Emotional reactions to the French colonization in Algeria: The normative nature of collective guilt

Fifty years after the end of the Algerian war of independence, French colonization in Algeria (1830–1962) is still a very controversial topic when sporadically brought to the forefront of the public sphere. One way to better understand current intergroup relationships between French of French origin and French with Algerian origins is to investigate how the past influences the present. This study explores French students’ emotional reactions to this historical period, their ideological underpinnings and their relationship with the willingness to compensate for past misdeeds, and with prejudice. Results show that French students with French ascendants endorse a no-remorse norm when thinking about past colonization of Algeria and express very low levels of collective guilt and moral-outrage related emotions, especially those students with a right-wing political orientation and a national identification in the form of glorification of the country. These group-based emotions are significantly related to pro-social behavioral intentions (i.e. the willingness to compensate) and to prejudice toward the outgroup.

Social psychology research on colonization has shown that the way the colonial past is collectively remembered and represented affects the current relationship between the formerly colonized or immigrants from former colonies and the former colonizers (Volpato & Licata, 2010). Likewise it has been shown that individuals' emotional reactions to their nation's past actions during the colonial period (and particularly, harmful, violent actions) influence their attitudes and behavior toward the formerly colonized and their descendants (Branscombe & Doosje, 2004;Doosje et al., 1998).
A review of studies on present-day attitudes and feelings about colonialism revealed that, with the exception of the French colonization, most of the European colonization in Africa and Asia has been investigated (for reviews see Leach et al., 2013;Volpato & Licata, 2010) i.e. Dutch colonialism in Indonesia (Doosje et al., 1998(Doosje et al., , 2006Figuereido et al., 2010), Belgium colonialism in the Congo (Licata & Klein, 2010), Portuguese colonialism in Mozambique and Angola (Cabecinhas & Feijo, 2010;Figuereido et al., 2010), Italian colonialism in Ethiopia and Libya (Leone & Mastrovito, 2010;Mari et al., 2010) and British colonialism in Kenya (Allpress et al., 2010;Morton & Postmes, 2010). The French colonial experience has so far been neglected in socio-psychological research, and the present study aimed at filling this gap, studying more in detail the ideological factors linked to collective emotions. In particular we took into account, along with political orientation and family implication, two other factors that are strongly dependent on the socio-political context and that have been neglected so far: the normative nature of collective emotions and the type of national identification (i.e. glorification vs attachment).

French colonization in Algeria: Historical and political context
French colonization in Algeria was one of the longest colonial periods of European colonization in Africa, lasting from 1830 to 1962. It was also a very specific one, being the only settlement colony organized as a French province (Stora, 2004a(Stora, , 2004b. The French colonial period was marked by numerous dramatic events, from the 'enfumades' (i.e. filling caves with smoke to asphyxiate rebels) during the conquest to the massive use of torture during the war of independence , which killed at least 400,000 Algerians and 30,000 French, as well as several massacres over the whole colonial period. During this time, colonized people were the target of workforce exploitation, massive land expropriations leading to the destruction of the rural world (dramatically reducing the local political influence of the indigenous) and marginalization in town. They were subjected to the Code de l'Indigénat (since 1881), a code for native-born Algerians that ratified status inequalities between French colonizers and indigenous colonized inhabitants, and legalized institutional violence (Merle, 2004). It was not until after the Second World War that the indigenous colonized inhabitants were finally granted French citizenship, although they were still treated differently by being officially called French-Algerian Muslims.
Several of today's historians trace discrimination and racism against Maghrebians and black people to the legacy of colonial racism (e.g. Bancel et al., 2005;Coquery-Vidrovitch, 2009). According to House (2006), '[o]nly by studying these policies [conducted after the war of independence by former colonial police and welfare officers] can we understand the origins of the social and ethnic segregation in France's poor outer suburbs (banlieues) today'. Thus to understand the enduring conflictual relationship between French natives and French of Algerian descent and Algerian immigrants, it is important to treat this intergroup relation as historically situated and to examine the extent to which the historical past, through collective emotions, echoes in the present (see Licata et al., 2007 for a discussion on collective memory). Social psychology research has neglected to take into account the historical roots of intergroup relations between French citizens of different origins (for an exception, see Haas & Vermande, 2010). The aim of the present study was to examine the collective emotions this historical period triggers in French students and whether such emotions are linked to pro-social action intentions and prejudice. We were also interested in the ideological antecedents of these emotions, and particularly in the role culture-specific emotion norms (i.e. a collective-guilt norm) might play in inducing emotional reactions regarding the French colonization of Algeria.

