Giulia Evolvi. Information, Communication, and Society. Volume 22, Issue 3. 2019.
Twitter holds democratic potential in eliciting social change and helping liminal publics express their viewpoints through storytelling (Papacharissi, 2014). However, Twitter can be used also to spread hate speech against minorities (Graham, 2016). Indeed, Twitter permits the formation of opinion-based groups that do not necessarily articulate their identities in a way that leads to constructive confrontations. On the contrary, these groups might refuse to acknowledge the claims of minorities, which are represented as ‘other’ to social and national identities.
This paper shows that group polarization on Twitter can be explained through Chantal Mouffe’s (2013) distinction of agonistic and antagonistic politics. The argument is discussed with reference to an empirical example: anti-Muslim discourses on Twitter in relation to the British referendum. On 23 June 2016, UK citizens decided, by means of a referendum, to leave the European Union (EU), an event which is commonly referred to as ‘Brexit’. In the aftermath of the referendum, a number of racist episodes were recorded throughout the country, with a spike in anti-Muslim attacks. The decision to leave the EU is not directly connected with Islam, as it will actually prevent citizens of European countries with a predominantly Christian tradition migrating to the UK. However, Islamophobia became particularly visible during the referendum debate, being viewed as symptomatic of social anxieties about the foreign ‘other’ and of a xenophobic need to define British and Western identities in non-inclusive terms. By undertaking a qualitative analysis of tweets about Islam sent in the aftermath of the referendum, this paper shows how Islamophobic tweets framed group identity not only in terms of religion, but also as being connected to ethnicity, politics, and gender. Anti-Muslim tweets show how Mouffe’s (2013) model of antagonistic politics helps problematize this type of group polarization on Twitter, which fails to create constructive agonistic confrontations.
Twitter: Identity, Homophily, Emotions
Twitter was founded in 2006 as a microblogging platform for real-time interactions. Twitter users send ‘tweets’: messages with a maximum of 140 characters that can include links, videos, and pictures. They can follow pages, quote other users (via ‘@username’), and retweet existing Tweets. The use of hashtags (‘#’) allows certain concepts to be emphasized as keywords and facilitates finding tweets about the same topic, thus spontaneously creating networked discourses and drawing mass attention to specific trends (Naaman, Becker, & Gravano, 2011; Wilkinson & Thelwall, 2012). Not only do hashtags have the potential to create collective conversations in times of crisis, conflicts, and controversies, they also mark and declare identities in distinction to other groups and opinions (Giglietto & Lee, 2017).
Identities on Twitter are often articulated around shared opinions. The term ‘opinion-based groups’ refers to identities that are based on social opinion in a way that entails being for or against a specific issue within broad social categories (Bliuc, McGarty, Reynolds, & Muntele, 2007). In analysing conflicting discourses about the ‘Cronulla riots’ in Australia, Bliuc, McGarty, Hartley, and Hendres (2012) assert that opinion-based groups use rhetorical stratagems to debate and manipulate the terms of national identity and claim dominant status in society. The groups involved in the riots shared common anti-Muslim feelings, but the surrounding debates were framed not only in terms of religion, but also in broader social categories of race, ethnicity, and socio-political boundaries. Opinion-based groups can thus facilitate collective actions as they usually form around key social issues (McGarty, Bliuc, Thomas, & Bongiorno, 2009; Thomas, Mavor, & McGarty, 2012). McGarty, Thomas, Lala, Smith, and Bliuc (2014) observe that the formation of opinion-based groups in relation to collective action movements is often witnessed on Twitter.
While they help defining opinion-based social identities, media platforms such as Twitter do not necessarily facilitate exchanges between groups. As for the results of Gruzd and Roy’s (2014) study on Twitter and the Canadian Federal Elections, it is more likely that Twitter use reinforces people’s political viewpoints rather than shifting their allegiances. In this respect, Twitter can lead to political polarization, a phenomenon that can be understood in terms of homophily: ‘the principle that a contact between similar people occurs at a higher rate than among dissimilar people’ (McPherson, Smith-Lovin, & Cook, 2001, p. 416). Twitter’s users are exposed to a high number of public discussions, but the platform’s structure facilitates conversations with like-minded people rather than improving dialogue between different groups. This happens because Twitter ‘is insufficient for reasoned discourse and debate, instead privileging haste and emotion’ (Yardi & boyd, 2010, p. 325).
