#ShamimaBegum: An analysis of social media narratives relating to female terrorist actors

This study explores the constructions of gender in social media narratives regarding Shamima Begum, a British born woman who travelled to Syria to join Islamic State of Iraq and Syria (ISIS). Using the Twitter hashtag #ShamimaBegum – developed in response to Begum’s expressed interest in returning to the United Kingdom to give birth to her third child – we employ critical discourse analysis to examine social media users’ responses to Begum’s case across a 3-week period. We portray how the vernacular narratives constructing femininity, gender, and their relation to terrorist activity are built on the expectation that the female actor should express remorse for her actions and is judged according to certain perceptions of maternalism, religion, and victimhood. We also explore the absence of considered agency in the narratives about women engaging in and disengaging from violent activities, demonstrating the weight of race, religion, and gender in shaping narratives surrounding perceived violent women.


Introduction
There is a growing interest in women who engage in political violence, their motivations, the causes, and consequences of their actions (Bloom, 2012;Dissanayake, 2017). Simultaneously, some scholars have recognised the existence of misunderstandings or lack of interest in how gendered narratives contribute to overall understandings of the construction of female terrorist actors (FTAs) (Åhäll, 2012;Gentry, 2009;Sjoberg, 2009). Gender performativity, a widely accepted concept in poststructural feminist theory, poses that all language, discourse, and behaviour surrounding women, men, masculinity, and femininity contribute to a societal understanding of gender (Steans, 2013). Poststructuralism's influence on late 20th century feminism was profound, with many scholars embracing the notion that gender is not a fixed or stable identity category, but instead is a production of societal discourse and performance (see Brown, 2020;Butler, 1988). This can be clearly seen through the gendered and gendering knowledge reproduced by Western narratives regarding FTAs. It departs from dominant configurations of masculinity and femininity, which serve the ideal of Western rationalist hypermasculinity and contribute to the subordination of femininity. According to Narozhna and Knight, Such discursive deployment of gender hierarchy and sexual difference has practical effects across socio-political life, not least of which are discursive reproduction of gendered power relations globally through ideological reification of Western hegemony, as well as legitimation of coercive and structural violence. (Narozhna and Knight, 2016: 5) This article argues that this can be seen in the case of Shamima Begum, a 19-year-old Londoner who joined the so-called Islamic State (IS) and was discovered, in early 2019, in a refugee camp in northern Syria. Begum had already lost two children born in Syria and was pregnant again. She wanted to return to the United Kingdom. In a quick succession of events, Begum's British citizenship is revoked a few weeks before the death of her new-born baby. This story quickly became headline news around the world and led to thousands sharing their views across social media, mirroring heavily gendered narratives regarding FTAs. This study analyses these everyday narratives through critical discourse analysis (CDA), contributing to enrich our understandings of how lay members of the public perceive FTAs and help make sense of the broader discursive environment beyond elite and media discourses (Jackson and Hall, 2016). It answers the overarching research question: RQ. What are the gendered narratives present in Western social media discourses surrounding Shamima Begum?
And it shows how social media narratives of FTAs construct gender as a binary, fixed concept where femininity is valued if understood as peaceful, maternalist, and submissive, which helps maintaining patriarchal gender formations and hierarchies. Thus, we focus on women because, as defended by Narozhna and Knight (2016: 6), they are the ones who see 'their agency and political motivations marginalized or denied on the basis of gender'.
In the following pages, we begin by situating this study in the relevant feminist literature regarding gendered narratives of FTAs. We focus particularly on three gendered narratives related to FTAs -motherhood, religion, and victimhood -which will subsequently be used to discuss the case study under analysis. The choice of such narratives was guided by an attempt to integrate a top-down with a bottom-up approach, in which we first identified the themes in the literature regarding the analysis of gendered narratives behind the construction of FTAs, and second, we consensually selected the most prominent theoretically driven themes according to the specificities of Begum's case. We then outline the methodology used to collect and analyse social media narratives. Through CDA, we analyse everyday narratives of security and terrorism (see Lister, 2013, 2016;Jackson and Hall, 2016;Vaughan-Williams and Stevens, 2015) to examine how gendered narratives contribute to understandings of the construction of FTAs. This is important because terrorism is 'a socially constructed phenomenon, its meanings emerging from social interaction and analysable through studying language use and visual media' (Stump and Dixit, 2013: 108). Therefore, we focus on lay people's knowledge, beliefs, and attitudes regarding FTAs, which transmit their own experiences and realities. This originates an 'inter-subjectivity' (Jackson and Hall, 2016), that can create 'powerful language for articulating support or opposition for political projects, for exploring social cartographies and for expressing the diversity of an individual's needs' (Jarvis and Lister, 2013: 170). This is followed by a discussion of the case study under analysis and by concluding remarks highlighting the power of everyday narratives in shaping politics, which pervades not only the present case study but also a variety of other FTAs' cases.