Emotions for the past, reactions in the present
In the last decade, an impressive body of research has studied collective emotions, and particularly collective guilt individuals may feel for the harm committed by members of their group on some other group. For instance with regard to past colonization, Doosje et al. (1998) have shown that Dutch students indeed felt collective guilt regarding Dutch colonization in Indonesia (see also Zebel et al., 2007). Similar results have been found for (French-speaking) Belgians regarding colonization in Congo (Licata & Klein, 2010), for Italian (Mari et al., 2010) or for Portuguese colonialism (Figueiredo et al., 2011).
Collective guilt has been found to increase the willingness to compensate for past harmful actions (e.g. Doosje et al., 1998) and to reduce prejudice against the harmed group (e.g. Powell et al., 2005). However, other collective emotions, such as shame (e.g. Brown et al., 2008) or anger toward the ingroup (e.g. Mari et al., 2010), have also been found to be aroused when being reminded of the ingroup's past actions. Moreover when comparing collective emotions, guilt is not always the emotion that is the most disposed to promote willingness to repair for past misdeeds. Other-focused emotions such as moral outrage or sympathy are more likely to arise when confronted with past wrongdoings and are more likely to induce actions intended to redress inequalities than self-focused emotions such as collective guilt, which is associated with actions aimed at compensating for the past (e.g. Iyer et al., 2003;Leach et al., 2006). In the same vein, Mari et al. (2010) showed that collective shame predicted reparation intentions in the form of economic compensation, while anger predicted intentions to help immigrants from former Italian colonies. Collective guilt, however, did not predict any of these pro-social action strategies. This last result was explained by the absence of a public debate about colonization in Italy, which may have prevented the 'internalization of constructive collective guilt' (Mari et al., 2010). As in France public debate on colonization is also scarce, we hypothesized collective guilt and negative other-focused emotions such as moral outrage to be linked to greater compensation intentions and less prejudice albeit with different strength, with guilt being less predictive of these outcomes than other-focused emotions.

Ideological factors linked to emotional reactions
The second aim of the present study was to examine some ideological factors that may promote, or, on the contrary, immunize against the feeling or expression of these negative collective emotions. We investigated the role of group identification, political orientation and the normative nature of collective guilt. The existence of an emotion norm regarding the appropriateness for a nation to feel and express regret for its colonial past may indeed strongly influence individuals' feelings and behavior. To our knowledge no study has explored the question of collective emotion norms.

National identification: Glorification vs attachment
Most research on group-based emotions considers a minimal identification with the ingroup necessary for a collective emotion to arise (Wohl et al., 2006). However evidence for the association between ingroup identification and the experience of collective guilt is ambiguous. Some studies have shown that the more people identify with their group, the more they feel collective guilt for the ingroup's past wrongdoing (e.g. Doosje et al., 2006). Others have found that highly identified individuals report less guilt, supposedly because they are more motivated to protect themselves against the threat to their group identity and to preserve a positive social identity (Doosje et al., 1998). One way of reconciling the conflicting results concerning the identification-emotion link is to consider identification not as a unitary construct but as a twofold concept (Roccas et al., 2006). Roccas et al. (2006) suggested that collective guilt can increase or decrease depending on the type of identification. When individuals identify to the point that the national group is viewed as being superior to other groups, and its behaviors as being unquestionable and incontestable, identification takes the form of 'glorification' of the national group. As high glorifiers are motivated to perceive their group as better and more worthy than other groups, they are more likely to protect their group against any threat to its positive image by minimizing or denying the responsibility for the ingroup's harmful action or by legitimizing it. The second form of identification, referred to as 'attachment', is characterized by a positive feeling of attachment and a critical loyalty to one's nation allowing for taking a critical stance toward the ingroup's past and current actions.
Consequently the relationship between identification and collective guilt is supposed to be different depending on the meaning identification has for people (Roccas et al., 2006). Since defense mechanisms are more likely among those who glorify the nation, greater glorification should be associated with lower levels of collective guilt and outrage-related emotions. Those high in attachment identification are more likely to feel responsible for the ingroup's moral transgressions; higher levels of genuine attachment to one's national group should therefore increase both types of collective emotions. Few studies have taken into account these two modes of national identification (e.g. Bilali, 2013;Leidner et al., 2010), and only one has examined their link with collective emotions for ingroup wrongdoings (Sullivan et al., 2013).

