Post-Soviet migrants in Germany, transnational public spheres and Russian soft power

ABSTRACT When social networking sites based in authoritarian states expand transnationally, digital media environments create soft power resources for these states. Rather than being an intrusion into national cyber-spaces “from outside”, these communication networks are constructed by multiple actors, not all of whom are backed or coordinated by the authoritarian state. Diasporic actors are of special interest here. Using reposts as a direct measure of information flows on the Russian social network VKontakte(VK), this article reconstructs Russian-language networks of influence related to post-Soviet migrants in Germany. It shows that transnational communicative spaces evolve around collective resources related to these migrants by tracing how transnational and multidirectional content flow is organized between the post-Soviet space and Germany. Identifiable political sub-networks are found to be dominated by anti-liberal and anti-Western positions, ideologically and thematically compatible with official Russian state discourses. Right-wing radicals, another target audience of Russian soft power strategies in Europe, are shown to be interconnected with the publics of Russian-speakers in Germany and in the post-Soviet space. Communicative spaces on social networks are thus not purely metaphorical, but underpin the extraterritorial reach of authoritarian regime.


Introduction
Over the last decade, politicians, media and the general public in Western liberal democracies have repeatedly discussed the transnational reach of authoritarian regimes. Although it seems unlikely so far that such regimes would directly attack the territories of the core European or North American countries, a "phantom menace" of authoritarian influence has been taking shape: it is related, for instance, to digital election interference on Russia's part, international multilanguage broadcasting by RT and Sputnik, China's efforts to polish its image through investments in European media, and Iranian hackers targeting research facilities and personal data. Such prototypical threats were usually interpreted as penetrations of national communicative and political spaces coming from the outside. I argue that when authoritarian states seek to exert their influence, it is less about violating the enclosed space of a target nation-state, and more about constructing transnational communicative spaces. This process is only partly shaped by the authoritarian state itself. To explore this approach, I offer a case study of publics which involve Russophone post-Soviet migrants in Germany.
The extraterritorial reach of authoritarian regimes has become the subject of several academic discourses, focussing on transnational repression and other intelligence and security practices (Lewis, 2015;Moss, 2018), on the ways in which authoritarian control of the Internet is influenced by international politics or by the global diffusion of control tools (Kerr, 2018;Michaelsen, 2018), on exile and migrant communities from authoritarian "sending states" (Bernal, 2014), as well as on the efforts of authoritarian regimes to enhance their international influence and develop soft power (Barr, Feklyunina, & Theys, 2015;Rawnsley, 2015). Some of these discourses develop in parallel, while others are intertwined, for instance, when transnational repression against political exiles or policies related to co-ethnic communities are under debate. What stands out in these diverse fields is a new tendency to consider authoritarian rule not as a "territorially bounded regime type, but rather as a mode of governing people through a distinct set of practices" (Glasius, 2018, p. 179). Authoritarian states are, therefore, discussed as extending beyond their respective borders, without necessarily changing those.
Social media promote a variety of extraterritorial endeavors on the part of authoritarian states, including their efforts to influence the policies of other states via corresponding public spheres, known as soft power. As Michaelsen (2018) notes, the literature on networked authoritarianism in general and on the Russian case has so far provided only limited insights into extraterritorial advantages of social media for authoritarian regimes. Despite the fact that the soft power debates are dealing with transnational information flows, they are still shaped by the thinking in territorialized "containers". Further, both Russian political elites and scholars of Russian foreign policy have narrowed the concept to an instrumentalist interpretation. Under its banner, Russian political leaders have tried to influence public opinion abroad and achieve particular goals, often in the framework of the "Russian World", aimed at the reintergration of Russian diasporas and/or adjacent regions in the shared civilizational space (Makarychev & Yatsyk, 2017;Sergunin & Karabeshkin, 2015;Suslov, 2018;Yablokov, 2015). This has been mirrored by scholarly debates, which by and large deal with state-funded and centrally coordinated measures (Cheskin, 2017;Kaczmarska & Keating, 2019). Following an alternative path, the article discusses soft power resources as constructed bottom-up and focuses on networked communicative spaces as contexts of active soft power strategies.
I aim to contribute to the efforts to "de-Westernize" the soft power approach by focusing on Russian country-centered social networking sites (SNS) as soft power resources. Firstly, the article uses empirical evidence to consider in-depth a kind of resource, which has not received appropiate scholarly attention so far. Secondly, the case of post-Soviet migrants in Germany and the Russian state speaks to the broader comparative study of the extraterritorial reach of authoritarian states, as it casts new light on the domestic-foreign dimension of soft power and emphasizes the bottom-up perspective. In case of social networking sites, supposed recipients are indispensable producers of power contexts as prerequisites for state strategies.
The article proceeds as follows: First, I place the question of Russian soft power via countrycentered social networks in the context of the literature on extraterritorial effects and soft power by authoritarian states in network society, including their use of social media. Next, I discuss transnational migrant publics connected to such states and briefly introduce the case of post-Soviet migrants and Russian social networks in Germany. I then present the empirical data and methodology of the study, describing the discursive arenas and reconstructing related information flow on the platform, before discussing the key findings. My empirical analysis focuses on group types and repost behavior, on the structure of communication networks based on reposts, on their transnational nature and on most densely connected segments. The study shows ways in which networked forums of post-Soviet migrants evolve on Russian social media. It also provides insights into how both entertainment and political content flow from post-Soviet space to Germany-centered groups. Political sub-networks are highly compatible with official Russian state discourses and dominated by anti-liberal and anti-Western positions. Transnational communicative spaces on social media, the paper shows, provide resources for the extraterritorial expansion of authoritarian regimes.

