1. Introduction

In an intergroup context, the choice of words to use in the description of the behavior of the members of a group depends on the category in which the said group is classified (Assilaméhou-Kunz, 2013; Burguet, 2011). When it comes to the description of outgroup members, the choice to use descriptive terms corresponds to linguistic manipulation. These language variations observed in individuals when they are called upon to describe outgroups contribute to the intensification of the Perpetuation of Linguistic Intergroup Bias (PLIB), which is generally in favor of ingroup (pro-ingroup PLIB) and against the outgroup (anti-outgroup PLIB) (Assilaméhou-Kunz, 2013). These asymmetric variations in language are called Linguistic Intergroup Biases (LIB) (Maass et al., 1989).

The research on LIB is based on the Linguistic Category Models (LCM) (Semin & Fiedler, 1991) which analyze the semantic categorization of verbs according to their level of abstraction (Burguet, 2011). LCM have contributed to the classification of terms used in interpersonal relationships into linguistic categories, which refer to the set of verbs, adjectives and nouns that individuals use to describe the actions/behaviors and psychological states they have observed (Assilaméhou-Kunz, 2013; Chisango & Mayekiso, 2014). Thus, the use of LCM in the study of intergroup distortions has shed light on the role of linguistic abstraction in the expression of linguistic biases (Over & McCall, 2018; Otterbacher et al., 2019). It should be remembered that intergroup biases occur at the level of cognitions, attitudes and behaviors. Following this logic, the LIB model relies on the properties of abstraction, implemented within the framework of LCM, to predict the use of abstraction and its consequences in intergroup context (Assilaméhou-Kunz, 2013). It predicts that individuals make a differentiated use of abstraction, depending on whether the event to be described is desirable or undesirable and whether the use of linguistic abstraction subtly contributes to the maintenance of intergroup biases (Maass et al., 1989; Semin & Fiedler, 1991).

The literature highlights two types of factors that push individuals to use LIB. The first type is motivational (Assilaméhou-Kunz, 2013). In this case, the pressure to acquire a positive social identity and self-esteem during social comparison encourages the expression and maintenance of LIB (Gentet, 2014). The motivational explanation of LIB is based on social identity theory (Assilaméhou-Kunz et al., 2019). In this logic, the use of a particular language style depends on the strategies used to achieve a positive social identification (Marfaing & Tafani, 2011). Consequently, the level of asymmetric linguistic abstraction would be the object of social regulation, by valuing the positive ingroup’s image and denigrating outgroup’s image (Maass et al., 1996). The second type of factor is cognitive. It is linked to the differences between ingroup and outgroup’s expectations (Assilaméhou-Kunz et al., 2019). Concretely, individuals expect outgroup members to engage in more negative behaviors than ingroup members, and ingroup members to engage in more positive behaviors than their outgroup’s counterparts. In other words, it means that individuals make and communicate inferences from behaviors that match their expectations (Shulman & Clément, 2008).

In addition to the two types of factors mentioned above, certain aspects of intergroup context (group entity, status and power) vary the expression of LIB (Moscatelli et al., 2008). More explicitly, the increase in the homogeneity of the group pushes its members to resort to the asymmetrical description of ingroup and outgroup members’ behaviors (linguistic bias) (Chisango & Mayekiso, 2014; Otterbacher et al., 2019; Over & McCall, 2018). Likewise, when individuals face a situation of intergroup asymmetry (high versus low status), they express more discrimination through recourse to LIB than when they are in a symmetrical intergroup situation (equal status) (Malloy et al., 2011). Finally, linguistic discrimination would be higher under conditions of high and low power (asymmetry of power between groups) than under conditions of equal power (symmetry of power between groups) (Moscatelli et al., 2008). However, individuals value ingroup and outgroup when they adhere to and respect ingroup norms (Marques, 2004). They also directly verify the effects of the salience of ingroup norms. As a result, when they are not aware of the existing relationship between ingroup members’ behaviors and group norms, they adopt both favorable and unfavorable judgments (Assilaméhou-Kunz, 2013). Indeed, the results of several studies indicate that a user of LIB has pro-ingroup and anti-outgroup’s intentions and attitudes, concretely consisting in granting positive or desirable explanations and descriptions to ingroup members and negative or undesirable explanations and descriptions to outgroup members (Assilaméhou-Kunz et al., 2019; Valencia et al., 1998).

