Saying ‘Criminality’, meaning ‘immigration’? Proxy discourses and public implicatures in the normalisation of the politics of exclusion

This article explores political discourse in the context of an online-mediated 2021 rapprochement between Swedish ‘mainstream’ and far-right parties paving the way for their eventual 2022 electoral success and later joint government coalition. The article analyses specifically how the above political accord on the Swedish right – often seen as breaking the long-term cordon sanitaire around Sweden’s far right – would be legitimised via discourses that carried significant elaboration and deepening of the ‘criminality’ and ‘immigration’ connection later recontextualised into the broader Swedish public discourse and public imagination. Using social media analytics and qualitative, critical discourse analysis, we explore in depth a ‘discursive shift’ wherein the focus on criminality would become a key ‘proxy discourse’, i.e., a public-wide implicature, which, while referring to and debating a potentially genuine social issue would be strategically instrumentalised to effectively pre-legitimise ‘moral panics’ around immigration and cultural diversity. The analysis highlights that the emergence as well as the later recontextualisation of the ‘proxy discourse’ in question – implicitly suggesting that criminality, immigration, and cultural diversity are ‘somehow’ inherently connected – not only supported the political mainstreaming of the Swedish far-right’s anti-immigration stance but also normalised the wider tenets of illiberal, nativist ‘politics of exclusion’.


