Digital nationalism: Understanding the role of digital media in the rise of ‘ new ’ nationalism

While digital technologies were initially seen as harbingers of globalisation and cosmopolitanism, scholars increasingly acknowledge their role in the rise of nationalism and right-wing populism. Yet this surge of interest leaves at least two important questions unanswered. Where was nationalism before its apparent resurgence? Are contemporary forms of nationalism different from their predecessors, and can these changes be linked to digital technologies? To answer these questions, we argue for the importance of understanding the less visible ways in which digital technologies reproduce our sense of belonging to a world of nations. We discuss three such mechanisms: the architecture of internet domains, the bias of algorithms and the formation of national digital ecosystems. Next, we examine three characteristics of contemporary nationalism that can be partly linked to recent shifts in the global communication ecology: diversification, fragmentation and commodification. We conclude by considering the implications of our arguments for future research in the field. different from their analogue predecessors. We discuss three such qualitative changes: diversification, polarisation and commodification.

First, the participatory affordances of digital technologies have enabled a wider range of actors to contribute to public communication, thus making national imagination and nationalisms potentially more diverse but also more unpredictable. Second, this diversity had gone hand in hand with a greater fragmentation and polarisation of national imagination, with individuals and groups clutching to 'niche' versions of national identity, as well as mobilising more extreme forms of nationalism. Third, the prevailing political economy of digital platforms means that, in the online sphere, nations are increasingly imagined and communicated as communities of consumers-a trend that further enhances the already evident growth of economic and consumer nationalism, and the related practice of nation branding. In the second section of the paper, we discuss each of these changes drawing on a range of examples from existing research and our own observations. It is important to clarify that these changes cannot be reduced to technological shifts alone. Contextual factors in Europe, China, the United States and elsewhere all play an important role, and digital media can be and are used in ways that defy the logic of nationalism. Nonetheless, there is enough evidence to argue that changes in the communication ecology are sufficiently important in their own right to merit concerted attention. We conclude the article by considering the implications of our arguments for future research in the field.

| DIGITAL MEDIA AND THE RISE OF 'NEW' NATIONALISM
Existing research advances a range of explanations for the rise of nationalism and the associated retreat of liberalism, including demographic shifts, economic insecurity driven by the combined effects of economic recession and antiausterity politics, the jawing ideological rift between the winners and losers of economic and cultural openness, and the resulting 'cultural backlash' among growing segments of the general population (e.g., Eatwell & Goodwin, 2018;Hacker, 2006;Kaufmann, 2018;Norris & Inglehart, 2019;Piketty, 2014). Operating in different ways, these changes have arguably created fertile grounds for the proliferation of anxieties about national security, for the spreading of hostility against foreigners and for the popular embrace of the 'our nation first' rhetoric as a guarantor of stability.
Intertwined with these changes, the role of media and communications in the proliferation and appeal of right-wing populism and nationalist rhetoric is attracting significant scholarly attention as well. As several authors have noted, digital media-and social networking platforms, in particular-have allegedly encouraged the fragmentation of public debate by means of creating algorithm-driven 'filter bubbles' (Cardenal, Aguilar-Paredes, Galais, & Pérez-Montoro, 2019;Pariser, 2011) and 'echo chambers' (Dubois & Blank, 2018;Dutton, Reisdorf, Dubois, & Blank, 2017), within which people are selectively exposed to views that conform to their existing attitudes and beliefs. In the eyes of some observers, this has not only led to a fragmentation of the public sphere (Bennett & Pfetsch, 2018) but also encouraged growing disrespect for different views, tolerance of exclusionary nationalist rhetoric and hate speech more generally, as well as proclivity to extremism (Sunstein, 2017). At the same time, the participatory nature of social media, combined with the lack of editorial oversight and fact-checking procedures on social media platforms, has made it easier for individuals and groups to communicate extremist views and misinformation that 'support the spreading of far-right [nationalist] ideology online' (Fuchs, 2020, p. 248; see also Nichols, 2017;Rojecki & Meraz, 2016).
