the individual memories of Tutsi survivors to create the narrative driving genocide commemoration today.
In State Repression and the Labors of Memory, Elizabeth Jelin speaks of a selective, national "master narrative" created through the manipulation of historical collective memory. While such practices can contribute to peace in the aftermath of violence, repressive peace is fragile, and reconciliation is a messy process. Jelin warns against agents of the state abusing their power to create official, hegemonic histories, which both erase "errors and missteps by those who are defined as heroes" 7 and selectively centralize national identity and social cohesion in an oversimplification of postconflict dynamics. This article is not intended to be a deep dive into the current conditions of the

METHODOLOGY
This research comprises primarily desk research around the Rwandan prison system and collective and historical memory, discourse analysis of the state-led narrative around commemoration practices and postgenocide reconstruction, and interviews and participant observations conducted during 2017 field research in and around Kigali. 8 When analyzing that data, I continued to be interested by the increase in confessions, but there was no space in that article to analyze it much further. It was a point that was raised early into my fieldwork, so I made a point to ask my interviewees if they had heard about increased confessions and, if so, what their interpretation of such a phenomenon was. Overwhelmingly, I was told that yes, confessions increase because discovery and exhumation of bodies increase, but that no one knew why this was happening. However, because I primarily used the snowballing method of identifying interviewees, this data is nonrandom and speaks only to a small subsection of the population in question. The advantage to this method was that, since my time in the field was limited, I was able to gain informants' trust quickly by using existing networks and always having an intermediary introducing me to the next person or group. The exception to this process was the interviews I conducted with youth members of the survivor group AERG; those interviews were arranged by leaders within the group and happened one after the other in a single day.
While it is exceedingly difficult to obtain factual information about the realities within a state prison system (even if one has visited the prison itself), there is value in speculation around the observable effects of the goings-on within those systems, as well as value in attending to rumours and known unknowns perpetuated about what happens inside prison walls. For example, as I will expand upon further into this article, it is not possible for me to know the circumstances under which confessions are taking place, nor is it possible for me to know for sure that these confessions are, as I was told by government officials, actually being recorded during the commemoration period. However, I know that the government is releasing the details of those confessions during Kwibuka. 9 I also know that the Rwandan citizens I interviewed are aware of the confessions and that many have either been personally affected by confessions or seen their community members affected. 10 Likewise, there has so far been a dearth of scholarship linking prison conditions and confessions in Rwanda with the commemoration period. The only direct link I have been able to find is in a 2010 Amnesty International report, which notes the increase in confessions, as well as an increase in accusations and convictions of genocide ideology and denial, and the heightened perceptions of insecurity during Kwibuka very briefly. 11 Even still, the most extensive discussion of these spikes is a four-sentence footnote. 12 All in all, two months' worth of research with thirty-eight interviews in total is not representative of an entire country. It would be impossible to draw conclusions from my small sample about all survivors' experiences of finding their buried relatives. However, my interviewees overwhelmingly shared an attitude that bodies found during commemoration are only ever genocide survivors'that is, Tutsi bodies. During interviews, when discussing "found families"-a term seventeen informants used, unprompted, to refer to the bodies of genocide victims exhumed during Kwibuka-all but two of those informants, including one government official, would make a point to emphasize that bodies were found on the property of "perpetrator families," meaning Hutu land. Lee Ann Fujii has written extensively about taking interviews' metadata into consideration-the tone, location, facial expressions, rumours, and other nonconcrete, subjective data can tell just as much as, if not more than, direct answers to interview questions. 13 During commemoration, survivor testimonies are centralized in the public space. But when those public events include a burial, then confessions-a perpetrator's testimony-are centralized in the private space. Without the release of these confessions, whenever they have taken place, commemoration events might not include burials, and survivor families might not achieve the closure they say they experience as a result of those burials. After I began including questions about Kwibuka burials in my interviews, all but three Rwandan interviewees (twenty-eight in total) told me that while they believed it could be possible for burials to take place during the rest of the year, they knew only of concrete examples of confessions leading to burials during the hundred days of commemoration.