Social networking sites and soft power of authoritarian states
The paper argues that national social media platforms with a global reach are elements of soft power, understood as "the ability to affect others through the co-optive means of framing the agenda, persuading, and eliciting positive attraction in order to obtain preferred outcomes" (Nye, 2011, pp. 20-21). The concept of soft power as an alternative to the "hard power" of coercion or direct economic incentives, originally developed by Joseph Nye with US foreign policy in mind, has increasingly been applied to cases beyond the USA and the EU (Bakalov, 2019, p. 137). Opinions are divided as to whether a concept's liberal bias hinders its application to countries besides Western liberal democracies (Bakalov, 2019;Kaczmarska & Keating, 2019). While some scholars go so far as to deny the concept of soft power any heuristic potential in such cases, others seek to "de-Westernize" it to "better understand how nondemocratic regimes promote and manage their images" (Barr et al., 2015, p. 213).
International broadcasting by state-supported media institutions, such as the BBC World Service, CCTV or RT, has been generally accepted in scholarly debates as an important resource of soft power or public diplomacy, framing issues and providing specific worldviews (on RT and Russia, see Miazhevich, 2018;Rawnsley, 2015;Yablokov, 2015). Social media (SM) have also been discussed by journalists and the general public (and, to less extent, by scholars) as tools of political influence. However, the focus here is on particular statebacked efforts to influence the political process in other countries using the platforms' communicative venues (as in Rutenberg, 2017;Wiebe, 2018). Transnationally active SM platforms have not so far been studied as general instances of soft power in their own right. In politics, we can find several cases where whole SM platforms were deprecated as tools of foreign political influence: In September 2019, the Russian State Duma announced that Google and Facebook had been intervening in Russian elections and should be called to account. In May 2017, Ukrainian president Petro Poroshenko introduced sanctions on Russian digital products, including social networking services by Mail.Ru Group, Odnoklassniki (OK, also known as OK.ru) and VKontakte (VK, also known as VK.com). Such efforts correspond with the Internet sovereignty discourse promoted chiefly by China and Russia, according to which governments should control and limit the digital flow of information on their territory (Budnitsky & Jia, 2018;Michaelsen, 2018). Still, in the scientific discussion of soft power, there is a blind spot when it comes to social media.
This omission reflects the general Western bias of communication studies, where Twitter, Facebook or Youtube are implicitly seen as prototypical cases of social media platforms. All these US-based media companies are globally active and de facto multilingual and multinational platforms. The Russian networks OK and VK, in contrast, have a strong national focus: they dominate the social media market in Russia and have the great majority of users in their home country. Still, they continuously work in several other countries, mostly in the post-Soviet space.
The transnational expansion of an SM platform means, first of all, an opening of new markets by commercial structures. But even if media companies view the audiences in globalized media flows as "consumers" and not as "citizens" (Thussu, 2007), the venues they provide shape citizens' communication and political behavior; moreover, they can be used for mobilization on local, national and transnational levels. Private digital information intermediaries, including social networking sites, are enacting transnational Internet governance (DeNardis, 2014). They affect user experience, social interactions and information flows, and "have become powerful regulators of the online public sphere" (Budnitsky & Jia, 2018, p. 598). Globally expanding Russian social networking sites do, therefore, have powers over and act as regulators of online public spheres beyond Russian territory. Just like US-based private digital media, they face the challenge of "navigating between a company's values and national laws or norms" (DeNardis, 2014, p. 171). Likewise, Facebook undoubtedly influences digitalized publics in Russia and elsewhere. And still there are crucial differences: firstly, in the degree of national concentration of audiences and content, which makes VK more "Russian" than Facebook is "US-American"; secondly, the Russia-backed Internet Sovereignty ideology strongly supports the idea that the state should regulate information flows, in contrast with the opposing Freedom ideology (Budnitsky & Jia, 2018).
Border-crossing SNS make it possible to affect populations by setting agendas and framing issues in specific ways, with the aim of influencing political elites through public pressure, according to Nye's (2011, pp. 94-95) indirect model. Further, "soft power" cannot be reduced to active efforts to attract, but includes non-intentional influence as a quality of relations between countries. By operating beyond the borders of its "home country", a country-centered SNS creates a virtual soft power arrangement. Such expansion includes digital "information war" campaigns or other deliberate efforts by state actors to "promote its normative frame abroad" (Suslov, 2018, p. 340), but are not limited to those. Following the scholarly consensus on the difference between soft and hard power as difference-in-degree and not in kind (see Bakalov, 2019), this article argues that a country-centered SNS can be both a source of attractive identity offers and a tool of coercion or manipulation. When a social media platform with a strong national focus acts beyond the borders of its "home country", this means the extension of a specific digital media environment influenced by the home state's regulations, as well as hegemonic discourses and popular narratives. In short, such extension is directly related to culture as one of the central soft power sources identified by Nye.
The soft power debates reproduce, often implicitly, a nationally bounded conception of international communication (Volkmer, 2014, p. 129). They still fall into the "territorial trap" (Agnew, 1994), that conceives of modern states as "containers" for societies and fixed units of sovereign space, and of borders as the ultimate limits of state power. However, in a network society, multiple and overlapping transnational and translocal communicative spaces evolve (Couldry & Hepp, 2012;Volkmer, 2014). The analysis presented here shows that the soft power of authoritarian states must be conceptualized as the construction of transnational communicative spaces that are not solely produced by the state, but also from below by a range of different actors. This approach is rooted in the critique of methodological nationalism and in the idea of transnational fields "as systems of social relations composed of networks of networks" embedded in complex power relations (Glick Schiller & Çağlar, 2009, p. 180). Social media provide an excellent lens for testing this approach empirically.
The social media users are those who, by performing translocal communication via the platform, create relational communicative spaces and eventually expand platforms transnationally. As media systems become ever more hybrid, boundaries between "online" and "offline" media producing and distributing content become increasingly blurred, and users become indispensable agents of the creation and distribution of media content (Chadwick, 2013;Iannelli, 2016). As Miazhevich (2018) shows for the RT, participants in networked cross-media practices engage with the broadcasters' framing online. Still, this audience engages with an initial product by a broadcaster. By studying transnational networks on Russian SNS, I emphasize one of the core elements of the soft power concept, that its tools are not exclusively state-made, but constructed by multiple individual and collective actors (that obviously does not rule out that some actors are statebacked). The study of SNS as resources of soft power shifts the focus to multi-actor, horizontal, bottom-up soft power arrangements.
