Dog whistling far-right code words: the case of ‘culture enricher' on the Swedish web

ABSTRACT This paper uses the Swedish, once neo-Nazi expression culture enricher (Swedish: kulturberikare) as a case study to explore how covert and coded far-right discourse is mainstreamed, over time and across websites. A sample of 2,336 uses of the expression between 1999 and 2020 were analysed using critical discourse analysis. The findings illustrate how the expression works like a ‘dog whistle’ by enabling users to discretely self-identify with an imagined in-group of discontent white ‘Swedes’, while simultaneously showing opposition to the priorities of a generalised ‘establishment’. It shows how the expression is circulated in settings ranging from mainstream to far-right, and particularly, it highlights the potential role of semi-radical settings to act as gateways between mainstream and far-right ideas. Finally, the analysis shows evolving, ever more covertly hateful uses of the expression over time, illustrating the adaptability of far-right language online more generally, as a means to evade unwanted exposure by the far-right.


Introduction
As the far-right grows increasingly prominent, beliefs which were previously considered unspeakable are now progressively being normalised as mainstream political ideas. A number of factors have been raised as significant in the process of mainstreaming farright discourse, and among them, online communication has often been highlighted as especially important (Albrecht et al., 2019;Daniels, 2018;Klein, 2012).
The mainstreaming of far-right discourse online is enabled by more or less deliberate efforts by individual users and sites to disguise, launder and legitimise far-right ideas. In particular, coded, hateful humour has been identified as central to mainstreaming processes (May & Feldman, 2019;Schwarzenegger & Wagner, 2018;Winter, 2019). Although extant literature has provided insight into coded far-right language use online, this research has tended to explore individual sites over limited time periods (e.g., Malmqvist, 2015;Topinka, 2018). There is a lack of systematic investigation into these discursive processes diachronically and across digital settings.
The current paper addresses these gaps using the Swedish, once neo-Nazi expression culture enricher (Swedish: kulturberikare) as a case study, mapping its movement across the Swedish web, over an extended period of time. Specifically, using critical discourse analysis, this paper aims to provide deeper insights into how far-right, coded discourse is mainstreamed online over time and across online sites, by exploring (RQ1) how the culture enricher expression is articulated and (RQ2) where, and (RQ3) if, how and to what extent the expression is used in mainstream settings.
Because of these perceived issues, the far-right tends to express a nostalgic desire and longing to restore a lost age, however loosely defined, of native glory, which is discursively reconstructed as an idealised time 'before' immigration (Elgenius & Rydgren, 2017;Engesser et al., 2017). In Sweden, these constructions depict Swedishness as something homogenous, masculine, and importantly, white (Horsti, 2017;Merrill, 2020). This close interwovenness of whiteness and Swedishness originate in longstanding ideas about 'the Swede' as racially superior (Hübinette & Lundström, 2011;Kjellman, 2014).

The mainstreaming of far-right discourse
The mainstreaming of far-right ideas take place on a discursive level as the creation of a 'new normal' (Kallis, 2013). Notably, the far-right has successfully rebranded their ideas into more palatable packaging (Feldman & Jackson, 2014;Fielitz & Laloire, 2016;May & Feldman, 2019), and are employing their exclusionary discourse in ways and around issues commonly associated with mainstream European politics, in attempts to appeal to a broader audience (Cammaerts, 2018;Feischmidt & Hervik, 2015).
However, mainstreaming is not only about the normalisation of the far-right, it also involves the radicalisation of the mainstream, as Mondon and Winter (2020, p. 113) note: No matter how hard a far-right party tries to reform itself, if the broader political system is not open to welcoming it and its ideas, it is bound to remain at the margins. The breakthrough of far-right politics has thus required the actions of mainstream actors such as politicians, of course, but also of the media and academics, to legitimise, if not its cause, then at least its presence in the political debate.
In particular, news media's extensive coverage of, and attention to, provocative farright ideas and actors grants the far-right widespread exposure (Cammaerts, 2018;Vidra & Fox, 2014). Such reporting in turn, also helps ensure the success of media companies (Feischmidt & Hervik, 2015). The far-right is also aided by a rightward shift in the political mainstream, making ideas and issues previously associated with the far-right less radical (Fielitz & Laloire, 2016). This is both because of an increased focus on immigration related issues by mainstream politics (Kallis, 2013), but also sometimes due to actual appropriation of far-right rhetoric by mainstream politics (Vidra & Fox, 2014).