Gendered narratives of women and violence
Essentialist gender assumptions predominate in the study of terrorism (Brown, 2020;Sjoberg and Gentry, 2011). Mainstream conceptualisations of terrorism present accounts and analysis where actors are assumed, sometimes overtly but often subtly, male (Sjoberg, 2009). Masculinity's supposed monopoly on violence is commented on by various feminist scholars (Martini, 2018;Steans, 2013), but it becomes easy to refute when presented with obviously violent women who are members of violent groups. However, the discussion surrounding their role and motivation looks very different from that centred on their male counterparts, despite research showing that women decide to join violent organisations for much the same reasons as men (Alison, 2009;da Silva, 2019;Nacos, 2008). The political nature of terrorism is accepted when discussing White male actors because constructions of masculinity encompass assumptions of rationality. Hegemonic masculinity depicts White male bodies and masculinity as strong, dominant, and rational (Connell, 1995). Equating visions of hegemonic masculinity with rational military violence sheds light on why the construction of male rationality in relation to any other form of violence, including terrorism, exists. Therefore, when it comes to portraying a female agent of terrorism, much work goes into undermining this rationality, which is intimately connected to notions of agency. As explained by Brown (2020: 13), from a mainstream philosophical perspective, '[s]omeone's agency is realized when they carry out activities that are in their rational self-interest [. . .] agency is about realizing autonomy and the ability and capacity to liberate oneself from oppressive contexts'. Thus, FTAs joining violent organisations perceived as traditional and oppressive are never portrayed as the liberated type or as acting in a rational and agentively manner.
In the remainder of this section, we explore these issues through the most prominent gendered narratives related to FTAs according to the specificities of Begum's case, which are motherhood (Begum conceived three children during her time within IS); religion (despite being born in the United Kingdom, Begum was automatically categorised as non-Western due to her perceived Muslimness); and victimhood (Begum has been mostly portrayed as a perpetrator for all the reasons already mentioned, as well as due to her perceived lack of remorse when found and interviewed at al-Roj refugee camp despite her own claims of having been brainwashed). In this vein, we show how '[r]ace, religion, and gender intermingle in a static construction incompatible with reason, agency, self-determination, or political action' (Brunner, 2017: 968).

The role of motherhood in understanding female terrorism
Essentialist gender constructs reify the 'feminity/peace and masculinity/violence' binaries (Steans, 2013: 109), through which many behaviours in women are accepted if they can be explained in the pursuit of peace and the giving or preservation of life. Agency in political violence (i.e. a willingness to accept, encourage, or enact lethal violence) is therefore difficult to accept when it is in direct 'tension with understandings of female bodies "natural" associations with motherhood' (Åhäll, 2012: 104). Gentry (2009: 241) addresses this specific gendered narrative in her work on twisted maternalism: maternalist thinking assumes women's 'psychological compulsion to assist and support others' which trumps all other considerations in relation to any decision-making, including whether to engage in violence. Gentry (2009) then claims that women's agency is qualified as positive and morally correct in society if and when their decisions are made with motherhood or peace in mind. This implies that the capacity to care makes women somehow less threatening, definitely less violent. Åhäll (2012) explores a similar idea, arguing that motherhood is a discourse, a myth created to deny political agency to FTAs. Whether accepting the conceptual creation of motherhood or not, the power of maternalism remains significant in constructions of femininity, so much so that womanhood and motherhood are often understood in tangent.
The live to mother motif is troubling because it entrenches essentialist gender notions and means women's agency cannot be regarded as positive, but rather only in terms of the threat it poses. However, Gentry (2009) makes a useful point about women who actively decide to mother to support a politically violent cause. She suggests that the very act of living and procreating in violent circumstances can be considered political. In the case of IS, the women were invited to support the development of a state by mothering 'the cubs of the caliphate' (Horgan et al., 2017: 651), which challenges traditional gender lenses by bringing together motherhood and terrorist activities.