Introduction
The year of 2022 will probably be remembered internationally as one of the most tumultuous in European post-war history to date, and so also in the usually calm Swedish national political context.That year, namely, Sweden saw a complete paradigm change in many of the aspects of its, often internationally acclaimed, progressive social but also international and defence policy.Following several decades of military neutralitycoupled with pronounced humanitarian activity -Sweden eventually decided to join NATO military alliance in May 2022 in the face of, in particular, the ongoing Russian invasion of Ukraine happening in close vicinity of the Swedish territory.However, despite its huge significance, the declaration of joining NATO will probably still not be remembered as the main -and definitely not as the only -historically unprecedented occurrence in Swedish 2022 politics, especially if looking at the latter from a national-political point of view.Indeed, the fact that will probably remain the most notable would be, that, after years of its rather limited albeit more or less steadily growing presence in Swedish politics and the public sphere, 2022 saw the Swedish far-right being propelled to levels of public and electoral support that were unprecedented in Sweden and indeed even wider Europe.This became particularly evident when, following the national parliamentary election in September 2022, the far-right Sweden Democrats (Sverigedemokraterna, SD), 1 a strongly anti-immigration far-right party became the second largest in the Swedish parliament (Riksdag) with 20.54% support. 2 As it eventually turned out, a new Swedish minority coalition government was formed in October 2022 by three mainstream right-wing parties in close collaboration with SD.However, despite the party yielding formal support whilst gaining control over key parliamentary committees, SD would neither partake in any ministries nor receive any actual ministerial posts.
There are, to be sure, many aspects of political as well as social and economic dynamics that would have to be put behind, on the one hand, the gradual end of Sweden's once widely disputed 'exceptionalism' as far as the lack of wider presence and involvement of the far-right in the public life, and, on the other, the eventual rise of its political and public significance culminating in 2022 (see Krzyżanowski, 2018a;Odmalm & Rydgren, 2019;Rydgren 2022;Rydgren & van der Meiden, 2019).However, as this paper attempts to show, one of the main reasons to be considered here is certainly a very skilful mainstreaming of the wider nativist agenda and ideology of the Swedish far-right.This process was, as is shown here, significantly built on, as well as furthered by, political and ideological eagerness of the wider so-called 'mainstream' Swedish right to embrace the far-right agenda.
Indeed, as this paper argues, one of the key tenets of far-right ideology that have been strongly mainstreamed in Sweden in the last few years -eventually adding significant fuel to the far-right's recent electoral success -is, in particular, the anti-immigration, and multiculturalism-critical stance.Having gradually been growing in various types of Swedish media already since the early 2010s i.e. when the far-right re-entered Swedish parliament (see Bolin et al., 2015), the above frames have indeed escalated in recent years, and especially so during and in the aftermath of the critical period of 2015-2016 highpoint of the so-called 'Refugee Crisis' in Europe (Demker, 2019(Demker, , 2021;;Ekman & Krzyżanowski, 2021;Krzyżanowski et al., 2018).The above, coupled with the ongoing activity of the Swedish anti-immigration online uncivil society (Krzyżanowski & Ledin, 2017), eventually led to the significant 'spill over' moment with issues related to immigration and multiculturalism becoming a key focus of Swedish political discourses and agendas.These, however, became widespread not only on the far-right fringe of Swedish politics but also across the wider political spectrum with mainstream parties alike becoming ever more 'tough talking' with regard to immigration and cultural diversity (Krzyżanowski, 2018a).
However, as this paper elaborates, the wider negativised focus on immigration and diversity in the Swedish public sphere brought about, and progressively solidified, by the far-right has been in many ways not only temporally accompanied but also deepened and widened by the arrival of a related public discourse, specifically one on criminality.The latter, especially since ca.2018 onwards, erupted in Swedish politics and the media (Ekman & Krzyżanowski, 2021) and indeed beyond (Andersson et al., 2021;Milani, 2020) and became a dominating discourse in, in particular, the run up to the 2022 elections filled with criminality-related securitising arguments and debates.This eventually found a clear reflection in the post-election policy declaration of the 2022 new Swedish government formed by the right-wing parties with the support of the far-right SD. 3  That declaration of future goals and actions was, namely, dominated by the large number of recontextualisations of securitising language including when describing crime and criminals and, in a very similar vein, immigration.
It is indeed this connection that the current paper seeks to analyse and, insofar as possible, deconstruct, including in terms of its function in wider normalisationof the far-right anti-immigration and nativist agenda.Our argument is that the strategic coupling of the general focus on criminality as a social problem with implicit -albeit indeed ongoing and by now persistent -suggestion that criminality and immigration are 'somehow' deeply connected (see below) -has been at the core of recent debates in Sweden.It is that particular connection between criminality and 'otherness' (widely defined) and their effective, discursive-practical normalisation that have been, we argue, at the core of many recent public discourses -including in media and politics -eventually allowing for an even deeper, and wider, mainstreaming of the far-right's, anti-immigration-cum-securitisation agenda in the Swedish context.
As we explore and elaborate below, criminality has recently become in Sweden what we define as a 'proxy discourse' i.e. a discourse focusing on a potentially legitimate and genuine, vital social topic or issue (in our case, criminality and/or wider challenges to law & order in Sweden) which is, however, strategically coupled with, and hence effectively yielding legitimacy to, a set of other and in most cases exclusionary as well as nativist ideas and views.Thereby, 'nativist' refers to the discourse, actions as well as policies that promote the interests of 'natives' (however imaginary and elusive such category is) over the interests of an unspecified 'other' -in most cases embodied by immigrants and immigration -whose exclusion is thus also effectively normalised and 'pre-legitimized' (Krzyżanowski, 2014).It is also operationalised by means of the tools (such as, e.g. the aforementioned securitisation actions etc.) deployed allegedly to tackle the main, or 'proxy', social problem in question yet while de-facto also targeting issues that are strategically viewed as related (in our case cultural diversity).
Our analysis below looks specifically at the online discourse in general, and Swedish 'official' party-political profiles on Facebook (FB) in particular.We focus on the FB presence of the three key parties that, by now, have become the central players in the 2022ff Swedish government, i.e.The Moderate Party (Moderaterna, M), the Christian Democrats (Kristdemokraterna, KD) and the far-right Sweden Democrats (Sverigedemokraterna, SD).However, our focus is not on the tumultuous year of 2022 which saw the electoral victory of the coalition above but, specifically, on an earlier period, in late 2021, when the key moment of open, ideological rapprochement between M, KD and SD took place by way of alignment of their discourses and policy ideas performed on, and communicated via, social media, and most exponentially Facebook.
We analyse FB discourse using a two-tier model of critical discourse analysis of social media (Krzyżanowski, 2018b) which allows relating interactive and discursive aspects of examination.In doing so, we want to point empirically to the close interdependence of two simultaneous processes.On the one hand, we look into political-ideological processes of mainstreaming of the radical right and its ideologies, while on the other we also cover opportunistic radicalisation of the political mainstream including the reliance of this process on a deeper logic of normalisation of exclusion via discourses which not only argue for the securitisation but also criminalisation of immigration.Indeed, as we argue, the emergence of the major, carrier 'proxy discourses' that pre-legitimise exclusionary views by attaching them to debates about vital social problems and issues points to the fact that, as such, mainstreaming of the far-right builds on long-term normalisation processes and, indeed, multiple 'discursive shifts' (Krzyżanowski, 2018c(Krzyżanowski, , 2020a(Krzyżanowski, , 2020b and below) and below).These occur over time whilst making anti-immigration views and attitudes acceptable as well as legitimate elements of public imagination and political action.
The second half of the 2010s has meant a new stage in the process of mainstreaming of radical right populist (RRP) ideas eventually resulting in the fact that, today, the political debate in most of the Western world is dominated by sociocultural issues and identity politics, in particular in relation to immigration (Mudde, 2019).These issues have recently been mobilised by different actors whose goal has been to bring to the mainstream those ideas which, at least previously, would be reserved to the catalogue of far and radical right and would be largely considered as uncivil, politically-incorrect or as overall directed against the catalogue of values of liberal democracy (Krzyżanowski & Krzyżanowska, 2022;Moffitt, 2021).Accordingly, it has been shown that far-right parties can impact whether they and their ideas are perceived as mainstream, including by the way they behave and present themselves.This might be a way of broadening the political appeal and maximising voter support as well as a strategy to appear as possible and serious coalition partners to mainstream parties (Moffitt, 2021, p. 9).As Cammaerts (2018, p. 8) argues, many far-right parties have focused their strategies, notably in the media, to increase their respectability by making them sound and look ever more reasonable and legitimate.At the same time, however the far-right would often chose to retain its 'original' anti-establishment profile to attract voters sympathising with the more radical views (Akkerman, 2016, p. 278;Moffitt, 2021, p. 9) who still are a vital part of the farright's electorate.
In order to be perceived as mainstream, the far-right parties largely depend on other actors (Mondon & Winter, 2020, p. 112), and in most cases 'mainstream' parties, to potentially legitimise far-right agenda via common actions or even coalitions (Moffitt, 2021).This is at the core of the process of 'mainstreaming' of the far-right that, simultaneously, often entails mainstream parties becoming more radical and moving to the right, especially 'on the back' of adopting tough stance on standard far-right policy foci like immigration, integration, and law and order.Still, at the same time, despite spreading its image as acceptable, the far-right does not become any more 'mainstream' in the process and instead remains the standard bearer of radical political and social ideas and views.This is what Mudde (2019) refers to as radicalisation of the mainstream parties or 'the incorporation of populist notions into the political mainstream' (Curran, 2004, p. 38).
Looking, on the other hand, into the logic of normalisation which emerges as a wider process that underpins and to large extent precedes as well as follows the aforementioned politico-ideological mainstreaming, scholars like Krzyżanowski (2020a, p. 508) have argued that anti-immigration opinions and views are not only expressed livelier and more widely than ever within the public domain but that also the resulting, persistent anti-diversity opinions become increasingly seen as acceptable.This builds on a recurrence of discourses wherein immigrants are often framed as a 'threat' or 'danger' to (however imaginary) native 'people' and 'society' and/or viewed as a 'cost' and an 'economic burden' (Buonfino, 2004).Further, there are also evident patterns of politicisation and mediatisation of immigration (Krzyżanowski et al., 2018) thatpave the way for an 'obvious radicalisation of immigration-related voices and views and a perpetuation as well as eventual normalization of often discriminatory views in the public domain' (Krzyżanowski, 2020b, p. 508).
In such a view, normalisation tends to be seen as part of a multistep process of strategic discursive shifts (Krzyżanowski, 2018c(Krzyżanowski, , 2020b)).Such discourses, especially those characterised by calls to discrimination, are enacted, perpetuated, and eventually normalised in line with pronounced far-right strategies.Consequently, this creates path dependencies for the even deeper practices of not only rhetorical/discursive but also physical or systemic exclusion.The analysis of discursive shifts hence makes it possible to identify how and when public and political discourses transform and become politicised and mediatised (Krzyżanowski, 2018c, p. 79) and how, down the line, they pre-legitimise the everstronger expressions of discrimination and exclusion.
To some extent irrespective of the context in which these processes occur, discursive shifts also rely strongly upon construction of various imaginaries (see Jessop, 2012;Krzyżanowski, 2019;Taylor, 1993) wherein the often deeply untrue and unrealistic visions and ideas of the 'imagined' other can eventually be mobilised -and effectively perpetuated -as stable elements of public discourses pre-legitimising the wider politics of exclusion.Crucially, imaginaries central to normalised discourses of exclusion combine elements of the real and the unreal and promulgate more or less objective facts and processes identifiable in the social reality with imagined visions of how society was, allegedly, functioning in the past or how it apparently could/should be functioning in the future.
A particular inspiration for the understanding of the above process as the core element of discursive shift logic -which we view as underlying for the emergence of 'proxy discourse' we explore below -derives from the work of Graham (2019) who focussed on normalised -and in particular negativised -utopias of the social and the political as central to the dynamics of contemporary public discourse.Graham argued that enactment and recontextualisation of negative visions of society -such as, e.g.finding itself in constant 'danger' or 'decline' -has been central in institutionalising a process whereby social control can be effectively pre-legitimised via a number of prohibitive measures or 'thou shalt nots (…) expressed in legislation, literature, news or multimodal texts of any kind at all' (2019, p. 3).
Hence, as Graham showed, traditional combination of the real and the unreal, typical for the construction and legitimation of wider utopian visions (see also Levitas, 2011Levitas, , 2013)), is often operationalised within normalising practice not in order to find constructive ways forward for the wider social development and/or improvement of social relations.It is, instead, a powerful tool in institutionalising power of selected political actors -in our case both mainstream and far-right ones -whose aim is to thus govern society by way of 'moral panics' (Cohen, 1972(Cohen, /2011;;Krzyżanowski, 2020b) and as if 'logically' related measures to securitise and control enforced via diverse 'public pedagogies' (Graham & Dugmore, 2022) that effectively perpetuate normalisation of exclusion.
Incorporating the above is hence vital for our perception of the focal proxy discourses as profoundly reliant on both negative vision of the (imagined) 'other' -seen as one 'outside' and 'within' the community -as well as on the simultaneous profoundly negative perception of the developments in the wider society and social reality.We argue that, through proxy discourses, namely, the strategy is to introduce a discursive shift which not only emphasises but also exaggerates social problems -in our case below: criminality -yet does so not in order to find ways of addressing such problems but to stigmatise the other as both explicitly and implicitly responsible for the negative developments in the wider society.All of this relies on wider utopias of the social that promulgate 'retrotopia' (Bauman, 2018) that implies the once 'positive' state of society -which should thus become good 'again' -with the vision of its negative present and future which can only be, as is assumed, avoided through control of the 'other' who allegedly is the reason for the current, and potentially also future, negative state of affairs.