Collective Memory and Trauma in Contemporary Rwanda
The concept of collective memory was first named by Maurice Halbwachs in 1925 and is broadly defined as memories attributed to more than one person. 14 While Halbwachs writes that the idea was novel, arguing that humanity is accustomed to thinking of memory as inhabiting and manifesting only in an individual, Nicolas Russell points out that the concept is one that has existed for much of human history, with scholarly work evolving to articulate it over time. 15 Collective memory has been observed across poetry, literature, and mythology, but Halbwachs introduced the modern concept that has since been molded and expanded upon for application to, among other things, concepts of national memory.
While historical and collective memory are conceptually similar, a distinction must be drawn between them. While exact definitions vary, for the case of Rwanda and the purposes of this article, historical memory will refer to observable facts, or semantic (objective) memory. Collective memory will refer to a collective manifestation of highly personal and subjective remembrance. Scholars such as Phillippe Joutard and Marie-Claire Lavabre, developing these definitions over time, have expanded upon Halbwachs's work to link the concept of individual episodic memory to the collective memory and identity "which exists outside of or persists through time." 16 In Rwanda, much of the publicly acknowledged collective memory around the genocide is linked to historical memory by the mechanism of trauma. Trauma affects "the ways we represent and communicate historical experience," 17 and is an emotional mobilizer rather than an objective one. Stefano Costalli and Andrea Rugerri justify the operationalization of individual emotions as mechanisms by which individuals are connected to a larger collective action. 18 While they were referring to armed mobilization, the logic can be applied to other cases as well. Emotions connect logic to actions, which in turn construct memory of lived experience in rich (albeit not always factual) detail. 19 Elizabeth Jelin states that "for the individual subject, the imprints of trauma play a central role in determining what the person can or cannot remember, silence, forget, or work through." 20 In conflict-affected environments, the "master narratives" Elizabeth Jelin writes about can be a tool for maintaining a delicate peace, but the sustainability of such practices is questionable. Memory is used by states, individuals, institutions, and various collectivities to narrate their pasts, investigate futures, and make sense of (or justify) the present. When linked to a collective trauma, memory can have the effect of reinforcing group belonging and subduing political dissent, as happens in Rwanda. The RPF has used collective memory of the genocide to build and bolster its "new Rwanda" nationalist project, 21 which subsumes the Rwandan population into one collective identity, relying (especially during Kwibuka) on "symbols and rites that enhance… empathy and identification." 22 Phil Clark characterizes the RPF as a "deeply divided, fragile, paranoid party." 23 He points out that periods of "perceived uncertainty," like elections, are prime for the Rwandan government's ruling party to lash out, making extreme, often repressive decisions that it perceives as protecting its power. 24 I would argue that Kwibuka also represents uncertainty for the Rwandan government because reports of violence and convictions of genocide ideology and denial increase, thereby disrupting social calm and introducing an atmosphere of unease and traumatic memory for quite a long period. 25 In response to (or perhaps in spite of) the uncertainty, the RPF and its bureaucratic arms shore up national memory to continually reinforce the collective, rather than the individual, to emphasize the "group's vision of its past by means of cognitive learning and emotional acts of identification and commemoration." 26 In Rwanda, collective memory has been curated, in the aftermath of a mass atrocity and widespread political violence, within education systems. 27 Societywide, postgenocide practices have been pulled from Rwandan history and imposed by the government in the present day, reinforcing a long, national memory that depends on a time before the genocide-both gacaca courts, which relied on accusers' memories of accused genocidaires in order to dole out justice, relying on memories "strengthened or diluted with the passage of time," 28 and ingando solidarity camps exist for political and cultural reeducation to fit a unified narrative rooted in a collective historical memory. 29 Some point to deliberate government manipulation of collective memory deployed in order to maintain social and political control. 30 Tamara Hinan argues that Rwanda has actually placed an emphasis on forgetting, rather than remembering, an emphasis notable both in the government's society-wide erasure of ethnicity and in its enthusiasm for one-on-one, individual-level reconciliation, which necessitates a "chosen amnesia" as a coping mechanism for coexistence. 31 Timothy Longman and Théoneste Rutagengwa have written on the constraints placed on Rwandans and the role of history and memory in those constraints. In this analysis, Longman and Rutagengwa frame the RPF's postgenocide actions as an "ambitious program of social engineering," 32 predicated on suppression of dissent and unfaltering dedication to the official historical memory. Public remembrance has grown to be "regarded as a duty and one of the main elements on the road to reconciliation." 33 Postgenocide public memory in Rwanda has been studied quite extensively, but little has been written on the specific role that the Kwibuka commemoration period plays in the making and remaking of genocide memory.