Among scholars of soft power, there is a consensus on the crucial role of non-state actors, first and foremost civil society, in making a country attractive to foreign publics (Bakalov, 2019, p. 136;Nye, 2011, p. 83). These actors' capacity to frame issues and form discussions is not restricted by a potentially low degree of institutionalization (Brüggemann & Schulz-Forberg, 2009). In comparison, individual transnational communication on the SNS, which takes place without forming new arenas, does not provide novel mobilization structures and therefore does not indicate the development of soft power resources sui generis. That is why my study focuses on the collective level of transnational communication arenas.
In the context of digitalized communication, broadcasters and political actors potentially address multiple international and domestic audiences at the same time, and inconsistency between messages poses a risk to the source's credibility (Rawnsley, 2015). Managing such dissonance is a challenge for public diplomacy in any regime. In the Russian case, the relational character of soft power discources is strongly pronounced: first, they include negative Othering of the West in terms of values and, second, they are driven by deliberate efforts to undermine Western narratives, while Russia and its broadcasters are presented as underdogs (Kaczmarska & Keating, 2019;Kiseleva, 2015;Rawnsley, 2015;Yablokov, 2015). The notion of the "Russian World" is a centerpiece of Russian public diplomacy, which has been articulated for audiences both in the country and abroad and exploited for internal legitimation and mobilization (Feklyunina, 2016;Laruelle, 2015;Suslov, 2018). This article contributes to the discussion of the domestic-foreign dynamics of soft power by reconstructing transnational information flows on a social networking site in multiple directions and testing empirically how far German migrant publics have audiences in both Russia and Germany. Domestic-foreign dynamics have special relevance for country-centered SNS: the dominance of one language (here, Russian) means less restrictions for border-crossing information flow and makes "domestic" discources operable abroad, while making the compatibility challenge more acute. At the same time, the networks' soft power potential will be focused on or even restricted to Russophone and/or post-Soviet migrant audiences, although not necessarily in adjacent countries only.
Russian "diasporas" and transnational communicative spaces Russian foreign policy strategies addressed Russian-speaking minorities in Europe even before the concept of soft power was explicitly adopted by the Kremlin. These heterogeneous populations were constituted, expressed by Brubaker's (2000) bon mot, both through the movement of people across borders (the largest groups in recent history being co-ethnic migrants to Germany and Israel) and through the movement of borders across people (mainly affecting Russian minorities in the former Soviet republics). The "Russian World" concept and policies aim at constituting the Russian diaspora while referring to it as a given entity (see Brubaker, 2005). While it has been shown that the Russian diasporic project has changed its subjects and tools over time (Suslov, 2018), its perception among the putative members of the "Russian World" in different countries has not recieved enough attention (Kallas, 2016). This is in spite of the empirical evidence that people's subjective sense of belonging to the "Russian world" and their relationship with Russia vary even between the heavily Russified post-Soviet regions (O'Loughlin, Toal, & Kolosov, 2016) and within one community (Kallas, 2016). Moreover, Russian "diasporas" seem not to be easily mobilized for political causes, and a sense of belonging is separated from political allegiances (Pieper, 2018;Suslov, 2018).
This asymmetry in studies of Russian soft power and its Russophone subjects affects how digital media are discussed in this contextnamely, as an area of aggressive policies, information warfare, and other largely top-down, instrumental approaches, including state-sponsored efforts to mobilize Russophone populations in adjacent states (partly Kostyuk & Zhukov, 2017;Simons, 2015;Suslov & Bassin, 2016;Zelenkauskaite & Balduccini, 2017; see Golovchenko, Hartmann, & Adler-Nissen, 2018;Makhortykh, 2018 for some recent exceptions). What the dominant account lacks is, first, the understanding that states and their control strategies produce transnational spaces beyond containers (Collyer & King, 2015), even those states where political elites are strongly inclined to think and act in terms of territories and their protection. Second, it misses the construction of diasporic transnational networks on social media by non-state or oppositional actors, which can pursue very different cultural and political identity projects (Adamson, 2019). Addressing this gap, my paper focusses on the construction of transnational communicative spaces by Russian-speaking minorities on country-centered SNS, which provide mobilization structures and therefore soft power resources.
To put an emphasis on connections that are constituted through actual networks and not on the construction of belonging to a "homeland", I use the term "transnational public sphere" instead of "diaspora" (see Bonnerjee, Blunt, McIlwaine, & Pereira, 2012, p. 12). The transnational public sphere is considered as network of forums, "constituted by different interconnected areas of public communication" (Brüggemann & Schulz-Forberg, 2009, p. 694). Such forums, or discursive arenas, have a variety of relationships with each other, with an arena operating as a player in various external arenas (Jasper, 2015). In my empirical case of post-Soviet migrants in Germany, not all of those who are active transnationally on social networks are engaged in Russiarelated identity projects. The terms "Post-Soviet migrants", "Russophones" and "Russian-speaking diaspora" mean intersecting, but not identical populations (Panagiotidis, 2017). The central category this study relies on is the most neutral term, "post-Soviet migrants in Germany" (PSMG) in the first or second generation, who came to Germany from the Soviet Union or a country in the post-Soviet space, or were born in Germany to at least one immigrant parent. Not all of them are Russophones, as the second generation became increasingly assimilated in terms of language (Vogelgesang & Kersch, 2017). The subjective notion of belonging to the Russian cultural and political space is constitutive for the least inclusive group, the "Russian-speaking diaspora in Germany".