The literature reveals that ingroup members use LIB as a tool that allows them to express their loyalty to the said group and to assert its superiority over outgroups (Assilaméhou-Kunz et al., 2019). Indeed, individuals make language choices that comply with LIB when they wish to be perceived as good ingroup members (Assilaméhou-Kunz & Testé, 2013b). This behavioral tendency suggests the possibility that the use of LIB could be analyzed as a norm. The normative approach of LIB is based on two assumptions. The first predicts that speakers can use linguistic abstraction strategically. They can use LIB when they want to present themselves as members of a standardized group, in order to obtain the approval of the other members of that group. The second stipulates that recipients can judge speakers on their use of linguistic abstraction. The use of LIB is therefore presented as an ingroup differentiation criterion that is used to distinguish its loyal from its disloyal members. It means that being a good member of the group involves using distortions shared by the group and its members. With reference to Asch’s (1951) observations, it is specified that group norms includes fidelity, loyalty and ingroup favoritism.

From a normative perspective, Assilaméhou-kunz et al. (2019) analyze LIB by assessing the effect of the approval of group members who use them on their tendency to perpetuate them. These authors predict and observe that intragroup approval (vs. disapproval) of an ingroup member who uses LIB increases the likelihood of such biases being expressed and perpetuated. In other words, people use LIB more when their use is encouraged by their group. Likewise, the perceived relevance of ingroup or outgroup’s description increases the intention to express pro-ingroup LIB in future intragroup interactions.

The normative aspect of LIB opens a path for the present research which is in the perspective of the analysis of factors that push individuals to comply or not with group norms. These factors can be situational, groupal (Allen & Levine, 1969; Asch, 1951; Moreland & Levine, 1982) and linked to personality characteristics (Barron, 1953; Di Palma & McClosky, 1970; Irish, 2016). This study focuses on these characteristics. The literature suggests that the variability of personality traits impacts on compliance with group norms (Hyldegård, 2009; Martin & Alaphilippe, 2005; Walker & Andrade, 1996). This is why studies have looked at the characteristics of individuals who conform to norms, compared to those of people who do not. These characteristics include gender, skill/intelligence (Hyldegård, 2009), age (Martin & Alaphilippe, 2005; Walker & Andrade, 1996), dependence/independence of judgment (Barron, 1953) and social desirability (Crowne & Marlowe, 1960; Madern, 2014). Thus, following group’s prescriptions depends on the personality characteristics of each member (Hollander, 1960). This raises questions about the personality characteristics likely to incline members of a group who approve or disapprove the use of LIB to use them or not in their descriptions of ingroup or outgroup members’ behaviors. As a response to this question, the present study proposes the inclination to collective narcissism, a personality characteristic whose link with conformist behavior has not yet been established by the literature. Its relevance as a factor making it possible to modulate the link established between the approval of using LIB and their perpetuation (Assilaméhou-Kunz et al., 2019) is due to two main reasons: the relationship that this personality characteristic could have with LIB and the possibility that collective narcissism (CN) is a source of nonconformity.