Finally, the internet is considered crucial for mainstreaming far-right discourse (Schwarzenegger & Wagner, 2018). Like for other fringe movements, the internet has been important for far-right community building, and rapid and inexpensive dissemination of ideas (Adams & Roscigno, 2005;Burris et al., 2000;Caiani & Wagemann, 2009;Gerstenfeld et al., 2003;Perry, 2001). Importantly, the internet has been used to help increase the far-right's perceived legitimacy (Klein, 2012), and enabled access to less radical audiences who might not have consumed far-right content offline (Ekman, 2014;Winter, 2019). Covert, coded, and ever-adapting language use has been especially important for reaching beyond a far-right fringe.
Coded far-right language use online Like far-right organisations and movements offline, radical far-right sites online often attempt to convey a somewhat palatable, less blatantly hateful image (Gerstenfeld et al., 2003;Meddaugh & Kay, 2009). For instance, Daniels (2009) has identified how extremist sites 'cloak' their true political intentions by deceiving and disguising authorships, information and domain names in attempts to shift what ideas can be publicly discussed, sometimes referred to as the 'Overton Window' (see Daniels, 2018). Relatedly, Klein (2012) has shown how far-right ideas are 'laundered' to become legitimate discourse by far-right sites borrowing aesthetics and content from more credible sources, and leverage the interconnectedness of the internet to attract users and spread their ideas beyond their sites. For individual users, subtlety when expressing hateful ideas can be a means of avoiding human and automatic content moderation, and the potential repercussions that such detection might bring (Bhat & Klein, 2020;Merrill, 2020).
Coded language use plays a particularly important role for the far-right online, both in terms of claims to legitimacy as well as in appearing less hateful to moderators and internet audiences more broadly. For instance, previous research shows how users substitute letters or words to decrease the searchability of hateful content (Magu et al., 2017), how they avoid explicitly hateful expressions to appear politically neutral (Åkerlund, 2020), and importantly, how they use humour to express hateful ideas without explicitly stating them (Hervik, 2019;Marwick & Lewis, 2017;Winter, 2019).
While previous research shows that this humorous, yet hateful content tends to reiterate well-known far-right tropescriticising the establishment and political correctness, and othering and stereotyping immigrants (Malmqvist, 2015;May & Feldman, 2019;Merrill, 2020) it does so in ways that engage users to co-construct and build on each other's jokes (Schwarzenegger & Wagner, 2018;Topinka, 2018). Consequently, helping in their spread and mainstreaming, while providing distance and deniability as users can always claim they were 'only joking' if faced with criticism (Munn, 2019;Schwarzenegger & Wagner, 2018;Tuters, 2019).
Oftentimes, these humorous, hateful expressions are articulated in specific coded ways which are understood in this paper as dog whistlesspeech acts that appear inaudible or at least easily refutable towards the general public, while simultaneously conveying hidden meaning to fellow far-right sympathisers (Haney López, 2014;Kien, 2019). As such, dog whistles denote different meaning depending on audience, something that has elsewhere been defined as 'multivocal communication' (Albertson, 2015). It is a well-known and long-used technique among electoral politicians to appeal both to 'regular' voters as well as those holding more extremist views (Fish, 1994;Khoo, 2017).
For someone to 'hear' a far-right dog whistlelike the anti-Semitic meaning of the three parentheses signalling an echo '((()))' (Tuters & Hagen, 2020), the white supremacist meaning of the now famous Pepe the frog meme (Daniels, 2018), or any of the coded euphemisms constantly created and reinvented online (Bhat & Klein, 2020)they must be familiar with the specific shared knowledge within the far-right community. Because dog whistling is deployed with the objective to deceive parts of the audience, it can sometimes unintentionally be appropriated and repeated by unknowing outsiders, however, with the same potentially harmful effects (Saul, 2018).
Exploring culture enricher as a coded expression online Humorous, ironic and sarcastic expressions have a long tradition within Swedish farright settings, and often, these originate with progressive anti-racist, pro-immigration ideas. The culture enricher expression, which is known to still circulate in its sarcastic form in far-right settings in Sweden, has a background in neo-Nazism. In the 1980s and 90s, it became widespread in these settings as, a coded expression for everything they dislike about the immigration and refugee politics and as a joke on the anti-racists, who have often cited culture enrichment, exotic food and folk dance as arguments for increased immigration (Lööw, 2016, p. 503 Author's translation from Swedish).