The normative underpinning of collective memory and collective guilt
The second ideological factor of interest within this framework is related to the way a nation deals with its history. Work on collective memory has frequently shown that historical events are mobilized in order to cement a sense of positive collective belonging and national identity, and in this sense have a strong ideological value (see for instance Coquery-Vidrovitch, 2009;Pennebaker et al., 1997;Sibley & Liu, 2012). In agreement with these premises, acknowledgement of responsibilities for the injustices and harm done during the colonization of Algeria and the offer of apologies still generates intense and passionate debates and controversies in contemporary France whenever this period is -sporadically -brought up (e.g. Bancel & Blanchard, 2008). For example, the difficulty of dealing with the dark side of colonization became evident when, in 2005, the French right-wing government intended, finally unsuccessfully, to insert in the law a directive making it mandatory to recognize, in history textbooks, the 'positive role' of colonialism, especially in North Africa (Law n° 2005-158 of 23 February, 2005). This attempt to idealize the colonial past, as well as the superficial handling of this period in school textbooks and teaching (Corbel & Falaize, 2004), and more generally the lack of public amends regarding this historical period exemplify the amnesia surrounding it.
Only recently has the left-wing French President (Hollande, 2012) expressed recognition (albeit nothing more) of the bloody repression by French security forces of a demonstration in favor of Algerian independence in Paris on 17 October 1961, saying merely that 'The Republic acknowledges with lucidity these facts'. This declaration prompted virulent reactions from its opponents (e.g. see a compilation of reactions in Le Monde, 2012).
As a result France seems to oscillate between periods of relative silence and ones of occasional reminiscence when salient events occur (i.e. films, commemorations, bills …) that gave rise to controversial debates and sometimes even to quite violent reactions (e.g. threats made during the organization of a 'Franco-Algerian history' symposium; Abécassis & Meynier, 2008: 17). Considering this context, we were interested in whether French people feel collective emotions for what happened in the past.
In an attempt to explain the relatively high levels of collective guilt among Belgian students, Licata and Klein (2010) discussed the 'normative nature of collective guilt'. They argued that the high levels of collective guilt related to a country's inglorious past may in part be due to the normative pressure to feel and to report feelings of remorse for what was done. However, neither Licata and colleagues nor other researchers tested this assumption. Some national contexts may indeed allow for and even encourage feelings and expressions of negative group-based emotions. Others on the contrary may consider feelings of guilt and shame as not necessary, as counterproductive and inappropriate. The political speech given in July 2007 by the former French President Nicolas Sarkozy (2007) at the University of Dakar suggests that France's current normative context is not conducive to emotions such as collective guilt or more generally negative group-based emotions. Indeed Bancel and Blanchard (2008: 149) note that in his speech Nicolas Sarkozy explicitly rejected repentance: No one can ask the current generations to atone for this crime committed by past generations … The message is clear: a crime was committed -the culprit(s) is (are) not clearly specifiedbut the current generation cannot be held responsible for it, there cannot be any individual or collective 'recognition' (on the part of the French State or Republic), there is no obligation for 'repentance'. It is necessary to move beyond this past, and to advance toward the future.
Recently Sibley (2010) argued that the postcolonial national political discourse that permeates society and shapes public opinion is based, among others, on the ideology of historical recognition versus negation of the relevance for the present of past injustices committed during the colonial period. The former French President's political speech in Dakar, quoted above, can be taken as a good example of historical negation that relieves the current generation of the obligation to address past injustices (see also Augoustinos & Lecouteur, 2004, on the political discourse in Australia about Aboriginal peoples).
In the present study we investigated whether French students endorse this official anti-repentance position. The original contribution of this study is to directly assess participants' normative beliefs and their role in the expression of collective emotions. We hypothesized that, in the context of French colonization in Algeria, a no-remorse norm would prevail, preventing students from expressing negative group-based emotions, and particularly collective guilt. Furthermore collective guilt was measured in two ways. In addition to the classical collective-guilt scale, which contains a strong normative component, we also used a list of more than 20 emotion words, which allowed us to assess participants' different emotional reactions regarding French colonization. Indeed the classical scale may contribute to artificially inflating the level of collective emotions reported, especially when guilt is nearly the only emotion under scrutiny that is available for participants to report. Measuring collective guilt among other group-based emotions may indeed diminish the normative pressure to report guilt.

