This is an open access article distributed under the terms of the Creative Commons
Attribution License (A Diverse Landscape
Today's Europe is the home of refugees and descendants of first- and second-generation immigrants. In Belgium, and in other parts of north-western Europe, guest workers, many of whom came from Muslim countries (Voas & Fleischmann, 2012), have been recruited since the 1960s to boost local economies (Phalet et al., 2015; Strabac & Listhaug, 2008; Zick et al., 2008). As a result, Muslims are emerging as the largest religious minority in Europe, and Islam as the second most popular religion within the Belgian population (Sealy & Modood, 2020). Despite their long-standing presence, Muslim immigrants encounter considerable levels of discrimination1 (e.g., Bayraklı & Hafez, 2018). In 2019 Belgium, 88.1% of the complaints filed for religious or philosophical discrimination referred to Islam, with mostly Muslim women as the victim (Unia, 2021). At the same time, the presence of Islam and immigration has been the focus of attention of many European political parties. In this context, examples abound of politicians invoking gender equality arguments to advocate their anti-immigration policies. That is, associating Islam with the oppression of women, they imply that women’s rights are a core value of the ingroup that deserves protection (Akkerman, 2015; Farris, 2017). Interestingly, such rhetoric has been predominantly held by right-wing politicians, the same who often failed to support progressive policies including women's rights. As Akkerman (2015, p. 56) writes: "This defense of liberal principles seems to sit rather uneasily with the conservatism that these parties tend to display when formulating their policy agendas in the domain of family relations (…)”. Notably, right and far-right figures such as Marine Le Pen, Geert Wilders, or Matteo Salvini claimed to seek to protect their country from Islam in order to defend women (Akkerman, 2015; Farris, 2017). This collusion between racist views with gender equality ideas was recently coined “femonationalism” by Farris (2017), singling out political actors. However, one question of high interest is whether such rhetoric also prevails within a general European population. In other words, do right-wing anti-egalitarian individuals invoke gender equality when facing the question of Islam?
From a psychological point of view, this co-occurrence of anti-immigration and pro-gender equality discourses may come as a surprise. Indeed, historically, the two types of rhetoric have long tended to be associated with opposing poles of the political spectrum as anti-egalitarian views regarding immigration usually go hand in hand with antiegalitarian views on gender equality. However, the weaponization of egalitarian ideologies for antiegalitarian purposes echoes the concept of “malleable ideology” (Knowles et al., 2009), which allows one to reconcile this apparent incompatibility. Egalitarian ideologies would be exploited by people who wish to maintain the status quo to assert inegalitarian claims, such as legitimizing an insensitivity to racial inequalities on the grounds of colorblindness (Knowles et al., 2009).
Building on the concept of malleable ideologies, the present work examines whether individuals rely on gender equality to differentiate from and express negative views towards Islam. Indeed, while the weaponization of gender equality by politicians has received some attention (Akkerman, 2015; Farris, 2017; Norocel & Pettersson, 2022), to our knowledge, no study has yet looked at the existence of this phenomenon among lay people and even less so by relying on a quantitative approach. Using textual data analysis, we examine whether and how this weaponization of women’s rights emerges among lay people and how it relates to their ideological positions. In view of the normativity of gender equality and the mobilizing potential of this ideology, it is crucial to uncover such a treacherous phenomenon.
Malleable Ideologies
Although the diversity climate in today's Europe may be less welcoming than one would hope (e.g., differential treatment given to Afghan and Ukrainian refugees, see De Coninck, 2023); egalitarian traditions and non-discrimination standards often prevail in Europe and the notion of race superiority remains inadmissible for a majority of people (Bamberg & Verkuyten, 2022; Bratt, 2022; Monteith et al., 2010). In our post-holocaust world where the pressure to appear egalitarian is widespread (Bamberg & Verkuyten, 2022; Plant & Devine, 2009), people will tend to express their racist attitudes only if there is an acceptable reason for doing so, i.e., a reason that can be seen as unrelated to prevailing negative stereotypes and prejudice2 (Crandall & Eshleman, 2003; Monin & Miller, 2001; Snyder, 1979). In this context, weaponizing egalitarian ideologies, such as gender equality, allows individuals to express negative attitudes towards minorities in a socially acceptable way. Indeed, in 2009, Knowles and collaborators proposed that ideologies, widely held to be clear and non-negotiable, would in fact be exploited and interpreted according to one's personal and social motivations. Knowles and collaborators coined the concept of “malleable ideology”. These authors argued that the colorblind ideology, historically promoted as egalitarian (see Wolsko et al., 2000) and aimed at reducing racial discrimination, is in fact frequently endorsed by anti-egalitarian white people who feel threatened by diversity. By appraising this ideology as resting on procedural rather than distributive principles (i.e., as aiming at equality of process rather than equality of outcomes), anti-egalitarian white people find a socially acceptable way to deal with the threat of diversity and the possible demands from minorities (Knowles et al., 2009).