For nationalism scholars, as well as for media and communication scholars with a more long-standing interest in nationalism, these arguments will not come as a surprise. Almost two decades ago, Halavais (2000) demonstrated the persistence of national borders on the World Wide Web, showing that the vast majority of US websites link to other websites hosted in the United States rather than crossing national borders. Writing in this journal, Eriksen (2007) likewise argued against the idea that the non-territorial character of internet would make it impossible to maintain a shared sense of national identity and instead drew attention to evidence from a wide range of nations-from the Tamils of Sri Lanka to nations with large diasporas overseas-that suggested that 'the Internet has in the space of only a few years become a key technology for keeping nations (and other abstract communities) together'. Conversi (2012) pushed this argument further and pointed to the multiplicity of uses that the internet enables-from fostering unprecedented levels of global connectedness to imposing exclusion, erecting new intercommunal barriers and promoting aggressive patriotism and xenophobia. Works on media and nationalism have likewise highlighted the persistent national architecture of the digital world and highlighted the diverse uses of social media platforms, ranging from exemplary cases of cosmopolitan, cross-cultural engagement and the revival of minority languages to far more exclusive, aggressive and virulently nationalist interventions (e.g., Mihelj, 2011, pp. 37-38, 183-184). The link between nationalism and digital media beyond the West, especially in China, has also attracted significant attention well before the recent rise of interest in digital nationalism and has by now generated a significant body of research (e.g., Jiang, 2012;Liu, 2019;Schneider, 2018;Wu, 2007).
Yet the nature of the debate has certainly shifted: while the existence of digital nationalism used to be discussed only among a small circle of experts in media and nationalism, or with reference to cases beyond the West, it has now become a widely debated phenomenon, acknowledged by mainstream media and communication scholars, and of relevance to major Western powers. Why this sudden shift? The immediate answer to this question is obviousthe victory of Donald Trump in the US presidential election and the success of the Leave campaign in the United Kingdom's EU referendum, as well as the growing influence and popularity of nationalist rhetoric and right-wing populism around the world, have turned nationalism into a palpable political force that is no longer restricted to the margins of the political spectrum, or limited to seemingly remote corners of the globe, but has taken pride of place in the very centre of our politics, societies and culture. This apparent resurgence of nationalism demands explanations and a popular way of explaining the phenomenon has been to describe it as 'new' and as somehow fundamentally different from earlier forms of nationalism-for instance, because of its emphasis on cultural as opposed to biological markers of nationhood or due to the centrality of anti-immigrant and antiglobalisation rhetoric (Fuchs, 2020;Goodhart, 2017;Kaufmann, 2018). Yet, as Maleševi c (2019, pp. 6-9) argues, explanations that emphasise the 'new' elements of nationalism tend to lack an engagement with long-term historical developments that have shaped nationalism and are underpinned by the erroneous perception of nationalism as a transient phenomenon that will eventually disappear. It is worth noting that this perception of nationalism as transitory is far from new. As Chernilo (2007) showed, already Karl Marx thought that nation states were a transitory phenomenon associated with capitalism. Labelling nationalism 'new' therefore 'does not stand for a novel phenomenon but for the analyst's surprise that nationalism has not gone away' (Maleševi c, 2019, p. 6). Arguably, the surge in interest in nationalism among communication and media scholars stems from the same misconception and a similar surprise at the sudden resurgence of nationalism. This is not to say that there is nothing new about 'new' nationalism at all but rather that the discussion of what is new needs to be more carefully situated alongside long-term developments.
To move away from this misconception of nationalism as a transitory phenomenon, we need to reflect on the way we understand nationalism. There is no scope or need here for a detailed engagement with the long-standing debate on the definition of nationalism (see, e.g., Hearn, 2006;Özkırımlı, 2005, for overviews), but it is important to acknowledge that nationalism can be defined in at least two distinct ways. The first, more conventional understanding (particularly common in popular or political discourse) focuses on clearly evident, often fairly extreme manifestations of nationalism, such as separatist movements, institutionalised forms of exclusion along national lines, or xenophobia and right-wing extremism. This conventional understanding is also common in several of the classic studies of nationalism (e.g., Gellner, 1983;Smith, 1991) and is adopted in the majority of the literature on digital nationalism quoted so far (but not, we should add, in work published in this journal). The second view approaches nationalism from the perspective of everyday life, mundane practices and taken-for granted assumptions and defines it as an ideology or as particular way of seeing the world and acting in the world-a way that assumes that humanity is self-evidently divided into nations and that the world is naturally divided into separate national homelands (Billig, 1995;Fox & Miller-Idriss, 2008;Maleševi c, 2019;Skey & Antonsich, 2017). This definition of nationalism is explicitly adopted only in a small minority of existing literature on digital nationalism (e.g., Soffer, 2013;Szulc, 2015) and is almost absent in studies that focus on the recent rise of nationalism in the online world. As a result, the existing discussion of digital nationalism is conceptually bifurcated, with one (larger) body of literature adopting one definition of nationalism and the other (smaller) body adopting another. This bifurcation prevents us from understanding how the two modes of nationalism relate to one another and hence prevents us from fully understanding the role of digital media in recent wave of nationalism.
