This is an open access article distributed under the terms of the Creative Commons
Attribution License (After 1994, Rwanda was confronted not only with survivors of a genocide that had claimed nearly one million victims, but also with some 130,000 prisoners accused of participation in the massacres. Making matters even more complicated, survivors and perpetrators often came from the same hamlet. To deal with this challenge and to judge those accused of participation in the genocide, the Rwandan government introduced a modified version of a traditional Rwandan community-based conflict resolution system called Gacaca. In every community of the country, Gacaca tribunals were composed of persons of integrity elected by the inhabitants of cells (a local government division above the level of a village), sectors, districts, and provinces (Clark, 2010, 2012). Prisoners were brought before the tribunal in the community where they allegedly had committed a crime. In their presence, survivors and the entire community discussed the alleged acts, providing testimony and counter-testimony. Prisoners who confessed before the proceedings and asked for forgiveness received important reductions in penalties. Gacaca represented a Rwandan version of a truth commission. Truth commissions are temporary entities set up to investigate an episode of violations of human rights in a country (Hayner, 2001). Their purpose is to document the truth about the period of collective violence and to acknowledge and validate survivors’ suffering (Bloomfield, Barnes, & Huyse, 2003). Through these functions, truth commissions are expected to prevent revenge cycles, and the return of collective violence (Sikkink & Walling, 2007). As of early 2011, some 40 official truth commissions have been created to provide an account of past abuses (International Center for Transitional Justice, 2014).
From a psychological viewpoint, a truth commission represents a situation in which past traumatic events are recalled, sufferings are verbalized, and related emotions are expressed for the sake of psychological and social healing. Such a linkage of expression, truth telling, and healing was made central in the post-apartheid South African "Truth and Reconciliation Commission" (TRC). President Nelson Mandela and Archbishop Desmond Tutu initiated this commission in order to "heal" national wounds (Lund, 2003; Minow, 1998; Posel, 2008; Ross, 2003). Using mottos such as "Revealing is healing", they promoted the view that truth telling is healing and thereby leads to reconciliation (Lund, 2003; Minow, 1998; Posel, 2008; Ross, 2003). However, the sparse empirical evidence available to date has failed to support the assumption that truth telling is healing. At best, no effect was found (Kaminer, Stein, Mbanga, & Zungu-Dirwayi, 2001). At worst, testifying in a truth commission was experienced by a majority as a painful and disappointing experience (Byrne, 2004) or even led to new traumas (Brounéus, 2008; Hayner, 2001).
The postgenocide Rwandan Gacaca relied upon the South African TRC experience although each Gacaca was also a court handling out custodial sentences. In two different studies that were conducted in the context of Rwandan Gacaca (Kanyangara, Rimé, Philippot, & Yzerbyt, 2007; Rimé, Kanyangara, Yzerbyt, & Paez, 2011), we examined the question of effects of participation in a truth commission from a psychosocial perspective. We relied upon Durkheim’s (1912) classic model in which he differentiated between the emotional and social consequences of participation in a collective ritual. This model predicts two major consequences of participation in a collective ritual. On the one hand, rituals are expected to trigger a strong reactivation of the emotions associated with the commemorated event, due to the reciprocal stimulation of emotions they elicit among participants and the ensuing collective state of emotional communion. On the other hand, rituals are predicted to enhance participants’ collective identity as an effect of the shared emotional states.
These two predictions were addressed in our two studies. As regards emotional reactivation, our findings fully supported the prediction of the model. Both survivors and prisoners who participated in the Gacaca manifested a considerable increase in negative emotions after their participation. In addition, survivors who participated subsequently reported a sharp increase in symptoms of post-traumatic stress disorder. Such findings thus run against the healing perspective of the expression of emotion in social context. As regards collective identity, results strongly supported the view that participation in Gacaca enhanced social cohesion. After participation, a reliable improvement of intergroup variables was observed in both survivors and prisoners across three of the four indicators of the study.
The results of our two studies on Gacaca thus provided strong support for the predictions of Durkheim’s model. On the one hand, and in contrast to the healing hypothesis that is often associated with truth commissions, participation increased both negative emotions and distress symptoms among survivors. On the other hand, this participation generally resulted in positive effects on social integration. In both studies, the analyses reported were limited to those variables that allowed us to test Durkheim’s model. However, in the second study, data collection included the measurement of several additional variables addressing the question of dispositions to reconciliation among members of both groups who took part in a Gacaca. The data that resulted from these measures are the focus of the present article.