Neo-Nazi papers from this time show uses of culture enricher as a synonym to immigrant, and as a commentary of the issues that immigrant related crime, violence, attitudes and culture bring to Swedish society (e.g., Balder, 1991;Ramell, 1996).
The 1980s and 90s was a time when the Swedish far-right became more organised, and with the emergence of the commercial internet in the 1990s, far-right activity would, in Sweden like elsewhere, come to take place online to an increasing extent (Lööw, 2015). At this time, Swedish far-right sympathisers and organisations started websites, chats, blogs, and electronic newsletters. Since then, far-right social media accounts and 'alternative' (far-right) news sites have also become increasingly visible on the Swedish far-right scene.
This evolution has happened alongside the growing success of Sweden's primary farright party, the Sweden Democrats (SD). While SD remained insignificant during the 1990s, it slowly started gaining traction in the 2000s (Widfeldt, 2008), and by 2010, SD entered into parliament. Since then, the party has seen continuous electoral success.
In Sweden's last general election, held in 2018, SD received nearly 18% of the votes, and became Sweden's third largest political party in parliament.

Data collection
While the mainstreaming of far-right discourse can be understood in terms of redefining what is considered acceptable or common sense, it can be studied more concretely in terms of how visual or textual expressions and ideas are (re)framed and circulated (e.g., Ben-David & Matamoros Fernández, 2016;Kien, 2019;Merrill, 2020;Topinka, 2018;Tuters, 2019).
To map the uses of the culture enricher expression over time, Google Search results for all instances 1 of the suffixed variations (-er,ed, -es, -ment, -ing) in Swedish were localised, manually read, dated 2 and downloaded, along with associated URL's. Searches for the expression from November 12, 1999, when it was first recorded by Google, until May 2, 2020, resulted in a sample of 2,336 instances of text 3 , posted on 287 different websites. Instances of culture enricher used in a non-coded, non-far-right way were disregarded after reading. These instances tended to describe things like music, festivals, art, travel and tourist attractions in a literal sense as 'enriching culture'. Examples of these include a post on small town's official website where the local museum is described as 'cosy', 'well-planned' and 'culturally enriching', and a post on a travel blog that states 'it is not always necessary to go abroad to have fun, exciting and culturally enriching experiences'.
While the data collection did not actively exclude social media content, the sampling method found very few instances of text from such settings throughout the sample period. This is because Google's search function is quite inefficient at picking up social media discussions. There might also be cases where the expression is used in even more covert or intentionally coded ways, that cannot be picked up by this sampling method. However, this approach to data collection takes search engines at face value. Search engines, and Google in particular, make up a considerable part in how people experience the web, and have been described as important funnels leading unsuspecting individuals to radical online content (Daniels, 2009;Klein, 2012).

Methodology
'The far-right' and 'the mainstream' are socially constructed, fuzzy and historically situated concepts (Mondon & Winter, 2020), and the mainstreaming of far-right ideas is a discursive creation of a 'new normal' through language use which contests dominant, mainstream, political ideas (Cammaerts, 2018). This is because language use is socially conditioned but will at the same time have social effects that contribute to maintaining or challenging existing hegemony (Fairclough, 2003(Fairclough, , 2010. Language use has the potential to reinforce or challenge power relations, impact our knowledge, values and identities, and even affect societal structures (Fairclough, 2003;Wodak, 2015). At the same time, existing power relations, knowledge, values, identities and societal structures constrain language use (Fairclough, 2010). Critical discourse analysis (CDA) is aimed at scrutinising such struggles over the power to define meaning (Wodak, 2009) and their relationship to wider social and political structures (Fairclough, 2010).
