Harmful Memories—Present Dynamics: The Heroic Helper’s Effect on Collective and Individual Responsibility and Prejudice

This study was aimed at testing if exposure to a narrative about a heroic helper, can increment responsibility—taking about past ingroup wrongdoings and reduce prejudice and intergroup hostility in the present. We used the narrative of a Hungarian hero in an experiment who acted for targets of the Holocaust in Hungary, and measured if this narrative might increase collective responsibility for the Holocaust, decrease Hungarians’ hostile attitudes towards the Jewish minority, and this effect could be expanded to ongoing conflicts with other minorities. We used an experimental group (N = 99) ex-posed to the narrative, and a control group (N = 101) that was not. Both groups completed a test-battery measuring national identification, empathy, responsibility-taking, and prejudice. Data were analyzed with SPSS, and open-ended questions were content—analyzed by four independent coders. Results show that learning about a heroic helper increased acceptance of responsibility for the Holocaust and empathic abilities, whereas these effects were not generalized to current intergroup relations.

tions and ways of living.
As the history of past intergroup conflicts influences intergroup relations in the present, present day experiences influence what and how is remembered from the past. Heroes as products of this interplay between the past and the present represent morality change and gaps between normative frames.
In this way the reduction of nowadays intergroup hostility depends on how remembered and elaborated is the past, what we learn about past wrongdoings and victimhoods, are we able to reconciliate, and further to collaborate. Emotional needs arise, however these needs are based on the accepted, often dominant interpretations of the past. We assumed that acknowledging alternative narratives of heroic individuals, disrupting the sense of collective and sharping individual decision-making and responsibility can help to reframe the past, change the way of elaboration and the arousen emotional needs as well. These narratives could affect both victims and perpetrators while the quality of their identification with the group could play a role on it.

Reframing Past Conflicts by Alternative Narratives
In recent years, research on intergroup conflict and reconciliation focused on the emotional needs of victims and perpetrators [7] [8]. In consistence with the idea that the past weighs on the present [1], the basic tenet of this approach is that harms suffered in past intergroup conflicts may pose continuous identity threat to adversarial groups, which perpetuates mutual distrust and hostility [9] [10], which in turn prevents intergroup contact and reconciliatory efforts. Several alternatives have been proposed how to reduce long-term adversaries' mutual hostility and distrust, which mainly focus on structural interventions fostering trust based on cooperative actions (e.g. trade) and on various forms of public and political gestures to satisfy emotional needs [11]. These interventions mostly concern large-scale international involvement and/or political actions [12].
Another way to promote reconciliation by psychological interventions at a more personal level is presenting ingroup members with alternative, unconventional narratives that reconstruct the historical roles of the adversarial groups in the conflict and also acknowledge alternative perspectives [13] [14] [15]. As Bar-tal notes, "The psychological aspect of reconciliation requires a change in the conflictive ethos, especially with respect to societal beliefs about group goals, about the adversarial group, about the ingroup, about intergroup relations, and about the nature of peace" [16]. However, leaders and the majority of a society often approve and propagate dominant narrative social representations of the past [17] that justify and support ongoing conflicts [18] [19]. The construction and propagation of alternative narratives may enable groups to reframe past and present adversities posed by intergroup conflicts. One of these alternative ways of reconstruction is to show personal roles and stories about distinguished individuals who acted in different ways, which could undermine "the entitative perception of groups" [20]. Bilewicz

