Brexit as ‘politics of division’: social media campaigning after the referendum

ABSTRACT The UK’s referendum on EU membership that resulted in a narrow minority in favour of leave was followed by a leadership vacuum and intense debate about the implementation of the result. The politicization over Brexit resulted in the development of ‘Brexit identities’ of Remainers and Leavers that superseded party identities. We argue that in order to understand how this politicization took place despite a leadership vacuum we firstly need to look beyond the arena of formal party politics to more informal arenas of political contestation on social media, especially Facebook, and secondly understand the linkages between EU and national level politicization that polarised the country around new British-specific identities. Through this, we analyse the ‘politics of division’ not simply as a form of contentious politics driven by political parties, but as a social conflict driven by non-institutionalised groups, grassroots campaigns and ordinary citizens. We find evidence of significant mobilisation that extends beyond the realm of party politics but argue that this mobilisation cannot necessarily be considered entirely ‘grassroots’. Rather, it is driven not just by citizens but also shaped by mainstream and alternative media platforms. The debates cannot, however, be considered purely a form of EU politicization, rather, analysis of Facebook comments shows that politicization over Brexit through these campaigns is primarily contestation over the nature and legitimacy of British democracy. Because of this, we argue that social media is an essential site for the study of EU politicization and political campaigns in general.

that resulted from the government's lack of planning for the Leave eventuality and the incompatible promises made during the campaign. The ensuing polarisation saw the development of new 'Brexit identities' of 'Remainers' and 'Leavers' which divided not only the political party landscape but also became the most dominant cleavage in British society (Evans & Schaffner, 2019). This cleavage became the receptacle for broader political, cultural and identity-based divides that underpin what Wheatley (2019) describes as the 'politics of division', shaping not just partisan contestation over Europe but also division over British democracy itself. We argue that in order to understand how this politicization took place we need to: 1) look beyond the arena of formal party politics to more informal arenas of political contestation, specifically, online movements on social media, and 2) understand the linkages between EU and national level politicization that polarised the country around new British-specific yet at the same time EU-related identities. Through this, we analyse the politics of division not simply as a form of contentious politics driven by political parties, but as a social conflict driven by non-institutionalised groups, grassroots movements and ordinary citizens.
Firstly, we ask how and to what extent the 'politics of division' on social media reflects, in the absence of effective leadership, a form of 'bottom-up' politicization through (online) social movements beyond the sphere of formal party politics. The referendum campaign itself had demonstrated the importance of social media campaigning, revealing a high level of mobilisation on the Leave side (Heft et al., 2017). The Leave campaigns were subsequently implicated in one of the biggest cases of voter manipulation and disinformation, the Cambridge Analytica scandal (e.g., Cadwalladr, 2020). Yet, the period following the referendum saw significant and unprecedented pro-European mobilisation by Remainers (Brändle et al., 2018). Mobilisation through new Facebook pages representing 'the 48%' -the proportion of voters who opted to remain -and continued activity by existing campaigns resulted in more traditional social movement activity such as large-scale anti-Brexit marches, protests, and local pro-EU groups. This mobilisation, in turn, kept campaigning by Brexiters alive who sought to defend their vision of Brexit as the 'will of the people'. These social media campaigns cannot simply be understood as the continuation of traditional partisan politics online nor are they solely driven by social movements through bottom-up mobilisation and innovative forms of collective and/or 'connective action' (Bennett & Segerberg, 2013). Instead, via social media, both Leave and Remain campaigns displayed a dynamic of their own that mixes centralised political organisation with spontaneous protest (see Milan, 2015). Yet, there is little systematic research that explores these polarizing dynamics of online politicization after the referendum that reaches beyond party politics and into the domain of social movement studies.
Secondly, we ask how and to what extent politicization over Brexit on social media shifted from contestation over Britain's EU membership to a fundamental conflict over democratic legitimacy. While party contestation focused mainly on the form Brexit should take -mostly solidifying around the idea of a so-called 'hard' or 'soft' Brexit, social movements raised more serious questions about the British polity, not least, about the role of referendums in a system in which parliamentary sovereignty is the core constitutional principle. In this article, we explore these forms of bottom-up politicization through social media campaigns' use of Facebook platform affordances and contestation over Brexit in Facebook user comments. We provide evidence of significant mobilisation that extends beyond the realm of party politics but argue that this mobilisation cannot necessarily be considered 'grassroots' in the tradition of (online) social movements. Instead, this mobilisation is driven in large part by mainstream and alternative media platforms. Furthermore, as our analysis of user comments indicates, the 'politics of division' over Brexit relates not just to the question of EU membership, but is primarily contestation over the nature of British identity and legitimacy of British democracy.