Twitter’s homophily is often emotion-based (Colleoni, Rozza, & Arvidsson, 2014; Himelboim et al., 2016), because emotions are considered to be predictors of social and political actions for opinion-based groups (Thomas, McGarty, & Mavor, 2009a, 2009b). Building on the concept of homophily and in-group and out-group affiliation, Papacharissi (2016) describes Twitter as a liminal space that enables users to produce affective news streams, creating ‘networked publics that are mobilized and connected, identified, and potentially disconnected through expressions of sentiment’ (p. 5).
As Yardi and boyd (2010) comment, notwithstanding Twitter’s tendency to facilitate homophily, there has not been sufficient analysis of its group polarization, extremism, and hate speech. This paper shows that these phenomena can be further explored through Chantal Mouffe’s (2013) concept of agonistic and antagonistic politics, which help explaining how groups articulate their identity on Twitter.
Twitter: Agonism or Antagonism?
According to Chantal Mouffe (2013), identity articulation is always relational, as it entails ‘the constitution of a “we” which requires at its very condition of possibility the demarcation of a “they”’ (p. 5). In contemporary societies, certain subjects are constructed as subordinate in power relationships. This tendency was allegedly visible in Margaret Thatcher’s strategy to win over a part of the working class by identifying certain groups, such as feminists and immigrants, as scapegoats for social issues (Carpentier & Cammaerts, 2006). In the Brexit debate, and in the Twitter discourses that accompanied it, a similar strategy emerged through the identification of European and non-European migrants as a threat to the British culture and economy.
Mouffe addresses inequalities by promoting the acknowledgment of differences, expressed not only through class belonging, but also through characteristics such as religious identification. Rejecting the Habermasian idea of homogeneous consensus, Mouffe asserts that conflict is central to democracy. Indeed, social differences can lead not only to negative ‘antagonism,’ where different groups lack equality of participation, but also to positive ‘agonism,’ where hegemony is characterized by a plurality of actors. According to Mouffe, social and political groups should aspire to reach an agonistic model that promotes a democracy based on ‘consensual conflict’ and grants equal participation in agonistic confrontations ‘between adversaries who recognize the legitimacy of the demands of their opponents’ (2013, p. 138).
Agonistic democracy has been used as a framework to study political discourses and their affective dimension, as in the case of feminist blogs (Shaw, 2012). However, Mouffe expresses reserve about the potential of new media—and the Internet in particular—in encouraging the agonistic model, because media practices tend to reinforce pre-existing ideas rather than eliciting confrontations with antithetical positions (Carpentier & Cammaerts, 2006). Homophily would thus hinder the insurgence of agonistic confrontations, facilitating instead the creation of polarized groups that reinforce their identities by selective exposure to like-minded users.
Therefore, this article proposes to understand group polarization on Twitter not as agonistic democracy, but rather as antagonist confrontations that exclude the claims of part of the population. The examples of Islamophobia can meaningfully explain this model because Twitter has been identified as a particularly fertile venue for creating anti-Muslim discourses (Magdy, Darwish, & Abokhodair, 2015): The association Tell Mama (Measuring Anti-Muslim Attacks) reported that the majority of anti-Muslim acts in 2015 in the UK were expressed online, with more than half (57%) taking place on Twitter (Littler & Feldman, 2015). Tell Mama, commenting on data from the UK Government Home Office, also reported a spike in hate crimes against Muslims after the British referendum, which the association considers as having lasting impacts on faith-based and racially aggravated crimes (2016). While Islamophobia on Twitter is constantly increasing and could reach the same symbolic violence of street-level aggressions, anti-Islam discourses remain under-researched (Awan, 2014). This article looks at tweets sent within the context of hate crime surge after the British referendum as an example of how antagonistic discourses can be emotionally articulated on Twitter to negotiate terms of group identities.
Antagonism on Twitter: Post-Brexit Islamophobia
The context of Brexit helps in understanding antagonistic dynamics because the campaigns to leave the EU were often centred on questions of identity and exclusion of minority groups, such as Muslims. This reflects a European trend, where Islam is often considered as problematic by virtue of its alleged incompatibility with secular and Christian values (Arfi, 2010; Betz & Meret, 2009; Roy, 2013; Savage, 2004). Such anxieties are closely connected with the growth of Islam in Europe: In the UK, for example, the Muslim population currently represents the third biggest belief group after Christians and non-believers (Office for National Statistics).