Increasing Intergroup Trust by Narratives about Outgroup Heroic Helpers
It has been proposed that narratives about heroic helpers that disrupt the conventional perception of intergroup conflicts based on the victim and perpetrator roles may contribute to reframing the conflict situation and fostering reconciliation [7] [10] [20] [21] [22] [23] [24].
In two studies, Čehajić-Clancy and Bilewicz (2017)  Interventions also decreased intergroup anxiety and increased belief in humanity, but these two dimensions had no effect on the belief in reconciliation. As the authors summarize their findings: "narratives including outgroup moral behaviour can help to restore broken relationships by creating a common space in which reconciliation can occur" [21]. These studies tried to increase participants' willingness to reconcile by reframing intergroup conflicts in an intergroup situation with reminders of moral actions of outgroup members, harmonizing the ingroup's and outgroup's positions.
These studies focused on the importance of outgroup heroic helpers in influencing victimized groups' aversive emotions towards a perpetrator group.
Moral exemplars increased trust towards the outgroup, which in turn facilitated the willingness to interact [25] [26]. Cross-cultural research also found trust to be one of the closest correlates of forgiveness and reconciliatory intentions [22],

Increasing Perpetrators' Moral Responsibility by Narratives about Ingroup Heroic Helpers
According to Vollhardt & Bilewitz (2013), "National identities are built around symbolic commemorations of the past and the narratives of victims as well as of perpetrators. Motivated denial of these memories sometimes serves to restore moral self-image among national groups that were once involved in a genocide as bystanders or perpetrators" [24]. According to the needs-based model of reconciliation [7], victims need empowerment as social actors, while perpetrators experience a need for acceptance by others to restore their moral image. However, perpetrator or bystander groups could not establish commemoration of their past wrongdoings, or they could justify them or avoid remembering them, if the normative societal (present institutional ingroup, intergroup or international) frame did not explicitly condemn the actions and events. How people deal with past negative events committed by their group is framed and reframed continuously in the present normative societal frame according to their own continuously changing present identification with the group itself.
When a perpetrator group has a need for restoring their moral image, narratives are usually used as exonerating cognitions [30]. These narratives place them in a passive role, thereby reducing their importance in the aggression. The role of bystanders also needs to be considered. As Staub points out, passivity may be bystanders' response to collective guilt. "Passivity in the face of others' suffering makes it difficult to remain in internal opposition to the perpetrators and to feel empathy for the victims" [31]. Assuming passivity and keeping distance from what is happening allow groups to perceive themselves as victims rather than as bystanders, and this role indirectly contributes to the aggression. In the above mentioned study of Bilewicz & Jaworska (2013), learning about the heroic helper's actions enabled Polish participants to cope with the weight of passivity presumably assigned to them as bystanders, which helped them feel Jews as more similar and closer to them, diminishing psychological distance [20].  [21] and normativity of the group would sufficiently decrease the weight of the moral damage of the group image to cope with past wrongdoings.
Coping with past wrongdoings means assuming moral responsibility in both the perpetrator and bystander roles [41], and then being motivated to take reparative actions [42] [43] rather than legitimizing the harm done. Thus, instead of asking participants about perceived similarity and acceptance as Bilewicz and Jaworska (2013) [20] did according to the needs-based model for perpetrator groups [8], we asked them about past and present moral responsibilities. between group identification and group-based guilt by distinguishing between attachment and glorification as two different modes of identification with the nation. As the authors conclude, these "different modes of identification have opposing relations to feelings of group-based guilt" [30]. While attachment to the nation is part of the self-concept as members of a nation, glorification regards the need to assign moral superiority and distinctiveness to the national ingroup as opposed to other nations, thus strengthening the imagined community.