Politicization beyond Euroscepticism
To understand how the Brexit referendum created a deeply divided UK we need to theorize EU contestation beyond the framework of Euroscepticism. In the existing literature on EU politicization, it is divisions over identity that are expected to create a constraining dissensus that fundamentally changes the rules of the political game (Hooghe & Marks, 2009). Understood as a 'politics of identity', EU politicization is not just limited to times of crisis but denotes an axial transformation of European integration. While we share some of these concerns, we also believe that the existing accounts of EU politicization have focused too narrowly on the 'mainstreaming' of Euroscepticism as a force that undermines EU legitimacy (Startin, 2015). We argue that EU politicization remains incomplete if the opposing side, that of pro-European mobilisation, is not considered. Euroscepticism not only needs an external target (which is the EU), it also needs a domestic opponent, which is found in political actors who run pro-EU campaigns. Eurosceptic dominance as it happened before the referendum would be a case of incomplete politicization, while truly polarised debates with high levels of mobilisation by both opposing camps were only triggered afterwards. Following De Wilde and Trenz (2012), Euroscepticism can be understood as EU polity contestation in which Eurosceptic arguments about EU legitimacy are confronted in public discourse with their pro-European counterparts. Arguably, pro-Europeans before the referendum had an interest in de-politicising the EU in domestic politics (Fawcett et al., 2017). With the Brexit referendum, the dynamics of contestation changed as depoliticization for the losing side was no longer an option, requiring enhanced mobilisation by Remainers. We thus analyse politicization by looking at mobilisation and polarisation on both sides of the pro-/anti-EU divide.
However, we suggest that focusing solely on the pro-/anti-European divide misses an important dimension of EU politicization. Rather, the possible shift in polity contestation from the EU to the domestic level should also be considered. EU politicization has important repercussions on domestic politics (see e.g., Hoeglinger, 2016). EU debates contributed to delegitimizing democratic institutions and elites at the national level (Galpin & Trenz, 2019). Especially during Brexit, the effects of politicization cannot be properly understood if we do not take into account the way in which Eurosceptic mobilisation has fundamentally changed the rules of the democratic game domestically. Brexit campaigns only had a minor impact on EU technocratic governance, but a major impact on national-democratic government. It is not simply the legitimacy of the EU polity that is at stake, but the democratic constitution of the nation-state. EU politicization is therefore not a zero-sum game that weakens EU legitimacy while strengthening the arena of re-nationalised politics. EU politicization instead can seriously undermine the legitimacy of both EU and national politics. We would expect that this shift was promoted by pro-European mobilisation as a counter-movement to the dominance of populist mobilisation during the referendum debate (Roth, 2018). At the same time, this continuation of contentious politics at domestic level would imply even higher levels of polarisation leading to hostility and eliminating the in-between space where different alternatives could be developed. As we argue in the following, such polarisation effects are advanced by social media as an arena of contentious politics.

Social media as a site of politicization
Democratic deficits, a lack of political representation and low levels of trust in government and established political parties are often accompanied by the formation of protest and social movements trying to gain visibility in the public sphere (see e.g., Della Porta, 2020). This has been illustrated not just by progressive movements such as the Arab Spring, anti-austerity movements and Black Lives Matter, online far-right and white supremacist movements (Daniels, 2009) but also by Brexit. In the digital age, so-called online social movements have gained momentum, enabling broader mobilisation through less rigid forms of organisation and affiliation (McCaughey & Ayers, 2013). Digital activism results from, on the one hand, the efforts of existing social movements to use social media sites for the purposes of progressive, racist or anti-systemic protest (e.g., Della Porta, 2020; Dencik & Leistert, 2015). Improved communications and wider reach become, on the other hand, an opportunity to raise broader awareness of issues of shared concern and to assemble supporters of different backgrounds from the bottom up. These groups are loosely connected, but digital media technologies help make different levels of engagement possible, such as sharing opinions and engaging in forums, signing online petitions and participating in decision-making processes.