Muslims’ identities often occupy an ambiguous position in British society, thus making Islamophobia an interesting case to study group polarization. Britain, indeed, witnessed a shift in the perception of Muslim migrant groups. While the first waves of immigration in the 1950s and 1960s resulted in a politics of integration and multiculturalism mainly focused on race and ethnicity (Abbas, 2007; Marranci, 2012), in the last two decades focus shifted from the politics of colour and race to those of religion (Ahmad & Evergeti, 2010; Hopkins & Kahani-Hopkins, 2006). This has resulted in a ‘racialization of Islam’ that focuses on the religiosity of minority groups and compels British Muslim to articulate their identities in religious terms rather than in terms of ethnic belonging. This shift has largely coincided with an acknowledgment of what has been defined as ‘the failure of multiculturalism’ (Werbner, 2009), which sees British Muslims remaining fairly unassimilated into mainstream culture and facing ‘some of the highest levels of social deprivation in Britain’ (Nagel & Staeheli, 2011, p. 440).
The ‘racialization’ of Islam has been worsened by issues such as the Rushdie affair in 1996 or the terrorist attacks on 9/11, which many people claimed to demonstrate the inability of Muslims to accept Western values (Marranci, 2012), and resulted in negative media representations of Muslims (Poole & Richardson, 2006). This contributed to the spread of Islamophobia (Hutchison, Lubna, Goncalves-Portelinha, Kamali, & Khan, 2015; Liepyte & McAloney-Kocaman, 2015; Mcdonald, 2011), a phenomenon defined by the Runnymede Trust as ‘unfounded hostility towards Islam’ (Runnymede Trust, 1997, p. 4). Often considered to be a form of ‘cultural’ racism on account of the racialization of Islam and the framing of Muslims as culturally unable to accept Western values, Islamophobia has been compared to British anti-Semitism between the two world wars (Linehan, 2012; Meer & Noorani, 2008). Islamophobia, connected at European level with right-wing populism (Zick, Küpper & Hövermann, 2011), has been fomented in the UK by nationalistic groups (Allen, 2014; Kassimeris & Jackson, 2015; Wood & Finlay, 2008), which campaigned in favour of leaving the EU. Because of Muslims being portrayed as ‘other’ to national identity and represented in religious, ethnic, and cultural terms, Islamophobia became a way to discuss British identities in the context of the EU referendum. Therefore, the analysis of Tweets connected with Brexit helps in understanding how opinion-based groups can antagonistically build their identity claims at the expense of minorities.
#Brexit: Exploring Twitter Hashtags
In the aftermath of Brexit, a number of hashtags connected with the referendum started trending on Twitter, associated with both the ‘Leave’ and ‘Remain’ positions. Howard and Kollanyi (2016) identified in #Brexit the first most popular hashtag that was used by supporters of the Leave position. Because this research aims primarily at providing an empirical example for a theoretical model rather than exploring and comparing different keywords, the single hashtag #Brexit—relevant because of its popularity—has been analysed to explore Islamophobia associated with the Leave campaign.
A dataset of 2005 tweets written by 1789 users and mentioning the hashtag #Brexit in relation to the keywords ‘Islam’ and ‘Muslim’ was collected. When referring to ‘the tweets,’ the article talks about this specific dataset. The tweets were all written in English, but in order to capture global discourses about Brexit the search was not limited to the UK. The dataset includes tweets written between 23 June 2016—the date of the referendum—and 30 June. This week witnessed a surge in hate crimes: quoting data from the National Police Chief’s Council, the UK Government Home Office reports 3076 hate crimes and incidents in England, Wales, and Northern Ireland in the second half of June 2016, a 42% increase in respect to the previous year (Corcoran & Smith, 2016).
The analysis of tweets written the week after the referendum allows to capture the majority of Brexit-related discourses, because the attention for a topic on Twitter tends to trend and decrease in a short timeframe. Indeed, Magdy et al.’s (2015) study on Twitter responses to the Paris terror attacks in 2015 focuses on a span of 50 hours, and Giglietto and Lee’s (2017) article on the #JeNeSuisPasCharlie hashtag identifies three phases of discursive production in a timeframe of four days; Jackson and Welles (2016) analyse the #Ferguson hashtag by taking into account the week of tweets after the murder of Michael Brown, considering it as the timespan in which Twitter catalysed the national response to the event.