The Role of Identification on the Exposure to Heroic Narratives
As the national identification can be considered as a "multifaceted construct" [30], and although these different facets are correlated, they have opposite connections with group-related phenomena such as using exonerating cognitions including denial of responsibility and blaming the out-group to avoid negative  [49].
Since the belief in national superiority hinders reconciliation processes by increasing prejudicial attitudes [49], we expected that the heroic helper story would increase high glorifiers' (i.e. competitive nationalists') exonerating responses such as denying responsibility and would strengthen the perceived distance from out-groups. By contrast, we predicted that the moral exemplar narrative would have a different effect on participants highly attached to the national ingroup, since attachment is not associated with the need for superiority, therefore it does not inhibit critical assessment of the ingroup's past actions [46].
Both of these effects can be attributed to a heightened sense of collective shame or guilt [50], but these aversive emotions have different effects on glorifiers and attached individuals. The two groups give different focus and weights to group membership, and their emotional responses to past wrongdoings may be different accordingly. Although shame and guilt may be considered similar emotional responses [51], glorification of the nation as a sense of global inferiority accompanied by a need for reinforcement, is more closely associated with shame, which is elicited by facing failures and negatively evaluated ingroup behaviours questioning glorifiers' group identity as a whole and thus motivating avoidance strategies. By contrast, attachment is more likely to be associated with guilt, since highly attached individuals question the morality of specific actions of their group rather than the value of their group identity, and thus they are more open to constructive reparation [52]. Accordingly, Roccas et al. (2006) found that glorification and attachment to Israel among Jewish Israelis were inversely associated with collective guilt for the harm done to Palestinians, and exonerating strategies were only used by high glorifiers [30].
Following the work of Roccas et al. (2006,2008), our study focuses on how the present weighs on the past, and how present perception of, and identification with, the ingroup influence the effects of a moral exemplar narrative [30] [53].
However to make these considerations it is also important to take account of both the past normative context in which the hero acted and the present normative context in which the actor may or may not be considered a hero.

The Historical Contexts of the Study
Between 1941 and 1945 near 600,000 Hungarian Jews lost their lives during the Holocaust where anti-Jewish laws and decrees (1920,1938,1939,1941,1942) were deeply rooted in the antisemitic traditions and in the authoritarian ideolo-  rians' role in the Holocaust and found a particularly high correlation between defensive representations of the Holocaust and antisemitism among nationalists [59]. The term defensive representation refers to a "need to modify the group's narrative with regards to its culpability in past atrocities committed against another group" [59]. By defensive representations, the group reduced its own responsibility: they emphasized their victim status or stressed that the group acted under constraint. It is called secondary antisemitism the phenomena when

Past Conflicts and Current Intergroup Relations
Holocaust perpetrators in their narratives blame the victim to avoid responsibility and guilt [60] [61]. In an experimental study, Imhoff and Banse (2009) found that descendents of perpetrator groups facing outgroup suffering caused by the ingroup experienced increased negative feelings and prejudice towards the victim group [61]. In contrast with our perspective, the authors conclude as follows: "Although it may appear logical to emphasize victim suffering, our findings caution against such an approach. The data suggest that it may be counterproductive in many settings to emphasize victim suffering in an effort to evoke sympathetic reactions and reduce prejudice." [61]. There is a huge difference whether intergroup contact is promoted before or after a perpetrator group faces its past actions and responsibilities. Recognition of, and insight into, past roles and wrongdoings may facilitate identity elaboration and eventually to reframe general intergroup relations. As Cehajic & Brown (2010) express "acknowledgment of responsibility by collectives entails a notion of 'never again'-the hope that exposure of the past should prevent its future repetition. Through acknowledgment of collective responsibility and provision of punishment, the occurrence of future atrocities can be discouraged" [62].
There are different approaches to the generalization of prejudice across different target-groups. Some of these approaches explain prejudice generalization by personality or individual differences [

Research Method
In this study, we examined the impact of a historical moral exemplar of the members to deal with their group's wrongdoings [32].
3) We predicted that effects of exposure to the heroic helper would be influenced by the identity processes (GLORIFICATION & ATTACHMENT), since the story would activate a qualitatively different identification that serves as a perspective to interpret and deal with the narrative about the heroic helper.

Research Sample
Participants were recruited via online social media platforms. The total sample No statistically significant differences were found between the experimental and the control group either on the sociodemographic variables or on the input variables of glorification and attachment 10 , that is, differences in the tested variables between the two groups may not be explained by such differences.