Social media, therefore, also results in the development of 'digital opinion movements', understood as online social movements that comprise 'spontaneous online mobilisations of the general/attentive public, which temporarily turns into an active public' (Barisione & Ceron, 2017, p. 81). The leadership vacuum following the Brexit referendum arguably extended the arena of politicization by widening opportunities for these informal, spontaneous and grassroots online movements. Citizens can spontaneously express emotion, react and respond quickly to events, reach high numbers of people and convey messages quickly and cheaply. They can even organise protest activities such as marches and street actions, thus enabling the formation of more traditional offline social movements without limiting them to their respective political systems (e.g., Bennett & Segerberg, 2013). Looking at social media sites of EU politicization is all the more relevant because social groups can connect different campaigns and networks on their 'side'. Social media is therefore not simply a side-arena of partisan contestation, but, as in the case of post-referendum campaigns, is used to bypass partisan politics by facilitating the emergence of informal and bottom-up digital movements. While British mainstream parties wished to de-politicize the EU membership issue, social media campaigns allowed for continued or even intensified mobilization.
However, scholars are increasingly doubtful that social media is a passive platform empowering progressive social movements. The architecture of social media platforms promotes polarized communities. Digital movements of opinion are understood to be cross-cutting and ad-hoc, distinguished from more organised, long-term social movements that mobilise particular social interests or groups (Barisione & Ceron, 2017, p. 81). Facebook architecture is well suited to this development of polarized communities beyond formal party politics, creating so-called 'filter bubbles' in which users communicate in ideologically segregated communities (Flaxman et al., 2016). Platform features (or affordances) contribute to this polarisation by standardising how people can communicate with each other (e.g., through requiring specific ways of reacting to posts). Facebook's features such as posting, buttons for liking and other emotional expressions, sharing, or commenting enable users to express political opinion and provide content that can be replicated and modified, searched, and subsequently be recorded or archived (Bond et al., 2012;Boyd, 2011, p. 46). Platform affordances are consequently not passive enablers of political mobilisation, rather 'they significantly contribute to structure modes of interactions and relationships' (Milan, 2015, p. 57). Via mediated, monitored and curated campaigns on social media, Facebook presented a practicable solution to assemble like-minded Leavers and Remainers in separate groups, to share content from pro-or anti-EU media sources and construct narratives about Brexit.
Increasing criticism has also been levied at social media companies who have failed to prevent the spread of especially far-right disinformation. The Brexit referendum and the election of Donald Trump in November 2016 were supported by powerful campaigners making use of new digital technologies such as social media bots, mass data-harvesting and targeted ads (see e.g., Bastos & Mercea, 2019;BBC News, 2018;Boczkowski & Papacharissi, 2018). The hybridity of media platforms also enables the spread of not just mainstream legacy media content but also 'alternative' sources as a way of mobilising people and promoting ideas (Chadwick, 2013). 'Dark money' originating with US billionaires such as the Koch Brothers was found to lie behind the pro-Brexit alternative media platform spiked (Monbiot, 2018). Millionaire Leave.EU donor Arron Banks 1 coowns Westmonster, a UK-based news site modelled on Steve Bannon's Breitbart which launched in 2017 (Rajan, 2017). Furthermore, the House of Commons Digital, Culture, Media and Sport Committee Report (House of Commons, 2019) found evidence of Russian interference in the referendum, with anti-EU stories by Kremlin-aligned media circulating online at the time. While acknowledging that the actual impact on voting choices is hard to measure, the report makes clear that social media platforms potentially allow democratic processes to be manipulated by powerful internal and external actors.
We argue that to understand Brexit in relation to a form of polity contestation not just over the EU but also over British sovereignty and democracy, we need to conduct more in-depth research into not only how official online campaigns and non-institutionalised actors and platforms used social media platforms to target individuals (such as, for example, Heft et al., 2017;Usherwood & Wright, 2017), but also how users engage in campaigning. Such a bottom-up perspective of social media politicization does not deny that division can also be steered by powerful interests. Social media campaigning sites are, however, also used for exchanging knowledge on a plurality of issues as well as for expressing opinion and emotions (see e.g., Fanoulis & Guerra, 2017). Through new information feeds and through emotions, the public sphere is kept in a state of alert, even if formal political action discontinues. In order to investigate EU politicization as a form of identity politics, we therefore need to understand community-building through exchange of information and opinion between campaigns and followers that contributes to the construction of distinctive identities and closed online communities (see e.g., Brundidge & Rice, 2009). It is from this perspective that we can analyse how social media movements contribute to social divisions.