Collective narcissism is understood as an individual emotional investment in the unrealistic and unequal belief in the exaggerated greatness of a group, which depends on external validation (i.e. from the outgroup) (Golec de Zavala et al., 2009). It is a belief that leads to conceive ingroup as being exceptional and to benefit from a right to privileged treatment on the part of outgroups (Golec de Zavala, 2018). Thus, collective narcissism is when individuals exaggerate the importance of their group and are convinced that its true value is not sufficiently appreciated by others. This tendency is positively associated with variables relating to the idealization and perceived superiority of the group. However, it is not shared by outgroups members; hence the problem it poses. Indeed, at the heart of collective narcissism is the resentment that the exceptionality of ingroup is not sufficiently appreciated by outgroups (Golec de Zavala et al., 2019). As a result, this tendency generates hostility towards the outgroup and impacts on intergroup cognitions. Concretely, collective narcissism predicts prejudice, hostility, and retaliation in response to past, present, real and imaginary offenses against ingroup (Golec de Zavala & Schatz, 2013; Dyduch-Hazar et al., 2019).

Collective narcissism has a theoretical connection with LIB. Indeed, concepts such as ingroup, outgroup, ingroup favoritism, outgroup derogation, pro-ingroup or pro-outgroup biases, ingroup’s protection, social hierarchy or the social identity of the group are found both in the definitions and characteristic processes of collective narcissism and LIB. These two constructs also have the same sources. For example, collective narcissism is motivated by the belief in the group’s right to a privilege (Golec de Zavala et al., 2019), the desire to distinguish ingroup very positively from outgroups and the fear that the realization of this desire is threatened (Golec de Zavala et al., 2019). Likewise, the motivation to maintain and protect the positive ingroup’s image and to devaluate the outgroup inclines individuals to collective narcissism (Dyduch-Hazar et al., 2019). It is practically the same factors that push them to use LIB. Just as the approval of using LIB increases the likelihood of expressing and maintaining distortions between groups, and the feeling of being encouraged by other members pushes individuals to view their actions as accepted by group norms (Assilaméhou-Kunz et al., 2019), collective narcissism is a robust predictor of intergroup hostility and is associated with a biased perception of intergroup realities. In this vein, individuals with a collective narcissism tendency view ingroup hostility towards outgroup as justified and defensive (Dyduch-Hazar et al., 2019). They also perceive ingroup’s violent acts as a form of protection of its position (social identity) against the hostility of the outgroup. In the same vein, it should be noted that the people who use LIB and those who have an inclination for collective narcissism have in common the fixation with the group. It is in this logic that it has been observed that people who are identified as having an inclination for collective narcissism define themselves according to their membership in a group and expand their self-concept to include the said group (Roccas et al., 2006 cited by Golec de Zavala, 2019). They feel emotionally attached to that group and want to contribute to it.

A set of research findings lead to the conclusion that efforts to acquire a social identity defined by collective narcissism include intergroup hostility and internal tensions, as does support for policies and actions that, while affirming a social identity, harm the group’s well-being. It involves rejection of ingroup members who do not conform to a narrowly interpreted social identity (Golec de Zavala & Keenan, 2020). In other words, collective narcissism can be conceived of as an extreme form of attachment to ingroup. It implies not only hostility towards outgroup members, considered to be threatening to group identity (Golec de Zavala & Bierwiaczonek, 2020), but also the rejection of ingroup members considered as moderate in their attitudes and behavior towards the outgroup and who would thus call into question the group ideology to which they are particularly attached (Klar & Bilewicz, 2017; Marchlewska et al., 2019). On the basis of these elements, the present study proposes that because of the link between collective narcissism and a tendency to vengeful intergroup hostility, especially in situations where ingroup is the object of criticism or its image is tarnished, individuals overly believing in the greatness and superiority of the said group are likely to resort to LIB, even if group norms disapprove it. It is therefore predicted that this characteristic of the personality is likely to push people who have it not to comply with the group norm of non-recourse to LIB, due to the fact that they would be convinced of the need to protect group identity, threatened by outgroup members.