Stimulus Narrative about a Heroic Helper
Hereupon the experimental group read the story about Ocskay (while the control group continued to the next section). The stimulus used in the study was taken from the book Real Heroes by Krisztián Nyáry [75]. The sample size for this research was determined by computing estimated statistical power (β > 0.8), based on the results of prior experiment about the relation between defensive representations of the Holocaust, nationalism and antisemitism [59]. 9 Other party (33%), MKKP ("Hungarian Two-Tailed Dog Party"; 20%), Jobbik (far-right party; 17.5%), Fidesz (current ruling party; 16%), LMP (green party; 6.5%), MSZP (socialist party; 3.5%), Együtt-PM (socialist-liberal party; 2%), DK (democratic coalition; 1%), Magyar Liberális Párt (liberal party; 0.5%). The story was about Captain Ocskay, one of those Hungarians 11 who hid and saved Jewish individuals when the general public attitude was to active collaboration or to be passive. Captain Ocskay was a WWI veteran of an old aristocratic Hungarian family. In 1944, some Jewish veteran friends asked him to enroll in the army again with the aim to command the labour service unit No. 101/359 in Budapest. At the beginning, they were working in a few hundred but when their number grew to more than two thousand, women, men and children and since they pursued illicit activities they moved to a place less visible to public scrutiny.
Ocskay and his staff gave identity cards to new arrivals whenever they escaped from, updated the supposedly official lists, obtained food, medicine and other sustenances for all. He also delegated members of the unit to work with Wallenberg at section T of the International Red Cross saving children from orphanage at risk of deportation, and he ensured safety for his unit against the intrusion of the Arrow-Cross asking help from German military forces 12 . The research team comprising seven members developed the coding scheme in a bottom-up method using a small sample of responses, constructed a thematic categorisation, and then they trained the four independent coders. The calculation of inter-rater reliability (IRR) was based on the concept of Percent agreement for multiple raters. Inter-rater agreement exceeded the benchmark of 75% in each case (ranging from 76.5% to 95.6%). All cases of disagreement were discussed by the research team, and final judgments were based on consensus. time, the leaders, they/them); 3) did not refer to anyone; 4) denied Hungarians' responsibility.

Responsibility at Different Levels
• Attributed responsibility 15 : This code measured the attributed responsibility for the Holocaust. We identified 5 categories: 1) the people; 2) the government; 3) not us; 4) Germans; 5) single individuals (this category consisted of those responses that referred to the responsibility of single individuals, but not of others.) • Internal/External causes 16 : If the responsibility was explicitly acknowledged, then it was coded as "Responsibility;" if external contextual justifications appeared (e.g. political necessities) it was coded as "Constraint." This distinction is related to our hypothesis that the locus of control has importance in the dynamics of collective victimhood, since constraints and events are viewed to have happened without any kind of control; this indicates a lack of agency and inhibits the psychological elaboration of traumatizing events. • Subject expressed 20 : With this code, we measured participants' acceptance of responsibility for the current conflict. We distinguished between the following subjects in the responses 1) personal responsibility indicated by the pro- 15 Attributed responsibility IRR = 92% 16 Internal/External causes IRR = 95.6% 17 Attributed responsibility IRR = 89.6%. 18 Victimhood IRR = 87.4%. 19 Quality of prejudice IRR = 83.9%. 20 Subject expressed IRR= 87.8%. subject (e.g. government, politicians). • Responsibility-taking 21 : We focused on the individuals and investigated participants' acceptance of personal responsibility: 1) accepted personal responsibility; 2) denied personal responsibility; 3) made no reference to personal responsibility.
• Object of responsibility 22 : With this code, we marked the group for which participants felt responsible as follows: 1) Roma minority; 2) non-Gypsy majority; 3) both groups; or 4) none of them.