This malleability framework sheds light on a number of other principles mentioned in the literature and that seem to be exploited in similar ways. As a case in point, freedom of speech can be used to justify hateful speech, both by citizens and politicians (Pettersson, 2019; White & Crandall, 2017). Along similar lines, liberalism and freedom may come across as Western or domestic values that one needs to defend against illiberal Islam, thereby serving as exclusory principles (Gustavsson et al., 2016; Verkuyten, 2013). Finally, diversity has sometimes been construed in broader terms so as to include other types of heterogeneity and, as a result, overlook the issue of racial homogeneity and white dominance (Petts, 2020; Unzueta et al., 2012).
In this general context, the notion of “laïcité” (a rough equivalent of secularism) as it is known in France would seem to be particularly relevant. Laïcité traditionally combines freedom of worship, equality of all citizens before the state – irrespectively of one’s belief –, and the absence of State and Church interference in each other's affairs. Interestingly enough, this definition evolved in recent decades to give way to a so-called “excluding” laïcité, which strongly limits religious expression in the public space, most notably for Muslims (Allievi, 2012; Barthélemy & Michelat, 2007; Baubérot, 2012; Flood et al., 2012; Sibertin-Blanc & Boqui-Queni, 2015). Some scholars even noted some form of “exceptionalism” towards Islam (Allievi, 2012; Jacquemain, 2014; Schreiber, 2014) driven by a perception of violent, primitive, and fanatical character (Lyons, 2014; Saïd, 1978). Empirical research by Roebroeck and Guimond (2016) showed that people with a higher level of egalitarianism supported the laïcité principle more than individuals with low levels of egalitarianism. However, when exposed to intergroup threat, low egalitarian participants endorsed laïcité more strongly, suggesting that secularism can be promoted by people with opposing motivations and attitudes (Roebroeck & Guimond, 2018). Building on the work of Knowles et al. (2009), the present study aims at looking specifically at the strategic use of gender equality as a malleable ideology.
Anti-Sexism to Justify Anti-Islam Stances
The strategic use of gender equality against Islam rests on two assumptions. First, there is the construction of the East and Islam as particularly patriarchal, with sexism being an illustration “of the backwardness of non-western cultures” (Akkerman, 2015, p. 40; Saïd, 1978). Second, one finds the construction of Europe and Christianity as exempt from sexism. Although gender equality is indeed a declared priority of European instances (European Union Commission, n.d.), the phenomenon at hand can be characterized by the construction of gender equality as solely or primarily European and Western, and of Europe as comparatively virtuous on this front (Akkerman, 2015; Benelli et al., 2006). In this context, sexual violence in racialized communities is depicted as a reflection of immigrants’ culture, whilst overlooking the cultural aspects of sexual violence when perpetrated by Western men (Volpp, 2001). Gender issues become salient only in the context of immigration, with the condemnation of violence against women in Islamic countries or by immigrant groups (Akkerman, 2015). Presumably, the integration trainings designed to instil the values of gender equality supposedly lacking among immigrants who come from an “uncivilized” East are a clear illustration of this state of affairs (Afshar, 2008; Delphy, 2006; Duits & van Zoonen, 2006; Farris, 2017; Gianettoni & Roux, 2010). To be sure, individuals (and political actors) may well differ in their conception of gender rights or even have only a vague idea of what gender equality or even gender should mean (i.e., conceptualize gender as binary or not). However, they also share the view that gender issues are more advanced in the West and that the West should be the single model for women’s emancipation (Benelli et al., 2006; Delphy, 2006; Duits & van Zoonen, 2006; Farris, 2017; Fernandez, 2009; Gianettoni & Roux, 2010; Howard, 2012).
In contrast to this posture, a number of scholars point out that this construction of Islam as patriarchal and of the West as exempt from sexism (Benelli et al., 2006), is at best false and at worst hypocritical. Lyons (2014) explains that, as early as the time of colonialism, Eastern Christian countries have held an anti-veil discourse by framing it as a form of subjugation of women and used it as a justification for colonialism (Flood et al., 2012; Saïd, 1978; Volpp, 2001) while explicitly opposing advances for women's rights in their own nations. In the same vein, many scholars have argued for a critical and decolonial consideration of Islam (Al-Saji, 2008; Delphy, 2006; Joosub & Ebrahim, 2020).
Nevertheless, studies have shown that Europeans view Muslim immigrants less favourably than they see other immigrants (Heath & Richards, 2019). A study conducted in Germany showed that Muslims come across as a particularly problematic immigrant group, a stance that is associated with a desire to restrict their freedom (Van der Noll, 2014). In Belgium, about one in two citizens would like a more restrictive immigration policy for Muslims (Lafleur & Marfouk, 2017). Among Belgian women, Muslim men trigger feelings of disgust, fear, and anger, and are mentally associated with harassment and sexual violence (Kuppens & Yzerbyt, 2012). This attitude towards Muslims in Europe is also reinforced by a process of racialisation (i.e., the process by which a person, group, or social practice is assimilated to a specific racial identity) of certain religious identities, particularly among Muslims in white-majority countries (Amer, 2020; Dawes, 2021; Garner & Selod, 2015; Meer, 2013).