Using constructions of gender in Islam to explain female terrorism
Narratives surrounding Muslim women as passive victims of their circumstances are common and rooted in modern and historic forms of Islamophobia, often generated through tabloid headlines and media representations (Martini, 2018; for more on Islamophobia, see Allen, 2010). Here, 'the discourse on gender intersects with the discourse on Neo-Orientalism' (Martini, 2018: 459) -Western values are depicted as universally aspirational, especially when compared to Islamic values, as there is an 'assumed conflict between Muslimness and a Western identity' (Brown, 2020: 3). However, it becomes challenging to maintain essentialist and orientalised gender constructs when Muslim women choose to enact violent acts, most notably suicide bombings (see Narozhna and Knight, 2016). Female suicide terrorism is a much-covered topic in the literature on gender in terrorism, which tends to put religion on centre stage and to reflect gendered narratives. Most scholars developing theories to explain the behaviour of female suicide bombers tend to rely on instances of personal grief and strife, rather than rational political or religious choice (Bloom, 2012;Dissanayake, 2017;Von Knop, 2007). As an example, Wafa Idris, who died in a suicide bombing on behalf of the Al-Aqsa Martyrs' Brigade, was the subject of worldwide media attention. Idris' case was met with speculation surrounding how her infertility led to her poor mental health and thus her willingness to commit extreme violence (Plaza et al., 2017). Narratives generated around Idris seek to explain her violence by relying on gender roles, specifically her inability to mother. However, Idris belonged to a Palestinian Liberation group, which leads to the assumption that her motivations (though complex) might have been political.
Thus, Muslim women who engage in suicide terrorism undermine both essentialist and orientalist narratives. This explains what Von Knop (2007: 398) called the 'public outcry' in response to these sorts of attacks. If we accept that narratives surrounding political violence committed by women are created to re-stabilise a familiar Western and patriarchal order, the difficulty in a satisfying explanation of violent Muslim women is multifaceted because their agency is already undermined by the fact that they belong to an othered religion. It seems, therefore, that in response, gendered and orientalised narratives are more fervently pedalled to redress the imbalanced representations catering to a non-Muslim audience in Western geographies, which historically have projected Islam as violent and evil (Kirk, 2020).

Perceptions of victimhood in the construction of female terrorism
Women as victims of terrorist violence is a theme frequently discussed in the literature. Bloom (2010) considers that women may even, through various traumatic personal experiences (e.g. rape, rejection, revenge), be drawn to terrorism themselves. However, we concur with Kirk (2020) that Bloom's work tends to perpetuate the gendered narratives highlighted by Sjoberg (2009), Gentry (2009), and Steans (2013 by assuming that women are victims before anything else and by portraying women who take up terrorist violence as a-political. Polletta (2009: 124) highlights how the portrayal of women who commit violence is often limited to 'stock characters' -they are either represented as 'powerless and incapacitated victims' or as 'unapologetic and provocative'. The latter being treated more harshly in court and mainstream media, and the former being examples of good victims. Once again, the essentialist gender assumption that places women as passive in their behaviour results in a more forgiving, more authentic-looking portrayal of victimhood. Therefore, conceptualising victimhood as an exclusive experience for women who fit into certain narratives is problematic. Enacting the good victim identity requires FTAs to rescind their rationality in relation to their violence to access the label and the reaction it affords. It seems witnesses to female violence are unable to accept their actions unless they believe the FTA acted without an agency they can accept. They are a good victim if they were groomed or forced, but not if they rationally choose their violence. Equally, it warrants audiences to ignore or devalue the traumatic experiences that often accompany the commission of political violence. It is pertinent to consider here the cases of several women of the Colombian People's Army (Revolutionary Armed Forces of Colombia (FARC)). The FARC was a communist guerrilla movement active during Colombia's deadly civil war. The group contained thousands of women who fought using the same tactics as their male counterparts. Since the end of the civil war, many women who engaged in the armed struggle have discussed issues such as sexual assault and forced abortion they experienced while fighting against the state (Piňeros, 2018). However, in displaying agency and willingness to kill for politics, these women are not perceived as good victims. Their experiences have been left unprosecuted, despite the numerous cases made public and the evidence against their attackers. This suggests that in showcasing a dual identity -perpetrator and victim -women forfeit their claim to representation as a good victim, falling into the camp of provocative, rather than passive.

Case study
Considering that this study aims to analyse how gendered narratives contribute to understandings of the construction of FTAs, we selected a case that reveals considerable public engagement. In addressing this study's overarching question -What are the gendered narratives present in Western social media discourses surrounding Shamima Begum? -we focussed on Twitter users' responses to Begum's story in English. At the age of 15 years, Begum and two of her school friends joined 145 other British women by travelling to Syria to join IS. Through their propaganda, recruitment tools, and manifestos, IS sought to draw women to the caliphate by likening them to followers of the Prophet in his initiation of the first IS. In February 2019, after significant IS defeats, Begum was discovered by a journalist in the al-Roj refugee camp in northern Syria, where she was interviewed. She expressed her desire to return to the United Kingdom to safely give birth to her third child; her two previous children, born under the caliphate, had died. This interview took place on 13 February 2019 and was followed by Begum's British citizenship being revoked on 19 February 2019, and the death of her new-born baby on 8 March 2019.