Swedish 'Mainstream' and far-right
Swedish right-wing parties spent most of the twentieth century in the shadow of the Social Democrats (S), i.e. the country's major mainstream left-wing party.Having been for decades in power, the Social Democrats also laid foundations for most of the Swedish post-war policies including those related to immigration, refugees and humanitarianism (Krzyżanowski, 2018a).The mainstream right-wing, for its part, has at the same time been rather dispersed and since the late twentieth century effectively consisted of four political parties: M and KD mentioned above, the Centre Party (C), and the Liberal Party (L).Of these, M has usually been the dominant one and often also the second largest party in the Swedish parliament.The right-wing parties have since the 1990s formed coalition governments for altogether three terms: in 1991-1994, 2006-2010, and 2010-2014.The key contemporary Swedish far-right party, the SD, was founded in 1988 and originated in different far-and radical-right groups (Widfeldt, 2008, p. 267).From the very beginning, it was a solitary party with an outspoken anti-immigration and pro-nationalist stance (Salo & Rydgren, 2021, p. 38).For decades, SD was viewed as illegitimate by most voters and mainstream parties (Rydgren & van der Meiden, 2019, p. 450).However, in 2005, a new party leader, the then 26-year-old Jimmie Åkesson, was appointed with a goal to speed up the modernisation of SD and of the party's ideology.Since then, at least nominally, SD attempted to move away from its radical roots with the ambition to fill a nationalist ideological niche under the banner of Swedish 'Folkhemmet', or 'the peoples home', a concept originally established by S in the late 1920s and still carrying a significant mobilising force in Swedish politics (Widfeldt, 2008).
In the 2010 election, SD had its electoral breakthrough and entered Swedish parliament.Nevertheless, its political influence was still heavily limited as all mainstream parties joined a cordon sanitaire around the SD (Widfeldt, 2015;Rydgren & van der Meiden, 2019).While SD went on to advocate repatriation of immigrants and remained opposed to multiculturalism, PR-attempts were also made to simultaneously soften the party's image.Probably as a result of that duality, SD significantly increased its share of votes in the elections in 2014 and 2018, with especially the latter electoral success often ascribed to the increased politicisation of the immigration driven by SD within and outside parliament on the back of the 2015 'Refugee Crisis' (Rydgren & van der Meiden, 2019).
As SD's share of votes has progressively increased, so has also the interest in approaching SD and its policies among the other mainstream right-wing parties.Hence, especially M and KD, and eventually also the Liberal party (Liberalerna, L), gradually adopted more benevolent attitudes towards SD.However, the rapprochement between M, KD, and SD became evident only during the fall of 2021: at first via a debate article on gang criminality in Sweden published by the three parties in the daily Dagens Nyheter in October 2021, and secondly, via a joint budget proposal in November 2021.Partially as a result of by then well advanced mainstreaming of the SD agenda, the party reached 20,54% of votes in the 2022 national elections thus becoming Sweden's second-largest parliamentary party.Despite the fact that M and KD suffered a decline of support at the election, together with L and under the 'backseat' support of SD they could still form a minority right-wing coalition government.