Exploring discourse involves analysis on the three interconnected levels of text, discursive practice and social practice. First, on the textual level, formal, grammatical features of the texts are explored. Second, on the level of discursive practices, focus lies with the production and consumption of texts. Finally, the social practice level explores the wider social structures, as well as ideological and political contexts in which the text is situated (Fairclough, 1992). CDA helps expose the underlying meanings behind the coded use of the culture enricher expression. First, the analysis explores how it is articulated (cf. RQ1). This involves analysis of wording and word co-occurrences on the textual level (Fairclough, 1992), identifying discourses, styles and discursive strategies as means of representing meaning and identities from particular perspectives (Fairclough, 1992(Fairclough, , 2003, and uncovering the underlying assumptions of common ground in the texts (Fairclough, 2003). Thereafter, the analysis explores the settings in which the expression is articulated (cf. RQ2) and lastly, if, how, and to what extent the expression is used in mainstream settings (cf. RQ3). This analysis focuses on discursive practices, primarily on identifying genres, but also on intertextuality (Fairclough, 2003).
The general patterns identified from the analysis are summarised in Figure 1 and further elaborated throughout the analysis. First, from the left, it shows the five discourses which were identified through the analysis: Sweden under threat, crime and violence reports, anti-establishment, othering immigrants, words/expressions, and criticising hate. Furthermore, Figure 1 shows the types of sites in which these discourses were used. Then, the categorisation of these site types into three discursive 'fields': far-right, gateway and mainstream, identified in terms of how different types of settings have 'a relative internal coherence and are relatively demarcated from others' (Fairclough, 2010, p. 175). Specifically, the discursive fields were identified in terms of the ideas and values that different site types promoted. Editorially run far-right 'news' sites and far-right blogs were categorised as far-right. Online discussion forums that host a variety of topics and users, but which are known for facilitating far-right discourse were categorised as belonging to the gateway sphere because of their 'eclectic' nature but also their potential bridging functions between the far-right and the mainstream. Uses in the Swedish discussion forum Flashback constituted most of these instances, but there were also some uses on for instance Reddit and 9gag. Finally, sites categorised as belonging to the mainstream were those which users cannot be expected to visit with the intent to specifically seek out far-right discourse, and which are not known to host far-right discourse. For instance, this included news media sites, and forums dedicated to specific (non-far-right) topics, like travel and hunting. Finally, furthest to the right in Figure 1, are the points in time when expression was used in these respective fields.

Articulations of culture enricher
Various abbreviations are identified in the dataset but the overwhelmingly most common one is 'culture enricher' in noun form, making up over half of all sampled instances (see Table 1). This means that when culture enricher is discussed online it is often in reference to things or people.
In discussions specifically about the definition of the expression, in line with how it was used in neo-Nazi settings in the 1980s and 90s, lie a common understanding of its sarcastic meaning. Users also claim that since the expression originated in mainstream ideas, it cannot be considered hateful: The establishment claims that immigration leads to multiculturalism and that multiculturalism is something positive and enriching. If a person is from a culture other than Sweden, he/she is a cultural enricher. I consider it perfectly okay to call immigrants and refugees this, as it is the politically correct establishment's own definition (Flashback member, 2005).
The left basically coined 'cultural enrichment'. Using it should be seen as a victory by the left, instead of as racist. It is actually ironic and humorous. To stand up to the damnation of society, one must be able to joke about this shit. Would they rather we resort to violence? (Flashback member, 2016).
Many also highlight the discrepancy between these ideas and what 'real' or 'ordinary' Swedes feel towards 'culture enrichment', 'MENA' (Middle East and North Africa) and 'African' culture: Such reasoning is a total mismatch with reality that ordinary people live in. The word is selfdefeating because it is such a flagrant lie which any real Swede can easily see through (Flashback member, 2010). As an ordinary Swede, it is hard to see how we Swedes are enriched by MENA or African culture (Flashback member, 2010). These quotes showcase how users self-position with an imagined 'people' (Hameleers & Schmuck, 2017;van Dijk, 1993;Wodak, 2009). However, rather than this people representing Swedish citizens in general, what is articulated is instead an exclusionary, homogenous us of white far-right sympathisers (Merrill & Åkerlund, 2018). This is closely related to a strategy depicting the establishment as betraying the country and its legitimate people. Through supposed lies, hypocrisy, cover-ups and infringements on freedom of speech, the establishment are seen as challenging 'white solidarity' (DiAngelo, 2011): They never have any issues evicting white Swedes from housing they've had for decades, but anytime it is about culture enrichers, it's a different story (Commenter far-right blog, 2015) Those in power ignore violence against women when 'culture enrichers' are the perpetrators, but they are keen to make sure that us ethnic Swedes shut up and don't show our discontent. (Commenter on far-right news site, 2018) Will the regime's criminal incompetence and carelessness result in fewer welfare benefits to Swedes who have fallen ill, while bearded children and culture enrichers continue to receive the money from Sweden at the same rate as before? Probably. (Flashback member, 2020) These examples are highly racialised and situate the needs of 'white' and 'ethnic' 'Swedes' in opposition to those of (non-white) 'culture enrichers'. This discourse of Sweden being under threat (see Figure 1) as the welfare system collapses and immigrants take over, often uses 'zero-sum' arguments (Kallis, 2013). These claim that if immigrants are allowed part in the welfare system, resources will be too few for native Swedes.