Political orientation
A third ideological factor we were interested in was participants' political orientation, which is closely intertwined with the normative context and emotion norms. As can be seen from the previous paragraph, there is political disagreement on how to handle historical issues, right-wing proponents being more inclined to endorse an ideology of historical negation (Sibley, 2010) and being more opposed to expressing regrets for what was done (Bancel & Blanchard, 2008). Consequently we considered political orientation as a factor potentially related to emotional reactions to this historical period. Those few studies that measured political orientation (e.g. Doosje et al., 1998;Licata & Klein, 2010) found mixed results. Some found no relationship between political ideology and collective guilt (e.g. Doosje et al., 1998), while others showed right-wing ideology to be associated with less guilt (e.g. Figueiredo et al., 2011;Klandermans et al., 2008;Sullivan et al., 2013), especially for older generations (Licata & Klein, 2010).

Family implication
Finally we examined the role of family implication in the French colonial past in experiencing collective emotions such as guilt. According to Branscombe et al. (2002), for collective guilt to be felt, individuals must not only self-categorize as ingroup members (i.e. member of the former colonizer group), they must also impute responsibility for past wrongdoing to the ingroup. However, individuals tend to distance themselves from events that occurred in the distant past. As the French colonial period is situated in the distant past (it ended 50 years ago), current group members may feel less connected to it and thus feel less responsibility, and in turn experience less guilt or shame for the ingroup's past misdeeds. According to Zebel et al. (2007), family implication in the colonial past is a means through which the present (us) is bound to the past (our ancestors). Knowing relatives or family members who were directly involved in the colonial past facilitates the feeling of psychological connectedness with and the feeling of responsibility for this past and consequently induces collective guilt.
In the present study we therefore asked participants whether or not they were, through their family, involved in that historical period. On the one hand, family involvement seems to constitute a link to the past, facilitating feelings of responsibility, self-relevance and guilt. On the other hand, family implication may be a source of social identity threat making defense reactions more likely. According to social identity theory (Tajfel & Turner, 1986), people are motivated to establish and maintain a positive social identity, which they derive from belonging to a positive group (a nation, a family). When individuals' positive social identity is threatened by the ingroup's past wrongdoing, they tend to protect it by either distancing from the ingroup or denying the ingroup's responsibility or legitimizing the ingroup's harmful actions. To the extent that the identity of participants with family implication in the colonial past is threatened, they may be motivated to legitimize it (for example by using exonerating cognitions, Licata & Klein, 2010;Roccas et al., 2006), which leads them to express less negative group-based emotions such as moral outrage or collective guilt.
As the present study is the first to investigate French students' relationship with their nation's colonial past, we also included several other measures that could influence their emotional reactions, such as their knowledge and beliefs about and representations of French colonization in Algeria.

Participants
Undergraduates enrolled in a first-year introductory social psychology class were asked to complete a questionnaire during a mass-prescreening session, in autumn 2010. The present study includes only the data of those students who indicated they were of French nationality with both parents possessing French nationality, that they were under the age of 24, and who either reported that one of their relatives was a soldier during the Algerian war of independence (SO); N = 92) or were students with no family involvement (WFI) (N = 140).

Procedure
Students completed the questionnaire entitled 'Memories of historical events' during their first social psychology classes, in groups of 20-40. Demographic variables were addressed at the end of the questionnaire. Completion of the whole questionnaire took about 30 minutes.

Measures
After a short free-association task concerning the word 'colonization' (with no mention of French colonization in Algeria), students were asked their perception of the French colonization of Algeria, a period which lasted for a little more than a century. Then measures were taken in the order that follows. When not specified, participants indicated their agreement with each of the items on a seven-point Likert-type scale, ranging from (1) do not agree at all to (7) completely agree.

Self-reported emotions
Participants' current emotional reactions regarding French colonization in Algeria were measured using 26 adjectives. Twenty-four adjectives were adapted from Elliot and Devine (1994; translated into French by Gosling, Denizeau, & Oberlé, 2006), and completed with two adjectives ('sad' and 'outraged'). Participants were asked to indicate to what extent each emotion term corresponded to their current feelings with regard to French colonization of Algeria using a seven-point scale ranging from 1 (does not correspond at all) to 7 (corresponds very much). Guilt was assessed among the other emotions.

Collective guilt
Collective guilt was measured using five more contextualized items adapted from previous research (α = .90). Four items were taken from Mari et al. (2010;e.g. 'Even if I have done nothing bad, I feel guilty for the behavior of the French toward Algerians during the colonial period'); a fifth guilt item was created for this study ('I feel guilt when I think that so far the French government has so little acknowledged its responsibilities in the colonial period').