The politicisation of immigration and of its (Implicit) connection to criminality in Sweden
From the 1980s onwards and up until the 2015 'Refugee Crisis', Sweden was one of the main asylum recipients in Europe (see Krzyżanowski, 2018a;Krzyżanowski et al., 2018) while also long remaining a country where immigration was not strongly politicised contrary to many other Western and Northern European countries (Widfeldt, 2015;Berg & Oscarsson, 2015).Additionally, except for one term during the 1990s, there was for long no RRP party in the Swedish parliament.While there are different explanations for this, it appears that, on the one hand, there was a consensus among the political parties in the parliament that Sweden's refugee policy should be characterised by humanity and generosity (Andersson & Bendz, 2015), while, on the other hand, Swedish politics was long overwhelmingly structured around socioeconomic issues with sociocultural ones remaining of a much lesser public concern (Rydgren & van der Meiden, 2019).Contrary to many other countries, Swedish voters were also late to abandon the traditional political distinctions on the left-right scale -which traditionally created space for the far-right elsewhere -and continued identifying with parties that traditionally represented their interests (e.g. S for the working class, etc.; see Rydgren, 2002).
It is, therefore, only with SD's parliamentary access in 2010 that the immigration issue was also moved onto the wider, general political agenda.However, this still happened when all mainstream parties continued to advocate a humanitarian and open approach to immigration, a trend which de facto changed only in the aftermath of the 2015 'Refugee Crisis'.At that time, namely, the political debate clearly began to increasingly revolve around sociocultural issues related to immigration, with, furthermore, parallels extensively drawn between migration and the rise of criminality.In the following years, this discussion would become even wider via many institutional analyses and reports about growing levels of criminality in the country (Svedin, 2021) eagerly picked up by the national as well as international far-right actors.These often alluded to 'lawlessness' now being widespread in Sweden as a result of immigration -with, e.g. one significant example coming from the former US President Donald Trump who back in 2017 stated that: 'Sweden.They took in large numbers.They're having problems like they never thought possible' (BBC, 2017).
The image of Sweden as 'lawless' soon also became reinforced in the general, nationalpolitical debate and contributed to making immigration, criminality, and especially organised or 'gang' crime, into the most important and alarming issues for Swedish voters in the election of 2018 (Martinsson & Andersson, 2021).This new sociocultural dimension in Swedish politics also caused a split between the traditional right-wing coalition parties since M suggested stricter regulations for immigration while, e.g., the Centre party (Centerpartiet, C) opposed those (Aylott & Bolin, 2019, p. 1506).
In the years that followed, crime, and especially shootings ascribed to criminal gangs attracted even further attention in Sweden.This came along with the arguments that lethal violence with firearms has increased in Sweden since 2013 while at the same time decreasing in most other European countries (BRÅ, 2021a).The increased instances of shootings have often been connected to 'gang' criminality and put into wider context of state failure including, prominently, as far as immigration and integration policies (Svedin, 2021, p. 590).The attention on the shootings increased after shooting of a 12year-old girl in a Stockholm suburb in August 2020 and intensified even further as of 2021, mainly due to a number of incidents.The above resulted in many calls for the situation to be 'finally' handled (Andersson et al., 2021, p. 334) with the debate, however, ongoingly drawing on implicatures that there exists a connection between immigration (or at least immigrant-populated areas) and criminality.For example, Sweden's Deputy National Police Commissioner claimed in an interview on national public radio that ca.forty clan-based networks had established themselves in Sweden for the purpose of committing crimes (Sveriges Radio, 2020).On the other hand, an official report about differences in levels of criminality between people with domestic and foreign backgrounds was presented (BRÅ, 2021b) effectively claiming that migrants were overrepresented in criminal activities (Svedin, 2021).
Interestingly, most of the above arguments eventually often proved to be to at least some extent incorrect or emerged as a rather slanted and non-nuanced way of reading Swedish statistics on criminality if/when connected to immigration.Indeed, even the aforementioned 2021 official report claimed that, there has been a decrease in the proportion of people suspected of crime within all groups of the entire Swedish population (BRÅ, 2021b, p. 10; see also Adamsson, 2020) including among those with an immigrant background.At the same time, several other authoritative studies (see, especially, Sarnecki, 2021) pointed that, despite frequent suggestions about a potential correlation, neither violent and serious nor even petty crime has increased in the aftermath of the 2015 'Refugee Crisis' i.e. the period when the number of foreign-born population has significantly increased in Sweden.Nevertheless, with the connection of criminality and immigration effectively already well established in the wider public domain and persisting in the public discourse, the 2022 election also eventually revolved around these issues.And, with M, KD, and SD apparently remaining traditionally among the most trusted parties regarding law and order, those parties unsurprisingly emphasised the alleged centrality of the issue of 'gang' criminality and called for stricter punishments for 'gang' criminals in addition to other securitising law and order measures.Equally, they at the same time often made election 'promises' concerning much stricter limits concerning immigration and asylum.

Analytical framework
This study draws methodologically on the wider Critical Discourses Studies -and specifically on the CDS' Discourse Historical Approach, or DHA (Krzyżanowski, 2010;Reisigl, 2017;Wodak, 2001) widely deployed to the analysis of political and media discourses in connection with the far-right and its public-wide influence (see in particular Krzyżanowski & Wodak, 2009;Wodak, 2015).The DHA is mobilised here within a framework geared specifically towards analysis of social and online media (see Krzyżanowski, 2018c), which, originally developed in the context of Twitter-oriented analysis, is reapplied here to the analysis of FB data.
The above, social-media-centred and DHA-based framework connects two areas of analysis.On the one hand, there is the interactive analysis (see 4.3.)which -more widely helping underpin our examination of the mainstreaming of far-right discourse among the other parties -reveals patterns and tendencies in communicative interactions that occurred within and via FB posts.This analysis depicts the level of interactive reliance on various actors, their agenda-setting role in the wider interactions and diversity of sources.On the other hand, connecting to the first area of the analysis, the framework followed here also looks more specifically into the discourse patterns and dynamics (see 4.4.) which are followed via the usual 'levels' of critical discourse analysis in DHA (see Krzyżanowski, 2010) wherein an entry level analysis of discourse topics is performed prior to exploring the discourse's in-depth features mainly from the point of view of the deeper structures of argumentation and related discursive strategies.
Whereas the focus on (a) interactivity analysis allows pointing to the mainstreaming as it happens and is additionally performed in the communicative domain (specifically on official social media) -including by means of recontextualisation of posts and their contents across the three parties in question, the related (b) discourse-centred analysis allows exploring the qualitative features along the logic of the focal 'proxy' discourse and its strategic construction as well as its role as evidence of the deeper/wider process of normalisation of exclusion (Krzyżanowski, 2020a(Krzyżanowski, , 2020b)).Analysing the former and the latter helps both exploring the key features and dynamics of the discourses forming the explored discursive shifts as well as allow pointing to how the proxy discoursesspecifically geared towards criminality -and the wider normalised discourses of exclusion are formed and reinforced at both thematic (entry) and argumentative (in-depth) level.