Kwibuka
Within recent work on collective memory and commemoration in Rwanda, 34 very little attention has been given to the entirety of the 100-day Kwibuka period, nor to the profound shift in rhetoric and attitude that exists in the country for those 100 days. News articles surface in early April, highlighting traumatic survivor experiences and presidential speeches, but international coverage all but vanishes after that. 35 All but four of my 2017 interviewees mentioned the widespread retraumatization that commemoration participants experience across the entirety of Kwibuka-not just during the first week of activity. After the first week, government representatives tend to highlight that the country largely returns to normal, though a Rwandan Broadcasting Agency executive confirmed that national television continues to provide "nonstop coverage" until the very last day of Kwibuka. 36 Despite that rhetoric, survivors have stated that they experience feelings of trauma and "heightened sensitivity" during the entirety of Kwibuka-not just during the first week. 37 While the Rwandan state characterizes Kwibuka as celebrating all Rwandans and the progress that the country has made, individual experiences of the commemoration period are in opposition with this. Survivors frequently say that they feel less safe during Kwibuka than they do during the rest of the year, and all of my interviewees who identified as survivors said that they perceive an uptick in violence toward their broad community. All of this creates a volatile atmosphere and increases a widespread perception of vulnerability. Evocation of memories of the genocide-a central component of Kwibuka activities through survivor testimonies and the exhumation and reburial of genocide casualties-exacerbate this sensitivity and cause many Kwibuka participants to experience PTSD episodes and mental breaks.
In 2010, Bert Ingelaere observed the following about public life in Rwanda: When Danielle de Lame conducted fieldwork in Rwanda in the late 1980s, she noticed that all public gatherings-whether festive religious events, ritualized public drinking activities or 'politico-private' gatherings-'serve to transmit meaning, provide the instruments of memorization, and create consensus' (2005: 303). What she saw as a cultural predilection for consensus was of course encouraged and enhanced after the 1994 genocide as part of the massive effort to restore order and maintain security. Sensitization campaigns, commemoration ceremonies, speeches by dignitaries, and reeducation programs-the socalled Inganda and Itorero-are intended not only to eradicate 'genocide ideology,' but also to promote a specific image of Rwanda. The weekly Umuganda communal labor activities, for example, are carried out at the local level, but they conclude with a speech delivered by centrally appointed leaders on a theme chosen by the government and published in the official gazette. All of these activities have naturally instilled a high degree of self-censorship among the Rwandan peasant population. 38 Both Ingelaere's 2010 observations and Lame's, many years earlier, align with my own, many years later, during Kwibuka. Government officials and other elites tout the decentralization of commemoration events as a net positive, a sign of progress, of power reaching the hands of the people, reconciling communities, that seems to be a continuation of the transitional justice period, during which the appearance of harmony was emphasized over longer processes of communal healing. 39 According to Ingelaere, the government in contemporary Rwanda serves as the centre from which not simply policy, but knowledge itself is "actively construed, managed, and controlled." 40 This is true especially of the Kwibuka period. The Kagame administration has, in recent years, though, outsourced commemoration planning to various individuals and committees and has ultimately left the majority of events themselves to be executed in and by smaller, more intimate communities (the periphery). However, the committees and organizations planning are still very much directed by the central government. 41 All Kwibuka themes, public lectures, and programming follow a thread that may be developed in the middle space (between centre and periphery) but that ultimately requires approval from the centre. 42 Programming is delivered and guided by "survivor organizations." Trauma counselors are deployed to all corners of the country to provide mental health support for traumatized participants; those counselors are trained and provided by survivor organizations and, according to interviews with organizers and staff at the National Commission for the Fight against Genocide (CNLG), almost all trauma counselors undergo training specifically developed and carried out by an extension of the Ministry of Health called the Rwandan Biomedical Centre.