What could be expected from participation in Gacaca as far as reconciliation is concerned? Staub, Pearlman, Gubin, and Hagengimana (2005) stressed that reconciliation must include a changed psychological orientation towards the other. According to Nadler (2002; Nadler & Liviatan, 2006), such a change may occur only once the parties have resolved their respective emotional needs. As long as these needs remain unsatisfied, the path to reconciliation is blocked. The needs-based model of reconciliation (Nadler & Shnabel, 2008; Shnabel & Nadler, 2008) proposed that victims and perpetrators have complementary emotional needs that result from their asymmetric deprivation of essential psychological resources. Thus, victims through their inability to protect themselves and to control the situation experience a disempowering loss of status, honor, and self-esteem (Shnabel & Nadler, 2008). They therefore come to desire retribution against perpetrators (Carlsmith, Darley, & Robinson, 2002; Mikula, 1993). On the other side, perpetrators who have committed immoral acts lose their social status as a morally upstanding group (Gausel & Leach, 2011; Gausel, Leach, Vignoles, & Brown, 2012). They are thus motivated to reconcile in order to regain social acceptance. Whereas perpetrators’ willingness to reconcile empowers victims (Gausel, 2013; Nadler & Shnabel, 2008), victims’ willingness to reconcile absolves perpetrators in return. Thus, reconciliation should be sought after because it allows both parties to regain their lost social status.
The needs-based model thus describes reconciliation as an act of social exchange in which emotions play a central role. Could such a social exchange of emotions occur in a truth commission? Findings from our investigation of effects of participation in Gacaca provided initial support for this possibility (Rimé et al., 2011). As a matter of fact, examining the emotional profile of survivors’ participants revealed a heightened level of antagonist emotions (disgust, anger) after participation, directly suggesting an effect in terms of empowerment (Lazare, 2004; Nadler & Liviatan, 2006; Nadler & Shnabel, 2008). In addition, self-reported shame decreased among survivors who participated suggesting that participation indeed contributed to restore their self-esteem and dignity. Conversely, the participation resulted in an increase of shame among prisoners, an effect likely to favor the internalization of social control. Thus, the social exchange of emotions envisioned by the needs-based model of reconciliation seems to have taken place among participants in the Gacaca process.
In the present investigation, we intended to further test the needs-based model in the context of Gacaca by examining the dynamics of guilt feelings of participating survivors and prisoners. Guilt has been defined as an unpleasant emotional state associated with possible objections to one’s actions, inaction, circumstances, or intentions (Baumeister, Stillwell, & Heatherton, 1994). Whereas shame involves a depreciation of the entire self, guilt denotes a depreciation of specific acts. Guilt seemed particularly worth of examination in the context of Gacaca because we expected both survivors and prisoners to be concerned with guilt feelings.
Regarding survivors, their social image as people worth of dignity and respect was considerably damaged up to the point that they could experience guilt feelings that have no rational basis, as is often the case among victims (Jehu, 1989; Kubany & Watson, 2003; Price, 1991). Leach, Iyer, and Pedersen (2006) showed that guilt is associated with the abstract goal of compensation but not at all associated with specific political actions designed to restore justice. In other words, guilt feelings experienced by survivors are likely to perpetuate their disempowering status. It would thus be critical for a process such as Gacaca to show its capacity to substantially lower survivors’ guilt.
As for prisoners, they are expected to experience guilt feelings that would be well justified by the crimes they committed as members of their group. However, individuals can avoid the unpleasant experience of guilt in a variety of ways (McGarty, Pedersen, Leach, Mansell, Waller, & Bliuc, 2005) and thus it would also be critical for Gacaca to allow confronting prisoners with their crimes and enhancing guilt feelings among them. If participation in Gacaca could result in a reduction of guilt among survivors and an enhancement of guilt among prisoners, then it would contribute to feed the social exchange process described by the needs-based model of reconciliation.