As Howard and Kollanyi (2016) note, the hashtag Brexit, while mostly employed by Leave supporters, referred also to neutral or Remain positions. This can produce interesting data on how opinion-based groups use the same hashtag to articulate competing discourses and to enter into conversation with each other. Therefore, the dataset was manually coded through a sentiment analysis in order to capture the meanings attributed to the hashtag. Usually applied to market research (Philander & Zhong, 2016), sentiment analysis has also been increasingly employed for measuring political attitudes and identity construction processes on social networks (Ceron, Curini, Iacus, & Porro, 2014), with specific focus on emotions (Dang-Xuan, Stieglitz, Wladarsch, & Neuberger, 2013). As Table 1 shows, tweets were coded into three categories—‘positive,’ ‘negative,’ and ‘neutral’—according to the way they frame Islam. Because the tweets tended to very clearly state anti-Islam positions, which were dominant in the dataset, the analysis did not raise concerns in terms of reliability. Furthermore, human coding has the advantage of capturing certain nuances that computer coding would not allow, thus being more accurate than machine-assisted coding (Murthy, 2015).
| Table 1. Number of tweets that contain positive, neutral, and negative framing of Islam and the percentage they represent in relation to the total (approximated to the closest whole number). | ||||
| Keywords | Positive | Neutral | Negative | Total |
| #Brexit+ Islam | 78 (10%) | 158 (20%) | 553 (70%) | 789 (100%) |
| #Brexit + Muslim | 284 (24%) | 156 (13%) | 776 (63%) | 1216 (100%) |
| Total | 362 (18%) | 314 (16%) | 1329 (66%) | 2005 (100%) |
After the sentiment analysis, an in-depth discursive analysis was performed to capture the most popular topics and feelings within the dataset, as well as the nuances of identity-related discourses. The analysis focused on ‘language, image, tone, and other discursive strategies […] with an eye to how this choice of language, in combination with specific affective or informational images, collectively constructed meaning’ (Jackson & Welles, 2016, p. 404). The discursive analysis was performed by manually coding the words more often used in the dataset in relation to negative, neutral, and positive frames of Islam. Table 2 shows the number of tweets that contained each keyword, both in the text, in the form of a hashtag, or in a link. Each keyword encompasses related terms in the same semantic sphere: for example, ‘migrant’ includes tweets mentioning ‘migration’ or ‘immigrants.’ This keyword analysis is not quantitative in scope, but rather seeks to understand inter-group relations by exploring rhetoric and discourse, which are direct reflections of social issues such as immigration and multiculturalism (Bliuc et al., 2012).
| Table 2. Number of recurrent keywords used in tweets that contain negative, neutral, and positive framing of Islam. | |||||
| Keywords | Positive | Neutral | Negative | Total | % of the Total Dataset |
| Migration | 43 | 39 | 202 | 284 | 14.16 |
| Donald Trump | 28 | 25 | 204 | 257 | 12.81 |
| Refugee | 14 | 4 | 106 | 124 | 6.18 |
| Sadiq Khan | 10 | 20 | 82 | 112 | 5.58 |
| Terror | 9 | 2 | 89 | 100 | 4.98 |
| Barack Obama | 2 | 2 | 83 | 87 | 4.33 |
| Racism | 53 | 9 | 24 | 86 | 4.28 |
| Rape | 1 | 0 | 74 | 75 | 3.72 |
| Global | 3 | 5 | 64 | 72 | 3.59 |
| Women | 25 | 3 | 43 | 71 | 3.54 |
| Radical | 4 | 0 | 65 | 69 | 3.44 |
| Invasion | 0 | 1 | 59 | 60 | 2.99 |
| Hillary Clinton | 0 | 3 | 54 | 57 | 2.84 |
| David Cameron | 7 | 13 | 31 | 51 | 2.54 |
| Angela Merkel | 0 | 1 | 49 | 50 | 2.49 |
| Girl | 10 | 0 | 25 | 35 | 1.74 |
| Nigel Farage | 5 | 3 | 20 | 28 | 1.39 |
| LGBTQ | 5 | 2 | 19 | 26 | 1.29 |
This qualitative methodological approach considers Twitter as a major collector of data and links connected within a network of users (Chiluwa & Ifukor, 2015). Manual coding makes it possible to exclude automated scripts and to study Twitter-related contents, such as YouTube videos, Internet memes, and blog posts, which have been analysed in conjunction with the 140-character texts. The dataset is limited to specific discourses, being based on a single hashtag and taking into account tweets that were most popular according to algorithms produced by Twitter Search API. Therefore, the dataset does not claim to be comprehensive of all discourses created around Brexit or to include ideologies that reach the whole British population, but provides an example of inter-group relations that can be used as a model for further research about agonistic and antagonistic politics on Twitter. The way debates were framed in terms of identity proves how Brexit-related Islamophobia leads to a group polarization that can be explained through Mouffe’s model of antagonistic spaces.