We therefore understand pro-and anti-Brexit campaigning sites as networked movements of opinion which are formed in part through the design features that make Facebook a social network or platform provider. We analyse such features, comparing how anti-and pro-Brexit sites' use them and how they influence the dynamics of such groups as they become networks for spreading information, sharing and constructing opinions and mobilising over a common cause or identity. Through such dynamics, proand anti-Brexit pages can create their own content, post and share content that supports their common goals or collective identity and express critique. In the following, we outline our method and case selection.

Methods and case selection
Our data collection proceeded in two stages. The first stage of our analysis involved identifying the Facebook campaigns active during the period prior to the trigger of Article 50. We compiled a list of public pro-and anti-Brexit Facebook pages active during the two weeks from 24th January to 7 February 2017 marked by the Supreme Court's judgment that Parliament must vote to empower the Prime Minister to trigger Article 50 and the subsequent passage of the bill through parliament culminating in the European Union (Notification of Withdrawal) Act 2017. On these pages, we used a snowballing search technique through a variety of keywords such as 'Brexit', 'Leave', 'Remain', '48ʹ, 'Article 50ʹ, 'EU citizens', 'EU', 'Britain', 'Europe'. By following links to 'suggested' pages, we expanded this sample. Only those pages with a specific interest in the UK's membership of the EU or the EU referendum were included. We excluded pages with less than 1,000 likes, pages that were not active at the time of data collection in 2017, and party websites -with the exception of 'Labour Leave' as the only remaining active page in favour of a so-called 'Lexit' (left-wing exit from the EU). After a comprehensive list of Facebook pages was compiled, we collected data on page activity. Through Facebook's 'engagement' indicator, which combines numbers of page likes, reactions, comments and shares, we were able to identify the most active in a set of rough categories for anti-Brexit and pro-Brexit pages.
We sampled a mix of pre-and post-referendum pages, including pages of official or semi-official referendum campaigns as well as ostensibly 'grassroots' campaigns that have no obvious or formal link to official campaigns or parties. We also incorporated a mix of pages representing different ideological leanings and, for the anti-Brexit campaigns, what we refer to as 'special interest' pages (Table 1).
In the second stage of the data collection, we identified three key periods of two to three weeks during the first phase of 'uncertainty' for our data collection, outlined below in Figure 1: We collected page stats, posts and comments for all our selected pages (outlined below) during the key time periods. Looking at key time periods allows us to manage a large amount of data at moments that potentially trigger high degrees of user engagement and mobilisation. While this strategy does not allow for identificationof more spontaneous activity, we are interested in how the hybrid media system facilitates both top-down and bottom-up mobilization. The data were collected in February 2017, which means that the engagement we captured was contemporary and not the result of interaction over time. We developed a codebook in SPSS to analyse both posts and comments on pro-and anti-Brexit Facebook pages which we supplemented with qualitative coding in NVivo. In summarising our qualitative findings we do not provide specific information on page and date for comments. Although the data is fully anonymised and Facebook comments are not indexed on Google, providing such specific data risks individual names of commenters being identified through active searches.

Platform affordances for bottom-up politicization
In order to answer our first research question on how bottom-up politicization reveals a 'politics of division' during Brexit, we investigate how Remain-and Leave-affiliated campaigns on Facebook used platform affordances to express their opposing views. We find that while there is indeed some evidence of citizen-driven mobilisation through Facebook campaign networks, these movements are not necessarily 'grassroots' or adhoc. Instead, campaigns and online movements are to a significant extent guided by mainstream and alternative media sources. The extensive sharing of newspaper articles and dominance of 'guided opinion formation' over contestation suggests that the campaigns to some extent reproduce existing political and legacy media divides. 'Liking' and other reactions to posts, the abilities to share, create or modify content are all part of Facebook's platform affordances which help Leavers and Remainers to set up their own, separated campaigns with opposing views and maintain politicization over Brexit from the 'bottom up'. Figure 2 shows that pro-Brexit pages are generally more active in terms of engagement: the number of posts being shared (on other pages, private messages, or their own pages) by users following pro-Brexit campaigns exceeds the number shared via anti-Brexit pages, as is the case when it comes comments on these posts (the Leave.EU page reaches almost one million reactions in the three sampled time periods). This matches the pattern of the referendum where the social media landscape was dominated by the Leave campaigns (Heft et al., 2017). Yet, we also see a high level of engagement on the anti-Brexit side, particularly with posts from the Scientists for EU and Keep Britain in the EU pages, as well as the new 48 and Beyond page. The data shows that the post-referendum period involved continued politicization of EU membership Politicization from the 'bottom up' also took place through grassroots content networks. This can be seen in the types of platforms from which campaigns take their content. We distinguish between institutional platforms (such as legacy media), alternative media outlets and blogs, and social media. Table 2 outlines the types of platforms used as sources from which the content was shared. This is crucial, given that mainstream media now also have social media pages which connect news platforms directly with social media users. (e.g., through Facebook pages of The Guardian). We find that the most common platforms campaigns connect to are other social media pages, mostly on Facebook. This shows that Facebook features facilitate a high level of cross-posting from other pages on the same platform. These features also help campaigns to create their own 'content universes'. Given this, we also explore types of content shared via social media. Table 3 shows the social media content produced by journalists, other campaigns, experts, parties, government and social media users. We find that just under a third (anti-Brexit) and a quarter (pro-Brexit) of shared social media content was created by the campaign itself or by other campaign groups on the same 'side'. This finding demonstrates that, through platform affordances, campaigns are highly connected amongst themselves which contributes to building distinct communities of Leave and Remain supporters beyond the specific page in question.