In stark contrast to this three-month emphasis on public memory of genocide, rhetoric outside of Kwibuka is extremely subdued. The other nine months of the year are characterized by the illegality of "genocide ideology." 43 Under this law, identifying Hutu, Tutsi, or Twa-except in a historical context-is outlawed under the guise of achieving homogeneity in the new Rwanda ("We are all Rwandan"). 44 The ambiguous legal definition of genocide ideology has drawn international criticism due in part to its use by the state for suppression of both antigovernment sentiments and historical concerns over possible war crimes committed by the RPF. 45 The genocide-ideology law reinforces the Rwandan state's master narrative of the RPF and Paul Kagame (a key member of the armed group when it declared the génocidaires to be vanquished in 1994) as the saviours of Rwanda, ushering in an era of peace and unity.
Timothy Longman and Théoneste Rutagengwa have written on the constraints placed on Rwandans and the role of memory in those constraints, all with a heavy historical focus. In this analysis, Longman and Rutagengwa frame the RPF's postgenocide actions as an "ambitious program of social engineering," 46 predicated on suppression of dissent and unfaltering dedication to the official historical memory. Though nominally rooted in history, this manipulation and standardization have significant implications for the reinforcement of collective memory. With all Rwandans publicly adhering to the collective history that says that ethnic divisions were never natural to their country and are now abolished, the state has constructed a baseline from which to enforce its legal mandate erasing ethnicity from Rwandan discourse. Public remembrance has grown to be "regarded as a duty and one of the main elements on the road to reconciliation." 47 However, Rwandans continue to privately acknowledge ethnic difference, and my research has shown two things regarding ethnicity during commemoration. First, coded language stands in for Tutsi and Hutu as "survivor" and "perpetrator," respectively. Second, the commemoration period stands as an exceptionalbeit selectively-to the public outlawing of ethnicity, meaning that during

V O L U M E 1 4 N U M É R O 2 A U T O M N E / F A L L
Kwibuka, Tutsi identity is regularly spoken about (though "survivor" is still used a great deal as a stand-in). In contrast, Hutu identity is still rarely named. When these identifiers are used, their reference point is always the genocide, which means that when Tutsis are named they are often valorized as survivors, while mention of Hutu identity frequently recalls their genocide-era status as perpetrators of violence. 48 During the rest of the year, memorials, museums, and preserved sites of violence (some of which shock in their explicitness) 49 exist as reminders of the conflict that include references to ethnic identity, but those memorial spaces are static and exist more for tourists and foreigners than for Rwandan residents. 50 Commemoration, on the other hand, is a dynamic process deploying and modifying memory through active participation.
Kwibuka points to the successes of the state and the gratitude that all Rwandans should feel toward the RPF for liberating the country from violent destruction. The consistent, "never again" rhetoric, while internationally popular, subtly threatens the possibility of resurging violence; much of Kwibuka literature, from presidential speeches to community-level "curricula," contributes to the trauma and creates a sort of political paralysis: if we do not remember the horrors of our past, then your home may once more erupt in violence, and you will once again witness or be subjected to terrible things. In his 2014 Kwibuka speech, President Kagame warned that "if people's choices are not informed by historical clarity, the danger is ever present." 51 Since taking power, the RPF has forged its own historical memory in the aftermath of conflict which has, in turn, dictated the public expectations of Rwanda's collective memory. Both of these state-driven memory narratives are essential to the RPF's remaking of a unified Rwandan identity, 52 transitioning from ethnic division now globally recognized as having given rise to violence so horrific that it is frequently referred to as unspeakable. 53 Kwibuka is meant to reinforce this unification, but often ends up exacerbating and highlighting difference instead.