Collective-guilt norm
The predominant collective-guilt norm was assessed with one item created for this purpose: 'I think that generally French people should feel guilty for France's actions during the colonial period.'

Perceived ingroup threat
Two items assessed whether French colonization was felt as a threat to the ingroup's reputation in the world (Mari et al., 2010), r(232) = .44, p < .001. Threat to values of the national group, threat to one's own personal values and whether looking back into the past might be a threat for national unity were measured with one item each.

Reparation/compensation intentions
Based on past research (e.g. Brown et al., 2008;Dufoix, 2005), willingness to compensate was assessed using five items (α = .77), including presentation of public apologies, support for financial compensation, for the organization of commemorations, for the diffusion of accurate and complete information concerning the colonial period, and willingness to acknowledge that institutional discrimination during the colonial period underlies today's prejudice toward Algerian immigrants.

Exonerating cognitions
Eight items (α = .71), adapted from Licata and Klein (2010), assessed the extent to which participants minimize the harmful character of past ingroup actions, for instance by moral distancing from the event (e.g. 'One cannot judge colonial actions by today's values').

Representations of French colonization
Participants' representations were measured with 11 items taken from Licata and Klein (2010) and adapted to the French colonial period in Algeria. Seven items assessed negative representations (e.g. exploitation of Algerian people, exploitation of Algeria's natural resources, massacres, expropriations, forced labor, racial segregation, the abandonment of the Harkis (pro-French Muslims) at the time of independence; α = .79). Four items assessed positive representations related to 'civilization' inputs (i.e. building of educational and health facilities, development of economic infrastructure and of transportation facilities; α = .70).

Knowledge transmission via school and media
Students' opinions regarding transmission of knowledge by the schools and media was measured with items adapted from Licata and Klein (2010). Participants were asked to indicate (1) whether or not they studied French colonization in school (Yes/No), (2) to what degree this teaching was detailed (from 1 'very superficial' to 7 'very detailed'), (3) whether the image of colonialism conveyed at school was (1) very positive to (7) very negative, (4) to what extent French colonization in Algeria is presented in the media (from 1 'not at all' to 7 'a lot'), (5) whether this media coverage is (1) very positive to (7) very negative, (6) to what extent they have the impression they are informed about colonization in Algeria (from (1) not at all to (7) a lot), and (7) finally, whether they wished to be more informed about this period (Yes/No).

National identification
Participants' identification with their national group was assessed with 16 items adapted from Roccas et al. (2006). Eight items measured national glorification (e.g. 'France is better than other nations in all respects', α = .80). The other eight measured national attachment (e.g. 'Being French is an important part of my identity'; α = .87).

Racial-prejudice scale
Participants' level of prejudice against foreigners in France, and especially Mahgrebians/Arabs (people from North African countries -Algeria, Morocco and Tunisia) was measured using a scale developed by Dambrun and Guimond (2001;15 items, α = .88).

Knowledge of the Algerian colonization period
Ten multiple-choice questions were created to test participants' general knowledge of the colonial period (six questions) as well as of the war of independence that ended French colonization in Algeria (four questions). The number of correct answers indicated their level of knowledge.

Socio-demographic questions
At the end of the questionnaire, students were asked to answer several demographic questions, among which were: sex, age, political orientation, nationality, their country of birth as well as that of their parents, and family implication.
Political orientation. This was measured with one item (Gaffié, 2006), asking participants to place themselves on a 10-point scale, ranging from (1) extreme left wing to (10) extreme right wing.
Family involvement. Family involvement was assessed by asking participants whether a family member (parents, grandparents …) was personally concerned by the French colonization of Algeria and/or the Algerian war of independence, and to specify his/her exact role (e.g. as colonized Algerian, as French settlers in Algeria -commonly called 'piedsnoirs', or as French soldiers during the war).

Knowledge and beliefs regarding French colonization in Algeria
Words spontaneously evoked when thinking about colonization. Students rarely evoked words referring to Algeria (i.e. Algeria, Algerian War, Front de Libération Nationaleone of the main organizations fighting for liberation of Algeria and still the main political party in this country), and in the same proportion for students with no family implication (WFI, 19.6%) and with soldiers ascendants (SO, 19.47%). Moreover very few students evoked France (6.4% of WFI group and 3.2% of the SO group).