Reports of specific criminal offences and acts of violence by immigrants, often with reference to current events and news, constituted the largest discourse category (35%) in the dataset (see Figure 1). In these texts 'immigrants' are differentiated from 'Swedes' in terms of their supposed violent nature. A culture enricher in these examples is an immigrant ' … who enriches us by robbing, raping, and spreading terror' (Flashback 2008). Here, culture enricher is frequently co-articulated with preceding descriptors like 'rock throwing', 'trigger happy' and 'knife waving'.
While the expression has been associated with criminal activity in earlier neo-Nazi pamphlets and magazines too (e.g., Balder, 1991;Ramell, 1996), it seems as though it has taken new forms online. As illustrated in Figure 2, the number of unique abbreviations increase over time, signalling a development in the understanding of the expression, and a continued co-creation of its covert and coded meanings over time online.
One of these changed uses is with culturally enriched, which in the 1980s and 90s was primarily used to describe an increase in immigrants in Sweden or certain living areas (e.g., Sverige-Kuriren, 1989). However, online it has also often come to serve as a coded way of signalling immigrant violence, physical assault and rape against 'Swedes': To this we are also 'culturally enriched' by more and more robberies, scams, assaults, rapes, and murders (Commenter far-right news site, 2019).
Swedish women are 'culturally enriched' with surprise sex in the streets. (Commenter farright blog, 2016) Like in the Middle East, in this new Swedenistan, women are considered whores if they are outside without burqa and male chaperon and they can be raped and beaten! So, for anyone who doesn't want to be culturally enriched, stay home! (Commenter far-right blog, 2018) Discursive strategies of 'othering' immigrants, as illustrated in the examples above, are a common way of portraying immigrants as incompatible with native culture and heritage. Also, underlying assumptions seen in the excerpts above, illustrate that culture enrichers are not just any immigrants. Instead, they are specifically Middle Eastern, Muslim immigrant men, stereotyped as a homogenous group of violent rapists (Horsti, 2017;Sakki & Pettersson, 2016). Overall, almost a third of the dataset (31%) contained different kinds of othering representations of immigrants (see Figure 1).
This othering is further enhanced through co-articulations with other sarcastic, coded expressions like 'pension saving', 'high achieving' and 'highly educated', but also with overtly hateful and dehumanizing terms like 'parasitic' and 'inbred'. Echoing previous research, there are examples of how covertly hateful posts have prompted overt and explicit articulations of hate in the comment sections (Åkerlund, 2020;Ben-David & Matamoros Fernández, 2016;Schwarzenegger & Wagner, 2018), with several calls to murder immigrants that have supposedly committed crime.
Often, as it was also used by neo-Nazis in the 1980s and 90s, culture enricher becomes a complementary expression to 'immigrant'. Frequently coupled with scare quotes, it signals opposition to a mainstream, non-coded understanding of the expression. There are even instances where this substitution seems entirely natural. As one Flashback user comments on a post about a court case, claiming 'according to the prosecution, the perpetrator is a 'culture enricher'', they casually imply this would be an actual expression used in court. It also turned out that the perpetrator in this case was not immigrant, further showcasing how quickly negative events are tied to immigrants among far-right sympathisers online, often without factual basis.
The neo-Nazi background of the culture enricher expression tends to be omitted in discussions of its origin, and there even seem to be active attempts to obscure this history. In two online (slur) encyclopaedias for example, culture enricher is defined as a 'a slightly condescending word for migrant or immigrant' (Typkanske, 2018) and as 'a nicer expression for foreigner ' (Slangopedia, 2013). And on Flashback, someone goes as far as claiming it is a 'neutral term used by tolerant individuals instead of using negative labels like 'wog''.