There are, however, key differences in the extent to which the campaigns promote bottom-up mobilisation. Pro-Brexit pages refer to user-generated content (also coded when content appeared amateur or when the content creator was not identifiable) almost twice as much as anti-Brexit pages (23.5% vs 14.0%) ( Table 3). This confirms the tendencies of Brexit supporters to also focus their attention on more alternative sources  of opinion-formation and on appearing as grassroots movements. Despite this, pro-Brexit campaigns are less likely than anti-Brexit campaigns to use their pages for traditional collective action. Over a quarter of posts on anti-Brexit pages include explicit calls to action, for example, asking followers to write to their MP, join a protest, sign a petition, or share content on social media, in contrast to just one in ten posts on pro-Brexit pages. This demonstrates that anti-Brexit pages, to some extent, use their platforms for traditional grassroots social movement efforts alongside information-sharing and opinion formation. Such mobilisation efforts online likely entrenched polarisation over Brexit after the referendum, as its legitimacy was immediately contested on the Remain side. Pro-Brexit campaigns, however, drew on alternative media and linked extensively to other social media pages, but did not seek to transform this into genuine grassroots mobilisation in the same way. Politicization of Brexit through the communities of Remainers and Leavers cannot, therefore, always be considered simply 'bottom up'. We see that both sets of campaigns are not guided by formal party politics but by mainstream media and journalists. Table 2 shows that around a third of the platforms linked to by both sets of campaigns is mainstream media (forming the majority of 'institutional' platform types). We can also see that, furthermore, just under half of social media content is some form of (traditional or non-traditional/mimicked) journalistic output (see Table 3). News media are thus an important driver of mobilisation through these more informal Facebook pages after the Brexit referendum. This high level of mainstream and journalistic content indicates the hybrid character of news media, as professional journalists and news outlets have moved towards building audience loyalty via social media, establishing their Facebook presence (Chadwick, 2013, p. 47). In so doing, they allow users to engage with their content via the platform features of sharing, posting, liking and others.
We also observe the role of mainstream media in shaping distinct 'content universes' (Figure 3). The division between traditionally right-wing and pro-Brexit and left/liberal and pro-Remain news outlets is clearly demarcated. While anti-Brexit pages rely mostly on the left-learning pro-Remain The Guardian (39.5%) or BBC News (20.9%) (which is expected to remain politically neutral but is highly contested on both sides for being 'biased' to the other side), pro-Brexit pages share content from the pro-Leave Daily Express (28.9%) or The Telegraph (26.3%). We further find that pro-Brexit campaigns generally share more tabloid news than anti-Brexit campaigns. This difference is important given that British tabloids strongly favoured Leave, with the exception of The Mirror (Deacon et al., 2016). Furthermore, as Chadwick et al. (2018) show, the sharing of tabloid news on social media predicts democratically dysfunctional behaviour related to dis-/misinformation. The mainstream journalistic output shared via anti-Brexit and pro-Brexit Facebook pages thus to a large extent reproduces broader dividing lines within the wider media landscape and exposes followers to news opinion broadly in line with their existing positions.