In short, collective narcissism and LIB are two concepts that are theoretically linked. They have in common their characteristics, determinants and goals. These are the means by which individuals seek to protect and maintain group identity. However, to our knowledge, no study has empirically tested the existence of the link between these two constructs. From this perspective, the scientific project of this research is to examine the potential role of collective narcissism in regulating the link established between intragroup disapproval of using LIB and their perpetuation. Indeed, Assilaméhou-Kunz et al. (2019) note that within ingroup, the refusal to use LIB mitigates their perpetuation. In this vein, and in connection with the results of these authors, this study seeks to know whether collective narcissism is likely to lead to the perpetuation of LIB in situations where the use of these biases is disapproved by ingroup; hence the following research question: what is the effect of collective narcissism on the link between intragroup disapproval of the use of linguistic intergroup biases and their perpetuation? In response to this question, it is predicted that in a situation of disapproval of using LIB, individuals who have a high level of collective narcissism will tend not to conform to group norm and therefore perpetuate LIB.

2. Method

2.1. Participants

The participants are 54 Francophone students of both sexes (31 women and 23 men), aged between 18 and 43 years (M. age = 22.94 years). They agreed to participate in the study on a voluntary basis by signing a free and informed consent form. The experimenters informed them that they are participating in an investigation concerning the crisis in the anglophone regions of Cameroon. In this country, relations between the two linguistic communities (Anglophone and Francophone) are marked by stereotypes, prejudices and discrimination; hence the fact that this study is based on it as an intergroup context (Messanga, 2018; Messanga et al., 2020).

2.2. Experimental Material and Procedure

This research is duplicating Assilaméhou-Kunz et al.’s (2019) experiment. In fact, the participants are installed in an amphitheater on the university campus. The experimenters introduce themselves as researchers associated to the National Commission for the Promotion of Bilingualism and Multiculturalism, whose mission is to collect their opinions on the anglophone problem in Cameroon. Participants are informed that their class has been chosen for this exercise, because it is part of a bilingual university, therefore comprising both anglophone and francophone students.

The experimental manipulation begins with informing the participants that a previous study was conducted among students enrolled in the Sociology unit of the same university. During the said study, a group of 15 Francophone participants was asked to watch and rate the video of a debate between anglophone and francophone students. At the end of this debate, a statement was made by one of the 15 students (fictitious aspect of the study). It contains descriptions using pro-ingroup LIB, that is to say the speaker uses positive abstract and negative concrete terms to describe francophone students’ behaviors (ingroup members) (Messanga et al., 2020). In contrast, negative abstract and positive concrete terms are used to describe anglophone students’ behaviors (outgroup members). The purpose of this scenario is to create an intergroup context among the participants. After presenting this statement, they are informed of the reaction of other ingroup members when they learned of the comments made in the statement. This information is intended to induce reactions of disapproval or approval of using LIB.

The participants were divided into 2 groups (diagnostic and non-diagnostic), in accordance with the treatments done by Assilaméhou-Kunz et al. (2019). One half is randomly assigned to the condition of declarations that manipulate Intragroup Disapproval of Recourse to LIB (IDRLIB). The other half is randomly assigned to the condition of declarations that manipulate the Intragroup Approval for Recourse to LIB (IARLIB). After reading the scenarios of the experimental condition to which they are assigned, participants must give their impressions on scales evaluating various constructs related to the object of the research. These scales can be grouped into two categories: those that measure the main variables of the study and those that assess the control variables; constructs with potential theoretical links with LIB. The first category comprises: collective narcissism scale (Golec de Zavala et al., 2009; α = .78); perpetuation of LIB scale (Assilaméhou-Kunz et al., 2019; α = .80) and group norm scale (Assilaméhou-Kunz et al., 2019). The second category includes: ingroup identification scale (Assilaméhou-Kunz et al., 2019; α = .80); political extremism scale (Dono et al., 2018; α ϵ [.68-.80]); dependence/independence of judgment scale (Barron, 1953; α = .81) and social desirability scale (Strahan & Gerbasi, 1972; α = .78). The filling of these psychometric instruments was followed by a post-experimental debriefing session.