Transnational spaces stretch beyond migrant networks (Collyer & King, 2015, p. 190). In relation to country-centered SNS as resources of soft power, this provokes the question of how, and to what extent, post-Soviet migrants and other potential audiences of Russian soft power interact on the SNS. Another frontier which should be researched is between domestic and foreign audiences of soft power resources. Therefore, the paper explores in how far migrant publics as networked arenas reach audiences located in both Russia and Germany. And finally, it seeks to operationalize and reconstruct the information flow between Russia/post-Soviet space and digitally mediated publics of post-Soviet migrants in Germany. This case allows to study the transnational expansion of a country-centered SNS and the development of migrants publics on this platform as soft power resources, extending the reach of an authoritarian regime.

Case study: post-Soviet migrants and Russian social networks in Germany
Germany is home to a first and second generation post-Soviet migrant population of appr. 3.5 million (Statistisches Bundesamt, 2018), 1 which comprises individuals and families with various migration trajectories, including co-ethnic migrants; labor migrants, among them highly qualified EU Blue Card holders; Jewish programme refugees; asylum seekers and students. Recent years have shown a generational shift from direct migration experience to indirect migration background. For instance, only around 19.000 people moved to Germany from Russia in 2017 (www. bamf.de). The newer migrants are heterogeneous in terms of their social and cultural background. In the overall post-Soviet migrant population, at the same time, younger people currently attain higher levels of education and have a better command of German (Panagiotidis, 2017;Vogelgesang & Kersch, 2017;Worbs, Bund, Kohls, & Babka von Gostomski, 2013). The most populous subgroup here are ethnic Germans from Kazakhstan, Russia and other former Soviet republics (Russlanddeutsche Spätaussiedler, or Russian German re-settlers) and members of their families, whose immigration has been traditionally supported by the Federal Republic of Germany and had its peak in the 1990s (between 1990 and 2015, over 2.1 million came to Germany). Whereas Russlanddeutsche do not dominate recent migration from former Soviet republics, they still constitute the largest subgroup and have considerable potential of political participation, because they have easy access to citizenship and the full political rights that come with it. For a long time, however, post-Soviet migrants in Germany in general and Russlanddeutsche in particular remained a silent minority, perceived as well-integrated if not particularly influential or politically active (Goerres, Spies, & May, 2018).
Given that they are one of the largest migrant groups in Germany, empirical evidence on their media repertoires is surprisingly thin. A representative survey from 2011 shows that a great majority of post-Soviet migrants do not live in a "parallel universe" of Russian-language media (Simon & Neuwöhner, 2011). However, relations between Germany and Russia deteriorated following the annexation of Crimea, the war in Eastern Ukraine and the international sanctions against Russia, so that the identity bricolage between the Russian and German "worlds" became problematic. As migrants' media consumption is immediately related to identity construction (Hepp, Bozdag, & Suna, 2012), the pre-Crimea findings cannot automatically be extended to the present.
It is plausible, however, that the share of post-Soviet migrants in Germany who use social networks has increased since 2011, when about half of them were found to use social networks, especially the Russian SNS Odnoklassniki (OK; Müller & Beisch, 2011). Its popularity seems to persist: the platform itself claimed to have about 1.3 million Monthly Active Users (MAU) in Germany in 2018 (Boyarskiy, 2018). Another Russian SNS, VKontakte (VK), does not publish country-specific user statistics, but its MAU in Germany can be roughly estimated as 430,000. 2 External statistics confirm the popularity of OK and VK in Germany, with OK being clearly ahead (www.similarweb.com).
Russian social networks enable post-Soviet migrants to maintain contact with those who stayed behind and with other migrants from Russian-speaking countries. They get access to a variety of cultural and political content in their (second) mother tongueaccess which requires few cultural or financial resources. 3 In contrast to smaller migrant communities in Europe, where active SNS usage (of Facebook and the host country's SNS) was associated with more outgroup ties, but not with more ingroup ties (Damian & van Ingen, 2014), Russian country-centered social networks hardly facilitate outgroup ties in general.
The platforms' different profiles affect their attractiveness for migrants. OK is oriented toward maintaining contact with acquaintances from different periods of life, and other people with a shared background. The search for "old pals" is made easier for people who grew up in Soviet times, because the system includes old toponyms. OK (founded in 2006) provided an opportunity to the largest group of Russian-speaking migrants in Germanythose who came after the fall of the Iron Curtain and before the development of digital media; they could digitalize and revitalize their prior contacts. In the terms developed by Hepp et al. (2012), OK suits origin-oriented and ethnooriented (i.e. diaspora-oriented) migrant identities best. Communication on VK, by contrast, is focused by the SNS interface on certain topics. Of course, this social network can still be used to maintain personal contacts in the country of origin and in the migration contextsimilarly to OK, where topics are far from being restricted to the past. Nevertheless, the topic-focused interaction on VK makes it better suited for politicization, including the transnational flow of information and knowledge beyond personal networks.
Russian country-centered social networks provide the infrastructure for communication between Russia or post-Soviet space and Germany, for the construction of transnational discursive arenas with the participation of post-Soviet migrants and for online-offline mobilizations. Such mobilization was exemplified by the "Lisa case", when a Russlanddeutsche teenager was allegedly abducted and raped by men of "southern" origin, sparking rallies in more than 40 cities and making Russlanddeutsche visible for the German general public (Braghiroli & Makarychev, 2017;Mitrokhin, 2017).

Data and methods
The empirical study combined a keyword-based search for Germany-focused collective accounts dedicated to public communication with the automated collection of their connections (see Supplementary Material). I considered an account as transnational and/or related to post-Soviet migrants in Germany and included it in the study as a basic discursive arena, if it met at least one of the following criteria: • Addressed and/or claimed to represent post-Soviet migrants in Germany, or some subcategories; • Officially represented or was dedicated to political actors in Germany, who were working toward mass mobilization of supporters and/ or addresses Russlanddeutsche or other PSMG; • Was dedicated to current German politics and society (the accounts that explicitly addressed Russian residents were not included); • Was related to a Germany-based group which was active in Russian-German relations and public diplomacy.