Knowledge about Algerian colonization
Participants were not very successful in answering questions about Algerian colonization and decolonization (M = 4.00, SD = 2.03). A 2 (period assessed) × 2 (family implication) ANOVA revealed that participants, independently of their group membership, did better on questions concerning the Algerian war of independence (53.98% of correct answers) than questions covering the whole colonization period (30.67%), F(1,230) = 172.82, p < .001. This result is consistent with the strong emphasis put on the Algerian war in high school history classes. The main effect for family implication and the interaction were not significant (both Fs < 1).

Knowledge transmission via school and media
Almost all participants reported having studied French colonization in school (95.3%, no difference between groups, χ 2 (1) = 2.22, ns). WFI students slightly more than SO students reported school learning to have been moderately detailed (M WFI = 3.89, SD=1.52 vs M SO = 4.26, SD = 1.55; t[227] = 1.79, p = .07). Overall, participants considered school as having transmitted a negative, albeit not extremely negative, image of French colonialism (M = 4.89, SD = 1.06). Moreover students from both groups reported media coverage of colonialism as being quite weak (M WFI = 2.88, M SO = 2.67, t[229] = -1.73, p = .08) and rather negative, albeit not extremely negative (M = 4.79, SD = 1.27). Finally they did not have the impression of being very informed about colonization in Algeria (M = 2.97, SD = 1.31), and a large majority of the participants expressed the wish to be more informed about this period, WFI students (79.3%) slightly more so than SO students (69.2%, χ 2 [1] = 3.00, p = .08).

Emotional reactions to colonization
Emotions were aggregated according to the results of a PCA that revealed four factors with eigenvalues superior to 1 (accounting for 58.59% of the total variance; KM0 = .90, χ 2 [325] = 2866, p < .001). Only emotions clearly loading on one factor and with a loading criterion superior to .6 were considered in the grouping. The four factors emerging were outrage-related emotions (disappointed, critical, angry, outraged and disgusted; accounted for 35.14% of the total variance), discomfort-related emotions (embarrassed, uncomfortable, uneasy, bothered; accounted for 12.69% of the total variance), anxietyrelated emotions (anxious, frustrated, tense, worried; accounted for 5.69% of the total variance) and positive emotions (e.g. friendly, happy, content, optimistic, good; accounted for an additional 5.07% of the total variance). The first and the second factors are similar to the negself and discomfort indexes, respectively, found by Elliot and Devine (1994).
Guilt did not load clearly on any of the three negative components (loadings falling between .28 and .36) and was therefore treated separately. Based on the PCA results, we computed an outrage-related emotion score (α = .87). As can be seen in Table 1, overall, participants reported quite low levels of emotions when asked to think about Algerian colonization. In fact all emotion scores were rated significantly below the midpoint (4) of the scale. The strongest/most expressed emotions and the only ones for which the two groups differed were outrage-related emotions. SO students reported less collective outrage-related emotions (M = 3.39, SD = 1.47) than WFI students (M = 3.80, SD = 1.49), t(230) = 2.04, p = .04. Guilt on the other hand was rated very low (M = 2.00, SD = 1.35) by both groups (i.e. no between-group difference).