Exhumations and Commemoration
One interviewee in 2017 told me, "I don't know how it happens, but […] the Kwibuka period is when most people find their relatives." 54 She recounted a recent story of a man's body that was uncovered during the commemoration period while a house was being built. His family knew roughly "where he had been thrown" but had been yet unable to find him. His body was eventually found when the builders were digging the toilet pit.
Another interviewee informed me that bodies are almost always found on "perpetrator" property, and flatly stated that she did not care whether the occupants of the land knew whether or not the bodies were there because she felt that genocide survivors would never move onto land where genocide victims were buried. 55 She likewise told me that anyone occupying land underneath which graves were located did not have a right to request that their property or infrastructure be disrupted as little as possible in the exhumation process. The only priority, she said, is the bodies. She, too, told me that every instance she knows of bodies being exhumed has happened during Kwibuka, and she believes that bodies are still being found because "those who killed" are still confessing. Finally, she concluded that she believes many perpetrators are now walking free, despite having never confessed, and feels resentful that this is allowed to go on. When I asked her to elaborate, she told me that all perpetrators have a responsibility to confess so that survivors can find their families and achieve closure through burial. She rejected the notion that those who are free may have already confessed all they know.
In line with this, still another interviewee told me an anecdote about a mother and her daughter who had just moved into a new home, when an imprisoned génocidaire confessed to burying bodies on what was now their property. 56 Despite the women's pleas that they were unaware of the grave, their house was demolished in the exhumation process (the interviewee did not know whether or not the confession and discovery happened during Kwibuka). I was also told shortly after arriving in 2017 that a petrol station near the popular Nyabugogo bus terminal had been razed just two weeks before (early May 2017, during Kwibuka) because a confession had led to mass exhumation of a number of bodies buried there. 57 In 2018, the survivor organization Ibuka was part of a massive search for and exhumation of bodies in Gasabo district launched on April 11-the start of commemoration-and homes had been or would be demolished in order to uncover many of the graves. This action reportedly came after a "defendant accused of participation in the genocide revealed information about the burial sites." 58 The number of informants for this research is not large enough to reach any conclusive stance about whether nonsurvivor families are treated differently from the way in which survivor families might be treated when lives and property are disrupted in the interest of finding and exhuming bodies. But the woman who stated that bodies are always found on Hutu-owned land was not alone; around half of my interviewees expressed a comparable sentiment in some way or another, though none quite as directly. More often, I was told that such things could not possibly happen where survivors live, an exceptionalism that was likewise expressed in statements that survivors could not possibly inflict violence against other people. Once again, Lee Ann Fujii's work is critical here to understanding why interviewers should pay attention to interview metadata in order to grasp the many layers of Rwandan social realities. In a society where individual identities are politicized by the state, it can be difficult to move beyond surface-level regurgitation of or interaction with the state's master narrative. 59 As Helen Hintjens has written, "one cannot expect frank answers to […] sensitive questions, yet these are the questions that need to be asked." 60 Whether there is a widespread disruption of nonsurvivor property during commemoration or not, these 2017 interviews indicate that there is a perception of disruption and destruction of property, and a belief that such disruption and destruction are deserved or just. Likewise, the perception is that these small-scale exhumations (I was told, for example, that fifteen bodies were unburied at the site of the razed petrol station) directly result from génocidaire confessions. In the case of the Gasabo district exhumations, an Ibuka representative told the Associated Press that these exhumations would lead to relaunched investigations and further prosecutions of anyone found to have participated in the killings who had not already been identified. 61  62 These same survivor organizations have an intimate power over the design and implementation of commemoration activities and curriculum annually, even now. It does not seem a far stretch to guess that these two tools of influence may enjoy some overlap.