This racialisation of Islam is fuelled by the media under the banner of the “Muslim problem” in some European countries. Through this process, Islam is perceived as being made for and associated with “brown bodies”, “foreigners” or “other” non-white citizens (Amer, 2020; Dawes, 2021; Meer, 2013). In sum, studies across social psychology and sociology show that many Westerners construe Islam as an illiberal, patriarchal religion that is difficult to reconcile with traditional European values, sometimes called the “Inassimilable Other”, while other religions are not perceived as such (Ali & Sonn, 2017; Chryssochoou & Lyons, 2011; Moss et al., 2019; Van Acker & Vanbeselaere, 2012; Verkuyten & Martinovic, 2012). This can be traced back to colonial views of Muslim communities (Ali-Faisal, 2020; Al-Saji, 2008; Cesari, 2002)
Europeans also perceive Muslim women in essentialist ways and construe them as a homogenous, monolithic group (Ali & Sonn, 2017). Often, Muslim women wearing the headscarf come across as passive victims of submission who act under the influence of their husbands or family and need rescuing (Haritaworn, 2012). Yet, although a large number of scholars have reported that the motives behind the wearing of the headscarf in Europe are manifold (religiosity, humility before God, sense of belonging the Islamic minority, resistance against rejection of Islam; Djelloul, 2013; Fernandez, 2009; Howard, 2012; Krivenko, 2012; Mullally, 2011), and although various Muslim feminist groups are gaining visibility, many Europeans still appraise Islam as being intrinsically at odds with feminism and women’s emancipation.
In Belgian and French politics, the convergence of gender equality and Islam came under the spotlight during debates on the headscarf ban. In France, this ban in public places has been justified by the depiction of this garment as an enemy of laïcité and by it violating the principles of gender equality, of which the French Republic claims to be the guardian. By seeking to protect and liberate young girls from radical Islamism, the law allegedly intends to establish the republican principles of freedom, while overlooking the principles of equality and freedom of worship (Baubérot, 2012). In Belgium, the use of gender arguments to justify the discrimination against Muslim women is also recurrent. In 2020 and 2021, the banning of headscarves in some portions of Belgian higher education was the focus of a heated debate and legal proceedings, with Muslim feminists calling for demonstrations and emphasizing that banning the headscarf was incompatible with women’s emancipation (Cheurfi, 2020; Howard, 2012)
In this context, one may wonder whether people, just like political leaders, mobilize gender equality to justify their negative attitudes towards Islam. To investigate this hypothesis, we surveyed respondents about their perception of the values of the European way of life, the distinction between Islam and Christianity on this matter, and the compatibility between Islam and feminism.
The Present Research
In light of the above considerations, the question of the link between prejudice and a specific use of gender equality comes as an urgent and important issue. Scholars have previously alerted on the strategic use of feminism among politicians using qualitative data. However, to our knowledge, no published research has yet investigated and documented this phenomenon in a lay population. Moreover, using quantitative methods allows us to tackle larger samples and achieve a greater degree of generality. In the present study, we investigate whether participants conceive gender equality as specific to their ingroup and as incompatible with Islam. We also examine whether these views depend on individuals’ ideological position.
To encourage participants to provide us with as much written material as possible, we relied on several strategies. First, we approached participants’ views via three different open-ended questions, each addressing the issue of malleability from a different perspective. We started with broad questions on the values associated with individuals’ culture (What are the values representing European way of life? Are Christianism and Islam different?), before moving to a more precise question pertaining to the articulation of gender equality and Islam. In addition, we relied on news events to make respondents feel more concerned. For example, we introduced the third question with the notion of “Muslim Feminism” because this notion, particularly relevant to our question on Islam and gender equality, appeared repeatedly in the media (e.g., Ali, 2012; Le Priol, 2020) and could thus be thought-provoking. We analysed their answers with the assistance of a text analysis software, enabling us to draw conclusions about the associations existing between certain discourses (words, classes of words) and individual variables (political orientation and justification of the economic system). It should be noted that is difficult to study rhetorical strategies using rating scales because they constrain participants’ freedom in the articulation of their arguments. As for purely qualitative strategies, they are extremely rich but run the risk of subjectivity and impose strong limits in terms of sample size. The present approach lets participants express their views freely while offering a robust quantitative basis for the analysis of their responses.