Moving from static to relational data, I looked for direct connections amongst the sampled groups 4 and with other accounts on the platform. As OK did not provide required opportunities for automated data retrieval via API, the project focused on VK from this stage on. The free software Gephi was used for visualization and Social Network Analysis (SNA), with the aim of identifying the structure of online communication, its focal points, subnetworks, and the bridges between them (Bastian, Heymann, & Jacomy, 2009). Connections between groups were operationalized as reposts, which are an essential mode of information flow on an SNS. Via reposting, an account identifiably re-publishes on its own wall a post, a photo or other content published by another account. Whereas some Facebook-centered studies reconstruct links between groups based on "likes", those arguably have less subjective weight and are less reliable as an indicator of a meaningful connection. The focus on reposts instead of likes makes it possible to account for an important quality of relations: Because links are valued according to a total of reposts, a stronger flow of information from one account to another will be detectible on resulting directed graphs. Starting with the sample of 97 groups, I identified 3204 accounts as sources of reposts for the period January 2015-June 2018, including the initial sample. For each repost, information was retrieved on the account A which reposts, the account B which is being reposted, and the repost itself, including metadata and possible textual content. Here, we get information on A (it is publicly positioning itself near B), on B (influence B is having on A) and on the flow of knowledge between the two. Influence is understood not as being intentionally exerted by an actor, but as a quality of relations in a communication-based network: An influential account is one whose posts are often shared by others; in this sense, its influence is created by other actors. While some users might follow an opponent to keep informed, by sharing the content they distribute the message further and put themselves in a particular discursive context (Froio & Ganesh, 2018, pp. 8-9;Halavais, 2014, p. 35). Their action is to some extent affirmative, regardless of the intentions, and the study does not take into account the possible positive or negative framing.
The aim here is to explore transnational communication flows and their reproducible structures, as they evolve on the Russian social networking site and provide resources for Russian soft power with the focus on diasporas. Transnationality emerges if German residents with a post-Soviet background communicate on the Russian social network, if content from Russia and the post-Soviet space flows to Germanycentered groups, or if the groups' audiences are in themselves mixed in terms of geolocation.

Group types and repost behavior
To provide a more differentiated view of basic discursive arenas included in the study, the groups found on VK and OK were manually categorized according to the dominant orientation of posts: • Translocal landsman groups (in Russian): focused on maintaining small-scale, originbased identities; typically they bring together former and current residents of certain locations, urban and rural, in the former Soviet Union. • Local groups and networking (in Russian): here users can discuss practical questions, share information and local news, ask for advice, and organize offline activities. Most groups focus on specific German cities and facilitate local connections among PSMG. • Identity-centered groups (in Russian or German): dedicated to the history and contemporary state of certain prominent groups, such as Russlanddeutsche. • Entertainment (in Russian): jokes, pictures, popular science and other content intended to provide amusement or infotainment, with explicit reference to Germany/PSMG. • Political PSMG groups (in Russian or German): political groups of PSMG or their subgroups that position themselves within German society and politics. • Political groups, not related to PSMG (in German): political groups that position themselves within German society and politics or are dedicated to its actors. Whereas these accounts do not identify themselves with PSMG, they may focus on topics related to the post-Soviet space.
The group types are represented differently on VK and OK. Whereas the translocal and local networking groups are much more visible on OK, the presence of Germany-centered entertainment groups is comparable across the two platforms. Further, political groups set up by Russlanddeutsche and other post-Soviet migrants are more widespread and popular on OK, especially the groups related to the new rightwing political party Alternative für Deutschland (AfD, Alternative for Germany), which tried, partly successfully, to mobilize Russlanddeutsche as activists and voters, and the xenophobic PEGIDA (Patriotic Europeans Against the Islamization of the Occident) movement. Still, on VK I found far more Germanybased right-wing populist and radical political groups in general, representing a broader range of actors. German right-wing radicals consider VK a free space and an alternative to Facebook; they are not affected by the growing persecution of activists in Russia for sharing politicized content online. Groups and parties of other political orientations are virtually non-existent, with the exception of the left-wing party Die Linke (The Left), which is known for its pro-Russian attitude.
In contrast to the group types, their repost behavior can be described only for VK. Whereas "diasporic" links connect the sampled Germany-centered groups to each other, "transnational" reposting from user and group accounts based in the post-Soviet space is very widespread as well. Some groups tend to reposts both political and entertainment content, and this mix create low-threshold politicized platforms, which are suitable for the incidental exposure and mobilization of users outside hardcore political scenes. Somewhat surprisingly, (semi-)state Russian media outlets are barely reposted by groups and play a marginal role in the distinctive communication arenas formed by PSMG. Finally, several accounts related to right-wing German political actors (such as the main AfDgroup), who were actively mobilizing PSMG for elections, work in the "sending-only" mode and not network-oriented. Although it has already been known that PSMG use Russian social networks, the analysis demonstrates actual plurality of related groups that provide basic elements of communicative spaces and, in contrast to migrants' individual communication, develop novel mobilization structures and soft power resources.

Communication networks on VK
For the reconstruction of areas where identified groups are potentially growing together, so that more complex communication arenas evolve, I look for overlapping repost networks. The chosen period between January 2015 and June 2018 encompasses several events and developments which are important for German politics and society, and which are linked to the post-Soviet space or Russian-German relations: the "Lisa case", the 2017 Bundestag elections when the AfD entered the parliament, protracted conflict and war in Eastern Ukraine, and the European migration crisis related to the internationalized war in Syria.