In conclusion, the culture enricher expression is used as a replacement for 'immigrant' and 'immigration', without acknowledgement of its neo-Nazi background. Specifically, the use of the expression here shows it enables two things which the words 'immigrant' and 'immigration' do not. First, it works like a 'dog whistle' by enabling a discrete way to self-identify with 'a people' (postulated as 'the people') while simultaneously showing opposition to 'the establishment' and their priorities. Second, as the expression has taken new forms online, and is co-articulated with other hateful, coded expressions, it can be used to pin racist stereotypes to Muslim immigrants in very subtle ways.

Sites of circulation
The expression was distributed over 287 unique sites in the dataset in total. However, 73% of the uses of the expression were limited to only 10 sites (3,5%). Seven of these are far-right sites and two are mainstream news media. Most uses though, over half (56%) of the sampled texts, were user-generated content posted in threads in the online discussion forum Flashback, and this concentration increases over time. Potentially this large share is in part due to how little content is ever deleted on the site, that users cannot themselves delete their own content, and that Flashback has been around almost throughout the whole sample period, since May 2000.
The analysis makes clear that the culture enricher expression is spread throughout a variety of different forum parts on Flashback, even those unrelated to issues of immigration. Seeing as Flashback is thought to be used by nearly a third of Sweden's population (The Swedish Internet Foundation, 2019), uses of culture enricher in forum sections about for instance cooking, cars, and relationships means that users not interested in far-right issues might still encounter such discourse. Furthermore, much of the discussion regarding the meanings of words and expressions take place on Flashback (see Figure 1). There is even a dedicated thread asking, 'what words have you learnt on Flashback?', where several users answer that culture enricher is one of the expressions they have come to know through the site. In light of these findings, Flashback can be considered an important gateway 4 , preparing Swedish users for more radical sites and ideas (Åkerlund, 2020; see also Munn, 2019).
Sites within the sphere of the far-right make up the second largest category of websites, 29% of the uses of culture enricher are published in dedicated far-right settings. Within this sphere, most uses of the expression are in far-right blogs. In total, 95, or 33% of the unique sites, and in total 602 instances (25%) of all uses of culture enricher are posted in far-right blogs. Blogs have often been considered important for circulating far-right discourse (Horsti, 2017;Sakki & Pettersson, 2016).
Even before encountering any actual posts on these blogs, their populist, far-right approaches are clear. Several of the blogs' descriptions claim to 'tell you what the PC media won't', and 'defend everyone's right to freedom of speech' even if it is not considered 'politically correct in society today'. This works to demonstrate their positioning on the side of 'regular people' against a lying, corrupt 'elite establishment', while simultaneously disguising far-right discourse as claims to freedom of speech (Gerstenfeld et al., 2003;Mondon & Winter, 2017). However, far-right sites lack credibility and legitimacy among the general public. To increase credibility, one large blog informs on its starting page that while they all share 'the same opinion', they caution commenters to be aware of their language use to make content more easily sharable, and to not reflect badly on certain political organisations. This shows the calculated and intentional efforts by bloggers to spread and normalise far-right ideas (see also Meddaugh & Kay, 2009). As is further illustrated in the use of blog names like 'The Government Agency', 'Current Politics' and 'The News Site', drawing upon legitimate genres to signal credibility, and thus, disguise their content as important public information (Daniels, 2009;Gerstenfeld et al., 2003;Klein, 2012).
Moreover, the far-right blog posts are written as objective, yet outspoken news items, often with reference to mainstream news articles. However, it is clear that they cherrypick articles that further their specific arguments (Merrill & Åkerlund, 2018). In total, almost half the uses of culture enricher in far-right blogs were in relation to reports of criminal offences and acts of violence by immigrants. This is considerably higher than the 35% this discourse category makes up of the dataset generally. Overall, these strategies seem somewhat successful. According to the online traffic analytics tool Alexa, many visitors to the largest far-right blog overlap with visitors to editorially run, farright news sites. This indicates that the blog might be used by readers as a source for news and information.
Finally, many of these sampled sites are interlinked through static page promotions, but also through intertextual references to each other's contentsometimes reusing entire blogposts. This enables the exposure and spread of expressions like culture enricher, and the creation of a 'network of hate' (Burris et al., 2000;Caiani & Wagemann, 2009) of easily accessible far-right discourse and websites.