For the purpose of clarity, it is useful to think of the difference between the two understandings of nationalism outlined above in terms of difference between form and content. 1 Nationalism in the banal and everyday sense is a stable and taken for granted element of social life; it provides the basic 'forms' for thinking about the world around us. In contrast, nationalism in the first, conventional sense is about the actual 'content' of the nation-about the distinguishing characteristics or identity of a nation, about who belongs to it and who does not, about what the national interests are and so forth-and about the uses the national 'form' and 'content' are put to. This nationalism is not evident all the time-rather, it goes through ebbs and flows and becomes evident only in extraordinary moments of collective celebration, conflict and war, or during major changes and disruptions such as regime changes, elections, natural catastrophes or economic recession. Such extraordinary moments temporarily upend the taken-for granted forms of our national lives and invite contestations over the 'content' of our nationhood-over our national identity and boundaries and over our nation's past, as well as its future. The rise of nationalism we are currently experiencing is one of such extraordinary moments. If we want to understand how the digital environment might potentially be affecting this extraordinary moment, and moulding the contestations over national 'content', we first need to understand how it shaped the national 'form' that underpin these contestations.
driven by a combination of factors, ranging from the ownership structure and political economy of individual digital platforms (for instance, the extent to which they are governed by commercial imperatives or specific political agendas) to social and individual factors such as personal search preferences and interests, demographic factors and social networks we are part of. In short, search engines are far from neutral 'windows' onto the world-much as any other medium, they act as filters that conform to a range of commercial, political and socio-demographic factors and therefore offer rather selective visions of the world (Gillespie, 2014). The algorithms guiding some of the most widely used digital platforms have been shown to reproduce established gender, racial and ethnic biases-for instance, a study of online ads showed that women were considerably less likely to receive ads for high-income jobs (Datta, Tschantz, & Datta, 2015), while another study showed that searching for names associated with African Americans generated a higher proportion of adds that suggested arrest (Sweeney, 2013).
Evidence of nationalist bias is more limited but nonetheless instructive. The Chinese case, most systematically examined by Schneider (2018), offers a good example. For instance, a parallel search for the Chinese words for 'Tiananmen' and the 'Gate of Heavenly Peace' produced rather different results if one used Baidu or Google; while Google provided links to generic information about the square as well as links to information about the Tiananmen Square 1989 protests, a search on Baidu conducted at the same time did not produce any links to information about the 1989 events at all (Schneider, 2018, pp. 68-69). The choice of language, script and geography appearing in such searches is noteworthy as well-China's search engines tend to return sources in the simplified Chinese script that dominate in mainland China and not the traditional script that is used in Hong Kong, Taiwan and among the Chinese diaspora overseas (Schneider, 2018, p. 77). In sum, different search engines carry different biases and act as gateways to rather different imaginings of the same nation. In the Chinese case, the biased digital windows provided by search engines 'guide perceptions in subtle, often banal ways towards a particular vision of Chineseness, for instance by privileging authoritative domestic sources, reproducing simplified character materials, or flagging the mainland Chinese context through national services like weather and travel applications' (Schneider, 2018, p. 79).
Arguably, the architecture of top-level domain names and the algorithmic nationalism combine to form what is best seen as network of national digital ecosystems. These national digital ecosystems are of course interdependent and often rely on the same digital moulds but nonetheless offer citizens rather different experiences of the world online. Although companies and platforms such as Google or Facebook emphasise their global credentials, both by potentially offering unprecedented reach to information across the world and by seemingly ignoring or bypassing national structures (van Dijck, Poell, & de Waal, 2018), in practice, they have adapted to the national architecture of the internet and have developed nation-specific-and often incompatible-versions of their products, as in the case of Google or Amazon, both of which automatically direct users to a version of their website that is subordinated to the country-level top domain (e.g., www.google.ac.uk in the United Kingdom or www.google.com.br in Brazil). These national versions produce different search results depending on which national version of the platform is used, thus contributing to the formation of a national digital ecosystem. Furthermore, due to commercial and/or legal reasons, particularly in the case of audiovisual content, users cannot easily switch from one version to another, therefore effectively building clear digital national boundaries between them (Stewart, 2016). This ecosystem is also supported by the webs of hyperlinks between websites; as existing research demonstrates, the majority of websites link to other websites in the same country or even direct users to other content provided by the same website or affiliated providers (Bharat, Chang, Henzinger, & Ruhl, 2001;Dimitrova, Connolly-Ahern, Williams, Kaid, & Reid, 2003;Halavais, 2000;Schneider, 2018). Even where hyperlinks transgress national borders, the influence of national histories and established geopolitical links is clear-a good example being Brazil's hyperlinks to Portugal and Portugal's to Brazil (Bharat, Chang, Henzinger, & Ruhl, 2001, p. 5).