We formulated several hypotheses. First, when asked about the values representing the European way of life, participants who are less egalitarian, as evidenced by their political orientation and their justification of the economic system, should refer to women's rights more than other participants should. Second, regarding the difference between Christianism and Islam, participants with a higher level of economic system justification, and participants with right-wing political orientation should perceive these religions as more different and should refer to gender-related concepts to justify this perceived difference. Finally, participants with a higher level of economic system justification, and participants with right-wing political orientation will perceive them as more incompatible, explaining this by the fact that Islam is too illiberal, both on issues of treatment of women (notably the veil) and on issues of freedom of thought and belief.
Method
Participants
We recruited 504 Belgians (316 women, 182 men, and 6 “other”, Mage = 45.08, SDage = 15.61) by means of a Facebook ad for a study about “social issues”. Regarding educational status, 64.8% attended higher education, 32.2% attended secondary school and 2.9% attended primary school or another type of education. The average political orientation was 3.06 (standard deviation = 1.46) on a scale from 1 (far left) to 7 (far right). All participants were living in Belgium, with 92.26% of the participants of Belgian nationality, 3.96% of French nationality, 1.79% of Italian nationality and the remaining 1.99% had other nationalities (e.g., Bhutanese, Bulgarian, Chinese, Spanish, Greek, Uruguayan). Among the participants, 56.15% declared themselves agnostic or atheist, 29.96% reported being Catholic and 13.89% selected the “other” option (the details ranged from paganism to spiritualism, to ecumenism, etc.). Participants who indicated that they were Muslim or held a nationality of a country with a majority Muslim population were removed from the dataset.
Procedure
Participants had to complete an online questionnaire lasting about twenty minutes. Upon giving consent, participants answered a number of demographic questions including political orientation and an economic system justification scale. They were then presented with the core of the questionnaire, consisting in three parts: the European lifestyle, the (dis)similarity between Islam and Christian religions, and the (in)compatibility between Islam and feminism. There was also a fourth question related to secularism that we do not examine in this paper. We collected and analysed the data in French and translated the results in English (e.g., the lexicon highlighted by the results of the analyses). This project has received the approval of the Ethics Committee from the UCLouvain Psychological Sciences Research Institute, reference 2019-10.
We are aware that as part of the white and non-Muslim scientific community, we, as authors, run the risk of continuing to use “concepts, methods, and beliefs about normality that are rooted in the WEIRD realities that inform scientific and epistemic imagination” (Adams et al., 2015, p. 214) and to reproduce hegemonic forms of research (Decolonial Psychology Editorial Collective, 2021; Lazard & McAvoy, 2020). This tendency to be influenced by the interpretive and argumentative background of our researchers is present at all stages of the research process. It occurs from the evaluation of the literature (mainly produced by privileged WEIRD individuals), through the choice of methodology and data analysis, to the writing of the manuscript.
Materials
Political Orientation
We measured participants’ political orientation with a scale ranging from 1 (totally left) to 7 (totally right). For textual analyses, we grouped the very few participants (N = 29) with scores 6 (N = 18) and 7 (N = 11) in a single category, resulting in a total of 6 levels of political orientation. For correlation analyses, we relied on the original 7 levels.
Economic System Justification Scale
We measure participants’ level of egalitarianism with the economic system justification (ESJ: Jost & Thompson, 2000). Although previous research on malleability of ideology typically relied on Social Dominance Orientation to examine how anti-egalitarian people mobilize egalitarian ideologies to serve their political goals (see Knowles et al., 2009; Unzueta et al., 2012), we opted for ESJ to avoid floor effects and non-normal distributions often encountered with the SDO scale in French-speaking samples. The ESJ scale comprises 17 items, examples of which are “Most people who don’t get ahead in our society should not blame the system; they have only themselves to blame” and “Economic positions are legitimate reflections of people’s achievements”. We divided participants into decile groups, with the lowest decile referring to the lowest economic system justification scores. Together with political orientation, the measure of economic system justification thus serves as a proxy for ideology (Jost et al., 2009), with political orientation covering a more deliberate and explicit aspect of ideological views in comparison to ESJ.
European Lifestyle
Next, participants read “Six months ago, the European Commission created a new commissioner position with the main task of “Protecting the European way of life. Not everyone agrees on the meaning of the European way of life”. Participants then read “In your opinion, what would be the values attached to the European way of life? (Specify why)” and were to answer by means of a textbox with no word count limit.
(Dis)similarity Between Islam and Christian Religions
Participants then had to assess the (dis)similarity between Islam and Christian religions with the question “In your opinion, are the Muslim and Christian religions rather similar or rather different?”. Responses options ranged from 1 (similar) to 4 (different). Then, they had to explain their answer using a textbox.
(In)compatibility Between Islam and Feminism
Finally, participants read “In the newspapers, some articles mention women claiming to be Muslim feminists. In reaction, some people point out that these are two contradictory terms, that feminism and Islam are totally opposed”. They were then asked to indicate to what extent they thought that Islam and feminism are (in)compatible on a scale ranging from 1 (compatible) to 4 (incompatible). Again, they were asked to explain why with another textbox.