When victims and perpetrators are involved in a truth commission such as Gacaca, their respective roles are necessarily emphasized in a most extreme manner. The solemnity of the situation, the mutual exposure of opponents, their testimonies and counter-testimonies contribute to intensify their identification with the group of victims or with the group of perpetrators. Such enhanced identification to a role may well entail important psychosocial consequences. This was particularly well documented by recent studies showing that a higher focus on the role of victim could jeopardize intergroup reconciliation. As a matter of fact, the perception of ingroup victimization was shown to be associated with a reduction in intergroup trust (Noor, Brown, Gonzalez, Manzi, & Lewis, 2008; Noor, Brown, & Prentice, 2008), with a decreased willingness to forgive (McLernon, Cairns, & Hewstone, 2002; Noor, Brown, Gonzalez, et al., 2008; Noor, Brown, & Prentice, 2008) and an increased likelihood of revenge (Lickel, 2012; Lickel, Miller, Stenstrom, Denson, & Schmader, 2006). When victims focused mainly on their sufferings as victims, they were found to have less concern for the suffering of others (Mazziotta, Feuchte, Gausel, & Nadler, 2014; Noor, Brown, Gonzalez, et al., 2008).
The theoretical propositions and empirical findings just reviewed allowed us to formulate a number of hypotheses for our investigation of the effects of participation in Gacaca upon attitudes toward reconciliation. In this investigation, before and after their participation in Gacaca, survivors and prisoners rated scales assessing their beliefs about the effectiveness of Gacaca, as well as their sense of personal and collective guilt. In addition, before and after Gacaca, survivors who took part, but also a control group of survivors who were awaiting participation in Gacaca, answered questionnaires assessing variables indexing attitudes toward reconciliation, that is, their perceptions of the fairness of prisoners’ excuses, their feelings of revenge, their readiness to forgive, and their actual contacts with members of their own group and with members of the other group.
The predicted effects were as follows. First, in Gacaca, survivors’ sufferings were acknowledged and validated in a very formal way, in the presence of the entire community and of the accused prisoners who were expected to confess crimes. This situation should have empowering effects for survivors and we expected these effects to manifest themselves in a significant reduction of their guilt feelings. Second, where prisoners’ acknowledged their crimes and publicly confessed in the presence of their victims and of the entire community, this was expected to heighten their awareness of their crimes and thus to enhance their guilt feelings. Third, despite the two previous predictions, the extreme focus of the truth commission situation on survivors as victims and members of the group of victims together with the limitations of Gacaca in granting redistributive justice (i.e., weak punishments; lack of reparations) and restorative justice (i.e., victims view perpetrators’ confessions and expressions of regrets as insufficient; see Brounéus, 2008) led us to expect poor results with regard to reconciliation. In other words, we expected victims’ participation in Gacaca to lower their perceptions of the fairness of prisoners’ excuses, their readiness to forgive, and their level of contact with perpetrators and to heighten their desire of revenge.
Method
Participants
In total, 755 volunteers who were able to read and write Kinyarwanda and were at least 18 years old at the time of the genocide took part in the study. Among them, 395 were survivors (238 females, or 60.25%) and 360 were accused of being perpetrators (162 females, or 45%). The experimental group comprised 384 participants of whom 200 were survivors (Mage = 30.60, SD = 7.70) and 184 were accused (Mage = 42.70, SD = 7.83). The control group involved 371 participants of whom 195 were survivors (Mage = 28.89, SD = 6.54) and 176 were accused (Mage = 44.31, SD = 7.78).
Procedure
The study was conducted between February and April 2006 in four of the five Rwandese provinces with citizens who had been involved in the 1994 genocide, either as a survivor or as an accused perpetrator now in custody. Survivors and prisoners were matched for their residence in the same neighborhoods during the genocide. They participated either in the experimental or in the control condition. Control participants came from a neighborhood where no Gacaca trial had yet taken place and where no such trial was being planned for another year. Also, control participants had not taken part in any other Gacaca trial outside of their neighborhood. Both experimental and control participants responded twice, once before and once after the Gacaca trial that took place for the experimental participants. The two sets of judgments were collected over a period of 10 weeks. Participants lived in hamlets based on agricultural economy. No relevant differences exist in socioeconomic indexes between regions. No gender differences were found in data analyses.
In order to recruit the survivors, we first selected specific, accessible neighborhoods as April-June is a rainy season in Rwanda and as our data collectors (students) had limited means of transportation. We then contacted the local authorities to ask permission to conduct the study within their district. In the presence of the authorities, the research assistant identified each survivor and explained the purpose of the study. It was stressed that they could refuse to participate or, if they accepted, that they could terminate their participation at any time during the study without any consequence. The selection procedure for the prisoners was identical with the exception that written permission was secured from the Minister of Interior and from the head of the penitentiary in order to meet with the convict in prison.