Empirical data
The empirical data analysed in this study stems from posts published on M's, KD's, and SD's official Facebook pages between 3 October and 1 December 2021.This period is of particular interest for this study since M and KD actively collaborated with SD for the first time during this period.The collaboration specifically oscillated around three events, i.e.: (a) Publication of a common debate article in nationwide daily Dagens Nyheter on 10 October 2021; (b) A common cross-party SD, M, and KD budget proposal presented on 16 November 2021 (Wikén and Knutson, 2021); and (c) The passing of the above budget proposal through parliament on 24 November 2021.
All posts from the official Facebook pages of M, KD, and SD published between 3 October and December were collected manually by the researchers.Posts that did not contain thematically relevant information were not included in the data.Additionally, the data did not include any published live posts, mainly as it was the direct and concise communication on Facebook that was the focus of the analysis.The analysis of the posts took place on the original Swedish language versions, and only selected examples below were translated into English (all translations below are hence ours).Images were not part of the analysis due to limitations of space, and were analysed only whenever they included significant textual elements.
The result of the data collection was 155 posts for M, 92 for KD, and 303 for SD.While the entire corpus above was put under the interactive analysis, in order to prepare the sample for the qualitative analysis a down-sampling strategy was applied while limiting the posts eventually put under examination to those containing topics and arguments usually associated with far-right discourse on immigration and wider social issues (see especially Reisigl & Wodak, 2001).This allowed far-reaching disambiguation of some parts of the corpus and excluding posts which, while qualified via keyword sampling, often concerned topics that were not of relevance here such as, e.g.electricity prices, benefits of nuclear power, statements about political parties from beyond the rightwing, and results of opinion polls.In the end, the sample which underwent the discourse-oriented analysis consisted of 85 posts by M, 44 by KD, and 183 by SD -which also emphasised the SD's traditionally very strong focus on FB communication; see Larsson, 2018, 2022 4 ).
Looking into the time-based distribution of posts over the chosen period, it is possible to reveal certain peaks in the publishing of posts connected to the three events of particular interest for this study.This is most obvious regarding SD (Figure 1).

Interactive analysis
Our interactive analysis seems to confirm the above data which showed that, of the three parties in focus of our analysis, it was the SD that remained the most active in general as well as in terms of initiating interactions or exchange of ideas in particular.It was also, accordingly, the SD that retained the agenda-setting role that effectively enabled mainstreaming of its ideas and views in FB interactions with other right-wing parties and by way of their eventual picking up on proposals and views furthered by SD.
As is presented in Table 1, FB posts of SD differed significantly from those of M and KD.First, almost every third of SD's posts was shared from/by party representatives, and almost half of those were shared by the party leader.This indicated that the posts of SD to a significant degree consisted of personalised messages, or at least such that strategically were supposed to appear as such.This was seen as, indeed, a typical feature of populist discourses online (see, e.g.Forchtner et al., 2013) which usually follow the pattern whereby a specific 'polyphony' is created by way of parallel communication: the official one 'by the party' and the somewhat informal one 'by the leader' (though with both of these in most cases being run by party officials/employees and only appearing to be of different type/character/source etc.).Notes: All shared posts are not shared Facebook posts.In the cases of M and KD, screenshots from party representatives' Twitter accounts that have been published on the parties' official Facebook page are counted as shared posts.This is also the case with a few of SD's shared posts.
Secondly, links to external websites and chiefly news sites were far more present in SD's than in other parties' posts. 5By linking to different sources of news, SD was, however, not only extensively commenting on issues/events but also laying out its own perspective on current events and news -thus to large extent also solidifying its opinion-making and agenda-setting role. 6Therein, the SD very often put its policy in the wider context while effectively legitimising and pre-legitimising the party's proposals as if 'responding' to wider news and ideas circulated in the public sphere (for example, growing anxiety with criminality, etc.).
The most linked site, however, was still Riks channel on Youtube which has been developed as SD's own hyperpartisan tool for spreading 'opinions' and 'news' while also effectively entering interactions with established media outlets.Looking, on the other hand, at the cross-party interactions online, the analysis below uses 'Circos plots' (Krzywinski et al. 2010) covering our period of investigation (Oct-Dec 2021).Exploring development over time, and looking into how/which of the parties retained a leading voice based on their posts and their interactivity scores (in particular in mentioning of the other parties; see Figures 2-5), 7 our analysis mainly focusses on explicit mentions of both specific parties and their leaders.Therein, each line in the figures represents one post that mentions another party or that party's leader, and ends with an arrow indicating the party that is mentioned.The position of the start of each line represents the sequence of the Facebook posts arranged clockwise from the first to the last Facebook post in the time period.The length of each circle segment represents the total number of Facebook posts by each party during the time period.Each tick marks 10 Facebook posts.
As a result of our interactivity analysis, we can see in the plots that SD posts comprise of the largest proportion (55%) of the total number of Facebook posts during the period 1 October 2021-31 December 2021 (SD = 303, KD + M = 247), followed by M (28%) and KD CRITICAL DISCOURSE STUDIES (17%).Focusing at first on KD (Figure 2), one can see that KDs posts relatively frequently referred to the other two parties mentioning both parties together, specifically from post 56 (16 November) to post 88 (1 December).However, while KD was still relatively eager to mention the other two parties, seems the strategy by M (Figure 3) was significantly different as it, effectively, was very rarely mentioning the other two parties directly in their Facebook page posts.This, however, would be typical given the then already emerging synergy between M and SD and that in some matters M preferred to remain mute while SD took a stand for, implicitly, both parties (see, e.g.those parties' corresponding behaviour during the later televised pre-election debates in August/September 2022).Finally, focusing on the SD (Figures 4 and 5), one can observe that there seem to be more direct mentions of other parties, in particular of M, across SD's Facebook page posts.Notably, there is a closely packed series of posts over just 2 days that mention both M and KD together between posts 192 (16 November) and 205 (17 November).