PRISON
The Rwandan transitional-justice process has come under increased criticism in recent years, and, while much of that critique is well-placed and important when considering transitional justice after future atrocities, it is likewise important to note that all states struggle to identify the "right" ways to pursue justice and reconciliation. Kristin Doughty writes that much of the "violence of social repair" is not necessarily dependent on the state-driven reconciliation process and its flaws, but that any reconciliation efforts of this scale will be inherently fraught. 63 The Rwandan approach was, at least, a divergence from the global norm. In particular, its use of gacaca community courts-a return to traditional, grassroots forms of justice in response to the sheer numbers of accused in the aftermath of the genocide-is closely linked to the ways in which memory, perceptions of justice, and punitive mechanisms are deployed as part of commemoration today.
Anu Chakravarty's research on the "nuts and bolts" of the gacaca courts reveals the uncertainty of that period; she points out that ordinary citizens were encouraged to prove themselves to be "patriotic Rwandans" by undercutting local-level social ties, denouncing neighbours and family members in the service of a greater good, in the service of a grand unification project. 64 She further points out that social ties in reconciliation processes have long been important to how justice is or is not carried out in postgenocide Rwanda; Chakravarty's research shows that gacaca judges were quite capable of impartiality if they had no prior links to the accused, but that they did often give more lenient sentences to people whom they knew, to whom they were related, or from whom they had something to gain. 65 In the early days of Rwandan transitional justice, génocidaires who admitted to their crimes before they were identified by survivors were typically given more lenient sentences. One problem with this, highlighted by Timothy Longman, was that the incentive for false accusations was quite high. Indeed, an incredi- ble number of prisoners in Rwanda are estimated to have been falsely convicted, often in the service of accusers' personal or political vendettas. 66 Following the genocide, thousands of people at a time were rounded up-often, their names would be written on pieces of paper, indicating that someone had named them genocidaires-in the middle of the night, thrown into prison, and never given a proper trial. 67 Over several years of interviews, many prisoners in Rwanda told Carina Tertsakian that they were optimistic that everyone would someday get a trial, or that someone would vouch for them in a gacaca process. However, many also told Tertsakian that they had no faith in the system; without lawyers, they did not see a path to justice. 68 Gacaca ended in 2004, at which time prisoners were given a final opportunity to "confess their guilt and ask for forgiveness" before a state-determined deadline. 69 From gacaca onward, there have been no (publicly known) incentives offered for confession; this fact was reinforced in 2017, in an interview with the aforementioned politician active in prison outreach. Indeed, since gacaca ended, the justice process in Rwanda has shifted from "a system designed to facilitate confessions to a process driven by accusations." 70 Though my research provides preliminary evidence that this accusations-driven process is amplified during Kwibuka, the lack of large-scale and long-term projects on the commemoration period represents a serious gap in the literature on ethnic identity and interethnic trust in postgenocide Rwanda.
Gacaca was a means by which to gather information about the locations of bodies countrywide, and finding those bodies often meant a great deal to survivors, particularly the families of those exhumed and subsequently reburied. 71 However, as some have pointed out, the process of exhuming and reburying, particularly as part of an established, public memorial space or commemoration activity, can be extremely traumatizing for survivors. 72 Additionally, exhumations and continued confessions and reburials can widen the divide between social groups. The Rwandan daily The New Times published an editorial during Kwibuka 2018 that stated explicitly that the continued discovery of mass graves suggests that many perpetrators "never bothered to reveal [their locations] to bereaved families so that they can get closure." This same editorial prods at existing social fissures, ending by saying that "some very cruel people still live in our midst." 73 Such discourse places the responsibility of reconciliation through confession on one group, Hutus, while simultaneously positioning them as threats to the survivor population.