An initial group of 40 senior students of the clinical division of the psychology department of the National University of Rwanda were familiarized with techniques and rules of survey research. A final team of 24 assistants was retained on the basis of skills and motivation. They were paid 15 Euros per data collection day.
Measurements for Survivors and Prisoners
Where scales were available in English they were translated to Kinyarwanda. There was no back translation.
Personal Guilt
The Trauma-Related Guilt Inventory (TRGI; Kubany, Haynes, Abueg, Manke, Brennan, & Stahura, 1996) is a 32-item instrument with six subscales measuring both emotional and cognitive aspects of guilt associated with a traumatic event (war, genocide, death of a loved one, rape, and sexual abuse). For the purpose of this study, we adapted the TRGI subscale "global guilt" in order to tap both the guilt feelings experienced by perpetrators because of their crimes and the guilt feelings experienced by victims in view of their incapacity to have prevented the crime and to have coped with it, as is often the case among survivors of traumatic events (e.g., "I could have prevented what happened"; "I am responsible for what happened"; "I blame myself for what happened"), to be rated on 6-point scales anchored with 0 = not at all and 5 = very much (Cronbach’s alpha = .72 at test and .80 at retest).
Collective Guilt
The extent to which group-based guilt was accepted by participants was assessed using a five-item measure (e.g., "I feel regret for my group’s harmful past actions toward the other group"; "I feel regret about things my group did to the other group in the past"; "I believe that I should have repaired the damage caused to the other group") adapted from a scale originally developed by Branscombe, Slugoski, and Kappen (2004). Groups with a history of exploitation were demonstrated to score reliably higher on this measure compared with groups with a history of being exploited. The five items had to be rated on 6-point scales anchored with 0 = not at all and 5 = very much. In the current sample, the reliability of the scale was somewhat low at test (alpha = .61) and more satisfactory at retest where it formed an internally consistent scale (alpha = .79).
Doubt About the Effectiveness of Gacaca
Prior observation suggested that negative expectations about Gacaca were common among survivors and prisoners. We measured these concerns with six items formulated for this purpose: "Gacaca trials will have no positive impact on those who have emotional suffering"; "Gacaca trials will not be helpful in the healing of emotional pain"; "It would be better to forget the past: The trial will only cause more emotional suffering"; "The trial will not help Rwanda people to deal with their emotional suffering"; "The trial will not bring justice for those responsible in the genocide"; "The trial will not bring justice to the survivors of genocide". These items were rated on 6-point scales anchored with 0 = not at all and 5 = very much (Cronbach’s alpha = .66 at test and .68 at retest).
Measurements for Survivors
Distrust
To examine the extent to which the expression of apology by prisoners was perceived as faked or sincere, we adapted the 12-item questionnaire developed by Schmitt, Gollwitzer, Förster, and Montada (2004) assessing how survivors react to expressed excuses and justifications. The short version adapted to the context of Gacaca had four items rated on Likert-type scales (from 0 = completely false to 5 = completely true): "I do not believe that they actually expressed regret for what they did"; "I think that by apologizing, their purpose is to protect themselves from being punished for what they did"; "I think that they feel no remorse for that they have done"; "I think that they feel no guilt for what they did" (Cronbach’s alpha = .67/.79)
Feelings of Revenge
The Transgression Related Interpersonal Motivations Inventory (TRIM) developed by McCullough et al. (1998) was adapted and translated in order to measure feelings of revenge. To assess forgiveness in the interpersonal context, the TRIM first asks the participant to remember a situation that was experienced and during which someone else was offended. Respondents are then asked several questions about their motivation to seek revenge on the part of (Subscale 1) and avoid (Subscale 2) the author of the negative act. We used and adapted the revenge subscale using the following items: "I want to pay them back for what they have done"; "I would like to see them suffer and be miserable"; "I wish that misfortune falls on them"; "I would like them to experience for themselves what they have done to others"; "I wish they would be punished". Each item was rated on Likert-type scale ranging from 1 = not at all in agreement to 5 = strongly agree (Cronbach’s alpha = .78/.80).
Forgiveness
We translated and adapted the Transgression Narrative Test of Forgiveness (TNTF) developed by Berry, Worthington, Parrott, O’Connor, and Wade (2001) which contains five items that assess the extent to which the respondent forgives the offender. Several validation studies (Berry et al., 2001) showed that this scale measures general forgiveness. One item was removed from our scale in order to improve its reliability (Cronbach’s alpha = .80/.76): "I think people should try to abandon their negative feelings towards them"; "I think those who have remorse for what they have done should be forgiven"; "I think those who admit and confess to harming others should be forgiven"; "I think that those who have repaired their wrong for what they did should be forgiven". Participants responded to all items on a Likert-type 5-point scale (1 = not at all in agreement, 5 = strongly agree).