Data Analysis
We analysed participants' answers to the open-ended questions with the IRaMuTeQ software (Loubère & Ratinaud, 2014; Ratinaud, 2009). This statistical tool allows studying textual data by performing lexicometric analyses (Leblanc, 2015). For the present study, we relied on four types of analysis: descriptive analyses, analyses by descending hierarchical classes (DHC), similarity analyses, and specificity analyses. To perform its various analyses, IRaMuTeQ relies on different dictionaries, identifying “full words” (i.e., nouns, adjectives, verbs, and adverbs) and “tool words” (i.e., pronouns, articles, and conjunctions). Main analyses performed by the software focus exclusively on “full words”.
First, the software lemmatizes the “full words”, such that each form of the corpus corresponds to the root of the word and independently of its syntactic category (e.g., the form “rac+” will group together the nouns “racism” and “race”, as well as the adjectives “racist” and “racists”, Heine et al., 2007). Then, the program proceeds to divide the corpus into “elementary context units” (ECU), based on the number of words analysed and the punctuation of the text. This provides an analysis of the frequency of words to highlight the words used most frequently by participants when answering the questions.
Secondly, the similarity analysis proposes a graphical representation of the textual corpus (Degenne & Vergès, 1973). By selecting the most significant links between different lexical forms, this method provides an overview of the thematic structure of the textual corpus (Marchand & Ratinaud, 2012). To reveal the distribution and the association of terms within the text, the software computes the connections as well as distances between the most salient terms of the corpus and produces a graphical representation. Therefore, the more frequently a word appears in the individuals' speeches, the more prominent it is in the graph. Also, the shorter and thicker the line between two words, the closer they are associated in participants' answers.
Thirdly, the DHC consists in a successive fractioning of the text to constitute classes of ECUs based on their lexical content. This descending hierarchical classification allows the identification of the main lexical fields or themes within a textual corpus (Ratinaud & Marchand, 2015). The associated chi-square determines the degree of association between the various words and the classes defined by the software to highlight the most representative forms for a given class. It is then possible to consider these classes in relation to additional variables previously defined by the analyst (i.e., which in the present case comprise gender, age, political orientation, ESJ score, etc.). This method allows processing large text corpora in a statistical and mostly hypothetical-deductive way. Indeed, it is only when the software has created the classes that the researcher is expected to provide an interpretation. In other words, the software performs the initial analyses, thereby reducing the interpretive bias on the part of the researcher and increasing the objectivity of the analyses.
Finally, the specificity analysis provides insights into the distribution of terms in different parts of a text corpus and checks if words are used more frequently (i.e., score above 0) or less frequently (i.e., score below 0) within parts of the corpus (Leblanc, 2015). The positive or negative score created for each term is based on a threshold relative to the expected frequency of occurrence of each form in the corpus, by comparison with the other forms or variables. Practically, we considered here that an association was significant when the value was equal or superior to 3, with a score of 3 corresponding to p-values between 0.001 and 0.009 (a score of 4 corresponding to p-values between 0.0001 and 0.0009, etc.).
Results
In this section, we report the results for each of the three questions addressed to the participants. For the sake of brevity, we only report the results of the analyses pertaining to our hypotheses. Therefore, for each of the questions investigated, the four types of analysis provided by the software will not be systematically presented. The corpora and a dictionary of the coded variables used for the analysis are available on OSF at the following link: https://osf.io/3dy5c/
European Lifestyle Values
For the open question about the values associated with the European way of life, the most frequently and consensually cited values are: freedom (N = 147), rights (N = 109), respect (N = 89), equality (N = 88), social (N = 71), democracy (N = 65) and solidarity (N = 60).
In addition, the specificity analysis shows that the word “woman” is more associated with decile 10 of the ESJ scale (p < .001), as well as with political orientation scores 6 and 7 (p < .001). Looking at the “characteristic text segments” (i.e., participants' answers considered the most representative of the class by the software), we understand that the participants belonging to these ideological categories refer to the word “woman” to refer to women’s rights or gender equality. Among the most representative answers, examples are:
“A Christian religion, respect for work and traditions, respect for laws and private property, respect for women, freedom of speech, freedom, democracy, national identity” (participant 453);
“Human rights, equality of men and women, democracy, laïcité” (participant 336).
In other words, and in line with our hypothesis, participants showing a higher level of ESJ and a marked right-wing political orientation perceived women's rights and women's issues as more integral to the European way of life.
Perceived Similarity Between Muslim and Christian Religions
Participants’ ratings of the (dis)similarity between Islam and Christianism showed that greater perceived dissimilarity between Islam and Christian religions was associated with a more right-wing political orientation (r = .283, p < .0001) and a higher level of ESJ (r = .095, p < .05).