Collective-guilt scale
When considering the collective-guilt scale (see Table 1

Collective-guilt norm
Overall, participants did not think their compatriots should feel guilty, indicating the endorsement of a no-remorse norm (M = 2.77, ET = 1.67; compared with 4, the middle value of the scale: t(231) = -11.21, p < .001). There were no significant differences between groups for any of these variables (both ts < 1). It should be noted that the collective-guilt norm is more strongly related to collective guilt when measured with the guilt scale (r[232] = .53, p < .001) than when embedded in the list of emotions (r[232] = .27, p < .001), z = .4.53, p < .001.
Associations between emotions regarding French colonization in Algeria, ideological factors, prejudice and compensation. Pearsons' correlations were similar in both groups so they will be considered at the level of the entire sample (see Table 2). Higher levels of collective guilt were significantly related to more knowledge about colonization ( . Note that, although the emotion norm item focused on a single emotion (i.e. collective guilt), the no-remorse norm was also associated with less moral-outrage-related emotions (r[232] = .36, p < .001).
We computed partial correlations in order to assess the unique associations of ideological indicators with both collective emotions. As correlations did not differ between groups, we report the partial correlations based on the entire sample (see Table 3). Results revealed that, when controlling for all other ideological factors, the collective-guilt norm was still significantly related to collective guilt (r[227] = .51, p < .001) and to outragerelated emotions (r[227] = .336, p < .001). Furthermore national attachment was marginally related to collective guilt (r[227] = .12, p= .07) but not to outrage-related emotions (r[227] = .07, p= .25). On the contrary, national glorification was linked to outragerelated emotions (r[227] = -.14, p= .03) but not to collective guilt (r[227] = .03, p = .34). Political orientation was found to be significantly associated with both collective guilt (r[227] = -.16, p = .01) and outrage-related emotions (r[227] = -.30, p < .001).
Finally we conducted two regression analyses using the two sets of negative emotions we hypothesized to affect attitudes toward reparation and prejudice. Since our correlational data and past research have shown links between political orientation and prejudice (e.g. Luguri et al., 2012) as well as between glorification of the country and prejudice (e.g. Falomir-Pichastor & Frederic, 2013), we included national attachment and glorification, and political orientation as well as the collective-guilt norm in the analyses. We primarily ran regression analyses including the family implication variable and all interactions with the other predictors (centering the predictors on the mean in the first place).  .14* -.28*** .29*** -.23*** -.44*** -.41*** .66*** -.14* .36*** .56*** a p < .10, *p < .05, **p < .01, ***p < .001.
There were neither significant association with family implication nor significant interactions, except for a family implication by collective-guilt norm interaction on prejudice (β = .12, t[218] = 2.13, p = .034). This interaction revealed that, while the more students with family implication endorsed a no-remorse norm the more they expressed prejudiced attitudes, no such link was found for students with no family implication. As including these interactions did not change the results, we report results of the regression for the entire sample (see Table 4).
First, concerning reparation intentions, the model accounted for 38% of the variance, F(6,225) = 25.07, p < .001. As can be seen in Table 3, all variables were significant predictors of compensation will. Secondly and even more important for our hypotheses, when controlling for these variables, collective guilt (β = .13, t[225] = 1.95, p = .03) as well as outrage-related emotions (β = .29, t[225] = 4.43, p < .001) were both significant predictors of the desire to compensate for past wrongdoing.
Finally, our set of predictor variables accounted for 50% of the variance in prejudice (F[6,225] =40.39, p < .001). National glorification and political orientation were the only two ideological indicators that were significantly related to prejudice. In other words the more participants glorified the nation and were right-wing oriented, the more

Discussion
The present study was designed to investigate how French people's colonial past affects their current relationships with a historical outgroup by taking into account the emotional reactions to the French colonization of Algeria. We hypothesized collective guilt and outrage-related emotions to be linked to willingness to compensate for past wrongdoing as well as to less prejudiced attitudes toward French of Maghrebian origins. Considering the particularity of the French context relative to these questions, we expected various ideological variables to be related to participants' emotional reactions. In particular we expected French students to endorse a no-remorse norm that prevents them from feeling strong levels of collective guilt for what happened in the past. What can be learned from French students' emotional reactions toward past colonization of Algeria?

Family implication
First of all, family implication with regard to the past does not seem to play a significant role in the present-day context. In fact we almost never found significant differences between students with soldier ascendants and those without family involvement, either in terms of emotions (except for moral outrage) or in terms of their relation with ideological variables or attitudes toward the Maghrebian outgroup. Spontaneous evocations, knowledge about and perceptions of knowledge transmission of French colonization in Algeria were also similar in both groups. Groups differed mainly in terms of national identification, representations of colonization and exonerating cognitions, with SO students being more identified with the national group and more prone to rely on positive aspects of colonization and on cognitions lessening collective responsibility. Future research would do well to compare other sub-samples of participants. For instance, descendants of repatriated 'pieds-noirs' deserve greater attention since they might be especially prone to exonerating cognitions and sensitive to the normative context.

Expression of collective emotions, prejudice and willingness to compensate
Taking the entire sample, French students reported reasonably low levels of collective emotions regarding this historical period. Moreover when students were given the opportunity to express various emotions, guilt was one of the less intense negative emotions to be reported. Even when guilt was assessed using the collective-guilt scale, its level was again quite low. Furthermore and consistent with past research, the two group-based emotions of interest predicted the willingness to compensate and prejudice toward the outgroup in both samples. The finding that collective outrage-related emotions were more strongly related to compensation intentions than collective guilt is also consistent with past research (e.g. Leach et al., 2006). However, contrary to our expectations, the strength of the emotion-prejudice link was similar for collective guilt and outrage-related emotions.
One interesting, and quite unexpected, result was the weak correlation between the level of knowledge about the Algerian colonization and collective guilt, on the one hand, and outrage-related emotions, on the other hand (the latter did not reach statistical significance). Considering participants' low scores on the quiz, their poor knowledge of this historical period might prevent them from reacting emotionally to it, which in turn would explain their relatively low levels of reported emotions. This leaves also more room for the normative context to influence participants' emotional reactions to historical events.