Contact With the Ingroup and the Outgroup
Intergroup contact was measured by a translation of the questionnaire developed by Tzeng and Jackson (1994) comprising seven items related to various factors that make intergroup contacts a way to improve the evaluation of the outgroup. Participants responded on a 5-point scale ranging from 1 (not at all) to 5 (very much). The translated and adapted version had three items (Cronbach’s alpha = .70/.72) measuring contact with survivors (e.g., "To what extent are you in contact with your neighbors who are survivors of genocide?"; " To what extent are you in touch with your old friends who are survivors of genocide?") and three items (Cronbach’s alpha = .76/.69) measuring contact with perpetrators (e.g., "To what extent are you in contact with your neighbors who confessed their responsibility in the genocide?"; "To what extent are you in touch with your old friends who have confessed their responsibility in the genocide?").
Results
Measurements for Survivors and Prisoners
As regards the measurement of personal guilt, a 4 (groups) × 2 (before and after Gacaca) mixed ANOVA demonstrated highly significant effects for groups, F(3, 751) = 66.0, p < .001, ηp2 = .21. Tukey post hoc comparisons (p < .05) showed victims to evidence a significantly lower personal guilt (M = 1.18, SD = 0.43) than the two control groups (for control victims, M = 1.36, SD = 0.63, and for control prisoners, M = 1.45, SD = 0.55) who both had a lower score than prisoners who took part in Gacaca (M = 1.89, SD = 0.36). A significant main effect was also observed for measurement time, F(1, 751) = 191.32, p < .001, ηp2 = .20. It evidenced an overall reduction of personal guilt after participation (M = 1.33, SD = 0.80) compared to before (M = 1.59, SD = 0.57). In addition, the interaction of the two factors was highly significant, F(3, 751) = 492.77, p < .001, ηp2 = .66. Examination of mean values (Figure 1a) showed that both control groups remained stable from one measurement to the next. In contrast, important changes, but in diametrically opposite directions, were observed for the two groups that took part in the Gacaca. Whereas survivors manifested a considerable decrease in personal guilt after their participation in Gacaca, prisoners presented an increase in their personal guilt feelings.
For collective guilt a significant effect was again found for groups, F(3, 751) = 171.68, p < .001, ηp2 = .41. According to Tukey post hoc comparisons (p < .05), both groups of victims manifested a lower collective guilt (for victims who took part in Gacaca, M = 0.82, SD = 0.51, and for control victims, M = 0.95, SD = 0.65) than control prisoners who themselves (M = 1.47, SD = 0.82) were significantly lower for this variable than prisoners who took part in Gacaca (M = 2.14, SD = 0.49). A significant effect for measurement time, F(1, 751) = 517.62, p < .001, ηp2 = .41, revealed that collective guilt was higher before Gacaca (M = 1.49, SD = 0.80) compared to after (M = 1.20, SD = 0.92). Finally, the interaction of the two factors considered in this analysis was again found highly significant, F(3, 751) = 467.16, p < .001, ηp2 = .65. As can be seen in Figure 1b, no change in average values of collective guilt occurred at the second measurement time among control groups or among prisoners who took part in Gacaca. The latter showed higher values of collective guilt than any other group, even before their participation in Gacaca. More specifically, before Gacaca, the collective guilt of the prisoners who took part in the Gacaca was already significantly higher than that of prisoners who did not take part, t(358) = 8.94, p < .001. In sum, the only observable change from the first measurement to the second one was found in survivors who took part in the process. Among these participants, the average value of collective guilt showed a drastic decrease.
Figure 1
Level of personal guilt (1a) and collective guilt (1b) reported by the respondents of the four groups before and after participation of two of these groups in Gacaca.