Discussing Islam on Twitter: Ethnicity, Politics, and Gender
The analysis shows that Brexit played a role in fomenting Islamophobic discourses in an antagonistic way, as suggested by the aforementioned reporting of anti-Muslim episodes in the aftermath of the referendum. As Table 1 shows, the majority of tweets—1328—frame Islam negatively, while 362 tweets frame Islam in a positive manner. A total of 315 tweets are neutral commentaries or retweeting news, and are thus framed neither negatively nor positively. Because the hashtag #Brexit was used mainly by supporters of the Leave position, this overview confirms that the choice of leaving the EU has been framed also in terms of excluding Islam from processes of national identity articulation.
Tweets in the dataset often contain multiple hashtags and links to external material, including Facebook content, YouTube, or other shared video websites, blogs, news sources, memes, or images. Furthermore, 211 tweets quote other tweets, and 173 are responses to other users. Some users appear up to 10 times in the dataset, suggesting that they frequently tweeted about Islam and Brexit during the period in question and were retweeted by other users, thereby emerging as opinion leaders in this specific context. Some tweets use the hashtags #Islam or #Brexit in a decontextualized manner: for example, the hashtag #Brexit, a trending topic in the analysed timeframe, is sometimes employed to give relevance to discourses about Islam rather than talking about the referendum. This use of hashtags suggests connections between various political and social discourses that Twitter users viewed as either related or as giving legitimacy to one another. Therefore, according to Bliuc et al.’s (2012) analysis, Islamophobic tweets cannot be simplistically reduced to an inter-group conflict between British non-Muslims and Muslims, but they touch upon broader socio-political categories. In this case, there are three predominant themes that emerged in relation to anti-Muslim feelings: ethnicity, politics, and gender. A reflection on tweets that framed Islam in a positive way follows the analysis of these three themes.
Islamophobia and Ethnicity
Islamophobic tweets frame Muslims as non-British foreigners that are ‘different’ from white British. The Muslim presence in the UK is described in terms of nationalism and cultural clashes due to Islam’s supposed rejection of secular institutions and Christian values, thus existing only in a ‘radical’ form (Table 2). This portrayal of Islam is closely connected to ethnicity because it equals Muslims with Middle-Eastern immigrants; the hashtag #Islamexit, for example, was often employed to imply that Muslims, as ‘foreign others,’ need to be expelled from the country.
Islamophobic tweets perpetuate symbolic violence against Muslims through derogatory terms and disparaging images, probably facilitated by the possibility of remaining anonymous on Twitter. An example of Islamophobic narratives can be found in an image, frequently retweeted and commented upon, which shows a map of continental Europe shaped as a Muslim man praying, and states that after Brexit the EU will have to deal with a Muslim problem that supposedly no longer interests the UK.
This use of visual and textual material connects Muslim migrants to terrorism, as illustrated by references to terms such as ‘migration,’ ‘refugee,’ ‘terror,’ as well as ‘invasion’ (Table 2). Tweets describe migrants coming from Africa and the Middle East as Muslims who enter Europe with the aim of spreading ISIS and terrorism. These discourses display anger or fear by commenting on news of terrorist attacks committed by Muslims in Europe. External content which is brought in to support Islamophobic positions is often either produced by anti-Islam actors or forces a link between unrelated facts: for example, instances of violence for which ISIS has claimed responsibility outside the UK are recounted in conjunction with refugees in order to call for the expulsion of British Muslims from the country. While many tweets claim that social networks are vital in unveiling truths about migration and terrorism that the mainstream media allegedly suppress, their narratives are often factually wrong because they confuse migrants with European-born Muslims and do not acknowledge the diverse religious backgrounds of asylum seekers. These stereotypical and incorrect narratives discuss the terms of national identity by excluding ethnically diverse populations.
Such attitudes are explicitly anti-Muslim and closely connected to the question of ethnicity, but often reject any association with racism. Even if some tweets perpetuate a white supremacist rhetoric that connotes Muslims as different from ‘white British,’ their symbolic violence is explicitly directed at religious belonging more than ethnic belonging, echoing the aforementioned ‘racialization’ of Islam. This suggests that Islamophobia carries different nuances from other forms of discrimination and racism; because Islam is framed as ‘ideology’ that can be chosen or rejected rather than as a race or ethnicity, Muslims are blamed for willingly adhering to a set of values that are fundamentally opposed to Western culture. In this way, ethnicity plays a central role to define inter-group identities but, at the same time, is often not explicit in Twitter discourses. Discourses on ethnicity are connected with two interrelated topics: politics and gender.