The discovery of four mass graves just outside of Kigali prompted the abovementioned editorial. These graves were found "hardly a week" into Kwibuka, which means that the divisive language used by the publication called doubly upon the heightened tension that my interviewees discussed. The piece makes it sound as though these graves were found by chance, which is certainly possible. However, the serendipity of a spontaneously discovered mass grave (at the time of the editorial's publication, over 200 bodies had been exhumed and more were expected) just at the start of commemoration is suspect. Timothy Longman writes: "Rwanda's genocide trials contradicted the public rhetoric of the regime in another important way. Although the rejection of ethnicity was a central component of the RPF's public rhetoric, the genocide trials actually served to reinforce the centrality of ethnicity within Rwandan public life." 74 The trials have been over for over a decade, but Longman's observation, like so many around the distribution of postgenocide justice, holds for commemoration activities. During each 100 days of Kwibuka, ethnic designations are selectively revived ("Tutsi" much more freely than "Hutu"), which in turn does the work of reinforcing the divisions that caused the genocide despite an official policy of ethnic nonrecognition 75 -the result of which is that no one discusses ethnicity at all-during the rest of the year. 76 While commemoration events look different throughout the country and certainly take place in villages that would be all Hutu (were Hutu still an official identity group), for communities made up primarily of survivors, an essential component of Kwibuka programming is the burial of found families. This practice relies heavily on the Rwandan prison system and the state's decision to extract and release confessions.
The Rwandan prison system is, like so many, fairly impenetrable. Several reports have been released in recent years that detail use of torture to extract confessions. The assertion is that these confessions keep arising anew year by year; multiple interviews in 2017 held with high-level government officials, Kwibuka organizers, and memorial archivists revealed that those in power maintain a somewhat outlandish suggestion that prisoners are simply choosing, of their own volition, to confess more frequently during commemoration than they do during the rest of the year. If we are to understand legal systems as exerting the coercive power of the state that controls them, 77 then it would be unwise to accept the state's claims of prisoner goodwill-being moved to confess, en masse, by some sort of spiritual pull, as asserted by the parliamentarian-without evidence beyond the claims on their own.
The interplay between the government's control over the commemoration programming and the government's control over the prison systems within which people confess but raises ethical questions about the retrieval and deployment of certain people's memories. If prisoners are confessing only during Kwibuka, why is that? To be emotionally moved by the power of the commemoration period alone, as I was told, is a nice sentiment that at face value fits neatly into a unity-and-reconciliation narrative, but it lacks evidence and empirical explanation. On the other hand, if prisoners are confessing year-round or have already confessed years earlier, then we must ask why the vast majority of found bodies are located and exhumed during the same hundred-day period of every year, rather than being exhumed immediately after confessions are made.

CONCLUSION
In his 2016 memoir, Hisham Matar writes: "One of the frustrations of prison life, which is also one of its intended consequences, is that the prisoner is made ineffective. He is unable to be of much use. The aim is to render him powerless.
The frustration builds up until he takes an unreasonable risk." 78 The point is well placed and likely true of many prison settings. But what happens when the prisoner, both as an individual and as a symbolic entity, is rendered ineffective to his original cause or to himself, but at the same time is being instrumentalized by the state? What happens when a prisoner's presence, words, and inaccessibility allow the imprisoner to manipulate all three to formulate a component of the master narrative? When do a prisoner's words stop being his own? Do memories cease to belong to the memory-bearer after they have been shared?