Empathy and Prejudice
Empathy was measured with the Interpersonal Reactivity Index [76]. According to Davis, empathy can be defined as the "reactions of one individual to the observed experiences of another". The IRI measures 28 intrapersonal processes that mediate responses to others' behaviour in interpersonal relations. The test consists of four subscales: Perspective taking (7 items), Empathic concern (7 items), Fantasy (7 items) and Personal Distress (7 items). Participants indicated on a 5-point scale how accurately each item characterizes them: (e.g. "I can truly feel the emotions of a novel's character.") Finally, we measured explicit prejudice with the Bogardus social distance scale [77]. Participants used the following items to indicate the closeness of the relationships they would be willing to accept with members of the following groups: Germans, Gypsies, Muslims, Jews, homosexuals: "as a citizen of your country; as a colleague; as your neighbour; as a close friend; as a family member".

Results
We explored data on the overall sample without taking account of possible ef- = 0.050). Participants' political orientation was related to national identification processes and to social distance from various context-relevant Others. The more right-oriented one was, the higher attachment to, and glorification of, the Hungarian national ingroup one reported, and the greater psychological distance one kept from relevant outgroups and minorities. 21 Responsibility-taking IRR = 76.5%. 22 Object of Responsibility IRR = 79.2%.
We predicted (H1) that attachment and glorification would show opposite correlations with ingroup responsibility for past conflicts, individual responsibility for the current intergroup conflict, empathy, and prejudice. The results indicate that while both attachment to, and glorification of, the national ingroup were related to political orientation, they activated different processes. While attachment a relationship with empathic abilities, glorification was associated with social distance from others and with acceptance of collective and individual responsibility-taking for intergroup conflicts.   high glorifiers used more defensive representations or exonerating cognitions concerning the Holocaust, depicting it as one of our respondent argued: "A stranger power forced us from the first anti-Jewish laws of the '30s'" [59].
We also revealed differences in identification according to the responses to the question concerning the acknowledgment of the Gypsy-majority intergroup conflict. Results presented in Figure 2 Table 2 also confirm that glorification of the national ingroup is strictly related to responsibility avoidance, both concerning past ingroup conflicts, current intergroup conflicts and recognition of the individual's own activity and personal responsibility.    Table 2.

Content codes, presented in
Perception of the Gypsy-majority intergroup relation as conflictual negatively correlated with empathic concern (r(n = 200) = −0.228**p = 0.001), positively correlated with social distance from Gypsies (r(n = 200) = 0.293**, p < 0.001) and Jews (r(n = 200) = 0.162*, p = 0.022), and showed a low negative correlation with responsibility of Hungarians in the Jewish Holocaust (r(n = 200) = −0.143*, p = 0.043). Thus, a small tendency emerges in the data which shows that those perceiving current intergroup relations with Gypsies as conflictual attribute less responsibility to their own group concerning past conflicts (see Table 2).
Acceptance of personal responsibility for the ongoing intergroup relation showed a moderate positive correlation with the responsibility attributed to the In sum, higher personal responsibility is associated with higher perceived ingroup responsibility, better empathic abilities and reduced social distance from others (see Table 2).
Summing up, our first hypothesis was confirmed. While attachment to the nation was in relation only with the empathic abilities, glorification was positively connected to prejudice, negatively to individual and collective responsibility, as those who are using exonerating strategies for past group wrongdoings, blame Minority and feel to be victim in current conflict are significantly higher glorifiers.