Islamophobia and Politics
Tweets that connote Islam as a religion entangled with migration and terrorism often contain references to political parties and emerging political identities. As Table 2 shows, rather than being limited to the UK, these references discuss worldwide politics, especially in relation to the United States and Germany. Indeed, the then United States presidential candidate Donald Trump appears more prominently in the dataset than British politicians, such as David Cameron and Nigel Farage.
Barack Obama and Hillary Clinton, at the time president and presidential candidate of the United States, as well as German chancellor Angela Merkel, are often negatively alluded to as supporting the Remain position. These politicians are blamed for being unable to stop Muslim migration or even for supporting it. Barack Obama, for example, is often mentioned as a Muslim who aims to create a powerful Muslim presence in the West: a frequently tweeted article from American conservative news website Freedom Daily uses derogatory language to frame Brexit as an action against Islam and Barack Obama. The British politician named most often is Sadiq Khan, the first Muslim mayor of London; usually referred to as ‘the Muslim mayor,’ Khan is often accused of supporting the Remain position as a means of spreading Sharia law in the UK.
The negative attitude to certain political leaders results in support for anti-Muslim and anti-immigration politicians, in particular Donald Trump. Tweets tend to describe Brexit employing a language that resonates with Trump’s electoral campaign, for example, advocating ‘Making Britain great again.’ These tweets perpetuate a notion of nationalism and the superiority of Western culture that goes against the principle of globalism. While ‘global’ and ‘globalism’ are not negatively connoted words per se, they are often found in the dataset framed in negative terms (Table 2); indeed, tweets assert that globalism is hindered by the violent character of Islam and prevents Western countries from preserving their identity, thus necessitating the step advocated by Trump of ‘taking the country back.’
Such discourses suggest the existence of opinion-based groups that form around shared political identities. They show dissatisfaction with global political elites and mainstream media, which are accused of being sympathetic to certain political ideologies. Nationalist and xenophobic tweets identify Brexit as an opportunity for citizens to gain agency over the country’s decisions. The fact that characters such as Donald Trump are favoured as potentially disrupting an unsatisfying political status quo is symptomatic of a global attitude that goes beyond Brexit. Indeed, the British referendum is symbolically cited by British and non-British citizens alike as the victory of a political mentality often referred to as ‘alt-right’ (short for alternative right), a type of conservativism that rejects mainstream ideologies; these tweets often blame left-wing parties, dubbed the ‘regressive left,’ for favouring an alleged reactionary Islamic mentality. Negative frames of Islam are connoted not only in terms of preserving national identities, but also of national security, in particular with relation to women.
Islamophobia and Gender
The analysis reveals a gender dimension in tweets discussing the threat potentially posed by Islam for women, as shown by the recurrence of the terms ‘women’ and ‘girls’ (Table 2). Twitter discourses often frame Islam as a misogynistic religion that rejects the Western principle of gender equality, thus understanding Western identities also in terms of shared gender norms. The sexuality of Muslim men is described as uncontrollable and exaggerated: firstly, they allegedly take multiple wives and father numerous children, thus contributing to the spreading of Islam in the West; secondly, because they cannot ostensibly control their ‘base’ instincts, they sexually assault non-Muslim women. The hashtag #rapefugee and references to ‘rape’ (Table 2) suggest that sexual violence is seen as a primary problem of Islam and Muslim immigration. Such narratives sometimes also accuse Muslims of paedophilia, thus framing Islam as a threat to children, too. The sexuality of Muslim men is further problematized as being directed against LGBTQ people (Table 2): the terrorist attack in Orlando on 12 June 2016, committed by a self-identified Muslim in an LGBTQ club, is used to support the claim that Muslims are unable to accept Western gender norms.
The analysis shows that gender discourses are based also on a different understanding of the women’s body: the Muslim veil is problematized as incompatible with the Western culture, while secular-based freedom is framed in terms of freely exposing the female body. Some tweets comment upon the decision of London mayor Sadiq Kahn to ban body-shaming ads from the London Underground as an example of Muslims’ unwillingness to accept European dressing codes. In demanding Khan’s resignation after Brexit, such tweets implicitly state an understanding of women’s attractiveness that is allegedly threatened by the presence of different female bodies.