Data collection
In this study, we used Twitter as a data source, as it presents researchers with important opportunities to study a range of topics in a naturally occurring setting, facilitating the understanding of what, and how, people communicate in particular situations (Ahmed et al., 2018). Our data collection periods were informed by three significant dates in Begum's experience: All 11,368 tweets published in English in the above date frames featuring the hashtag #ShamimaBegum were extracted using a Java application. Twitter settings prevent tracking the locations of tweets, unless users opt in; however, 51% (5843 of 11,368) of the collected tweets featured language overtly identifying users' home nation as the United Kingdom, allowing the assumption that they were written and published in Britain. From the total of tweets, a randomised sample of 200 tweets corresponding to each date window was pulled for in-depth analysis. In addition, every top tweet from these date ranges was also analysed. According to Twitter's Chief Scientist, 'Top Tweets is a new algorithm we developed that finds tweets that are catching the attention of other users. The algorithm looks at all kinds of interactions including retweets, favourites, to identify the tweets with the highest velocity' (Sir Patrick Vallance quoted in McGee, 2010). Using this second data collection strategy means the sample also features 324 tweets across the time period that gained the most interaction from other users on the platform. As a result, a total of 923 tweets were read, coded, and analysed. Of the 923 tweets in the sample, 183 contained images.

Data analysis
Each tweet -words and images -was coded according to CDA principles (Fairclough, 1992(Fairclough, , 2003, as language is seen as a social construction which shapes and is shaped by society (Machin and Mayr, 2012). Moreover, we aim to shed light on how language is used to construct and crystallise specific identities and maintain unequal power relations (Stump and Dixit, 2013). The coding process departed from a structured framework created by the authors, whose motives were twofold: (1) to add rigour to the identification of how gendered narratives contribute to our understandings of the construction of FTAs; and (2) to allow for a reliable replication of this study. To define and operationalise the analytical framework, we first identified the themes in the literature regarding the analysis of gendered narratives behind the construction of FTAs. Then, we consensually selected some of these theoretically driven themes according to the specificities of Begum's case, attempting to integrate a top-down with a bottom-up approach. Through this process, we reached three main themes: motherhood, religion, and victimhood. The coding procedure also incorporated the temporal dimension presented in the previous section to explore possible differences in the gendered narratives regarding Begum across the three periods of time.
When coding images, we took into consideration the context in which the image subjects were being assessed (see Banks, 2007;Pink, 2013). In the following section, we focus particularly on memes because they formed the majority of images linked to the hashtag under analysis, but also because memes are images that are intentionally altered by human interaction. Unlike still image or film, curated by only the director, memes offer Internet users the ability to interact with image, mutating it to change its meaning (Wiggins and Bowers, 2014).
Finally, our presentation of social media data follows Rodham and Gavin's (2006) guidelines on online consent, anonymity, and privacy. Personal consent was not sought since the online comments are publicly available; however, the quotes presented throughout the article have been slightly edited, so that comments cannot be traced to their online originators. They also do not refer to usernames/identities to protect the privacy of the members of the public whose comments we reference.

Results and discussion
Nearly 40% (359 tweets) of the data analysed feature gendered language and portrayed essentialist and orientalised constructions of femininity or masculinity. These tweets were consistent across the data collection periods, indicating their pervasiveness. Developments in Begum's case did not impact underlying or overt gender assumptions in user tweets. A total of 85 (23.68%) of these tweets are also composed of offensive language, which is used to imply either Begum's sexual promiscuity or a deviancy that involved her trying to steal from the British state: Fuck you!!! You little Jihadi Slut #ShamimaBegum Bye bitch! No NHS and benefits for you #IsisBrides #IsisSchoolgirl #ShamimaBegum The majority of these tweets' authors are men (66%), which is indicative of the trend which suggests that women face higher levels of gendered abuse online, usually from men (Cole, 2015).
The data also show various attempts to alienate Begum. Over 50% (279 tweets) of those who discuss Begum's re-entry to the United Kingdom do not support her return. This attitude is common across each of the data collection periods. To justify these feelings and to pressure institutional decisions, many Twitter users rely on conceptions of othering -Begum is othered as a mother, a Muslim woman, a bad victim, and a guiltless FTA. It is, therefore, difficult to analyse any reaction to Begum without clear consideration of each of these narratives. The following pages will address each of these themes using the sampled data to answer this study's overarching research question, contributing to the understanding of wider constructions of gender across social media narratives of FTAs.