In summary, the expression is mainly used in the large online forum Flashback, and in far-right blogs. On Flashback, the expression is widespread, and several users claim to have learnt it through the site, signalling its importance as a gateway setting to more radical far-right discourse. Because far-right discourse is not socially acceptable, these ideas are often expressed subtly. Therefore, far-right blogs attempt to borrow credibility from news media, and actively downplay overt hate. This, and the interconnectedness of sites and content enable the exposure and spread of coded expressions like culture enricher and therein, of far-right discourse more generally.

Uses in the mainstream
Primarily, there are two ways in which culture enricher makes its way into the mainstream. First, through counter-discourse and resistance to hate, and second, through appropriation of far-right ideas. In total, 316 instances (13,5% of uses) of the expression took place in mainstream settings.
Counter-speech takes place throughout the studied period, with most uses in 2008 and 2009, where news media reports call out Sweden Democrat politicians who have used culture enricher and other hateful expressions online. In total, almost 39% of mainstream uses are a means to criticise hate, and take place in news media, academic texts, and in anti-racist and political blogs (see Figure 1). These critiques aim to highlight the racist meanings and usage of culture enricher, and unlike far-right discussions referred in previous sections, these acknowledge the expression's racist history: Culture enricher is a deeply racist expression used frequently by radical nationalists like the National Democrats and the Swedes Party (The Swedish public service television company, 2010).
Immigrant as culture enricher might sound positive but is often used in the exact opposite way. And if you're missing the irony, you're not listening to the dog whistle […] it is a rhetorical strategy to make something negative appear neutral or even positive. One example of this is when immigrants are ironically called culture enrichers by immigration critics (Focus Magazine, 2020).
Several critics also highlight its frequent usage on Flashback, further indicating that this is an online forum through which also non-affiliated users encounter far-right expressions: Flashback is great as a news site in the sense that they are quick to figure things out, even before the media are able to. However, the constant arguing of culture enrichers and so on bothers me. (Family forum member, 2012) Not always when culture enricher is used in mainstream settings is everyone fully aware of its coded usage. For instance, when the expression is used in a travel forum to discourage others from visiting an amusement park where immigrants are 'causing trouble', someone replies: What do you mean by culture enrichment? Are there problems at the Tivoli? (Travel forum member, 2009) And in a hunting forum, as someone asks for advice on how to build a dog pen, another member claims that the pen needs to be 'enriched' with things that the dog could play with. This leads to a misunderstanding, as seen in another member's reply: Oh, that's what you meant! I thought you meant culture enrichment… (Hunting forum member, 2016) These examples show how the expression 'spills over' into mainstream settings, and when users incorporate (or think someone else incorporates) far-right code words, there is confusion. Specifically, users seem unsure as to what common ground they share in regard to far-right discourse.
Interestingly, while it is more common that the expression is used in the mainstream as a means of countering hate, in certain mainstream blogs and forums, the expression is used mostly for othering purposes: Incorrect invoices are often not mistakes, but rather deliberate cheating, often by welfare tourists […] There is an increased risk of cheating in the culturally enriched areas and that's why government agencies are now going to work more intensely in the multicultural paradises that the politicians have created. Now, culture enrichers too are going to enrich the welfare system, rather than just exhaust it. (Tax blog, 2018) Violence and criminality are all of a sudden not something bad, but a great culture enrichment! There are still some brave men who dare tell the truth about Sweden, but they risk being silenced by harassment as they are the primary threat to the indoctrination of socialist, multicultural Sweden. (Coaching blog, 2019) While these sites are dedicated to supposedly non-political issues, they echo the ways in which culture enricher is articulated in far-right settings. It seems as though these are examples of far-right sympathisers who also happen to be active in mainstream settings online, and leverage these to spread far-right ideas to unsuspecting others.
Finally, some instances of culture enricher in mainstream settings instead show an unknowing adoption of the expression as a substitute for immigrant: My thoughts are on food right now as it's past five o'clock. I don't have much at home, so I'll need to take a walk to the grocery store, which is no problem as it is just around the corner. Also, the culture enrichers have opened a shop in the neighbourhood so the range of things to choose from is huge. The use of culture enricher to describe everyday events in this routine way shows that the far-right's struggle contesting the hegemonic understanding of 'immigrant' is to some extent being normalised into everyday conversations online, also outside of its original sphere.