These discourses rarely appeal to a feminist public or talk about women’s agency. Some Twitter users do identify as women, but tweets are rarely written from a personal perspective and tend to talk about women as the subjects of conversations. As a result, these discourses create a dichotomy between Muslim women, who are seen as part of an alien culture and as the submitting subjects of a misogynistic religion, and non-Muslim women, who need to be protected from Muslim men. Such attitude unveils a willingness to socially and culturally control both Muslim and non-Muslim women within a monolithic and Western understanding of femininity that heavily informs the identity articulation of anti-Muslim groups. This approach hinders self-representations of female users, quite unlike tweets that frame Islam positively and provide more nuanced narratives.
Muslim and Pro-Islam Narratives
The hashtag #Brexit is used, in a minority of cases, to counteract Islamophobic discourses. Tweets that frame Islam positively include users who self-identify as Muslims and non-Muslim users who condemn the rampant post-referendum Islamophobia. The two categories of tweets could overlap, as not all users openly disclose their religious affiliation. These tweets are different from anti-Islam tweets because they tend to employ a less aggressive language. Also, as Table 2 shows, such discourses often touch upon the issue of ‘racism,’ connecting Islamophobia with the general expression of xenophobia against other ethnic and racial minority groups; in this way, they unveil a different understanding of ‘Islamophobia’ that was framed as unrelated to racism by anti-Islam users. Even if they constitute a smaller percentage of the dataset, these tweets tend to offer more nuanced and personalized accounts and rely less on the circulation of popular hashtags and repetition of keywords.
The analysis shows that many tweets describe the experiences of Muslims in the UK and their feelings about the post-Brexit Islamophobia from a personal angle. Indeed, tweets that frame Islam positively tend to use the word ‘Muslim’ more than ‘Islam,’ (Table 1), given that they probably underscore the experience of specific Muslims rather than talking about Islam in general. These tweets tend to have a gender dimension and frequently mention Muslim women (Table 2), who often feature as the target of Islamophobic abuse due to the headscarf worn as a visual symbol of religious affiliation. An often-shared news story published in the online edition of the daily newspaper The Independent, reporting that a Welsh Muslim woman campaigning for the Remain position was told to ‘pack and go home,’ shows that even non-veiled women with British citizenship have had to endure hate speech episodes.
Tweets describing Islam positively tend to disconnect Islamophobia from politics, sometimes asserting that the referendum does not necessarily need to be understood in terms of racism and intolerance because some Muslims did actually vote in favour of Brexit. Such tweets usually do not attack specific groups or political ideologies, as is the case with anti-Islam tweets, but rather advocate solidarity for Muslims and minority groups in general. In doing so, they tend to create statements more than seeking conversations with other users.
Discussion and Conclusion: Twitter’s Emotional Antagonism
Twitter mirrors society at large by reproducing discourses articulated in different venues, but also displays some unique characteristics in terms of identity articulation, inter-group relations, and emotions. This paper has shown that Mouffe’s model of agonistic and antagonistic politics helps to explain group polarization on Twitter. As the example suggests, post-Brexit tweets about Islam are antagonistic because they often rely on identities that entail a sharp distinction between ‘us’ and ‘they’ at the expense of Muslims and, in so doing, seldom recognize the claims of other actors. Three elements in particular emerge from my analysis.
Firstly, the way Twitter helps articulate identities in opinion-based groups tends to be antagonistic. Tweets allow individual and collective opinions to be stated by means of keywords and hashtags, facilitating the creation of groups based on agreement or disagreement with a specific issue. As Bliuc et al. (2012) noticed in the case of the Australian Cronulla riots, post-Brexit Islamophobia is not a simple ‘non-Muslim’ versus ‘Muslim’ conflict and it was not merely based on religion; on the contrary, it involves complex identities articulated in terms of ethnicity and anti-migration sentiments, shared ideas with certain political characters, and understandings of gender norms and women’s bodies. The example thus shows how articulating and reinforcing opinions on Twitter can lead to certain groups becoming marginalized not solely in terms of religion, but also by virtue of a complex and non-inclusive understanding of national identities.
Secondly, homophily on Twitter can be connected to antagonistic identities. While in theory Twitter can function as a public sphere in exposing users to different viewpoints, individuals and groups can use hashtags to enter into conversation solely with like-minded users and easily avoid conversations aimed at informing, challenging, or interacting with members of other groups. This tendency is shown by the lack of direct dialogue—through retweets or replies—between anti-Islam and pro-Islam positions after the EU referendum. When tweets do quote or retweet other users, they tend to state their positions or belittle others with different ideas rather than invite conversation. As a result, tweets create collective anti-Islam narratives that usually go unchallenged and hinder the possibility of agonistic exchanges based on equal participation.