A repost network represents groups as nodes and repost relations between groups as directed links from the "sink" to the "source" of the content, weighed according to number of reposted items, so that the graph takes into consideration intensity of relations and influence. The resulting interaction graph is not a complete network in terms of SNA, but a sum ofpartially overlappingego-nets of sampled groups without alteralter ties (where each single group is an "ego", connected by repost links to its "alters", see Supplementary Material). To identify the links between groups and the shared communicative arenas, I consider both direct connections between the sampled groups and indirect connections via reposts from the same sources.
The direct flow of information was reconstructed from repost links within the sample. Given its heterogeneity, several disconnected or sparsely connected sub-graphs and isolated egonets were likely. Figure 1 shows some scattered dyads and isolated groups, but reveals connected larger subnetworks as well, centered around local networking and entertainment in Germany (the lower part of the graph) and the mixed, political and entertainment subgraph (the upper part), which is related to issues concerning the post-Soviet space and to German right-wing radicals. Both subgraphs' cores will be analyzed in the section on repost coalitions.
The combined 2015-2018 graph (Figure 2), which includes repost links from sources beyond the sample, uses a broader interpretation of communication structures. If a sub-set of groups reposts repeatedly from the same range of sources, i.e. information flows from the same pool, then the "sink" groups are indirectly connected and have similar positions in a sub-graph. While textual information is often hard to evaluate (for instance, memes have a picture format not directly accessible to textual analysis), my structural, repostsbased approach provides a sensible way to estimate the semantic proximity of accounts.
The resulting graph, based on partially overlapping ego-networks, is highly segmented and comprises several loosely connected sub-networks. To identify the connected segments and therefore potential communication arenas, modularity classes (containing nodes which are better connected to each other than to the rest of the graph), were calculated and visualized using Gephi (with resolution 1.3, 19 communities, modularity score 0.638). The Force Atlas 2 algorithm was implemented for visualization. Whereas the orange "local networking" sub-graph does not change much due to the links beyond the sample, some otherwise isolated elements are connected to other sampled groups indirectly, by information flow from the same sources.
In the "politicized" subgraph in the upper part, subsets of sampled groups show extensive indirect connections, via their respective shared sources. As the comparison between the 2013-2014 and the 2015-2018 periods shows, specific nodes in a "source pool" are replaceable, and no single node accounts for a particularly large share of content. Thus, instead of influential nodes, there are locally influential positions. Two cases of such pronounced triads (Figure 3) are Freunde der AfD/ Unabhängiger Infokanal/Deutsche Wahrheit and Germania-Avstria-Shveitsaria/Sahra Wagenknecht/ Deutsche Wahrheit (for their description, see Repost Coalitions section). Their respective elements are connected not just pairwise, but also share a common source pool. Multiple connections between both triads and related groups establish, in addition, the information flow between the "transnational" and the "German right-wing" segments of VK networks. Here, transnational political integration has a strong anti-elitist and anti-liberal slant. The reconstructed information flows demonstrate that interconnected arenas evolve on VK. To assess if their partial similarities with Russian hegemonic discources actually indicate the continuity of digital media environment, the paper evaluates their transnationality.

Transnationality of communication networks
To determine the extent to which the reconstructed networks are transnational, I explored the geolocation of content sources and groups' membership structures. For geolocation, I refer to "post-Soviet space" rather than "Russia", firstly because the specific country information Node size acc. to in-degree (number of incoming links); edge thickness acc. to number of single reposts; curve direction: clockwise from the "sink" to the "source" node; node color: red for sampled groups, modularity class for the rest; edge color: color of "source" node.
can be identified reliably not for all group accounts, and secondly because some Russiabacked online campaigns like Antimaidan or SaveDonbassPeople were started by pro-Russian activists, but not necessarily Russian residents. A representative sample of source accounts for the 2015-2018 period was manually sorted into pre-tested categories according to both topical profile and declared geolocation (Table 1). From N = 3204 reposted accounts, a randomized sample n = 344 was selected (confidence level 95%, margin of error 5%).
User accounts are clearly a more popular source type than groups and public pages. A complementary and surprising finding is the relative insignificance of hybrid and online-only news media, including those from the post-Soviet space. 5 Whatever their role for individual media consumption by PSMG might be, they are not particularly relevant for the construction of migrants' collective communication structures. Among user accounts, those located in Germany dominate. Strikingly, about 13% of reposts come from untransparent (even in comparison to clearly biased media sources) accounts which cannot be reliably located in Germany or post-Soviet space due to the lack of retrievable information (because of privacy settings or "deleted" status).
Sources that publish local or Germany-wide information on events, news, services for Russianspeakers living in Germany and help to find and maintain contacts among community members, are responsible for slightly above 4% of reposts. . Node size acc. to in-degree (number of incoming links); edge thickness acc. to number of single reposts; curve direction: clockwise from the "sink" to the "source" node; node color: red for sampled groups, modularity class for the rest; edge color: color of "source" node. This classic function of social media for migrant communities seems to be underrepresented here. However, this interpretation is less convincing in light of the uneven distribution of reposts between the "seed"-groups (see Supplementary Material on ego-nets). If the few "heavy reposters" do not repost from networking groups in Germany, it does not mean that such sources are irrelevant for the remaining groups.
Roughly the same amount of reposts come from "entertainment" and from "political" groups. Whereas almost all "entertainment" sources 6 lack an explicit relation to Germany and address a general Russophone audience instead (which is located mainly in the post-Soviet space), German political groups make up almost one fourth of all political sources. German political groups which address or claim to represent post-Soviet migrants play a marginal role as sources, which probably reflects their marginality on VK. The political content, therefore, flows rather from the groups in the post-Soviet space or in Germany without the focus on migrants. These findings are illustrated by the visualization in Figure 3: the "autochthonous German" sector on the upper left side and the "post-Soviet" sector are well-connected, but distinguishable. An evident division marker is the dominant language, German vs. Russian. 7 The political profile of "source" groups from Germany is clearly right-wing, with the exception of Die Linke. They spread right-wing populist criticism of Western/German political elites, mainstream media and cultural liberalization, alongside a xenophobic discourse on the migration crisis. Their counterparts from the post-Soviet space focus on the topics of "Novorossiya"/Donbass and "Antimaidan", interpreted as a legitimate reaction to Western "intrusion" into Russian "civilizational space"; conservative moral discourses; anti-capitalist stances, sometimes flanked by the glorification of the Soviet imperial project. It is a highly compatible combination of German right-wing populism and Russian political mainstream.