National digital ecosystems vary considerably in the extent to which they are intertwined with other national ecosystems and in the extent to which they rely on the transnational digital infrastructure provided the 'Big Five' digital corporations, all of which are based in the United States (i.e., Google, Amazon, Facebook, Apple and Microsoft). Examples of more independent, self-sufficient national digital ecosystems can be found in countries that have developed digital platforms that rival the dominance of the 'Big Five'. The most successful among such alternative ecosystems is the Chinese one, which encompasses successful equivalents of all major Western platforms, ranging from the search engine Baidu and social networking platforms Weibo, WeChat and Tencent QQ to e-commerce sites such as Alibaba, Taobao and JD. The establishment of these platforms was often explicitly guided by nationalist considerations and concerns over national sovereignty. As Plantin and de Seta (2019, p. 1) show, the development of the WeChat platform model was shaped 'by markedly techno-nationalist media regulations and an increasingly overt cyber-sovereignty agenda'. While China represents by far the most complete, self-contained and successful national digital ecosystem, it is worth noting that other countries have at least attempted to develop distinct national digital platforms. Many of such platforms ultimately succumbed under the pressure of rivals; for instance, the Polish social networking sit Nasza Klasa (Our Class), which was used by over 50% of Polish internet users in 2008 (Malachowski, 2009), has since been displaced by Facebook. In other cases, nation-specific alternatives continue to fare well despite competition; in Russia, for instance, social networking sites Odnoklassniki and VKontakte still maintain their popularity over Facebook and Twitter.
Two caveats are important. First, while significant, the role of digital platforms in the everyday reproduction of the world as a world of nations should not be overstated. The preservation of the national structures of the digital environment is also a consequence of the fact that states remain significant actors that delineate the national legal and linguistic frameworks in which these companies and platforms operate (Kohl & Fox, 2017). Second, although digital media inevitably reproduce nationalism in its banal, everyday sense, they do not necessarily foster an exclusive and hegemonic form of nationalism. Indeed, in some cases, digital media can actively contribute to undermine socially dominant discourses or promote alternative forms of national imagination. As we argue in the next section, one of the distinctive elements of digital nationalism, which differentiates it from its predigital counterpart, is precisely its openness to a multiplicity of voices and hence to a diverse range of imagining and enacting the nation. Consequently, it is important that we do not restrict our view only to its most exclusive, aggressive or violent manifestations and instead also consider how they interact with more inclusive, politically progressive and liberal forms.

| HAVE DIGITAL MEDIA CHANGED NATIONALISM? DIVERSIFICATION, POLARISATION AND COMMODIFICATION
Having considered the ways in which digital media contribute to the reproduction of nationalism on an everyday basis, we now turn our attention to our second line of enquiry, namely, the nature of digital nationalism itself. Have digital media changed nationalism? We examine three sets of characteristics of contemporary nationalism that can be linked, in part, to digital technologies: participation and diversification, fragmentation and polarisation, and commodification and commercialisation.

| Participation and diversification
A key change brought by digital technology can be conceptualised in terms of greater participation, diversification, or even democratisation, all this without necessary challenging the national form or framework discussed earlier.
Classic works on the media and nationalism (e.g., Anderson, 1991;Deutsch, 1966) have often stressed the role of the media as centripetal forces that create a forum where dominant imaginaries, narratives or myths about the nation are developed, sustained, renewed and contested. These arguments have however been developed by studying 'old' or 'legacy' media such as newspapers, film, radio and television, which facilitated the construction of shared 'media events'-supporting or trying to impose particular readings of what the nation is or should be (Dayan & Katz, 1992).
Although audiences may effectively read mediated representations of the nation in different and contradictory ways (Madianou, 2005), their chances of producing and disseminating alternatives accounts about the nation with old or legacy media are limited.
The participatory design, ease of use and (relatively) low cost of new digital technologies have nonetheless produced an important shift, with a far greater number of actors contributing to public communication, hence making their views or their understanding of themselves, at least potentially visible to the public. This shift also led to a blurring of the boundary between media producers and media consumers, turning us all into 'prosumers' or 'produsers' (Bruns, 2008;Jenkins, 2008). This has then led to a marked diversification of voices visible in the public realm and to a kind of democratisation of public communication. Digital technologies can in fact complement the consumption of 'legacy' media, such as when audiences live tweet about television shows, a practice that contributes to the online reproduction of the national form (Stewart, 2019). The participatory design and accessibility also mean that voices that were previously stifled or censored could now more easily reach the public eye, turning digital media into tools of political opposition and resistance in authoritarian regimes (Cammaerts, 2018).