Regarding the text data, the similarity analysis (see Figure 1) shows how participants structured their answers for this question and we can see that the arguments are numerous. Of special interest to us is the position of gender and women. Interestingly, the word “woman” (N = 39)3 falls in a cluster of arguments associated with the notion of “Muslim” (N = 80). In this cluster (top left of the graph), we also find the notions of “place” and “status”, but also “submission” and “inequality”.
Figure 1
Similarity Analysis on the Question “In Your Opinion, Are the Muslim and Christian Religions Rather Similar or Rather Different? Why” (English Translation)
Next, the DHC analysis determined a breakdown of the content of participants' responses to the question of similarities between Christian and Muslim religions. At the end of the DHC iterative process, the software analysed 77.29% of the textual data content for this corpus of responses. Out of the total 546 text segments, 422 segments (or ECUs) could be divided into six classes. Figure 2 shows the dendrogram summarizing the division of the classes (as well as their associated forms), their importance in relation to the corpus of responses (in percentage), and the variables associated with each class. Here, we explore more specifically Class 1 and 2 because these seem to fall more readily within the scope of our hypotheses.
Figure 2
Descendant Hierarchical Classification (DHS) Analysis Performed on the Question “In Your Opinion, Are the Muslim and Christian Religions Rather Similar or Rather Different? Why” (English Translation)
Class 1 (14.93% of the analysed corpus) refers to the place of the woman in the Muslim and Christian religions. The term “woman” (χ2 = 139, p < .0001) is the most representative form, along with the notions of “position” (χ2 = 40, p < .0001), “society” (χ2 = 33, p < .0001), etc. The term “man” (e.g., χ2 = 59, p < .0001) also stands among the most popular forms, with respondents resorting to the comparison between man and woman. In this class, we also find the terms “submission” (χ2 = 29, p < .0001), “right” (χ2 = 23, p < .0001), “patriarchal” (χ2 = 29, p < .0001), “inequality” (χ2 = 23, p < .0001), “status” (χ2 = 17, p < .0001), etc. Looking at participant’s most representative answers, we understand that the respondents discuss the issue of women’s status in each of the two religions. Among the most representative answers, we observe for example:
“Islam invitation to jihad submission of the woman to the man Koranic text considered as divine therefore to be interpreted literally authorization of the slavery of non-Muslims Koranic penal code barbarian prophet conqueror cruel pedophile polygamist and conqueror Christianity” (participant 288);
“The roles the place of the woman etc. are strongly different I have the impression that nowadays the Christian religion has evolved more and accepts that we live it in a way that it does not go against the modern way of life” (participant 60).
Of note, this class is associated with the most right-wing levels of political orientation (i.e., 6 and 7 on the 7-point scale, χ2 = 14, p < .0001). For this question, individuals with the most right-wing scores are the most likely to stress this topic. Interestingly, the specificity analysis reveals that the form “submission” is significantly associated with a right-wing political orientation (again, with the rightmost political orientation: 6 and 7 on the 7-point scale), p < .001. Conversely, the term “patriarchal” is particularly associated with the third decile of the ESJ (p < .001), but also with a far-left political orientation (1 on a scale of 1 to 7, p < .001). The excerpts using this term show that respondents mobilize the term patriarchal to emphasize the fact that both religions are patriarchal, which contrasts with the use of the term “women” used to criticize Islam.
In class 2 (12.09% of the analysed corpus), one finds differences between Muslim and Christian religions in terms of rules and politics: “Islam” (χ2 = 86, p < .0001), “politics” (χ2 = 59, p < .0001), “impose” (χ2 = 37, p < .0001), etc. The typical answers illustrate this point:
“Islam is a law that regulates private social and political life in Muslim countries. In de-Christianised Western countries the state is not religious nor is private or social life” (participant 136).
The representative variables for this class correspond to decile 10 of the ESJ (i.e., the highest decile on the economic system justification scale), as well as right-wing political orientations (pol5, χ2 = 19, p < .0001; pol6.7, χ2 = 11, p < .0001) and the male gender (χ2 = 5, p <.05). Once again, participants who raise the idea of Islam being a more politicized religion or having a normative system that needs to be distinguished from that of Christian religions are the ones who are furthest to the right on the political scale and who show the highest levels of system justification.
Perceived (In)compatibility of Islam and Feminism
First, looking at participants’ quantitative evaluation of (in)compatibility between Islam and feminism, results show a positive correlation between the perception that Islam and feminism are incompatible and the score of justification of the economic system (r = .11, p = .024), and a correlation between the perception that Islam and feminism are incompatible, and right-wing political orientation (r = .282, p <. 0001).