How is maternalism portrayed in narratives surrounding Shamima Begum?
Maternalism is a concept familiar to and often rebutted by feminists because of its ties to gender essentialism and its focus on the biological functions of the sexes (Åhäll, 2012;Gentry, 2009;Kirk, 2020). Maternalism's claim, though rooted in biology, asserts that women have an innate need and duty to provide care that underlies their gender identity. Maternalist thinking is one of the most pervasive narratives to feature in the data collected, taking form in both the expectations levelled at Begum and internalised in the personal responses of Western Twitter users.
To identify maternalist thinking in the sampled social media narratives, a code was assigned to any tweet containing #ShamimaBegum alongside mention of heteronormative family roles. Specifically noted were the following words: mother, motherhood, baby, family, bride, wife, and husband. Of all the explicitly gendered tweets, 60.3% (321 tweets) contained language related to maternalism and the most cited theme was Begum's baby (157 tweets /48.9%). Users commented on Begum's baby more frequently after his death, with 60% (94 tweets) of these tweets coming in the third data collection period. Although the frequency with which these conversations increased in this period, the content remained consistent across the sample. Begum's capability as a mother was under judgement, usually through direct discussion of her children who died: How is it possible that three of her children have died of malnutrition and so many other babies have survived? Was she not breastfeeding? Was she a fit mother? #ShamimaBegum Her child didn't die because the government revoked her citizenship. He died because she moved to a warzone without a consular or diplomatic service. Those who say this conveniently forget that she had two other babies that died before denaturalisation. #ShamimaBegum Regarding Begum's decision to move and have children in Syria, the audience granted her the agency discussed by Gentry (2009). Begum's capacity to decide on her childbearing is not questioned because motherhood is assumed as a natural choice for all women, as agency in motherhood is an accepted norm in Western consciousness. Granting such agency allows users to deny any sympathy to Begum in relation to the death of her children by placing culpability directly on her. In this instance, the strong presence of maternalist narratives in relation to Begum's case serves a dual purpose: to highlight her calculated failure as a mother and to justify further othering.
The narratives in these tweets, however, do not reject the components of maternalism. The expectation that Begum be a mother figure somewhat mirrors IS's own intentions for women. Whether tweeting with or without sympathy, there are several examples of internalised maternalism that lead Twitter users to discuss Begum's case through the frame of maternalist expectations: Every baby is a mother's responsibility. As a Muslim mother of four, I wonder what makes a mother decide instead of raising my children in a nurturing way I'm going to become a killer for ISIS & put them and countless others in danger.
I am still thinking about #ShamimaBegum and wondering what a young mum who has lost three babies and who is now stateless must be going through.
Of the 198 tweets (21.45% of full sample) that explicitly comment on Begum's motherhood, only 31% (61 tweets) express sympathy, which is rooted in the loss of a baby. The suggestion that a lack of sympathy supports the othering of Begum as a bad mother holds here but, significantly, the weight placed on the value and expectation of motherhood remains. It seems that Twitter users who are unsympathetic to Begum's case do not reject maternalism, but instead discard the concept of mothering in a society that rejects their own: Bang goes her excuse to come back. Her baby is Syrian so take your medicine like a good girl! You are a member of ISIS and loyal to the cause you joined. Your baby has been born in ISIS so you should be happy to stay now #ShamimaBegum We see here an intersection between maternalist and nationalist and national security narratives. As suggested in the literature, there is worth in childbearing in a state building capacity, and some of the tweets acknowledge this through an expression of concern regarding state security should Begum and her child have been granted re-entry to the United Kingdom: #ShamimaBegum says she wants to return to the UK so she can bring her baby up in peace and safety. I want her to stay where she is so I can bring mine up in peace and safety.
Figure 1 seems to have been produced by the British division of the yellow vest demonstrators, denoted by the Union Jack image in the top right corner. The image uses mockery to make the point discussed in the literature regarding IS 'cubs of the caliphate'. Superimposing Begum's face into an image of a new mother and baby and replacing the child with a time-triggered bomb makes the point that Begum's child, under her parentage, is a deadly weapon. The image, alongside all analysed tweets in relation to Begum as a mother, makes no attempt to challenge her maternal role. Here, the issue is not her motherhood, but her ideology and the risk it poses to national security. In this instance, Begum is being othered through her IS affiliation, whereas in previous tweets, she was being othered as an incompetent mother. Either way, each case of othering provides an example of an identity that counters what is valued in Western constructs of gender.
Keeping in mind the power of national and gender identities, it becomes pertinent to consider how this intersects with maternalism at a micro-and macro-level. As discussed, many of the tweets speak to a narrative of maternalism that is wholly internalised by the Twitter users themselves. When cross-referencing users who discussed Begum's baby and her citizenship claim, 54.2% of them (10 out of 23 tweets) suggested Begum should be tried based on the rule of law. This contrasts directly with the majority's preference relayed in the whole data, that Begum stays away from the United Kingdom either through the revocation of her citizenship or through her death. The former indicates the desire that her baby should be protected, speaking to an internalised maternalism in Twitter users discussing Begum's children. Thus, it is arguable that in the face of a child at risk, Twitter users take on the role of caregivers by suggesting the United Kingdom should take some responsibility for the safety of her child: #ShamimaBegum should be allowed to come back and have her baby who must immediately be taken. She must never have access to them she would only corrupt her child . . . Protect her baby not her. Therefore, constructs surrounding women and motherhood are extremely pervasive in social media narratives surrounding Begum. The data have shown arguments made by Gentry (2009) andÅhäll (2012) hold weight in discourses online, as well as in wider behaviours in Western societies. Gender is constructed, through these social media narratives, as rooted in maternalist values that see women understood and valued first and foremost as mothers -much the same, actually, as IS's views on women.