Another way to evaluate the transnationality of the networks is by investigating whether the groups' members include large proportions of both German and Russian residents. As an indicator of geolocation, I used "country" information in user profiles collected for sampled groups. The users were classified as "Germany", "Russia", "Other" (countries) and "Unknown" (no information). Whereas it is plausible that some member accounts in the seeded groups are fake, lack of geoinformation ("Unknown") is not a sufficient indicator for this.
To represent the distribution of country subaudiences for each group and to evaluate how mixed the group audiences are, a Gini coefficient was used: the higher the number, the more uneven the distribution of members between the country categories. As demonstrated by Figure 4, a pronounced dominance of one geolocation is highly untypical, and a great majority of groups have balanced, mixed audiences. Thus, the seeded groups have a mostly transnational membership.
A few of the seeded groups have a majority of their members and followers in Russia, and not in Germany (see Figure 1). Still, each of them (with the exception of Sahra Wagenknecht -Budushchii Kanzler Germanii/Future German Chancellor) has over 1000 (up to 4000) members in Germany, and potentially reach as many users there as smaller Germany-centered groups. Even more importantly, these groups are densely connected to several "German" groups (see the next section) and therefore form part of a transnational sub-network. As the study shows using two indicators, the networked forums on the Russian SNS, which German residents co-produce, have transnational character and affect both domestic and foreign potential audiences of Russian soft power.

Repost coalitions
Using the concept of cycles and cyclic components in directed graphs (see Supplementary Material for details), I identified the densely knitted discursive arenas on the 2015-2018 graph, in which groups not only operate as players (Jasper, 2015), but grow together into repost coalitions: one "political", and one relating to "entertainment and local networking" (Figure 5). The political cyclic component 1 consists of 10 explicitly political and mixed-content groups: Deutsche Wahrheit -Gib Lügen keine Chance (in German and Russian, right-wing critique of political and media mainstream, strongly connected to the "autochthonous German" right-wing network on VK), Sahra Wagenknecht... (in Russian, dedicated to a populist, pro-Russian politician from the party "Die Linke", not her official public page), Bundestagswahl (in German, supportive of "Die Linke"), Unabhängiger Infokanal (German rightwing "alternative media"). An important area are Russian-German relations and "Volksdiplomatie" (public diplomacy) supportive of the Russian state policies and specifically soft power efforts (Drezden International and Golos Germanii 8 ), "Novorossiya" separatism and anti-Ukrainian discourse (Save Donbass People Deutschland and Antimaidan Germaniia, both regional branches of larger mobilization projects), and Spenden für Novorossiya (in German, international support for the Russia-backed insurgent pseudo-states in the Eastern Ukraine). The last one of these, Germaniia-Avstriia-Shveitsariia na russkom (mainly in Russian), publishes unpolitical memes and pictures, mixed with politicized content, compatible withor directly reposted fromthe other groups mentioned. Thus, the repost coalition in general is supportive of the Russian state in complementary contexts of Russian-Ukrainian and Russian-European relations. It stages popular support for the Russian cause by ordinary people and a few journalists in Germany, in the context of deteriorating Russian-German relations in the aftermath of Russia-backed military interventions and separatist mobilizations in Eastern Ukraine.
The groups in the nonpolitical cyclic component 2 address (in Russian) local networking and events, help to find practical information and meet new people, or provide infotainment with a focus on German and local news. With one exception (Germaniia), all these groups belong to the same family of public pages, moderated by same group of users. This diasporic and practical orientation is compatible with the finding that these groups, with one exception, have a majority of members in Germany.
In the repost coalitions, several groups influence each other reciprocally and effectively spread compatible messages to their audiences. One of these components, which includes the most actively reposting actors, is politicized and highly compatible with the Russian official discourses on international relations and, in particular, Ukraine. Moreover, as shown in Figures 1-3, this cluster is interactively connected with German right-wing radicals on VK. This repost coalition is not an isolated case, but fits perfectly into Russian soft power strategies in terms of subjects and audiences. There are no clear boundaries between the (passive) resources and the (active) strategies of Russian soft power.

Conclusion
In this article, I have analyzed the discursive arenas of post-Soviet migrants in Germany on Russian social networks, with the focus on the largest Russian SNS VK. Whereas opinions on the extent of transnationality of post-Soviet migrants' life in Germany are divided (Kaiser & Schönhuth, 2015), the transnational character of media repertoires had already been demonstrated (Simon & Neuwöhner, 2011). However, the effects of changes in the political contexts and sociotechnical developments of media systems remained under-researched. The study presented here deals precisely with these communicative spaces on the SNS as part of a larger "bridgespace" of transnational communication (Adams & Ghose, 2003) and shows how a migrant public evolves as a network of groups and forums. Online publics are regulated by the platforms enacting transnational governance (DeNardis, 2014); in the case of the country-centered SNS such as VKontakte (VK) or Odnoklassniki (OK) this means that a digital media environment influenced by the Russian authoritarian state extends far beyond state boundaries.
By addressing the Russian case, the paper dismantles the assumption that the online-mediated soft power of authoritarian states is restricted to disinformation warfare and mobilization campaigns. It is more than a discursive invasion of another state's container-like territory in order to influence its politics and policies, presumably via public spheres. The construction of transnational communicative spaces by multiple, not necessarily state-backed actors, on a country-centered SNS promotes its extraterritorial expansion and create soft power resources.