Entry-level & in-depth (Critical) discourse analysis
The entry-level analysis of the collected data focused on the identification of topics connecting criminality to immigration issues in RRP discourse, which, using abductive analysis, eventually were identified in the studied texts.Given that the topics were extensively interconnected, some of them have been considered as sub-topics though under assumption that in several cases even greater thematic complexity would have occurred between various themes in the studied material.
In any case, as the outline of the topics shows, the topics reveal the enactment of antiimmigration discourse and the use of arguments in the texts well-known from far-right discourse.As seen in the table, all of the topics were furthered by the SD, but the most were also present within the texts of M and KD.Common to the three parties was, equally, that they established a connection between immigration and criminality, especially 'gang' criminality as such (Table 2).In order to show how the above thematic connections were qualitatively established and furthered, the in-depth analysis below focuses on the key argumentative structures deployed within realisation of the topics above, wherein we point especially to a number of key topoi and other discursive strategies (Krzyżanowski, 2010;Reisigl & Wodak, 2001) used to structure and support the argumentation.The analysis focuses in particular on the discourse strand embedded by Topic 1 within which immigration and crime as such were initially connected, and hence wherein the 'proxy' focus was also established, before strategically, continuing discussions on those topics in isolation, i.e. while deepening them yet also while retaining implicit connection between the two focal processes/issues.In order to depict that tendency too, we also chose to highlight, insofar as possible, the in-depth level of realisations of Topic 2 ('gang' criminality) while showing how, despite focussing explicitly on criminality, it still retained many implicit cross-references to, and critical stance towards, the wider semantic fields of immigration, diversity and the like-indeed remaining a case of 'borderline discourse' (Krzyżanowski & Ledin, 2017) i.e. avoiding explicitness and retaining a quasi-politically-correct stance.
Exploring the topics and sub-topics above that were central to the establishmentand solidification -of the 'proxy discourse' in our focus, one observes the recurrence of the number of more or less classic argumentation frames -or topoi -usually deployed in anti-immigration discourse of the far-right (thus also emphasising the wider, normalised character of argumentation).Among the key topoi there were, in particular: (a) Topos of threat -Arguing that migrants pose a physical as well as otherwise understood threat to Swedish society (Reisigl & Wodak, 2001;Ekman & Krzyżanowski, 2021); (b) Topos of criminality -a variation of the classic topos of law and order (Reisigl & Wodak, 2001) which focuses explicitly on making a connection between immigration and criminality and on representing immigrants as involved in criminal activities, (c) Topos of securitisation -again a variant of the classic law & order framing (above) though with explicit arguments in favour of 'securitising actions' including investments in more police, increased police salaries, stricter prison punishments, expulsions, stricter border surveillance, etc..
Starting with the arguments constructed on the basis of Topic 1 (Immigrants and Foreigners as Criminals), it appears that the argumentation in the collected texts was related to different types of crime and has either been framed within the topos of criminality or the wider topoi of threat or of law and order.However, these topoi were closely overlapping, which is why arguments would be occasionally framed by all three arguments.In the collected texts, SD emerged as the one most explicitly connecting Table 2. Key topics and sub-topics related to the alleged connection between immigration and criminality in M, KD, and SD Facebook posts.
Topics and Sub-Topics Party 1. Immigrants and foreigners as criminals.M. KD, SD 1a.'Clans' targeting the welfare system to finance organised crime.
M, SD 1b.Increased reports of honour violence.SD 1c.IS-members returning to Sweden from Syria.
M, SD 2. 'Gang' criminality as the major challenge of the Swedish society.
M, KD, SD 2a.'Gang' criminality as a consequence of immigration.
M, KD, SD 2b.The need for investments to maintain law and order.
M, KD, SD 2c.The need for stricter prison penalties and expulsions.
M, KD, SD immigration to criminality, both regarding foreigners in general, and immigrants in Sweden in particular: Honour violence is an imported problem.
Over 4,500 honor crimes reported in two years.
(SD, 2021, November 1) Knife attack on a train in Germany -perpetrator with Arab background arrested.
(SD, 2021, November 'IS woman makes slave girl thirsty to death in the sun in Iraq.'It is these murderers the government flies to Sweden.The terrorist's citizenship must be cut.
(SD, 2021, October 27 The examples above were all expressions of discursive strategy of 'functionalisation' (Van Leeuwen, 2008) as they are associating foreigners and immigrants with various crimes as a specific form of activity.In all three cases, however, there were links in the posts to 'news' media, with two of those, though, specifically the hyperpartisan media of the SD (the newspaper Samtiden and the Youtube channel Riks).U sing these allowed SD to highlight connections between crime and immigration without expressing it 'themselves' and while claiming that the statements rely on 'facts', or reports/descriptions thereof, from external sources.This could also be considered a strategy of 'disclaimer' or 'denial of racism ' (van Dijk, 1991' (van Dijk, , 1997) ) or of a related 'blame avoidance' (Hansson, 2015), employed by SD to not to be accused of racism or xenophobia while still expressing their statements.It is obvious, though, that the linked news is also carefully selected and commented to spread the perception that immigrants are largely responsible for various types of crime.It is an example of 'constructing scapegoats and enemies -'others' -which are to blame for our current woes -by frequently tapping into traditional collective stereotypes and images of the enemy'.(Wodak, 2015, p. 4).Additionally, the two latter texts are also expressions of what Krzyżanowski and Ledin (2017) define as borderline discourses.In all probability, SD would not, in the first case, push that topic, and in the second case, use such an expression at all in speeches in different and especially more formal contexts (e.g. in parliament) whereas it does so online, i.e. at the borderline between the private and the public expression and within the usual site of production of civil/uncivil borderline.In this context, it is also telling that the 'news' that is quoted and linked to is presented not only online but in particular by SD's hyperpartisan media largely responsible for the enactment and the diffusion of borderline discourses and exclusionary views.
Similarly, texts in the posts related to 'gang' criminality include an increasing number of references to shootings in Sweden wherein immigrants are highlighted as responsible or at least complicit.This framing is, however, deepened by associating the shootings with, e.g.certain vulnerable areas which in most cases are chiefly known for their migrant populations.Similarly, however vague, the act of connecting of 'gang' criminality with unrelated 'gangster rap' ('relating unrelated phenomena' as a strategy; see Reisigl & Wodak, 2001) serves largely the same purpose: it implies that both criminal and cultural activities in question are in/directly connected to first or second-generation immigrant youth.
Since 2018, 17 young men have been shot dead in Järva. 8  (M, 2021, October 27) Every second year, the police release a list of areas in the country where criminals have a strong impact on the local community.These are areas in which the majority society is not sufficiently present.(Andersson et al., 2021, October 14) We want to commission BRÅ 9 to investigate the impact gangster rap has on crime and vice versa.Does it generate recruitment for criminal networks?(SD, 2021, October 29) Such strategies as the above allow establishing a connection between 'gang' crime and immigration.They pre-legitimise M, KD, and SD calls to stop this 'criminality' while de facto also constructing the (implied) role of migrants as multidimensionally active in the process.These suggestions are partly leading to various 'parallel' conclusionsindeed typical for the proxy discourse -about, inter alia, reduction of the number of immigrants in Sweden by 'strengthened border protection' (SD, 2021, November 17) on the one hand, and more expulsions of 'gang' criminals (M, 2021, October 10) on the other.The dynamics above -realising several topoi including topos of securitisation as an apparent 'solution' -derives from standard discursive elements of typical construction of 'moral panics (Cohen, 1972(Cohen, /2011;;Krzyżanowski, 2020a) whereby exaggerated mood of insecurity is formed in the public discourse, and a group of persons is defined as (vaguely) responsible for the said threat.Such threat and moral panic can -and often wouldoften be furthered and substantiated as is the case, when, e.g.M refers in its posts to 'gang' members as 'Sweden's domestic terrorists' (M, 2021, October 22).
Another group of immigrants that is turned into an explicit scapegoat of both M and SD discourse are the widely used, yet never unspecified, 'clans'.In several of the posts, hence, M and SD are also connecting clans to the new (immigrant) arrivals, while also amplifying statements with cross-references to welfare abuse/fraud and organised crime.
Recently, it was revealed that a clan has defrauded taxpayers of SEK 40 million.Every year, about SEK 20 billion are cheated away from welfare -money that is taken from welfare to finance organized crime.(M, 2021, November 12) It is obvious that organized crime in recent years has targeted the welfare system as a kind of basic income.Welfare crime is increasingly characterized by clan-related and organized crime.The government has failed to take control of crime, which has affected, for example, people who really need to take part in welfare.(SD, 2021, November 11) This argumentation is, as previously, articulating the topos of threat, and doing so in two ways.On the one hand, the 'clans' are identified (personalised, though still in an aggregated/unspecified form; Van Leeuwen, 2008) as a threat to the Swedish welfare system.
On the other hand, the 'clans' are also viewed as financing organised crime, which in turn is interpreted as a threat to the wider security of the country.Additionally, the texts are also examples of what can be defined as calculated ambivalence (Wodak, 2015) or conceptual flipsiding (Krzyżanowski & Krzyżanowska, 2022, in press; see also Krzyżanowski, 2016) that allows for the construction of 'humanistic' or 'patriotic' arguments in a way that expresses concern for native citizens while de facto creating a 'flip-sided' exclusionary argument against migrants and other (usually unspecified) 'them' rooted in the need to 'protect' the native population (Krzyżanowski, 2018a;Ekman & Krzyżanowski, 2021).Hence, M's and SD's statements regarding the 'clans' are designed to be seen as an expression of care for the Swedish 'taxpayers' (M, 2021, November 12) as the 'people who really need to take part in welfare' (SD, 2021, November 11).
On the other hand, arguments and strategies encompassed by Topic 2, related to 'gang' criminality (Gängkriminalitet 10 ) and deepening representations of its threatening character, emerged in the analysed discourse among the key frames that in/directly structured representations of diversity and multiculturalism and their mid-and long-term socio-political implications.Those issues remained in the centre of focus of M, KD, and SD including in their joint proposals as both those parties' common debate article in Dagens Nyheter in October, 2021 (Forsell et al., 2021) as well as the common budget proposal in November, 2021 largely pertained to how the 'gang' criminality should be curtailed and controlled (M, 2021, November 16;KD, 2021, November 16;SD, 2021, November 16a).
This is what an ordinary day in Sweden looks like nowadays: 13.30: Man killed after stabbing and explosion in Bagarmossen.Are you frustrated and maybe like me worried about the situation in Sweden?Maybe you're even scared?In no other country is firearm violence at such levels as here in Sweden and city after city is put into fear.Sweden needs a change.(KD, 2021, October 28) This Monday, German Bild, the country's most-read newspaper, published an article about Sweden in which our country is labelled 'Europe's most dangerous' and tells that murder statistics are declining throughout Europe -while at the same time increasing in Sweden.Bild is appalled how a role-model country now has turned into a 'nightmare'.(SD, 2021, October 27) 'Gang' criminality, as a form of unspecified 'them', was supported by argumentation based on the topos of criminality and topos of law and order but, strikingly frequently, also on the topos of threat.Facts in the texts above are constructed by way of claiming that shootings and lethal violence with firearms have increased in Sweden and that innocent citizens as well as policemen have been killed or hurt (Bengtsson & Samuelsson, 2020;Thambert et al., 2021;TT, 2021;Svensson, 2021).This, however, remains an overexaggerated element of the discourse as those are other 'gang' members -and far less frequently 'regular' members of the population or of the police -which are most commonly affected by, or become victims of, the criminal activities .Hence, not diminishing the severity of gang criminality, it appears that the theme is used to amplify the threat and scare logic via generalising claims of entire Swedish society being 'threatened' or affected by shootings and other 'gang-criminal' activities.
Crucially, the argumentation about 'gang' crime framed within the topos of threat was furthered by M and SD when highlighting Sweden as the most dangerous country in Europe.This is an example of intertextuality linking to an article in the German magazine 'Bild' in October, 2021 making similar claims (Raagard, 2021).Interestingly, despite Bild's known status as tabloid newspaper (Media Bias / Fact Check, 2022), SD referred to it as 'Germany's largest newspaper' (SD, 2021, October 27) while M, on the other hand, did not even refer the Bild's status as a source and directly recontextualised its ideas that, because of the high level of criminality, Sweden has 'been singled out as the most dangerous country in Europe'.(M, 2021, November 10).
By founding their arguments on the topos of threat, M, KD, and SD made the increased firearm violence between criminal 'gangs' appear as a threat to all Swedish citizens and the safety in the country.M and KD also, apparently, purposefully emphasised the emergency-like status of the situation to, effectively, pre-legitimise their own rapprochement with SD and the breaking of the wider, long-standing cordon sanitaire around the far-right party.M and KD expressed this in an almost identical way: 'Putting the political game aside' in this context should be understood as the pre-legitimatory statement assuming that, since 'gang' crime situation became so critical, M and KD now make themselves available to cooperate with all parties to manage the situation.This was pre-legitimised by arguments that 'soon it will be too late' characteristic for rightwing populist actors' efforts to take advantage of various crises (Albertazzi and McDonnell, 2008, p. 5).