3. Results

The inferential statistics presented in this section relate to the variations of tendencies in the perpetuation of LIB among participants. First, linear regression indices are calculated to test the effects of interactions of the independent, moderator, and control variables on the dependent variable (Tables 1 and 2). Second, comparison of means tests (ANOVA and t-test) are applied to analyze inter and intragroup differences (Table 3).

3.1. Testing the Study Hypothesis

Table 1 presents the results of the linear regression model, which analyzes the effects of interactions between variables under diagnostic condition. These results show that not all adjustment qualities are satisfactory. Indeed, IDRLIB very weakly and insignificantly predicts PLIB (N = 23; R² = .039, β = -.197; t = -1.116; p = .273). Concretely, these two intragroup and intergroup behaviors are asymmetrical. They are not located on the same slope because they have interactive effects. It reflects the fact that they do not act in the same direction, although they have a positive, very weak and insignificant connection. The recourse to collective narcissism as a moderator variable of this causal link is important. In fact, the data collected show that collective narcissism very weakly and not significantly predicted the tendency to IDRLIB (N = 23; R² = .034; β = -.185; t = -1.046; ρ =.304). This relationship is not satisfactory. The same is true when collective narcissism is linked to PLIB (N = 23; R²= .003; β = .067; t = .375; ρ= .710). It predicts and impacts negatively and not significantly on PLIB. With reference to the model of moderation of the link between IDRLIB and PLIB considered, statistics show that collective narcissism positively moderates this link, but not significantly (N = 23; R² = .036; β = .032; t = .176; ρ = .861). This modulation changes the direction of the link and the impact of IDRLIB on PLIB. This is a positive result. In conclusion, collective narcissism has a positive effect on the link between IDRLIB and their perpetuation.

Table 1. Control (CV) and moderator (MV) variables, Intragroup Disapproval of Recourse to LIB (IDRLIB, IV) and Perpetuation of LIB (PLIB, DV) (diagnostic condition)

It emerges from the analysis of Table 1 that social desirability has no significant effects on IDRLIB (N = 23; R² = .000; β = .019; t = .105; ρ =.917), PLIB (N = 23; R² = .014; β= -.120; t = -. 673; ρ = .506) and the link between these two variables (N = 23; R² = .014; β = .116; t = -.654; ρ = .518). We also observe the negative and not significant effects of ingroup norm on IDRLIB (N = 23; R² = .000; β = -.120; t = -.674; ρ = .506), PLIB (N = 23; r² = .003; β = -.061; t = -.338; ρ =.738) and the relationship they have with each other (N = 23; R² = .003; β = -.085; t = -.476; ρ = .638). There are positive effects, both non-significant and significant of identification on IDRLIB (N = 23, r² = .104; β = .322; t = 1.892; ρ =.068), PLIB (N = 23; R² = .295; β = .543; t = 3.60; ρ =.001) and the link between IDRLIB and PLIB (N = 23; R² = .449; β = .677; t = 4.728; ρ =.000). Independence of judgment has negative and non-significant effects on IDRLIB (N = 23; R² = .006; β = .076; t = .426; ρ=.673), PLIB (N = 23; R² = .005; β = -.074; t = -.411; ρ = .684), and the link between these two variables (N = 23; R² = .205; β = -.059; t = -.329; p = .745). Extremism has the same predictive effect on IDRLIB (N = 23; R² = .001; β = -.028; t =-.156; ρ =.877), PLIB (N = 23; R² =.002; β = -.047; t = -.263; ρ = .794), and the link between these two variables (N = 23; R² = .041; β = -.198; t = -1.108; ρ = .277).