How is religion portrayed in narratives surrounding Shamima Begum?
Religion is the fourth most prevalent theme to come out of the social media narratives surrounding Begum, with Islam being the religion most frequently cited. As predicted in the literature, Islam is regularly understood in direct contrast to Western culture and is referenced by users in an attempt to other its practice. This section will highlight how religion in narratives surrounding Begum is used to stimulate debate outside of Begum's case, sometimes unrelated to gender narratives. Extending the method used to identify maternalist thinking in individual tweets, words related to religion, race, and the identification of Islamophobia were collected and coded. It became evident that religion and race were often discussed interchangeably or in tangent. Tweets were coded to the religion and race themes if the content was actively derogatory (the same being applied to Islamophobia): I have pretty close to zero sympathies for her and her entire fucking family and tribe. When the lot of them publicly piss on the Quran I MIGHT cut them a bit of slack. Should #BanIslam #ShamimaBegum Within the tweets that perceive Begum's religion negatively, subtle gendered narratives remain. Most common are users who comment on the threat or rejection of Islam, rooting their discourse in the security threat posed by Islam and its representative (in this case) Begum. Narratives focus on Islam's ability to groom individuals' features commonly: She was groomed, her father was groomed, his father was groomed as well. How far do we go back? Around 1400 year I say. They were all groomed by Islam #ShamimaBegum She was groomed by Islam! Just like 6% of the #UK population to hate and to divide and to kill! In every #mosque every day! Evil Islam.
Begum, and others, is perceived as grooming victims of Islam itself. Claiming Islam has the capability to groom individuals on a large scale is indicative of a conflation of Islam with the terrorist group IS, born from the Islamophobia which Martini (2018) discusses in her work. Equally, of all 923 tweets analysed, only 27 discussed state security in relation to Begum, implying the majority of users do not feel threatened by her as an individual: I, for one, am not afraid of #ShamimaBegum because she poses no threat to me. I am afraid for her for what has happened to her and how the state will now choose to treat her. This is further corroborated in the fact that of the 359 tweets coded as explicitly gendered, only four tweets feature the word or label terrorist. In these social media narratives, the threat is not the woman who joined IS, but her religion.
The trend to discuss Islam and its impact ahead of Begum herself extends beyond narratives surrounding its perceived threat. Begum and her case are often used as a conduit to discuss other issues related to Islam. In some tweets, users directly call out Islamophobia: Recently there has been so many islamophobic and racist comments on #socialmedia about Islam. Could you imagine blaming and persecuting all Christians for KKK? #ShamimaBegum Using the platform to identify and call out Islamophobia while by-passing Begum's case highlights an alternate narrative but remains problematic. It challenges orientalised assertions but, through the comparison to Christianity, maintains the Western-centric attitude addressed in postcolonial feminist critique. Equally, omitting Begum's case in the condemnation of Islamophobia fails to acknowledge how she is a victim of this very attitude. Finally, it supports arguments related to constructions of gender that fail to acknowledge women's agency in political violence (Kirk, 2020;Sjoberg and Gentry, 2011;Steans, 2013). When Begum goes unacknowledged, it strengthens narratives that suggest women are not considered as capable of political violence.
In other instances, Begum's case is used to develop narratives that undermine Islam through a direct comparison to the supposed superiority of neo-orientalism: #ShamimaBegum would not be very happy in the UK with our rule of law and democratic system of governance. It's a country full of #infidels, no beheadings, women free to wear what they want, all that anti-Islamic equal rights nonsense. Surely she is fine where she is? This user's perception that equal rights are anti-Islamic indicates, once again, the conflation of IS and Islam, made all the more apparent by the reference to IS beheadings and language commonly associated with their propaganda. Here, we can see a subtle mockery that implies this user's perceived superiority to Begum, specifically in relation to women's freedoms. Much like the narratives surrounding maternalism, these narratives justify the rejection of Begum's re-entry to the United Kingdom through the othering of her religion. This trend extends across the data. A total of 68% of those (16 out of 27 tweets) who negatively discuss Begum's race or religion while also referencing her citizenship suggest she should stay in Syria, have her citizenship revoked, or die. Figure 2 shows a meme embodying almost all the trends already discussed in relation to narratives surrounding religion. The meme is rooted in neo-orientalism, using familiar Western pop-culture to juxtapose Begum in the image's foreground. Pictures of Begum in her hijab are common across memes attached to her hashtag and, in this instance, are directly comparable to the only other woman in the image -Carrie Fisher as Princess Leia. The comparison between the liberal lives of Western women and Begum's experiences is often made, contributing to the narrative identified by Martini (2018) that Muslim women are perceived in some way submissive in comparison to Western women, even when the Muslim women discussed are also born and bred in the West. Initially, Begum's appearance in this image implies a significance that counters that narrative. She features in place of the central protagonist in the film, she seems to be holding a weapon, and sits above all other characters in the poster. However, the meme exists to trivialise and mock Begum, with the creator editing the image to subvert its original meaning. The intended humour behind the meme undermines any implication that Begum is a real threat, while reminding the audience of Begum's otherness.
On the whole, the portrayal of religion in narratives related to Shamima Begum shows various attempts to solidify Begum as an other, relying on orientalised assumptions to justify keeping her away from the United Kingdom. Even users who identify the rife Islamophobia at play in these narratives undermine Begum's experiences and agency, indicating that powerful gender narratives are still at play.