Conclusions
The analysis above has aimed to reveal how diverse themes and arguments related to immigration, immigrants, and criminality have been deployed as well as connected in the political and discursive rapprochement between Swedish right-wing parties M and KD with their far-right counterpart SD in late 2021.As the analysis shows, discourses describing and legitimising the said rapprochement relied on the construction of the mood of 'crisis' around immigration by connecting 'dangers' related to criminality with immigration as well as wider multiculturalism and diversity.This is a process which has long been spotted among scholars (e.g.Albertazzi & McDonnell, 2008;Krzyżanowski, 2018aKrzyżanowski, , 2018cKrzyżanowski, , 2019;;Krzyżanowski & Krzyżanowska, 2022;Moffitt, 2016;Wodak, 2015;) arguing that taking advantage of crises, real or constructed, is frequent if not outright central in far-right politics and discourse.In both the former and the latter, construction of (in most cases negative) imaginaries and even utopias of the 'other' (Graham & Dugmore, 2022;Krzyżanowski et al., 2023) remains central for the pre-legitimation as well as the ensuing normalisation of anti-immigration and/or nativist opinions.They derive from pedagogy of negativity which, apparently, becomes indispensable part of the mainstreamed and normalised far-right anti-immigration scare (Graham, 2019;Krzyżanowski, 2020a).
As the analysis above reveals, the main strategy that emerged from the discourses entangled within the political reproachment between Swedish far-and mainstreamright has been the one resulting in establishment of a proxy discourse on criminality which, while using social concerns around criminality , used related anxieties to construct the wider scare of immigration while thus enabling criticism of -as well as atmosphere of hostility towards -migrants and immigration on the one hand, and issues of multiculturalism and diversity on the other.As indicated above, the establishment of the proxy discourse in question took place within a wider, multi-step strategy.The latter, only initially connected immigration and criminality in an explicit way, and eventually separated the issues while furthering -and, as is known further recontextualising -argumentation around different 'components' of the general claim.This allowed maintaining central the wider 'public implicature', that, as such, criminality and immigration are somehow' connected and, per se, all might/would, or already do, constitute a threat to contemporary Swedish society.
Our analysis explicates what Ekman and Krzyżanowski (2021), Wodak (2015) and others have highlighted i.e. that criminality becomes one of the central foci of contemporary right-wing populist discourse.The latter often aims to simplify the wider social, political and economic ontology of social anxieties and problems -including criminality -by arguing that the roots of criminal and anti-social behaviour can -and according to the far-and often mainstream right indeed should -be connected to immigration and migrants as such (see also Pelinka, 2013).However, as our analysis shows, this discourse now acquires a new set of dynamics wherein criminality is not only connected/related to immigration in the somewhat usually vague and generalised way, but, as has been shown, it becomes an overarching proxy frame for the general (mis)perception of cultural diversity even when the latter, or immigration for that matter, are not even anymore mentioned or made explicit.
As has been highlighted, the three focal parties in our analysis have deployed strategies as well as rhetorical and linguistic toolswell-known from far-right discourse, a trend that was clearly supported by the far-right SD's dominating and agenda-setting as well as, effectively, 'normalising' voice and the one 'mainstreaming' the far-right agenda.This, as was also indicated at the outset of the analysis, constitutes evidence that, during the rapprochements between far and the 'mainstream' parties such as the ones highlighted here, those are the radical-right, highly-negativised perceptions of immigration that effectively gain ground and become significantly mainstreamed and normalised.Obviously, and not unexpectedly, the far-right emerges from this connection as the political actor that gains the most: while remaining the most explicit in anti-immigration stance, it also becomes the agent behind the wider 'discursive shift' (Krzyżanowski, 2018a(Krzyżanowski, , 2018c(Krzyżanowski, , 2020b) that sees the general change and de-facto radicalisation of views as well as solidifying new exclusionary discursive proxies and public implicatures that are closely related.