Effects of the Heroic Story (H2)
Our or to social distance 31 (see Figure 3).
Concerning the content codes, we found differences between the experimental and the control group only in acceptance of personal responsibility. Those who read the story about Ocskay considered their own or their group's responsibility in current intergroup relations, using the personal pronouns "I" and "We" more frequently than those who were not presented with the story (M2 (n = 200) = 9.391 p = 0.025, V = 0.217) (see Figure 3). According to Pennebaker work on pronouns (2011) the increased use of self reference words refers to increased activity.
From Figure 3 we can see that the stimulus narrative was associated with in-  The results lead to the conclusion that reading the story of Ocskay had a general effect on participants' views on the Holocaust, in whose context the recounted events took place. Those who read the story attributed essential responsibility to the Hungarians for the Holocaust, used more first-person personal pronouns reflecting their acceptance of personal responsibility and reported better empathic competences in self-reporting empathy.
However, the ingroup helper story did not have more general effects: it did not influence participants' perception of a current conflictual intergroup relation (i.e. the Gypsy-majority conflict), did not clearly facilitate acceptance of personal responsibility, and did not decrease social distance from others. We hypothesized that this restricted effect was due to the mediation effect of glorification of the national ingroup, which acted as a blocking mechanism against the evaluation of current intergroup conflicts and relations.
Summing up, our second hypothesis concerning the effects of the heroic narratives on individuals was partially confirmed. Those who read the story increased ingroup responsibility, however did not generalized to the current situations. The exposure to the heroic helper narrative increased also empathic abilities and in an indirect way personal responsibility as well (increased use of first person personal pronouns).

Identification and the Heroic Helper Story (H3)
To test this hypothesis (H3), we ran the PROCESS macro [78] with glorification as a moderator variable, but no significant moderation effect was found for ingroup responsibility for the Holocaust, acknowledgment of the Gypsy-majority conflict, personal responsibility for this conflict, empathy factors or any of the Bogardus scales (i.e. Gypsies, Muslims, Jews, homosexuals). It seems that identification with the national ingroup did not moderate the effects of the story on dependent variables such as different levels of responsibility-taking, empathy or prejudice.
However, since these variables showed no differences between the experimental and the control group with regard to the mediation effect of the national identification variables, we examined the correlation matrix of the tested variables in the two subsamples, which revealed differences (see Table 3). In the found in the control condition (apart from the positive correlation with social distance from homosexuals and the negative correlation with responsibility attributed to Hungarians for the Holocaust). The negative correlation found between glorification and personal responsibility was not significant in the experimental condition as opposed to the control condition.
Results presented in Table 3 show that attachment was less related to the intergroup variables except for the perception of current intergroup hostility which covariated in the experimental condition. Furthermore, although attachment was related to the emotional competences of empathy in both conditions, it showed a lower correlation with empathic concern in the experimental condition. Thus, attachment also acted as a national identity activator not directly influencing intergroup relations but rather empathic concern for others in general.
We interpreted these data as follows. Although the two subsamples had no differences in means and variances of the tested variables, the story of the ingroup helper activated group identification processes. Glorification showed higher correlations with all intergroup variables in the experimental condition.
Attachment enabled individuals to identify with the ingroup, but it was less closely related to empathic competences in the experimental situation than in the control condition without the context constructed through the stimulus. Table 3. Correlations between the tested variables in the experimental and control condition.