Yet, we also find that pro-Brexit pages tend to favour connecting with alternative media platforms (23.5%) when compared with anti-Brexit pages (11.2%) (see Table  3). We see here that pro-Brexit campaigns link to 'alternative' or alt-right websites and blogs such as, for example, Arron Banks' Westmonster, brexitcentral.com and now defunct websites such as brexitbible.co.uk and yourbrexit.co.uk. With their dubious or opaque funding sources, the latter sites might be considered 'Astroturf', or fake grassroots organisations, that have wealthy funders and mimic journalistic or grassroots output (Bennett & Livingston, 2018, p. 129). Pro-Leave Facebook pages may also be investigated as 'Astroturf' organisations, for example, Get Britain Outone of the most active pro-Brexit pages we identified -is run by bookseller Julian Blackwell, who, according to a Channel 4 News report (Channel 4 News, 2013), gave UKIP its second largest individual donation since 2001, and Jayne Adye, who writes for Breitbart and the Daily Express. Such campaigns therefore lack authenticity through masquerading as bottom-up movements and hiding their elite support (Walker, 2014, p. 33). Anti-Brexit pages, when they do use alternative media, link to civil society platforms such as the New Europeans and Europa United. These differences between anti-and pro-Brexit pages seem to reflect the distrust of Leave supporters especially in institutional sources and the success of pro-Brexit platforms with opaque funding sources, backers and agendas in gaining social media traction. As such, the politics of division on social media is not only grassroots but can also be steered by powerful interests through not just mainstream media but also fake grassroots campaigns that receive funding by alt-right groups and companies.

EU or domestic polity contestation? Analysing posts and user comments
For our second research question regarding whether the campaigns are engaged in politicizing the EU or the national polity, we analyse issues raised by users in both camps and how they justify their opinions relating to EU and British democracy. We show how, while users do debate EU polity issues, they also contest the legitimacy of the British polity in the wake of the referendum shock. Both campaigns dedicate about half of their posts to Brexit issues specifically such as arguments for or against Brexit, the trigger of Article 50, or Brexit's impact on Britain. Anti-Brexit pages demonstrate greater mobilisation by raising campaign issues. Thus, as the campaign came to an end for the Leave side, a new campaign began for the Remain side. This is mirrored in the attention paid by pro-Brexit pages to general party politics (including questions about political parties and candidates in the UK, EU or internationally), as Brexit becomes a tangible reality and needs to be dealt with in UK, EU and international politics, particularly in the relationship to the US. Both campaigns also raised issues concerning immigration, social welfare, or foreign policy, which were however not always directly related to Brexit.
We further distinguish between three modes of justification for why particular issues raised in the main posts were relevant: problem-solving (utilitarian/efficiency), collective self-understanding (values-based), and justice/fairness (rights-based). Justifications were coded with regard to these three different understandings of Brexit as affecting our interests (utilitarian), our rights in relation to universal principles and/or fundamental norms or fairness, or our shared values and identity. For instance, posts that called for a stop to Brexit because of the expected negative impact on the UK economy were coded with a utilitarian justification, but posts that called for Brexit to be implemented in order to defend British national identity or the 'will of the British people' would be coded with a values-based justification, because it relates to justifications in relation to a bounded community. Posts addressing Brexit as a question of democratic legitimacy but did not include reference to a bounded community were coded as rights-based.  Figure 4 shows a large congruence in the use of justifications with campaigners on both sides focusing mainly on rights-based justifications. This emphasis on rights is even more pronounced amongst users in comments than in the main posts on Facebook, which indicates a process of bottom-up mobilisation of concerns with democracy. Nevertheless, 52.2% of the 209 posts and 54.7% of the 716 comments coded do not contain a justification as to whether rights, utilitarian or identity-related aspects underpin the message. The absence of justifications in around half of the posts and comments indicates a rather moderate level of persuasive, argumentative messages in the posts, mirrored by reactions in comments.
In order to understand the extent to which politicization refers to the EU and/or the domestic polity, we sub-sampled and analysed comments which specifically addressed the issue of the political and procedural aftermath of Brexit. Doing so resulted in 261 comments, 104 comments in anti-Brexit campaign sites and 157 pro-Brexit campaign sites. We paid particular attention to the themes that underlie their justifications. Table 4 provides an overview of the qualitative analysis of comments.