A significant consequence of this shift is the fact that digital media have weakened state attempts to sustain a dominant version of national identity. Scholars seeking to examine digital nationalism have noted the impact of participatory designs and the democratising effects of digital technologies, with more actors having the potential of constructing and communicating alternative versions of national identity. Writing about the case of Pakistan, Kalim and Janjua argue that 'the common man' (sic) is currently a more significant actor in nationalism, given that digital media facilitate 'the construction and dissemination of democratized bottom-up discourse' (2019, p. 69). Similar arguments have also been developed in studies about Russia and Central Asia (Glukhov, 2017), Iran (Yadlin-Segal, 2017) and Brazil (Eakin, 2017). Notably, some states have themselves embraced the more participatory nature of social media and sought to use it as a means of encouraging a more diverse sense of national belonging. A widely publicised initiative was the Twitter @Sweden, a project financed by the Swedish government and presented by the authorities as 'the most democratic Twitter account', which each week put a different Swede in charge of it. The campaign intended to embrace national diversity by showing that 'not one Sweden is conveyed, but several' (Christensen, 2013, p. 32).
However, we should be wary of assuming that the participatory affordances of digital media turn the new communication ecology into an even playing field for all actors involved or that the era of common national narratives is over. First, we should acknowledge that not every individual is equally able to take advantage of digital technologies and hence contribute to the creation of more diverse, bottom-up landscapes of national imagination participate in the digital nationalism. As the literature on the digital divides has shown, some segments of the population, and some parts of the world, are more likely to benefit, not simply due to lack of access but even more so due to the lack of skills and knowledge that are required to make use of digital technologies (DiMaggio & Hargittai, 2001). Second, as already noted, digital platforms are not neutral intermediaries, even though their owners tend to emphasise their neutrality and thereby seek to escape regulation (Gillespie, 2010). Rather, all digital platforms are involved in processes of content moderation and selection, which determine the kinds of choices we are offered online. As we have seen, search engines introduce important preferences into the information we have access to online and foster particular versions of the nation. The same is true for social networking platforms that individuals use to disseminate narratives about themselves and the world. Decisions over interface design and functionalities are beyond the control of individual users and yet exert profound influence on the kinds of identities, representations and narratives that individuals can construct online. Rather than seeing social media as stages for creative self-expression and online identity formation, they are best seen as 'sites of struggle between users, employers and platform owners to control online identities ' (van Dijck, 2013, p. 199).
We should also keep in mind that the ability to voice an opinion online does not mean that all opinions are also listened to, let alone being acted upon (Couldry, 2009). While individual users can indeed create a range of different imaginings of the same nation, their capacity to disseminate these narratives effectively, and turning them into shared anchors of identity, is limited-despite participatory designs, corporate and state players in the digital arena command much greater power than individual users, and are hence better equipped to promote their own narratives of national belonging. The tweets and posts of established political and economic actors, or those with greater means to invest in online promotion, will get greater visibility, and earn more reposts and retweets, than those by average citizens. The sense of empowerment of national masses through digital media is therefore to an extent illusory-the empowerment may well mean simply the ability to repost a tweet from a nationalist political leader (Fox & Miller-Idriss, 2019). Even cases of apparent democratic national participation, such as the Twitter account @Sweden, were in practice rather restricted. Curators received specific guidelines on what they could post, and some commentators questioned the supposed 'representativity' of those chosen to tweet on behalf of Sweden (Christensen, 2013). The apparent 'power' of average digital media users is therefore rather limited. While many actors can post content online, and thereby promote their own vision of the nation, they can only do so within the parameters of specific digital infrastructures and their affordances that are available to them and with little guarantee of being listened to.