In summary, there are two ways in which culture enricher makes its way into the mainstream: through counter-discourse and resistance to hate, and through appropriation of far-right language. However, as the analysis highlights, when coded far-right expressions are used in mainstream settings, uninitiated individuals risk encountering them, even though they have not actively sought them out. Whether as a form of resistance, or through appropriation, use of coded expressions in mainstream settings can enable the spread of far-right discourse.

Concluding discussion
This paper aimed to provide insights into how far-right, coded discourse is mainstreamed online, by exploring the neo-Nazi expression culture enricher as a case study on the Swedish web between 1999 and 2020. It contributes new insight, beyond that of previous research, into the longitudinal and multi-sited evolution of coded, far-right expressions.
The culture enricher expression works like a 'dog whistle' (Haney López, 2014; Kien, 2019) by enabling users to discretely self-identify with an imagined in-group of discontent white 'Swedes', while simultaneously showing opposition to the priorities of a generalised 'establishment'. The analysis showed users drawing on well-established far-right discourses and discursive strategies based in ideas about Swedishness and whitenessincluding practices of othering and criminalising immigrants, and portraying Sweden as under threat by immigration and irresponsible governmental politics.
The analysis illustrated the importance of far-right blogs and the large, Swedish online forum Flashback in circulating the culture enricher expression. Most notably, the analysis found that the Flashback forum worked as a particularly important gateway between mainstream and far-right discourse, also exposing users who had not intentionally sought out far-right content.
Furthermore, it showed how the expression was used in mainstream settings throughout the studied period, both through hateful appropriations as well as through resistance. The expression peaked in mainstream settings in 2008 and 2009 with news media reports calling out Sweden Democrat politicians for using culture enricher and other hateful expressions online. With uses of the expression subsequently not reaching the same high levels, it would appear as though news media's reporting worked educationally. However, previous research instead suggests that mainstream media reporting of farright activity or ideas inevitably works in favour of the far-right (Burris et al., 2000). Wodak (2015, p. 19) explains how dealing with far-right discourse is a 'no win' situation for news media: if they do not report a scandalous racist remark or insinuation […] they might be perceived as endorsing it. If they do write about it, they explicitly reproduce the prejudicial utterance, thereby further disseminating it.
Potentially, mainstream media's attention has prompted caution. Even subtler variations of the expression not sampled here (e.g., enrichers) and use of other coded words might have come to take its place. These above explanations seem plausible, in light of how the expression developed over time.
The analysis identified an expansion of unique culture enricher abbreviations over time, indicating the expression's continuous co-creation among users. The analysis also showed a practice of co-articulating coded expressions, which facilitate a broader vocabulary of coded language, and the findings identified deliberate attempts to mask and downplay far-right discourse. Furthermore, the analysis highlighted a shifting and ever more covertly hateful use of the expression that helped pin racist stereotypes to Muslim immigrants. Finally, the analysis identified attempts at normalising culture enricher into a substitute for 'immigrant' while cloaking its neo-Nazi background. Unpacking these uses requires a lot in terms of familiarity with subtle, coded language, what Hughey and Daniels (2013) call 'Racial Internet Literacy'.
However, unveiling coded language use online is easier said than done. The evolution of the culture enricher expression illustrates a broader theme in far-right language use onlineas constantly adapting and reinventing to avoid unwanted exposure (Åkerlund, 2020;Bhat & Klein, 2020). With coded, humorous expressions' ability to 'hide in plain sight' (May & Feldman, 2019;Topinka, 2018), they have proven hard to detect, and thus by extension also to moderate and regulate (Meddaugh & Kay, 2009). Yet, despite this subtlety, coded language use risks having severe consequences.
Phillips (2019, pp. 2-3) argues that hate can be laundered into the mainstream through seemingly innocent humour: 'by exploiting the fact that so many (white) people have been trained not to take the things that happen on the internet very seriously'. Despite the perception of the internet as harmless, previous research illustrates the potential stepwise progression of far-right radicalisation, starting in a normalisation of humorous, but nonetheless hateful content, and gradually moving towards exceedingly radical ideas and settings (Munn, 2019). With this, the potential harm of subtle, ever-evolving, coded expressions like culture enricher, should not be understated. Future research is needed to explore the movement of coded far-right expressions online in order to find ways to mitigate their potentially harmful effects. This includes further inquiry into how coded language use is perceived and appropriated by different types of internet users, in different online settings, and in different national and international contexts.