Normative context
Consistent with our main hypothesis, we found that a strong no-remorse norm prevailed in the French context, which contributes to explain the relatively low levels of collective emotions expressed regarding this specific historical period. Thus contrary to what Licata and Klein (2010) suggested in the Belgian context, and contrary to what some writers refer to as the tyranny of guilt (e.g. Bruckner, 2010), the French context does not seem to promote a 'culture of guilt' regarding France's colonial history. In fact when asked whether French citizens should feel guilty for past colonialism, French students thought they should not. This no-remorse norm had strong effects on self-reported collective emotions, whether collective guilt or outrage-related emotions: it prevented not only the expression of guilt but more generally the expression of any kind of emotional reactions to past inequalities and violence. The finding that collective guilt was higher when measured with the collective-guilt scale than when embedded in a list of various emotions, and the fact that the correlation with the norm was stronger for the latter than for the former suggest that self-reports of collective emotions are very sensitive to the normative context. This raises the question of the extent to which collective emotions are really experienced and felt or simply expressed or suppressed depending on the prevailing emotion norms.
The issue of the normative nature of collective emotions certainly deserves greater attention in future research. We are currently investigating whether and how different normative contexts promote or discourage the expression of collective emotions regarding various historical events. More generally the literature on the influence of group norms (e.g. Schultz et al., 2008) is of particular relevance for examining descriptive emotion norms (i.e. the prevalence of a specific collective emotion among people in the country) and injunctive emotion norms (i.e. the emotions considered as appropriate) with regard to specific historical events and outgroups, and their relation to emotions, attitudes (e.g. prejudice) and behaviors (e.g. discrimination). Cultural differences concerning emotion norms for expressing emotions might also be of importance (Frijda & Mesquita, 1994). For example guilt, contrary to pride, has been found to be less desirable in individualistic than in collectivist cultures (e.g. Eid & Diener, 2001). Culture may affect not only the display of individual emotions but also the motivation to display certain collective emotions more than others.

Ideological treatment of the past
More generally our results suggest that participants' collective emotions were strongly influenced by the French ideological climate. Indeed apart from the no-remorse norm, political orientation was also quite reasonably related to emotional reactions, reflecting the intense political debate surrounding these questions in the French public space (see Bancel et al., 2005). This finding departs from those of other studies (e.g. Doosje et al., 1998;Licata & Klein, 2010), which found no association between political orientation and collective guilt. Interesting, and somewhat unexpected too, was the finding that national identification was only modestly and inconsistently linked to collective emotions. Nevertheless, the distinction between two types of national identification (attachment and glorification; Roccas et al., 2006) still appears to be relevant, even in the French context, where patriotism is not a particularly popular value, especially among students. Indeed the two dimensions of national identification were found to be differently related to our measures of collective emotions and attitudes toward the outgroup (e.g. attachment was positively and glorification negatively associated with compensation intentions). However, as the question of French colonization of Algeria is an eminently political one, ideological issues such as the inclination to legitimize existing social systems should be explored in more depth, in addition to group identity-protection motivation (Starzyk et al., 2009). The ideological treatment of the past might also be rooted in the specific propaganda the French Third Republic used to legitimize its colonization agenda (e.g. as compared to the English colonialist ideology). The displayed ambition of France was to civilize the indigenous with love, to generously 'guide "people in the dark" rather than subjugate them' (Vergès, 2003: 196), an idea embraced in the expression 'a mission to civilize'. How this ideology concerning the past has spread into today's French ideologies and attitudes is an interesting question that deserves future investigation (for a similar inquiry regarding the influence of Portuguese colonial ideology, see Vala et al., 2008).
To conclude, Igartua and Paez (1997) reported a study in which students, after watching a dramatic film on the Spanish Civil War, expressed negative self-focused emotions such as sadness, guilt and shame, emotions which were in turn linked with consideration of how the past conflict has determined present-day politics and with the necessity of remembering. These findings point to the importance of linking the socio-psychological analysis of present-day relationships between French with French ascendants and French with Algerian ascendants with an analysis of their historical roots (e.g. Abécassis & Meynier, 2006;Bancel et al., 2005). There is no doubt that inquiring how the past is perceived and felt is useful to better understand how it echoes in people's present psychological functioning.