| Fragmentation and polarisation
While the discussion of the participatory nature of digital nationalism and its democratising potential is often infused with a sense of optimism and reflects the perception of digital technologies as forces for the good, more recent work tends to adopt a more sober, if not outright negative tone. As noted in Section 1, the bulk of recent research on digital nationalism focuses on the role of digital media in facilitating the rise of right-wing populist parties and leaders and in fostering more exclusive, even aggressive forms of nationalism (Fuchs, 2020;Lim, 2017;Pajnik & Sauer, 2018;Rao, 2018;Waisbord & Amado, 2017). Yet we should be careful not to replace the naively optimistic vision of the digital media as a progressive force conducive to cosmopolitan connections, with an equally one-sided pessimistic vision, which sees digital platforms as inherently parochial, and prone to amplifying exclusive visions of national belonging. Rather, we suggest that the digital communication infrastructures might be facilitating a growing fragmentation and polarisation of nationalism, leading to a disjunction between, on the one hand, the more liberal, inclusive forms of national belonging, which are open to diversity and immigration, and on the other hand, the more exclusive, locally grounded forms of attachment that emphasise the importance of ethnonational roots as a key criterion of national belonging. An important driver of this polarisation is, arguably, a more profound and affective individual engagement with nationalism; in a digital environment, individuals and groups can form a myriad of 'niche' or 'bespoke' nationalisms that can become objects of heavy personal investment (Fox & Miller-Idriss, 2019). While such affective investment may have positive consequences, as it may help encourage and sustain engagement with public debate, it could potentially also fuel divisiveness.
The link between the changing communication environment and political or ideological polarisation has been widely debated among communication and media scholars. Although selective exposure-namely, the tendency of audiences to follow sources congruent with their political attitudes-has a long history (e.g., Stroud, 2011), several authors have argued that the phenomenon has become more acute and widespread in the online environment, leading to an unprecedented fragmentation of the public realm and the rise of 'filter bubbles' or 'echo chambers' within which people are only exposed to material they like or agree with rather than benefitting from exposure to a variety of different perspectives (Pariser, 2011;Sunstein, 2017). These 'bubbles' or 'chambers' have likely been strengthened by political economic decisions, such as recent changes to the Facebook algorithm that emphasise posts by 'friends' over those of publishers and news agencies (Cornia, Sehl, Levy, & Nielsen, 2018).
In the more pessimistic interpretations, this has led to an emptying out of a shared public space where different opinions are confronted and contested, as well as to a growing polarisation of public debate, where more extreme opinions get greater visibility and support than then would have in an offline media setting. Recent cases like the election of Jair Bolsonaro in Brazil appear to support the fragmentation and polarisation argument. Bolsonaro's campaign was conducted primarily through social media, and their structures facilitated the creation of closed groups supporting him. Users were often invited by relatives, friends or acquaintances to join WhatsApp groups in support of Bolsonaro, and in these closed digital 'chambers', illiberal views about Brazil and misinformation about other candidates were shared, and opposing voices were silenced (Davis & Straubhaar, 2020;Nemer, 2018). Nonetheless, it is important to note that empirical evidence remains inconclusive and that some studies have suggested that digital media use can in fact counter polarisation or showed that the polarising effects of digital media use may be limited to rather small segments of the population (e.g., Dubois & Blank, 2018; see also Van Aelst et al., 2017, for an overview).
How exactly selective exposure and polarisation relate to nationalism is even less clear, largely because, as noted earlier, existing research approaches nationalism in a way that reduces it to its more exclusive, hate-filled variants.
For instance, research on populist communication often acknowledges the centrality of anti-immigrant attitudes in populist ideology and communication (e.g., de Vreese, Esser, Aalberg, Reinemann, & Stanyer, 2018;Engesser, Ernst, Esser, & Büchel, 2017) or investigates the impact of right-wing populist discourse on anti-immigrant attitudes in different countries (e.g., Van Klingeren, Boomgaarden, Vliegenthart, & Vreese, 2015) but pays less attention to how the communication environment may contribute to polarisation between those favouring or opposing immigration-and hence between citizens who embrace different visions of national belonging and nationalism. Even when research explicitly focuses on the link between polarisation and national identity, there is little attempt to distinguish between different types of national identification or acknowledge that the same nation can be imagined in more or less exclusive ways. For instance, one study conducted in the United States found that exposure to media coverage that makes one's national identity more salient tends to increase affective polarisation among those already opposed to immigration, making them even more likely to associate immigrants with negative traits and to report colder feelings towards them (Wojcieszak & Garrett, 2018). However, this effect did not exist for supporters of immigration-a result that may suggest that their conception of nationhood is more open to diversity. Indeed, a rare study that explicitly focuses on the link between polarisation and different versions of nationalism among US citizens found 'strong evidence of persistent partisan sorting of nationalist beliefs over the past two decades', with different varieties of nationalism becoming increasingly connected with different party preferences (Bonikowski, Halikiopoulou, Kaufmann, & Rooduijn, 2019, p. 2). Specifically, while different conceptions of the American nation had cut across party affiliation in 1996, the situation has changed considerably in 2016; Republican Party supporters have become predominantly ethnonationalist, while whose identifying with the Democratic Party largely came to embrace what the authors of the study describe as a more 'disengaged' conception of nationhood (Bonikowski, Halikiopoulou, Kaufmann, & Rooduijn, 2019, p. 2). While this study did not examine the role of the changing media environment in this context, it is feasible to suggest that the growing reliance on digital media as sources of information may have been a contributing factor in this 'partisan sorting' of nationalist attitudes.