On the measure of doubt about the effectiveness of Gacaca, the ANOVA showed a group effect, F(3, 751) = 16.49, p < .001, ηp2 = .06. Tukey posthoc comparisons (p < .05) evidenced the two groups who took part in Gacaca to express higher doubts (for victims, M = 2.69, SD = 0.47, and for prisoners, M = 2.64, SD = 0.52) than the two groups who did not participate (for victims, M = 2.35, SD = 0.69, and for prisoners, M = 2.41, SD = 0.56). A significant effect of time of measurement, F(1, 751) = 43.38, p < .001, ηp2 = .37, revealed that doubts were higher after Gacaca (M = 2.69, SD = 0.70) compared to before (M = 2.35, SD = 0.67). A markedly significant interaction, F(3, 751) = 388.28, p < .001, ηp2 = .61 was again observed. Examining the means showed that doubts remained stable from one measurement time to the other for three of the groups: prisoners who participated in the Gacaca (Mbefore = 2.61, SD = 0.59; Mafter = 2.67, SD = 0.49), the control group of survivors (Mbefore = 2.38, SD = 0.77; Mafter = 2.33, SD = 0.63), and the control group of prisoners (Mbefore = 2.38, SD = 0.60; Mafter = 2.45, SD = 0.54). In contrast, for survivors who participated in the Gacaca, the index of doubt manifested a considerable increase following their participation (Mbefore = 2.05, SD = 0.56; Mafter = 3.33; SD = 0.64).
Measurements for Survivors
The results for the five variables measured only for survivors are displayed in Table 1. Values observed at the two measurement times among the control respondents were relatively stable whereas they varied substantially for those who took part in Gacaca. Thus, survivors who took part in Gacaca initially manifested a higher level of distrust, t(393) = 4.12 (Cohen’s d = 0.41), much lower feelings of revenge, t(393) = -27.40 (Cohen’s d = 2.77), a higher level of forgiveness, t(393) = 16.81 (Cohen’s d = 1.70), a higher level of contact with the outgroup, t(393) = 5.39 (Cohen’s d = 0.54), and a lower level of contact with the ingroup, t(393) = -9.77 (Cohen’s d = 0.99), all ps < .0001. The 2 (group) × 2 (time) mixed ANOVA conducted revealed for each of the dependent variables significant main effects as well as a significant interaction. Among survivors who took part in the Gacaca, participation seemed to have enhanced their level of distrust toward prisoners as well as their feeling of revenge and to have reduced considerably their inclination to forgive. As regards contact, despite the fact that they reported a higher level of contact with members of the outgroup at baseline compared to control survivors, survivors who took part in Gacaca reported a considerably lower level after their participation. The opposite effect occurred for contact with the ingroup.
Table 1
Comparison of survivors who took part in Gacaca and survivors who did not yet participate (control group) for measurements taken before (Time 1) and after (Time 2) Gacaca.
| Survivors/Gacaca (N = 200)
|
Survivors/Control (N = 195)
|
Group
|
Time
|
Group x Time
|
||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Time 1
|
Time 2
|
Time 1
|
Time 2
|
|||||||||||
| M | SD | M | SD | M | SD | M | SD | F(1, 393) | η2 | F(1, 393) | η2 | F(1, 393) | η2 | |
| Distrust | 2.51 | 0.55 | 3.13 | 0.60 | 2.28 | 0.53 | 2.19 | 0.62 | 111.87*** | .22 | 236.56*** | .38 | 430.98*** | .52 |
| Revenge | 2.23 | 0.43 | 2.41 | 0.49 | 3.45 | 0.45 | 3.43 | 0.63 | 831.55*** | .68 | 5.70** | .01 | 9.19** | .02 |
| Forgiveness | 2.71 | 0.45 | 2.54 | 0.61 | 1.80 | 0.60 | 1.90 | 0.62 | 187.82*** | .32 | 7.02** | .02 | 96.43*** | .20 |
| Contact with perpetrators | 2.32 | 0.72 | 1.47 | 0.60 | 1.92 | 0.76 | 1.88 | 0.71 | <1.0 | - | 1111.22*** | .74 | 916.33*** | .70 |
| Contact with survivors | 1.28 | 0.61 | 2.08 | 0.69 | 1.96 | 0.76 | 1.90 | 0.71 | 13.46*** | .03 | 620.48*** | .61 | 809.03*** | .67 |
**p < 01. ***p < .001.
We explored the correlations between the various variables of the study as measured after Gacaca among the survivors who took part to Gacaca. However, the observed coefficients were generally of a small size and did not provide additional information about the data.