Table 2 presents the results of the effects of interactions between the variables of the study and the control variables in non-diagnostic condition. In this condition, we manipulate Intragroup Approval of Recourse to LIB (IARLIB), collective narcissism and PLIB. These results show good adjustment qualities when IARLIB predicts PLIB (N = 20; R² = .735; β = .857; t = 7.071; p = .000). IARLIB has strong significant predictive effect for PLIB. When we associate collective narcissism as a moderator variable of this effect, we observe a positive, strong and significant predictive power of PLIB by IARLIB (N = 20; R² = .940; β = .74; t = .25; ρ = .023). Collective narcissism works by changing the intensity of the effect of IARLIB on PLIB. The analysis shows positive, strong and significant interaction effects between collective narcissism and IARLIB (N = 20; R² = .870; β = .933; t = 10,95; p = .000) and between collective narcissism and PLIB (N = 20; R² = .803; β = .896; t = 8,57; p = .000).

Table 2. Control (CV) and moderator (MV) variables, Intragroup Approval of Recourse to LIB (IARLIB, IV) and Perpetuation of LIB (PLIB, DV) (non-diagnostic condition)

Following the logic of the effects of control variables on the link between IARBIL and PLIB, we note that social desirability interact significantly with IARLIB (N = 20; R² = .795; β = .891; t = 8,345; ρ = .000), PLIB (N = 20; R² = .897; β = .947; t = 12,53; p = .000) and the link between the two variables (N = 20; R² = .898; β =,890; t = 5,211; p = .000). These effects are positive and significant between group norm and IARLIB (N = 20; R² = .863; β = .929; t = 10,647; p = .000). Group norm acts significantly on PLIB (N = 20; R² = .895; β = .946; t = 12,380; p = .000) and non-significantly on the link between IARBIL and PLIB (N = 20; R² = .898; β = -.155; t = -.744; p = .467). Identification with the group does not significantly interact with IARLIB (N = 20; R² = .782; β = .884; t = 8,026; ρ = .602). However, it has a significant effect on PLIB (N = 20; R² = .744; β = .863; t = 7,239; ρ = .000) and on the link between IARLIB and PLIB (N = 20; R² = .785; β = .434; t = 1,805; ρ = .089). Identifying with the group by reducing the power of the effect changes the significance of that effect. As for the independence of judgment, it has a significant effect on IARLIB (N = 20; R² = .815; β = .903; t = 8,904; ρ =.000), PLIB (N = 20; R² = .944; β = .972; t = 17,467; ρ = .000) and impacts negatively on the link between these variables (N = 20; R²= .946; β = -.10; t = -.81; ρ = .425). This variable reduces the power and sense of prediction of PLIB by IARLIB. The tendency to extremism significantly predicts IARLIB (N = 20; R² = .824; β = .908; t = 9,181; ρ = .000) and not significantly PLIB (N = 20; R² = .851; β = .922; t = 10,125; ρ = .000). Rather, it reduces the predictive power of PLIB by IARLIB (N = 20; R²= .853; β = .818; t = 3,689; ρ =.002) and this not significantly.

3.2. Analysis of Intergroup and Intragroup Differences

Table 3 presents the results of the comparisons of the means of the scores of participants assigned to the diagnostic (IDRLIB) and non-diagnostic (IARLIB) conditions. The data collected indicate that there is a significant difference between these two groups (f (1, 49) = 23,254; Ŋ₁ = .957, Ŋ²₁ = .916; df = 22; ρ =.000). In other words, there is a difference in the tendency to perpetuate LIB under the two experimental conditions. This observation empirically supports the hypothesis of the study. Concretely, the participants having the same level of collective narcissism (MCN (experimental group) = 34,200> Average Score CN = 27; t_CN (experimental group) = 31,248, df = 19, ρ <.05; MCN (control group) = 27,250 > Mean Score CN = 27.00, t_CN (Control Group) = 15,652, df = 19, ρ <.000; f (2, 49) = 46,226, Ŋ₂ = .995, Ŋ²₂ = .990, df = 19, ρ =.000) have a similar tendency to perpetuate the recourse to LIB, both in a situation of intragroup disapproval and in a situation of approval of the said recourse (PLIB, tPLIB (experimental group) > t_PLIB (control group) = 24.30 > 5,88; MPLIB (experimental group) > MPLIB (experimental group) = 501,04 > 183,53; SDPLIB (experimental group) = 98,877 and SDPLIB (control group) = 159,03, ρ = .000). In addition, we observe that in a situation of IDRLIB, participants who have a High Collective Narcissism (HCN) perpetuate LIB more than those who have a Low Collective Narcissism (LCN) (f (3, 27) = 22,814, Ŋ₃ = .970, Ŋ²₃ = 940, df = 22, ρ = .000; t_PLIB (high CN, experimental group) > t_PLIB (low CN, experimental group) = 24,30 > 6,71; MPLIB (high CN, experimental group)> MPLIB (low CN, experimental group) = 501,04 > 147,17, df = 22, ρ = .000).