In both camps we find a strong tendency to engage with questions related to the domestic economy and British democracy. Indeed, the EU is not the main reference point. On anti-Brexit sites, commenters often use utilitarian justifications, balancing economic cost-benefit relations. Opponents of Brexit are generally concerned about the consequences of Brexit for the UK's economy, and implications for themselves and others in the labour market across different sectors (e.g., 'It is the hardest thing to understand why areas benefiting the most or with high profile success stories voted brexit. Wales when you read this. Sunderland with nissan. any clues'). Anti-Brexit commenters also express stronger concerns about the UK's political standing and its diplomatic and trade relations with other countries, mostly the US and the rest of the EU. Anti-Brexit commenters do not consider Brexit an appropriate solution for existing problems in the UK (e.g., 'You can't blame all of the woes of the NHS on Brexit! The NHS has been underfunded since God knows when!'). Many comments also deal with issues of efficiency and functionality, for example, seeing Brexit as a pointless exercise or believing there is no deal that could be better than the current arrangements. There is also a view that Brexiters believe in 'some magical utopian world . . . with prancing unicorns and fairies' rather than sound economic analysis. Regarding rights-based justifications, anti-Brexit commenters consider the referendum result to be based on too narrow a majority or to be non-binding, as one commenter maintains, 'referendums are never a "final" result because they are only advisory, albeit they can influence parliament towards a decision if they are overwhelmingly leaning towards one side'. In addition, this majority is considered to be the product of an either disinterested and/or uninformed majority of voters or of explicit disinformation involving 'deliberate lies' from fearmongering Leave politicians and Brexit campaigners, described by one commenter as a 'miscarriage of justice'. Another describes UKIP's campaign as 'waged with hatred at it's [sic] heart'. Others also speak disparagingly of Brexit voters, for example, 'May knows only the thickos voted for Brexit and those stupid idiots will believe any old rubbish, so she's saying "Brexit means Brexit" and doing nothing'. Finally, they also criticise politicians in the post-referendum period for failing to properly represent the '48%', perceived to be 'working on behalf of only the 52% rather than the interests of the whole country'. They are highly critical of MPs -especially Labour ones -preparing to trigger Article 50, for example, 'Tories are used to being turncoats but Labour MPs should be thoroughly ashamed of themselves'. In combination with arguments concerning political-democratic issues, comments in anti-Brexit groups tend to focus less on popular sovereignty, but rather on how to remedy the perceived illegitimate referendum result. As such, it transforms into contestation over British democracy and into values-based justifications that make claims on behalf of '48%' or 'remainers' as a new political identity.
Many pro-Brexit comments related to efficiency and functionality arguments often in relation to an understanding of the referendum as legitimate. They considered the political negotiations after the referendum as too long and complicated, asking to get Brexit done, e.g., 'And what is our government actually doing? Sitting on their hands hoping we will change our minds because everything the remoaners said is an absolute lie . . . ' This is interesting against the background of Johnson's 2019 electoral success with almost this exact slogan ('Get Brexit done!'). Furthermore, pro-Brexit commenters express the need to leave as soon as possible on the grounds that the EU is a failing organisation, for example, 'I dont think Europe needs to catch up . . . It needs to stop floundering and drowning itself', suggesting some element of EU polity contestation.
Generally, pro-Brexit comments understand the referendum as a legitimate and final decision and that democratic debate is illegitimate at this point after the referendum. The possibility of not implementing the result is expected to mean that, all of a sudden, democratic principles 'do not count'. Commenters support a model of direct democracy, for example, 'The highest court in the land has got it wrong we the people are sovereign body. Parliament and the members are there to do the will of the people'. Commenters often consider politicians as untrustworthy and corrupt when they speak out against Brexit or mention difficult negotiations, e.g., The UNELECTED House delaying the WILL of the people? Dare they? If they do, then even bigger change is coming and they WILL be swept away by it.
A delay to Brexit is considered a betrayal of the pure will of the people, 'The Snake MAY is planning to betray us all!!', with Remainers referred to as, for example, 'seditious traitors who should be stripped of public office', 'treasonous scum', or 'dangerous fascists'. Black Labour MP Diane Abbott, who abstained in the vote, furthermore appears in abusive or dehumanizing racist comments. Pro-Brexit commenters, far more than Remainers, also express a lack of trust in media organisations such as the BBC, which they consider to have a Remain bias, 'the BBC this morning with yet another panel loaded with Remainers banging on about a second referendum, it's time they were given a hefty kick up the backside'.
Some EU politicization is present through pro-Brexit comments relating to values and identity associated with Britishness or Englishness in which they claim exclusive national identities that exclude the European or supranational level in their defence of national sovereignty for example, 'Lets get back to English politics and rule first. Such a defence of Britishness often also involves defending the nation from immigrants (and by extension the EU) using racist language, for example, speaking of the current state of Britain as including run down NHS, dirty streets, litter strewn verges, street lights off, ideal for muggers, burglars, and murderers, which we now have plenty of disguised as so called refugees, a country now with thousands upon thousands of legal and illegal economic immigrants, every one a potential terrorist, a race of people invited into our country by our traitorous leader, and forced on us by the EU.