Communicative spaces on social networks are not purely metaphorical, because discursive control of imaginative spaces (Collyer & King, 2015) supports international legitimation of the regime among diasporic groups and beyond, facilitating its extraterritorial reach. Somewhat paradoxically, the popular geopolitical imagination in Russia which is deeply rooted in "container space" thinking thrives on the relational, networked cyberspaces (Suslov & Bassin, 2016). Such legitimation includes a re-territorialization of soft power concepts, as happened with the "Russian World" and the annexation of Crimea (Suslov, 2018). This paper demonstrates the plurality of collective resources, i.e. group and public pages, related to Russian speakers in Germany and dedicated to entertainment, practical or general information, construction of (trans)local belonging and/or political topics on Russian social networks. Whereas such groups provide communication structures in their own right, my analysis of how these "players" repost on VK indicates the development of overarching communication arenas. These arenas are constituted partly by directed repost links between several sampled groups, and partly by these groups' multiple connections to shared pools of other sources. The arenas evolve around "politics and entertainment" or "diasporic information and entertainment", with both "politicized" and "infotainment" subgraphs sporadically connected.
The analysis of repost sources shows a significant information flow from post-Soviet space to Germany-centered groups in the networked forums and beyond these. Both for entertainment and political content, such sources were more important than those located in Germany. Online or convergent news media, including such prominent projects as RT, had a surprisingly low influence in collective arenas, even if they can still be pivotal for the individual media repertoires. Additionally, the transnationality of group networks was evaluated based on the membership structures of the groups. A pronounced dominance of one geolocation was untypical, and a great majority of groups had mixed audiences in both Russia and Germany. This enables an empirical contribution to debates on the domesticforeign dimension of soft power: German residents with a migration background in the former Soviet area are receiving and distributing content on the same arenas as Russian residents. However, these arenas likewise create opportunities for the flow of information back to the territory of authoritarian "home state", where the content produced for or by nonresidents get consumed by residents as well. The extraterritorial reach of authoritarian regimes via social networks seems to be a two-way game.
Focusing on the horizontal and networked construction of digital communicative spaces as soft power resources, the study did not try to identify diaspora mobilization efforts by the Russian state directly. The politicized subnetworks' continuity with the Russian hegemonic discourses nevertheless became obvious: conservative and illiberal criticism of Germany and the "West", Euroscepticism, and support for Russia-backed mobilization projects and insurgent pseudo-states in Eastern Ukraine. A few of the "core" groups have, as the analysis of members' geodistribution shows, a majority of members in Russia rather than Germany. Still, they are included in the transnational multidirectional flow of content. Moreover, they present an "alternative Europe", friendly to the Russian state, to audiences in Russia.
Although VK is largely Russophone, its soft power potential in Germany is not restricted to post-Soviet migrants, because it has been used by German rightwing actors as a "free space" alternative to Facebook. This instrumental approach, easily detectable in VK discussions, is complemented by the fact that European right-wing radicals have long been attracted to Russian offers in terms of direct incentives and illiberal identity projects (Kaczmarska & Keating, 2019;Shekhovtsov, 2018). The soft power potential of a state depends on the acceptance of projected collective identity by key audiences on the receiving end (Feklyunina, 2016). Therefore, the right-wing AfD's current rise enhances Russian soft power in Germany. The domestic-foreign continuity of this development is shown by the paper, which demonstrates that right-wing radicals' digital communication is not isolated from the communication of Russian-speakers in Germany and in the post-Soviet space. Instead, these discursive worlds are interconnected by the content flow, bringing together participants in transnational illiberal publics. Such publics exist beyond the specific case of post-Soviet migrants in Germany on VK, and the implemented approach to reconstruction of transnational publics from information flows can be applied to other national cases on country-centered social networks. It would deepen our understanding of how social media facilitate extraterritorial reach of authoritarian regimes beyond their deliberate efforts. Notes 1. As the second generation tends to be underrepresented due to the methods used, real numbers are probably higher (see Panagiotidis, 2017). 2. Assuming that VK, with 800,000 user accounts in Germany in 2018, has the same percentage of MAU within the total of all German accounts as OK with its 2.4 million accounts. 3. For a long time, VK was well known for free-ofcharge, peer-to-peer access to commercial audio and video content. 4. Here and below, "groups" stands for all types of open collective accounts, if not specified otherwise. 5. A source was coded as a "News media" (professional or not), if it broadcast regularly and at least once a week, and focuses on news coverage and commentary on current events. 6. Description: publishing content designed to provide amusement, including (educational) infotainment. 7. An interesting language profile can be found in the Deutsche Wahrheit (German Truth) group, which (re) posts in both languages, has a considerable audience of both German and Russian accounts, is reposted by the seed groups and works as a "hub" between the Russian and German communities on VK. 8. Golos Germanii is a multi-platform social media project moderated by Sergej Filbert from Germany, who engages in "public diplomacy" projects. It offers a contrasting case to the filtering of Western media by Russian state-run web portals dedicated to translating the foreign press. As Fredheim (2015) has shown, these widely-used portals select articles that confirm a stereotype of reporting about Russia as tendentious and hostile. In contrast, Golos Germanii translates "Russia-friendly" content and produces original videos with right-wing political and media activists.

Acknowledgments
The author thanks Dr. Gregor Lämmel and Maryna Voitsekhovska for developing the Python scripts for the project and two anonymous reviewers for their helpful comments.

Disclosure Statement
The author declares that she has no conflict of interest.

Notes on contributor
Tatiana Golova obtained her doctorate in sociology at the University of Magdeburg with a thesis on the links between space and identity construction in the radical left-wing milieu. She worked at the University of Magdeburg and at the Freie Universität Berlin. Since 2016, Tatiana Golova has been working at the Centre for East European and International Studies (ZOiS) in Berlin on a research project on mobilization networks in large Russian cities and, further, on transnational communication networks of post-Soviet migrants in Germany. Research interests: collective action, political radicalism, social networks, social media, migration.