| Commercialisation and commodification
As noted earlier, we should be wary of assuming that the participatory affordances of digital media turn the new communication ecology into an even playing field for all actors involved or that the era of common national narratives is over. We have highlighted the persistence of hierarchical power relationships in the digital communicative environment and argued that different versions of nationalism found online are restricted by the parameters of digital infrastructures and their affordances. In this section, we go a step further and argue that these parameters tend to be underpinned by a rather specific, narrow nationalist vision, which sees the nation primarily as a community of economic actors interested in selling and/or consuming particular skills and products, rather than as a community of citizens with a variety of cultural, political, and social interests and needs.
Let us immediately add that there is nothing intrinsic to digital communication technologies that makes them prone to commercialisation and commodification. Rather, these traits are a result of the political economy of digital platforms, and specifically the fact that they operate as commercial entities, designed to make profit. According to van Dijck, Poell, and de Waal (2018), commodification constitutes a key mechanism that the currently dominant digital platforms such as Google, Amazon and Facebook inject into social interaction. In their analysis, commodification 'involves platforms transforming online and offline objects, activities, emotions and ideas into tradable commodities', which 'are valued through at least four different types of currency: attention, data, users and money'. This is a significant observation, not least because many of the institutions that are most profoundly affected by digital platforms-ranging from broadcasting and libraries to educational and healthcare institutions-are often either publicly funded and designed to serve the public good, as in most of Europe (van Dijck, Poell, & de Waal, 2018, p. 30) or controlled by the state or the governing political elites, as in China, Russia, Hungary, and other authoritarian or semi-authoritarian countries. Depending on the sociopolitical context, the spreading of digital platforms is thus likely to contribute to a growing influence of economic (commercial and consumer) nationalism over other variants of nationalism. Of course, the political economy of commercially driven digital platforms is far from being the sole factor contributing to this shift. Rather, it operates hand in hand with a range of other phenomena, including nation branding, consumer ethnocentrism and the commodification of ethnic identities, which have come to play an increasingly central role in the reproduction of nationalism in the 21st century (Aronczyk, 2013;Castelló & Mihelj, 2018;Comaroff & Comaroff, 2019;Wang, 2005).
The commercialisation and commodification of nationalism in the online environment is visible at several levels.
The most obvious examples can be found in commercial transactions online. Thanks to its anonymity and global reach, the web offers a perfect environment for selling far-right goods, such as the merchandise sold online by Thor Steinar, White Trash Rebel, Dirty Tees and Doberman's Aggressive. As noted by Cynthia Miller-Idriss (2018), these far-right brands often use global symbols and English language slogans imbued with coded nationalist and far-right messages-a testimony to their global reach, facilitated by digital media. The digital environment also plays an important role in facilitating the spreading of bottom-up consumer nationalism, including consumer boycotts aimed at brands that become tarnished by their perceived association with 'enemy' nations or whose owners engage in behaviour perceived as insulting to one's nation. Examples range from the #BoycottGermany Twitter campaign, which was launched in response to the new austerity measures that were felt to be imposed on Greece by Germany-dominated EU (Lekakis, 2017) to numerous examples of Chinese consumer nationalism in reaction to Western coverage of China (e.g., Jiang, 2012). However, examples of consumer and commercial nationalism online are not restricted only to such highly visible and extreme forms of nationalism. While internet users are addressed first and foremost as individual consumers, they are also treated as members of imagined national communities of consumers, and both global conglomerates and national companies routinely seek to cater to their national tastes. Netflix, for example, gathers data on user preferences in order to tailor its programming for different national audiences, including offering series about cricket in India or anime in Japan (Lobato, 2019). While Netflix users may not necessarily see their consumer choices as national, the national filtering of Netflix content means that they end up mindlessly perpetuating the national framework, albeit as part of a commercial transaction (see Stewart, 2016;Castelló & Mihelj, 2018, pp. 563-564).