Table 3. Descriptive statistics (Means (M.), Standard deviations (SD) and inferential validation of the hypothesis of the study (intergroup variances (f and t) observed)

Applying the Hartley Equal Variance Test gives calculated indices through summary data that help to further validate the hypothesis of the study. Although the Hartley indices of the hypotheses of equal and unequal variances of the observations made in the compared groups (PLIB-IDRLIB-HNCexperimental * PLIB-IARLIB-HNCcontrol) are appreciably close, the fact remains that they are significantly different and that the hypothesis of unequal variances is true (t_Hartley = 8,49, df = 42,41, ρ =.000 and f (4, 49) = 2,587, ρ = .0121). It is the same for the groups in which the level of collective narcissism differs (PLIB-IDRLIB-HNCexperimental * PLIB-IDRLIB-LCNexperimental) (t_Hartley = 11,75, df = 43,83, ρ < .05 and f (5, 49) = 1,131, ρ = .0385). These results provide empirical support for the hypothesis of the study.

The results of this study reveal that the effects of interactions between the control variables, IARLIB and their perpetuation are more positive and significant in non-diagnostic condition than in diagnostic condition. This could be due to the fact that in non-diagnostic condition, these variables had effects going in the same direction as the effect of the moderator variable, unlike the interaction effects observed between these variables in diagnostic condition (IDRLIB). In this condition, only collective narcissism (moderator variable) and group identification had a positive effect on the link assessed. We can conclude that collective narcissism generates positive interaction effects between both IDRLIB and IARLIB and their perpetuation. The effect is strongest when people approve the recourse to LIB. In this case, they are more likely to perpetuate them.

4. Discussion

This study examined the effect of collective narcissism on the link between intragroup disapproval of recourse to LIB and their perpetuation. Concretely, it wanted to know if the belief in the superiority of ingroup over outgroups could push individuals to perpetuate LIB, even if the use of these biases is not accepted by ingroup. The hypothesis tested proposed that in a situation of intragroup disapproval of recourse to LIB, individuals who have a high level of collective narcissism will tend not to conform to group norm and therefore perpetuate these biases. The data collected provides empirical support for this prediction. Likewise, linear regression index analyzes were used to assess and validate the effects of interactions between variables. We conclude that collective narcissism acts positively on the link between intragroup disapproval of recourse to LIB and their perpetuation in both situations of intragroup approval and disapproval of recourse to LIB.

The existing literature indicates that in a situation of intragroup disapproval of recourse to LIB, individuals are less likely to perpetuate these biases (Assilaméhou-Kunz et al., 2019). On the other hand, when ingroup encourages their use, their perpetuation is more marked. This tendency increases with the idealization or not of the group of belonging. Thus, recourse to LIB varies according to the beliefs of individuals and their assessment of the likelihood that these beliefs are or are not shared by other ingroup members (Stangor et al., 2001). Indeed, when shared by a large number of members of this group, the use and perpetuation of LIB is significant. The results of the non-diagnostic condition in the present study support this observation. In contrast, when these beliefs are not shared within ingroup, the use and perpetuation of BIL is less observable.