Discussion
Clearly, before their participation in Gacaca, survivors had a lower level of doubt about Gacaca than respondents from the other three groups. This suggests that the anticipation of their upcoming participation sparked positive expectations and feelings of hope. Such an interpretation is supported by the fact that, compared to survivors who were not participating in Gacaca, survivors anticipating their forthcoming court appearance also reported much lower feelings of revenge, a higher level for forgiveness, and a higher level of contact with the outgroup. That such hopeful expectations existed among Rwandan survivors in view of Gacaca does not come as a surprise in light of the widespread assumption that simply giving survivors a chance to tell their stories to an official commission would bring them healing (e.g., Hayner, 2001; Herman, 1992). As recalled in the introduction, the linkage of expression and healing was made particularly overt in the South African TRC (e.g., Lund, 2003; Minow, 1998) that had elicited a broad international attention. Unfortunately, the evidence coming from both clinical observation (Daley, 1997; Hamber, 2001; Hayner, 2001) and other empirical sources (Brounéus, 2008; Byrne, 2004; Kaminer et al., 2001) contradicted this conclusion. Our previous investigations of participants in Gacaca showed that, after their participation, survivors underwent a powerful reactivation of negative emotions and even an exacerbation of post-traumatic stress symptoms (Kanyangara et al., 2007; Rimé et al., 2011). In the current data, measures of doubts collected among survivors after Gacaca suggests that their participation engendered a considerable disillusionment. These results support the critics who stress the limitations of Gacaca in terms of its negative effect on survivors (Brounéus, 2008), especially as survivors who took part to Gacaca were exposed only to limited and potentially insincere apologies by perpetrators.
As far as the prisoners were concerned, participation had the ‘positive’ effect of increasing their personal guilt. This did not happen for collective guilt. However, as was mentioned in the results, it should be noted that the collective guilt of the prisoners who took part in the Gacaca was already significantly higher than that of prisoners who did not take part. This suggests that the mere anticipation of participation in the Gacaca was enough to significantly exacerbate the collective guilt of the accused.
The guilt feelings manifested by survivors possibly reflected survivor guilt, or guilt about surviving the death of loved ones and/or about being better off than others (Lifton, 1967; Niederland, 1961). Survivor guilt was previously found to be significantly correlated with submissive behavior (O’Connor, Berry, Weiss, Schweitzer, & Sevier, 2000). In sum, the present results show that a community courts procedure inspired by truth commissions has effects that nicely fit the social exchange of emotions described by the needs-based model of reconciliation (Nadler & Shnabel, 2008; Shnabel & Nadler, 2008). The large reduction of survivors’ personal and collective guilt observed after their participation suggests that the public hearings of the two parties in the presence of the entire community had the capacity to restore feelings of control among the former. At the same time, prisoners’ guilt feelings of guilt increased significantly after their participation. This exchange of guilt thus sets the stage for the reconciliation effects described by the needs-based model (Nadler & Shnabel, 2008; Shnabel & Nadler, 2008). Whereas prisoners’ willingness to reconcile in order to regain social acceptance sets the stage for the empowerment of the victims (Gausel, 2013; Nadler & Shnabel, 2008), victims’ willingness to reconcile allows for absolution of the perpetrators.
However, a truth commission process is a complex and multifaceted psychosocial event. Among others, it involves an accentuation of the respective roles of survivors and perpetrators. Abundant empirical documentation showed that accentuating the role of victim and of membership in the victims’ group may entail damaging consequences for the reconciliation process (Lickel, 2012; Lickel et al., 2006; Mazziotta et al., 2014; McLernon et al., 2002; Noor, Brown, Gonzalez, et al., 2008; Noor, Brown, & Prentice, 2008). The results of the current study clearly confirmed the reality of such costs. Compared to control group survivors still waiting for Gacaca to take place in their community, survivors who took part in Gacaca did not believe that prisoners expressed sincere apologies, were less inclined to forgive them, were more revengeful after participation, and opted more for intragroup contact and less for intergroup contact. Survivors in the control group did not vary markedly for these variables from the first wave of measurement to the second one.
These results should also be integrated with our previous observations according to which participation in Gacaca allowed an improvement in the intergroup relations both among survivors and prisoners (Kanyangara et al., 2007; Rimé et al., 2011), as was suggested by a decreased ingroup self-categorization, a reduced perception of outgroup homogeneity, and an enhancement of positive stereotypes about the outgroup. These observations were in line with Durkheim’s (1912) insights into the effects of emotional gatherings upon social integration – the mutual activation of intense emotional states favored the emergence of a shared emotional condition. They were also congruent with the decategorization and direct contact model of prejudice reduction (Pettigrew, 1998) – exposure to individual testimonies promoted a diversified perception of members of the other group.