Notes
University, Sweden, where he has also been Research Assistant in at Department of Informatics and Media since 2022.Since 2023 he is also a PhD Candidate in Media and Communications at the University of Gothenburg, Sweden.Info: https://www.gu.se/om-universitetet/hitta-person/hugoekstromMichał Krzyżanowski is one of the leading international experts in critical discourse research of normalization of the politics of exclusion and anti-immigration rhetoric in the context of far-right, illiberalism and neoliberalism.He holds a Chair in Media and Communications at Uppsala University, Sweden, where he is also Deputy Head & Associate Head (Research) at Department/School of Informatics & Media as well as Director of Research at Uppsala University Centre for Multidisciplinary Studies on Racism (CEMFOR).Info: https://www.katalog.uu.se/empinfo/?id= N20-1042 David Johnson specialises in social media analytics, data mining methods, and data infrastructures in social and life sciences research.He is an Associate Professor (Docent) of Information Systems at Department/School of Informatics & Media, Uppsala University, Sweden.Info: https://www.katalog.uu.se/empinfo/?id= N18-1553.

Figure 1 .
Figure 1.Time distribution of number of FB posts published per day by M, KD, and SD between 3 October 3 and 1 December 2021.(M = continuous/bold line, KD = dashed, SD = dotted).

Figure 2 .
Figure 2. Plot for FB-based Interactions of KD with M and SD in October-December 2021.

Figure 3 .
Figure 3. Plot for FB-based Interactions of M with KD and SD in October-December 2021.

Figure 4 .
Figure 4. Plot for FB-based Interactions of SD with KD and M in October-December 2021.

Figure 5 .
Figure 5. Combined plot for FB-based Interactions of/between KD, M and SD in October-December 2021.

16. 10 :
Bomb threat against a building in Ljungby.21.23: Shooting at an apartment in Haninge.01.15: 34-year-old father of small children shot to death in a terraced house area in Linköping.05.45: Man injured after explosion in a garden in Eskilstuna.(M, 2021, October 6) Now it's enough.The government has failed to stop criminal gangs.[…] We want to put the political play aside and really do what it takes to turn the tide for real.(M, 2021, October 10)

Table 1 .
Summary of Interactive features (images, native videos, shared posts, external links) of FB posts by M, KD, and SD in the period of investigation.