Contrary to what the literature reports, the results of the present research reveal that intragroup disapproval of using LIB does not prevent their use or their perpetuation in individuals who have an inclination for collective narcissism, that is to say those who invest emotionally in the unrealistic and unequal belief in the exaggerated ingroup’s greatness (Golec de Zavala, 2018). Indeed, in this study, participants with this characteristic used and perpetuated LIB, despite the intragroup disapproval of the use of these biases. Likewise, the data collected reveals that under the condition of intragroup disapproval of using LIB, participants who strongly identify with ingroup used LIB more than those who identify weakly. However, the existing literature affirms that individuals identify more with ingroup when they are placed in a situation of intragroup approval of recourse to BIL (Assilaméhou-Kunz, 2013; Assilaméhou-Kunz et al., 2019; Assilaméhou-Kunz & Testé, 2013b). In the present study, participants identified with their group more in the condition of disapproval than in the condition of approval of recourse to LIB.

The results of this study indicate that participants perpetuate LIB both in situations of intragroup disapproval and approval of the use of these biases. Therefore they do not point in the same direction as Assilaméhou-Kunz et al.’s (2019) observations, which show that the perpetuation of LIB is stronger in a situation of intragroup approval of the use of these biases than in a situation of disapproval. The present research suggests that the difference between these observations can be explained by the inclination of participants for collective narcissism; this characteristic of the personality being likely to push the individuals who possess it not to conform to group norm and thus perpetuate LIB. Thus, collective narcissism had a positive effect on the perpetuation of LIB in both IDRLIB and IARLIB situations. It is suggested that the motivation behind this tendency among participants is the maintenance and protection of the positive ingroup image (Tajfel & Turner, 1979, 1986; Wu et al., 2019).

Individuals with a tendency towards collective narcissism consider ingroup hostility towards outgroups as justified and defensive (Dyduch-Hazar et al., 2019). This personality characteristic is one of the sources of hostility towards outgroups and impacts on intergroup cognitions (Golec de Zavala et al., 2016; Lyons et al., 2010). In other words, it is a robust predictor of prejudice, hostility and retaliation against outgroups. It is associated with a biased perception of intergroup realities (Wu et al., 2019). Therefore, it positively moderates the relationship between approval/disapproval of the use of biases and their perpetuation. This positive moderation is related to the fact that data collected from participants with high collective narcissism indicate that they expressed more intergroup biases than their counterparts with low collective narcissism.

The main contribution of this research is to highlight collective narcissism as a personality characteristic likely to lead individuals to perpetuate LIB in a situation of intragroup disapproval of recourse to those biases. It means that in society in general, individuals who idealize their group will be able to stigmatize (label), stereotype, prejudge and discriminate against members of outgroups through language. In the case of Cameroon, for instance, Anglophones are victims of negative labels emanating from the use of LIB by Francophones. From the results of this study, it is important to note that these implicit language biases are involved in interactions between groups. Therefore, LIB explains the use of asymmetric language by Francophones when describing actions taken by Anglophones. Linguistic intergroup discrimination is rooted in the asymmetric language in which discriminatory expressions are used. Likewise, this research provides knowledge to explain certain behavioral differences observed in the political and sports fields, where biased language is used implicitly. These linguistic biases are catalysts for social crises. Thus, LIB contributes to the intensification of conflict and intergroup tensions in society. As an illustration, we can report the case of the Rwandan genocide where biased languages or speeches (from radio broadcasts) were used to maintain and intensify the crisis between Hutus and Tutsis. Finally, this study notes that the effect of collective narcissism on the link between intragroup disapproval of the use of LIB could also be analyzed within ethnic, political, sexual, religious categories, work teams, football teams, etc. Their particularity is that they are characterized by antagonistic, conflicting or strained relationships.

ACKNOWLEDGEMENTS

The authors would like to thank all the participants for agreeing to take part in the study.

Conflicts of Interest

The authors declare no conflicts of interest regarding the publication of this paper.

Funding

The authors did not receive any institutional funding to conduct this research.