How is victimhood portrayed in narratives surrounding Shamima Begum?
Victimhood is discussed in a somewhat binary way in relation to Begum. The two main narratives relating to this concept either accept her as a victim of grooming and IS violence: Being ideologically groomed is no different to be sexually groomed. In fact, the two often go together. Grooming is about taking the truth and perverting it which is why it is especially effective on younger impressionable people. #ShamimaBegum Or wholly reject Begum's victimhood status, citing Begum's choice as the reason for this denial: #ShamimaBegum you don't deserve #UK citizenship. You believe and belong to the scum of this Earth #terrorism. This was and still is your choice.
These narratives do not accept the possible dual identity of a person who is, simultaneously, a victim of ill-treatment and prejudice and someone who has made a choice which goes against societal approval. The gendered narratives discussed in the literature underscore these approaches alongside feelings of nationalism linked to masculinist protection.
Significantly, across the 923 coded tweets, the word 'victim' appeared only 9 times and often it featured to counter any argument that this label could be applied to Begum: #ShamimaBegum How about no? This 'victim' seems to have no qualms about joining IS so let her stay there.
The concept of victimhood was applied instead to groups of people or individuals that users felt sorry for, such as Yazidi communities or those who died in the Manchester Arena bombing in 2017: While the world wonders what to do about women like #ShamimaBegum thousands of #Yazidi families are desperately searching for their children kidnapped by IS. Are any agencies actually looking for them? Probably not! To the spineless pathetic do-gooders saying of #ShamimaBegum -'She was only a teenager'. Try saying that to the parents of the teenagers murdered at the #ArianaGrande #Manchester concert. Keep #Begum out of this country -and for good! Here, we see the denial of victimhood being used as a means to justify Begum's citizenship removal. The tweets demonstrate the intersection of these narratives, conflating the actions of Begum and IS, while relying on narratives surrounding female deviancy in relation to FTAs.
Across the sample, the phrase 'made her bed' appears 15 times. The expression exemplifies the wider attitude that Begum's choice to join IS denies her of a chance of coming home and, specific to this work, to be labelled a victim: #ShamimaBegum has made her bed now she must lie in it. There is no place in British society for extremist sympathisers. She wants to come home because it has all fallen apart.
Thus, it is in narratives related to victimhood that users are most inclined to accept Begum's agency. Often, these narratives rely on other plots of the same story, like neoorientalism or maternalism, to provide weight to their claim, but each happily grants Begum choice. However, in granting agency to Begum, individuals take on the role of the nation's masculine protector because, with agency, Begum is an actual security threat. In this instance, users are relying on the monster narrative that tells us women who engage in terrorism are in some way evil (Sjoberg and Gentry, 2011). Denial of victimhood under this narrative makes sense because acknowledging any potential harm to the adversary undermines masculine protectionism.
When it comes to Begum's UK citizenship claim, as discussed previously, overwhelmingly, Twitter users' preference is that Begum remains away from Britain. This further corroborates the view she is not a victim, implying that she deserves no state support.
However, some people consider that Begum should face British rule of law. Although not a direct comment on Begum's victimhood, it implies a softer response to her case, the most adamant of these discussing Begum's trauma at the hands of IS: Here's an unpopular thought. Why is this young woman not being treated as she deserves? As an unlawfully abducted raped victim who has been so brain-washed she believes the lies that twist her reality #bringherhome #ShamimaBegum Working in #CSE [Child Sexual Exploitation] has shown it is common for #victims to identify with their #abuser and accept responsibility for their #abuse. #ShamimaBegum's #unapologetic attitude is really #StockholmSyndrome reiterating this all happened to her as a child in her formative years of mental development.
Understanding Begum as an individual who has suffered at the hands of IS indicates the counter-narrative to the denial of her victimhood. These narratives follow a different plot, alluding to or directly commenting on the grooming Begum experienced in her radicalisation process, but also rely on gendered narratives in their understanding of her. However, such complex perpetrator perspective has not yet taken root in academic or policy circles, let alone in everyday security narratives. On the contrary, Figure 3 showcases the most common attitude -lack of empathy -portraying Begum in a black bin bag indicating the user's desire to disregard her and her experiences entirely, to throw her out. Begum's own bourka that moulds into a bin bag is evidence of reliance on orientalised, in this instance, Islamophobic narratives to distance Begum from users.
On the rare occasion that users relate to Begum's experience, it is usually in an attempt to further the perceived difference between them and her: Fifteen year olds get in trouble for drinking or smoking weed down the park! They don't join terrorist fucking cells. If you really can't see the jilted problematic nature of that then you're as fucking deluded as she is #ShamimaBegum The reliance on personal and Western-centric coming of age experiences shows the underlying neo-orientalism that pits Western experiences as superior, as well as a clear lack of understanding of the radicalisation process. Moreover, of all the users who expressed no empathy for Begum, 10% of them (81 tweets) explicitly attribute this to her lack of remorse. The percentage seems low, but the content of these tweets is vociferous: Put a bullet in her. Problem solved. She shows no signs of remorse whatsoever and allowed two babies to die by supporting a medieval barbaric death cult. She's a cunt. #ShamimaBegum The anger demonstrated in these responses to Begum is significant. The language is gendered, and reference to her children once again places weight on the gendered expectation to mother, but before mention of this comes reference to her lack of remorse. A 'good victim' is one who apologises for their wrongdoing. A lack of remorse in Begum is insulting because it indicates that not only did she subvert traditional Western patriarchal norms related to gender but she is also not sorry for doing it. As was the case for the women of ETA (Euskadi Ta Askatasuna) who claim they were punished more harshly because they were women engaging in political violence (MacDonald, 1991), the anger in tweets related to Begum's perceived lack of remorse hints at an important gendered expectation: women should not and do not reject patriarchy, especially through political violence.