Second, the DHC resulted in the analysis of 87.33% of the textual data collected for this question and produced three different thematic classes (see Figure 3). Class 2 gathers most of the content of participants' answers (with 38.68% of the corpus) and concerns (in)compatibility, containing arguments about the “values” (χ2 = 18, p < .0001) present in each of the concepts and about the “interpretation” (χ2 = 18, p < .0001) that one makes of Islam and feminism and of their principles. Responses range from arguing for the compatibility of feminism and Islam (e.g., “the values of the Muslim religion are for tolerance, therefore feminism is compatible” (participant 293), “considering that Islam is a religion and that any religion is first and foremost an ideology carrying values of tolerance, respect, love, etc. feminism aiming at equality”, (participant 24) to their incompatibility (e.g., “all major religions are based on archaic principles alien to the notion of feminism and are dominated by conservative currents incapable of reforming in this regard”, (participant 408); “One can be a feminist and practice a certain form of Islam, in that sense it can give them a feminist and Muslim identity but the texts of the Koran are quite explicit about the status of women and are incompatible with feminist values”, (participant 52). This class is associated with a centre-left political orientation (pol3 on a scale of 1 to 7; χ2 = 6, p < .0001).
Figure 3
Analysis by Descending Hierarchical Classification of Question 3: “To What Extent They Believe That Islam and Feminism Are (In)compatible? Why” (English Translation)
Class 1 (33.85% of the analysed corpus) points to the place of the Islamic veil (“port du voile” in French; “veil”, χ2 = 62, p < .0001; “wear”, χ2 = 39, p < .0001). Responses raise the possibility that the veil can be worn by choice, and is therefore a reflection of their wearers’ freedom, in line with the values of feminism (e.g., “Feminism is also about being free of one's body and choices, if Muslim women want to wear the veil, what's wrong with that”, (participant 270), “I think that if the woman does not experience her religion as a constraint and she is not obliged to do anything, that it comes from her own choice, there is nothing incompatible. The incompatibility comes from the fact that there is a family obligation to modify her behaviour, like wearing the veil”, participant 291). Associated with this class is the female gender (χ2 = 5, p < .0001) (thus compared to participants who identified as female or other), as well as a left-wing political orientation (pol2, χ2 = 6, p < .0001).
Finally, class 3 (27.47% of the analysed corpus) features comments around the comparison of the place of women with that of men in Islam. Individuals thus refer to notions of equality (“equal”, χ2 = 44, p < .0001), “inferiority” (“inferiority”, χ2 = 44, p < .0001) of women, in relation to men (“man”, “husband”, “male”, “father”, etc.). Most of the responses support the idea that women are placed as inferior or not equal to men within Islam: “the traditional status of women in the various interpretations of Islam, is not equal to that of men” (participant 455); “Islam advocates the superiority of men over women, how can one be a feminist (woman as superior) and a Muslim (woman as inferior)” (participant 289). This type of argument is strongly associated with right-wing respondent profiles (χ2 = 19, p < .0001), but also with the male gender (χ2 = 8, p < .0001) and with the highest ESJ score (decile 10, χ2 = 4, p < .0001).
Finally, the specificity analysis once again reveals that certain gender-related terms are associated with specific political positions. Indeed, the terms “submission” (p < .001) and “wife” (p < .0001) are more used by individuals reporting a political orientation of 6 or 7 (i.e., the most right-wing political orientations on the proposed scale). Conversely, the word “patriarchal” is mentioned more often by respondents placing themselves on the leftmost orientation proposed by the scale (pol1; p < .001).
Discussion
Building on the concept of malleable ideologies (Knowles et al., 2009), we examined how individuals rely on discourses pertaining to gender equality to articulate and justify their point of views on the compatibility between Islam and Europe. We hypothesized that less egalitarian participants seeking to elaborate and rationalize their negative messages about Islam would turn to the ideology of gender equality. They do so in light of the sexist nature that they attach to Islam and the normative nature of gender equality ideology. The data confirm our hypotheses. Across several questions, we found consistent and robust associations between the frequency of gender-related forms and word classes on the one hand and political orientation and system justification ideology on the other.
First, when asked to elaborate on the European way of life, right-wing participants and participants with a high level of ESJ referred to “woman” more than other participants did, even though this question did not focus on women’s rights. Second, when participants assessed the similarity between Muslim and Christian religions, the data show, as expected, that right-wing participants and participants with a high level of ESJ report more dissimilarity. As these respondents elaborated on this issue, the analysis confirmed that these participants were associated with a word class about the status of women (with words such as “submission”, “inequality”, etc.). As for the third question, when we questioned participants about the (in)compatibility between Islam and feminism, those with a more right-wing political orientation and higher levels of ESJ not only perceived more incompatibility, but also used word classes related to woman status in Islam (with words as “inferiority”, “equal”, etc.) as well as more words related to this issue (“submission”, “wife”).
Interestingly, for the two latter questions, the male level of the gender variable also tended to be associated with comments or arguments related to women’s status. It is worth noting, however, that the word “patriarchal,” also clearly related to gender equality, was used by a segment of participants, namely by left-wing participants. In all likelihood, these respondents wanted to highlight the patriarchal nature of religions in general.