Our findings show that the EU is rather marginal in these debates on both sides. We find evidence for strong polarisation and contestation over Brexit as a national-democratic rather than EU issue, paired with political frustration, yet for opposing reasons. Our qualitative analysis also points to a fundamental disagreement about how the decision to withdraw from the EU should be understood. Anti-Brexit commenters consider this decision to have been reached through undemocratic means and electoral manipulation. Pro-Brexit commenters, however, are concerned with the long negotiations, which they tend to consider as sign of political corruption, as an intentional delay to Brexit by government and opposition representatives.

Conclusion
Our analysis found evidence of a social movement-driven 'politics of division' over Brexit beyond party politics in the immediate aftermath of the referendum. These movements utilised new forms of social media campaigning and -in the case of anti-Brexit campaigns -incorporated calls for traditional collective action. Important drivers of these movements were, however, not always 'grassroots' but mainstream and alternative media platforms, in some pro-Brexit cases backed by entrepreneurs from the wealthy antiestablishment.
Through the concept of 'platform affordances', we describe such new forms of online mobilisation and the creation of grassroots and informal pro-and anti-Brexit communities. This mobilisation has resulted in the development of movement identities for the two opposing camps through campaign names (for example, those referencing the 48%), connecting with other like-minded online communities through shared content, as well as low-key political engagement through 'liking' and 'reacting'. The transition from online activism and 'connective action' to more traditional collective action, such as street protests or writing petitions to MPs, is also present primarily amongst Remainers. Further research is needed to consider the ways in which mobilization on social media connects with this more traditional civic engagement and mobilisation. Nevertheless, we expect that the large-scale anti-Brexit demonstrations witnessed since the referendum were primarily organised bottom-up via many of the social media pages we have analysed (see also Brändle et al., 2018).
The finding that EU-related campaigns transform into nationally focused collective action highlights a further contribution to the literature on EU politicization. The idea that EU issues transform into national ones through politicization can go some way to explain Dolezal et al's finding that mass-level EU politicization remains limited without increasing over time (Dolezal et al., 2016). Through a social movement and social media perspective, we show how post-referendum mobilisation involves redirecting the campaigning target. Facebook pages are used by both sides to spark debate not necessarily about the EU, but rather about the future of British democracy. While anti-Brexit commenters consider the referendum illegitimate, pro-Brexit commenters understand complicated negotiations as a delaying tactic by political elites. This demonstrates the way in which Brexit divisions are predicated not solely on EU membership, but on a wider conflict over democratic representation. The main concern that drives post-Brexit politicization is therefore the deprivation/assertion of rights and the future of democracy, which shows how debates are turned inwards and not solely about the UK's external relations in Europe. Against these findings, our research contributes to the understanding of the dynamics of social media mobilisation during heightened EU politicization in important ways. We observe that the political division over Brexit across ideological camps has transformed into a social and identitarian struggle across society.
For social movement studies, it is also important to emphasize that a neat distinction between bottom-up and top-down mobilisation cannot always be made. Content shared on Facebook pages takes a hybrid form, partly created by grassroots campaigners and partly by legacy and alternative media platforms. Furthermore, suspected 'Astroturf' content appears on ostensibly 'unaffiliated' social media campaign pages with little transparency regarding support and funding. Pro-Brexit campaigns and websites suspected of being financed by wealthy donors and potentially dark money 'borrow the repertoire of grassroots mobilisation' (Walker, 2014, p. 9). The post-referendum political vacuum might thus be filled by powerful influencers outside the sphere of formal party politics. Recent elections have been influenced by dubious financing structures, dark money, astroturfing and 'bot armies', thus putting grassroot movements at risk and threatening the further marginalisation of grassroots voices. The implications of 'astroturf mobilisation' on social movements, democratic politics, and civic participation need to be further investigated. We thus call for research into the way in which top-down interests meet with mass participation on social media, potentially in collaboration with computer scientists, security studies or investigative journalists, in order to explore the origins particularly of far-right or 'alt-right' social media campaigns. Nevertheless, through the case of Brexit, we have highlighted the importance of considering the hybrid media landscape and elite influence when exploring political mobilisation on social media.