As these examples suggest, the participatory affordances of the digital environment create opportunities for different actors to participate in the commodification of the nation online-from commercial companies and states to citizens-consumers themselves. While we should not overestimate the power of citizens-consumers, it is nonetheless important to acknowledge that this diversity can make the use of nationalism for commercial and promotional purposes online rather unpredictable. Instructive examples of such unpredictability can be found in state-sponsored nation branding and public diplomacy efforts disseminated through digital media, launched in the United States, the United Kingdom, Estonia, India and several other countries (Bjola, Cassidy, & Manor, 2019;Edwards & Ramamurthy, 2017;Tammpuu & Masso, 2018). While initiated by state actors or state-sponsored entities, these campaigns also rely on citizen participation for their success. In Colombia, for instance, the authorities successfully used the #itscolombianotcolumbia campaign to encourage Colombians to use that hashtag every time they saw the name of their nation misspelt in Twitter (Gómez Carrillo, 2018). On the other hand, in India, citizens have parodied nation branding taglines in order to voice their criticisms to current political, economic and social affairs in their nations, as in the satirical Facebook comic strip 'Inedible India', whose title references the 'Incredible India' campaign (Udupa, 2019).
Examples mentioned so far all involve explicitly economic activities, such as the selling and buying of goods or nation branding campaigns. Yet it is important to note that the principles of commercial and consumer nationalism extend well beyond explicitly economic activities. Thanks to the propensity to promotional discourse online-the fact that we are all, as individuals, encouraged to self-brand ourselves in the social media environment-these principles are visible in virtually all social spheres, from culture and education to politics and mundane interactions on social media. Platforms such as Facebook initially operated as tools for self-expression but have soon turned into instruments of self-promotion (van Dijck, 2013). The rise of micro-celebrities and social media influencers (Khamis, Ang, & Welling, 2017) as well as the proliferation of 'selfie diplomacy' that relies on political leaders using the selfpromotional tools of social media to negotiate diplomatic relationships and contributing to nation branding (Manor, 2019) offer further examples of how the logic of individual self-promotion and its underlying commercial logic have infiltrated various domains of social life and national imagination online, ranging from popular culture to politics.

| CONCLUSIONS
What do our arguments about digital nationalism mean for future research? At least four broad principles can be drawn from our discussion. First, future research should resist the temptation of restricting nationalism to its most exclusive, aggressive variants and instead examine how digital media contribute to the reproduction and spreading of different varieties of nationalism, including those that are more open to diversity or more compatible with liberal democratic values. Second, research on digital nationalism should ideally seek to combine the analysis of both the mundane reproduction of nationalism online and the dynamics of more explicit articulations of digital nationalism, ranging from nationalist consumer boycotts to competing articulations of national belonging during major political events. Third, research in this area could fruitfully combine institutional and infrastructural analysis with ethnographic approaches and examine how digital media users actively appropriate, adapt and perhaps even subvert the national grid that underpins our communication environment. With few exceptions, existing research on digital nationalism often privileges institutional and infrastructural analysis over the examination of audience practices and uses-indeed, our own discussion in this article reflects this bias-and there is certainly considerable scope for more bottom-up, ethnographic and qualitative work in this area. Fourth, the field would benefit from longitudinal, historical research that would seek to investigate how the interaction between media and nationalism changes over time. Especially the arguments we develop in the last section, where we seek to identify the key characteristics that make digital nationalism different from its 'analogue' predecessor, all imply change over time yet are largely based on evidence drawn from a single moment in time.
The adoption of longitudinal designs would enable us to gather more robust evidence and gain a clearer insight into the relative contribution of digital technologies-vis-à-vis other factors-in the changes we have outlined.
Furthermore, longitudinal designs could help us develop a better understanding of how and why the banal, seemingly inconsequential routines that reproduce nationalism on a daily basis can serve as jumping boards for the mobilisation of more evident and virulent modes of nationalism such as those that currently dominate political life globally.
On a more general note, our discussion also highlights the need for a more sustained dialogue between nationalism scholars and media and communication scholars, particularly in view of recent developments. All too often, debates on nationalism are lacking in the understanding and appreciation of the role of media and communication technologies and institutions in shaping both historical and contemporary forms of nationalism.
A significant part of the discussion about the impact of digital technology on nationalism-including several of the studies quoted here-includes little if any reference to existing research by digital media scholars. And conversely, discussions in media and communication literature often show little awareness of recent developments in the field of nationalism studies and as a result tend to adopt a rather narrow view of nationalism, often depicting it as a fringe ideology or a social disease that should be eradicated. As the arguments developed here suggest, our collective understanding of nationalism in the digital world cannot proceed without a closer engagement between these two academic fields.