The present results are the first to demonstrate, in the general population, that feminism can be a malleable ideology serving a right-wing, possibly Islamophobic, political agenda. They confirm that the ideology of gender equality is not only being used by politicians, in line with the concept of “femonationalism” (Farris, 2017), but also encountered among lay individuals. Following Knowles et al.’s (2009) work on malleable ideologies, the present evidence stresses the fact that gender equality, an ideology rooted in equality and inclusion, may be mobilized to promote exclusionary stances.
Our findings also highlight the critical role of the normative context. Specifically, less egalitarian individuals seize ambiguity in the principles held dear by the population to advance their agenda. Thus, the fact that gender equality is not only cherished by a large portion of the population in various European countries (Akkerman, 2015; Lyons, 2014), but has almost become part of the Zeitgeist, makes this malleable discourse particularly deceptive. It is worth noting that Muslim women and antiracist feminists actively dispute this association between Islam and sexism (Ali & Sonn, 2017; Sager & Mulinari, 2018).
Further, as much as we conducted an analysis at the individual level, one should note that this weaponization of gender equality (and with it, islamophobia) has implications beyond the individual level. Intergroup domination unfolds not only via individual bias and interpersonal relations but also at institutional levels by altering access to opportunities, goods, and services. In this respect, the perception that a culture is a serious threat to our ingroup values can lead us to favouring stricter integration policies (i.e., assimilation) that are deleterious for minority populations (Downie et al., 2006; Verkuyten, 2010), or to tolerate or even commit immoral actions (Reicher et al., 2008). In other words, the present research is part of a wider effort to study the expression of negative views towards minority groups, both in its individual functioning (i.e., impression management) and in its socio-cultural dimension (normative context, legitimizing ideologies, dynamics of power).
In parallel to the appropriation of gender equality themes described here, several scholars have denounced the emergence of a weaponization of LGBTQIA+ issues (termed “homonationalism”) in the UK, Ireland, the USA, and Israel (Ammaturo, 2015; Luibhéid, 2018; Puar, 2013). This opportunistic advocacy of gay rights and of progressive views of gender identity (i.e., supporting gays as victims of Islamic bigotry) stands in contrast with the exploitation of women's rights described here. While a small number of far-right parties in Europe seem to emphasize gay rights for their anti-Islam agendas, this strategy is much less common (Akkerman, 2015). Yet, in light of the fact that the issue of LGBTQIA+ inclusion gains attention, this is an important question to consider for future research.
Of course, the present research does not come without limitations. One may argue that the findings lack generalizability given the sampling method. Although this is a fair concern, one should stress that the ambition of the present endeavour is to identify a phenomenon that operates in a given context with a given population, i.e., one should thus see the present research as a “proof of concept”. To be sure, one interesting avenue for future research may indeed be to examine whether other respondents, in particular French, show similar patterns of response. Overall, it is important to keep in mind that statistical analyses of verbal material using software such as IRaMuTeQ are limited in terms of understanding and consideration of speech content (e.g., being able to fully grasp nuances in the subjects’ narratives, such as negations, irony, figures of speech, etc.). They are very useful tools for highlighting trends, but other, more comprehensive, approaches allow complementing the approach. Perhaps the most obvious limitation however is that the analyses building on DHC are subject to potential interpretation biases inherent to the researchers. This is because it is necessary to connect the different words included in the word class in order to assign meaning. Similarly, the words considered by researchers for specificity analyses can suffer from the same bias. Having said this, the complementarity of the analyses, thanks to the specificity analyses, the similarity analyses, but also to the examination of the closed questions, certainly contribute to minimizing this problem. As a set, these various tools offer the possibility to interpret the corpus in a more comprehensive and indeed less subjective way.
Second, we built on existing social categories that are generic (i.e., Muslims) in light of their widespread use and understanding and in order to encourage participants to generate responses with a large number of words. For example, we referred to the two major religions of Islam and Christianity without accounting for their respective heterogeneity. It is of course possible that a number of participants would distinguish between more moderate and more radical fringes within these groups. Future work may want to examine this possibility. Moreover, such reliance on generic social categories could induce a reified vision of the concerned groups (i.e., Muslims), that is, lead participants to believe that the latter are a homogeneous category, while the ways in which Islam is experienced are diverse (e.g., Joosub & Ebrahim, 2020).
Conclusion
The present study addresses the important question of how lay respondents may rely on viewpoints that would seem alien to their “expected” position in an attempt to justify exclusionary position with respect to minorities. We focused specifically on stances that are associated with cultural values seen as threatening in the eyes of portions of the majority group. Our study is but a first attempt at approaching this important phenomenon insofar as it concerns lay respondents and not the usual suspects such as politicians and opinion leaders. Despite their limitations, the use of more qualitative tools such as the ones mobilized in the present effort opens fascinating avenues for research. Intriguing as they are, our findings call for additional work that rests on such complementary methods as the reliance on archival data or even the use of experimental approaches. This is a clear agenda for our future research on these aspects.