To remember a location, a space where bodies were buried, is a memory interaction-a web of activity emerges, and one person's memories interact with those of others. If a génocidaire confesses a burial ground or a mass grave, that confession has a profound effect on the memories of the dead's still-living family. On a larger scale, the exhumations that follow constitute a piece of the greater, collective memory that links to the new Rwanda, a postgenocide nationalist unity reliant on invocation of a violent past and a future rooted in "never again." This research, as it currently exists, cannot shed light on what is actually happening in Rwandan prisons. It can, however, highlight that the deployment of confessions to uncover mass graves and exhume bodies relies on a prison system known for its abuse of ambiguous laws, its secret detainments, and its employment of torture. 79 A few possible explanations for the increase in confessions present themselves, though all is speculation given the already-discussed opacity and inaccessibility of the system in question. First, it is possible that prisoners are simply choosing to confess year after year, perhaps moved by guilt or personal trauma that either reaches a breaking point naturally or is exacerbated by Kwibuka activities, or perhaps following intraprison social norms that have developed over time. 80 Anu Charkravarty has pointed out that, counterintuitively, the number of confessions was very low in the early years of the gacaca courts. Prisoners took an unexpectedly long time to respond to incentives like temporary parole or release; Chakravarty's ethnographic work indicated that this low response to incentives resulted from a lack of trust in the government. Prisoners thought that confessions were simply a way they could be identified as guilty and subsequently punished indiscriminately. 81 Given this, perhaps over time, prisoners have, indeed, experienced reform and reconciliation on a personal level (which could in turn be influenced by the social influence of seeing other prisoners confessing over time) and chosen of their own volition to reveal burial sites. Carina Tertsakian's research describes how disconnected prisoners were from external society, and explains the unique social structures that arose in part as a result of this disconnect. 82 It is not outside the realm of possibility that late confession has become normalized as prison subcultures and social ties solidify over time.
Another, more sinister possibility is that the government is using torture or false promises (such as reduced sentences) to extract confessions. While there are credible reports of this happening generally, these reports often point to false confessions of more recent crimes like violation of the genocide-ideology law. 83 For torture to be used to extract truthful, provable confessions of genocide-era burials seems less likely, as génocidaires have now been imprisoned for many years; however, if there are perceived negative effects of confession-secretly extended sentences, perhaps, or in-prison ostracization or harassment following confessions-then to delay offering incontrovertible evidence of guilt by confessing grave locations would make slightly more sense. On the other end of that possibility, the prison system could be offering incentives, real or fake, to prisoners in exchange for confessions during Kwibuka, and simply is not publicizing that policy.
The final speculation that this research presents is that the Rwandan government is sitting on backlogged confessions and strategically releasing grave locations to coincide with Kwibuka every year. While this is a serious suggestion-that I must, once more, emphasize is speculative-it would align with the RPF's known track record of memory manipulation and social control. Considering that the state controls Kwibuka programming and commemoration messaging from start to finish, control over and instrumentalization of both memory extraction and deployment does not seem so far off.
Timothy Longman points out that "the confession process [during gacaca] encouraged perpetrators to seek reconciliation with survivors and provided important information about what happened in 1994. Learning the fate of specific relatives and, in some cases, being able to recover and bury their bodies was particularly important to many survivors. Yet many people were frustrated by a general lack of truthfulness in gacaca. Many Hutus complained of false accusations, while Tutsi survivors complained that too many Hutu refused to admit their real role in the genocide or implicate others." 84 This is still going on today, except that now face time between confessors and victims' families appears to be nonexistent. Confessions happen inside prisons (according to the government) and on the outside, while commemoration participants speculate about the truthfulness of accusations, 85 claim commemoration as a time for survivors only (rather than being a broader continuation of reconciliation), 86 and complain that "perpetrators" do not respect Kwibuka. 87 Very little is known regarding how people know who is buried where, how families are reached to reunite them with the dead, or who makes the decision to include found bodies in commemoration activities. 88 This article does not strive to be conclusive; the systems in question are far too opaque for that. Instead, readers and researchers should consider the information presented therein as, I hope, guiding future research into Kwibuka's link to prisoner treatment and interaction with both the state and collective Rwandan memory. What is the process of retrieving a confession and confirming it, and why are so many confessions being deployedwhether by the confessor or by the state-during the same hundred-day period each year? Over the course of seven weeks, I conducted thirty-eight informal interviews, primarily with people who self-identify as genocide survivors. Many, but not all, of those interviews contained questions and conversations relevant to the themes explored in this article, despite being conducted for a different project. All informants are anonymized and listed in Appendix A. I also carried out nine participant observations of commemoration activities during this time. This process was approved by the Columbia University Institutional Review Board, protocol number IRB-AAAR4442. 9