Thirdly, the emotional character of certain tweets is antagonistic. Displays of sarcasm, anger, and fear are often aimed at belittling minorities instead of confronting them. This is probably facilitated by the opportunity to express hateful sentiments in an anonymous and largely unchallenged manner. These emotions in relation to Brexit are accompanied by non-factual and irrational news streams. The linked images and articles create Twitter narratives—such as women being sexually assaulted by Muslim men—which are aimed at having an emotional impact on other users rather than diffusing actual information, thus perpetuating further divisions.
The articulation of group identities around shared feelings against a minority, the tendency to homophily, and the emotional character of Twitter are all antagonistic features inasmuch as they prevent the democratic participation envisaged by Mouffe (2013). Anti-Islam discourses after Brexit are antagonistic because, instead of being structured to recognize Muslims as adversaries in a constructive dialogue, they support the perpetuation of social inequalities. With Muslims being viewed as supposedly unwilling to accept Western culture by virtue of their ideologies and accused of choosing Islam over secular institutions, they are framed as the scapegoat for social problems in a manner that excludes them as actors in conversations. This contributes to the creation of anti-elite feelings that, in criticizing the political status quo, marginalize Muslims as a ‘common enemy’ instead of seeking to collaborate with them against economic and social inequalities.
The antagonistic character of the analysed tweets forces us to rethink Twitter in terms of emotional antagonism rather than as a public arena where all are free to participate; some of the characteristics of Twitter, such as the opportunity to share anonymous opinions with like-minded people, have in all likelihood helped make it a fertile venue for various types of hate speech, including Islamophobia. The present paper has proposed a single example in a specific situation, with a limited dataset. Further research using the same approach but focusing on different contexts and other minorities would, I believe, confirm the findings that Mouffe’s concept of antagonism can prove useful in understanding group polarization on Twitter. The peculiar character of Islamophobia—a form of racism that tends to view Islam as an ideology rather than a cultural or ethnical affiliation—makes the case of anti-Islam tweets a compelling example that can be used as a model for ethnic, religious, and racial minorities, as well as ideology-based minorities.
This does not mean that minorities cannot use Twitter as a venue for constructive agonistic participation, or that Twitter leads exclusively to group polarization. In the case of post-Brexit Islamophobia, it seems that Muslims—and pro-Islam people in general—avoided confrontations with Islamophobic users. While it is possible that Muslim actors did not come to the fore in the Twittersphere because of issues of access or linguistic and technological skills, or that they choose to use other hashtags and keywords to articulate their claims, the lack of Muslim participation may also be symptomatic of their inability to counteract Twitter’s antagonistic Islamophobia. While Islamophobic tweets tend to articulate complex but homogeneous understandings of British and Western identities in non-inclusive terms, Muslims seem to struggle to find a way of collectively negotiating their position within Western society. This could be explained by the argument that, in opinion-based groups, religion is not a homogeneous predictor of social identity, but can rather lead to heterogeneous understandings of socio-political engagement (Baysu & Phalet, 2017).
Minorities could counteract antagonistic politics on Twitter through a different model of participation. Affective publics on social networks are often based on connective actions, characterized by personalized and individual political actions rather than collective identity (Papacharissi, 2016). Because a more personalized frame of action could create stronger networks in social justice and protest movements (Bennett & Segerberg, 2011), it is arguable that minority groups such as Muslims in the West could adopt similar strategies to maintain the heterogeneity of their claims and their individuality within a larger anti-Islamophobia movement. In the aftermath of the Charlie Hebdo attack, for example, Muslims and non-Muslims employed the Twitter hashtag #JeNeSuisPasCharlie to criticize Islamophobia while maintaining the heterogeneity of their voices (Giglietto & Lee, 2017). A similar strategy could play a role in overcoming the fragmentation of voters that notably came out of discourses about Brexit: on the one hand, Muslims need to cooperate with other minorities, such as LGBTQs, women, and members of the working class, to counteract marginalization; on the other hand, anti-elite movements need to be inclusive and recognize the claims of different social groups instead of perpetuating social inequalities. In this way, the potential of Twitter in creating emotional narratives can be employed in a process of identity construction that, while criticizing and trying to change the socio-political status quo, grants equality in agonistic participation rather than resulting in antagonistic group polarization.