Discussion
Similarly to the research and conceptualisation published by Roccas [79] shared by participants with different group affiliations, the intervention seems to have successfully facilitated reconciliation and strengthened interactions between groups with a conflictual past. In the study of Bilewicz & Jaworska (2013) [20], an ingroup moral exemplar helped Polish participants, who assigned a bystander role to their group in a traumatizing past event, to feel more accepted by the victimized group, which in turn mediated perceived similarity of the outgroup and reduced psychological distance. The dependent variable was acceptance, since the needs-based model predicts that, reconciliation requires perpetrators (as well as bystanders according to the findings) to restore their moral image in order to feel more accepted by the victimized group.
In our view, however, bystanders' and perpetrators' must take responsibility for their past wrongdoings before being accepted in order to engage in the process of reconciliation, insofar as reconciliation is considered as intergroup cooperation that is also aimed at reducing the "imbalance that the harmdoing has created" [79]. If responsibility taking does not precede the satisfaction of needs to cooperate and to create shared objectives-as, for example, a shared and institutionalized historical narrative of what happened could be-advocates of the dominant positions may easily take control of these negotiation processes and promote historical revisionism, for example, which is more consistent with the dominant group's narrative.
What kind of tools can we offer for fostering reconciliation? As we learned from the above mentioned two studies, presenting past heroic actors can facilitate positive intergroup relations. Our study focused on the underlying process and its predictive relationship with present and future intergroup relations. Does it have a general effect extended to other intergroup conflicts with different outgroups?
We used the heroic helper story instead of satisfying the need for being accepted in order to encourage the acknowledgement and elaboration of past wrongdoings, which requires responsibility-taking a requisite for reparative actions. Exposure to the dissenting position represented by the heroic helper's actions sufficiently reduced perceived past ingroup homogeneity to enable group members to deal with guilt and take collective responsibility for what happened.
Likewise, perception of possible choices in each situation, even when norms are perceived as strict constraints, can help us to think about available or imaginable alternatives and consequently about our own responsibility. This is why we as- represented by the heroic helper enabled participants to question the perceived normative frame and face past misdeeds of the ingroup, which resulted in more active self-expression reflected in first person pronouns and higher general emphatical competences. However, responsibility-taking was not generalized to ongoing conflicts, nor did prejudice against minorities decrease. This lack of generalization may be due to various reasons from the current Hungarian normative frame and values to the acknowledgment of an implicit parallelism between the past genocide and currently arising conflicts. We expected that identification with the national ingroup holds the gap between past and present values and perceptions, and that it would mediate the effects of the heroic helper narrative. Glorification and attachment did not prove mediators of these effects. In the experimental condition, however, glorification showed higher covariation with intergroup variables such as acceptance of responsibility for past wrongdoings and for current conflicts, and with most social distance scales, while it did not show a negative relationship with individual responsibility as observed in the control condition. The heroic helper story contextualized the readers, activated participants' national identity, who expressed themselves accordingly. That is, stories may function as activators of identity, which provides a context that orients readers' thinking and behaviour.
An important limitation of the study is that participants' present values were not assessed (evaluation of Ocskay as a hero), which would allow for a better understanding of the observed lack of generalization. Furthermore, long-term effects of the stimulus narrative should be assessed as well.
Stories of heroic helpers of the past are interpreted in a network of possible identifications existing in the normative frame of the present. In general, the quality and extent of the impact of a moral narrative on its readers presumably depends on readers' current value systems and identity states. Knowledge of a past ingroup helper acting against dominant ingroup norms may have different effects depending on the relationship between the present and past value system and on the present forms of national identification. A hero of the past may be considered a hero in the present if her/his moral actions are positively valued in the present dominant social frame. Moreover, the past ingroup moral frame may be discussed and questioned if the present ingroup frame allows different forms and processes of identifications and supports alternative identifications rather than demanding blind conformity to the ingroup. The subjective experience of threat in intergroup relations restricts this possibility of self-questioning.
However, there is an effect of stories about ingroup helpers acting against ingroup crimes, whose extent varies with changes in the present frame. When intergroup hostility arises and there is a general perception of ingroup threat, the impact of such stories is confined within a strict normative frame and may not be generalized to other intergroup relations. In more peaceful social contexts, however, these stories may help group members discuss and question moral choices as well as individual and collective responsibilities.

Conclusions
In our study we find out that present type of identifications defines how we deal with the collective past as well as with present intergroup conflicts. Alternative narratives as heroic helpers stories acting against past group norms can help to increase collective responsibility for the wrongdoings of the group, probably because they render the group perception more heterogeneous and evoke individuals' own responsibility. Heroic helpers stories can also increase general empathic abilities. However the responsibility is not necessarily generalizable to nowadays conflicts; the empathic abilities with others are less mobilized when ingroup identification is activated.
In line with these results we suggest that using heroic helpers narratives in reconciliatory activities, in history teaching, in informal trainings etc, can be useful to increase empathic abilities, or for the reduction of prejudice and present intergroup conflicts; however other activities could follow to strengthen the extention of its effect. Expanding possible identifications and activating different group memberships, exploring the concepts of norms, conformity and values, and helping the abstraction and acknowledgment of perpetrator-victimhood dynamics and circularity can be among these activities.