Conclusion
This study shows that narratives that rely on essentialist and orientalised gender constructs pervade social media in relation to Shamima Begum. This case was chosen as an example of an FTA who gained significant media and public attention across the world. Drawn from a sample of 923 unique public tweets, it was found that narratives that explicitly address Begum's gender were common and were accompanied by references to maternalist expectations, religious criticism, and denial of victimhood. Such references do not just emanate from media and elite circles but are also present in the everyday narratives published online by members of the public in the West. Through CDA, we have shown how such narratives gendered, orientalised, depoliticised, and othered Begum, who was only perceived to have agency when such agency could serve to exclude her from Britain. This concurs with Brown's (2020: 15) assertion that '[u]nderstanding agency is about understanding the discourses and power that produce these subject positions'.
The narrativisation of Begum's plot according to gendered and orientalised standpoints represents what Wibben (2010: 108) called 'rigid framings [that] inhibit more imaginative approaches to conflict'. In this vein, instead of leading to new conceptualisations of agency regarding FTAs, taking on board the temporal and spatial aspects of their actions, Begum's case simply reiterates the hegemony of rational masculinity over irrational femininity and of the West over the orient. We have highlighted other FTAs' cases, such as the women of ETA and the Colombian FARC, whose experiences indicate how essentialist gender narratives resulted in harsher punishments and trauma inflicted by criminal action (e.g. rape, forced abortions) being left unprosecuted. The very same narratives were at play in Begum's case, contributing to actual policy decisions, characterised by an overall inability to understand her potential victimhood, the othering she faced, and the fierce desire to ban her from the United Kingdom. Begum as a security threat narrative comes to dominate the political landscape in Britain and has had awful implications in terms of the policies directed at women returnees and their children, which can also be seen in other contexts (see Toros, 2020).
Our findings are important because narratives serve to shape the conditions of possibility in politics; they legitimise 'certain dispositions and orientations while opposing and delegitimizing others' (Campbell, 1998: 10). By analysing online comments, we have demonstrated the utility of using digital methods to make sense of everyday narratives, and we have contributed to the literature by going beyond the interview and focus group methodologies that have been prevalent in research to date.

Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the British Academy Research Fellowship [GRANT NUMBER pf170092].