In principle, such effects should bring members of the two groups closer to one another whereas the results discussed in the previous paragraph point to the opposite direction. The divergence could be usefully examined in the framework of the distinction between macro-level and micro-level perspective justice that was first formulated by Brickman and colleagues (Brickman, Folger, Goode, & Schul, 1981) and then extended to the case of truth commissions by Lillie and Janoff-Bulman (2007). Macro-level justice focuses on the needs of the society as a whole, whereas micro-level justice focuses on the needs of the individual victim and concerns the relationship among individuals. Lillie and Janoff-Bulman noted that this distinction had a particular relevance to the findings from studies conducted about the South African Truth and Reconciliation Commission (TRC). Thus, on the one hand, in a survey of 3700 South Africans, the majority of them approved the TRC and believed it would contribute to and help maintain a new democracy (Gibson, 2004). On the other hand, on a personal level, lengthy interviews held with 30 Black South African victims who engaged in the TRC process indicated that for most of them, it represented a painful and disempowering experience filled with unmet expectations and promises (Byrne, 2004). Lillie and Janoff-Bulman (2007) concluded: "Although societal justice may ultimately be achieved, the need for personal justice may not be satisfied" (p. 222). These conclusions that were formulated from enquiries conducted after the South African TRC perfectly overlap with those that were reached from direct measurements conducted on Gacaca participants before and after their participation.
Thus, dissimilar investigation methods and dissimilar socio-political contexts consistently lead us to conclude that truth commissions positively fulfill macro-level justice purposes, but fail to deliver micro-level benefits. To be sure, the fulfillment of macro-level justice represents an important achievement. It contributes to change group categorization from a "they" to a "we" via the creation of a superordinate category, so that members of different groups come to think of themselves as members of a single overarching group. Gaertner and Dovidio (2000) showed that the building of such a new common identity ended up reducing prejudice, enhancing inter-group trust, and contributing to social readjustment.
That micro-level benefits failed to be met for participants in Gacaca might be due to a variety of factors. First, previous studies largely demonstrated that the evocation of crimes had revived painful emotions among survivors (Brounéus, 2008; Kanyangara et al., 2007; Rimé et al., 2011), which very likely resulted in an enhancement of their resentment toward those responsible. Second, the Gacaca procedure entailed reduced sentences for prisoners who acknowledged their involvement in the genocide and survivors generally viewed these lighter sentences as unjustified. Finally, as shown by the results of this study, survivors felt that the prisoners made insincere confessions in order to get lighter sentences.
Yet, the present results do not allow us to conclude on a totally negative note with regard to the capacity of Gacaca to contribute to the more personal, micro level, dimensions of justice. After their participation, survivors manifested a considerable reduction of their guilt feelings, both at individual and at collective level. In this respect, the level of the effect sizes was particularly high. As noted above, such results clearly suggest that, despite the activation of negative emotions they experienced in the process, the survivors left the Gacaca with a very significant gain in social power. According to Nadler and Shnabel (2008), such an evolution constitutes a critical condition for further developments towards reconciliation.
The present research has a number of limitations. The design of the study was quasi-experimental – participants’ assignment to conditions was not random but determined by geographical factors. The measures used suffer the usual limitations of self-reports. They were relatively transparent, thereby favoring the intrusion of experimental (there were no reverse scored items), social, or even political demands. But in our opinion, the major limitation was the lack of follow-up measurements in the long run. Participants completed the study scales in the immediate aftermath of their participation to Gacaca – in a window of a couple of weeks thereafter. It would of course be crucial to assess how far the dynamics and effects initiated by Gacaca would evolve in the longer term.
In spite of these limitations, the study has a number of strengths. It is one of the rare studies approaching psychosocial effects of Truth and Reconciliation procedures and, to our knowledge, it is the first one to assess predictions of the needs-based model of reconciliation in such a context. The study relied on a longitudinal design with a considerable number of respondents, in a part of the world in which an investigation of this kind is exceptional. The time frame that was adopted by the government of Rwanda in its implementation of the Gacaca ritual across the country allowed us to develop a classic quasi-experimental design, with effective control groups. The study was also unique in that both victims and prisoners were taken into consideration.
In conclusion, the present efforts confirm that establishing truth is a challenging exercise and that facing truth can indeed be harmful. However, "establishing who did what and why is essential for acknowledgment, for justice, and for collective memories or group narratives that can move groups toward a shared history" (Staub, 2013, p. 584).