The battle for truth: How online newspaper commenters defend their censored expressions

The presence of hate speech in the commentary field of online newspapers is a pressing challenge for free speech policy. We have conducted interviews with 15 people whose comments were censored for posting comments of a racist, discriminatory or hateful nature. What characterizes their self-understanding and enemy images? We found that central to their motivation for writing such comments was an understanding of themselves as particularly knowledgeable people. They see themselves as people who fight for the revelation of the truth, in contrast to the lies spread by politicians and the media. Furthermore, they regard politicians and the media as corrupt elites that are leading our society into destruction by their naïve support of liberal migration policies. By linking up to alternative news media, these individuals support various forms of racialized conspiracy theories, but also a form of radical right-wing populism in their concern that politics should be acted out by people themselves. As such, our study adds to the literature on conspiracy theories in general and racialized conspiracy theories in particular, but also to the literature on online far-right activists. Our contribution lies both in the newness of focusing on the self-perceptions, but also in opening up for a modification of existing literature on the far right.

understanding of people who believe in racialized conspiracy theories. There is a scarcity of such studies. One exception is two papers by Aupers (2015, 2017) on the self-understandings of people participating in a conspiracist milieu in the Netherlands. In contrast to their subjects, our interviewees are not part of a defined milieu, but rather people who take part in online debate on their own. Another difference to their study is that we examine people who adhere to racialized conspiracy theories.
According to Zia-Ebrahimi (2018), racialized conspiracy theories have existed since the 19 th century, but their contemporary manifestations have not been given sufficient academic attention. This is astonishing, taking into account the fatal consequences belief in such conspiracies might have. For example, the so-called Eurabia theory motivated the right-wing extremist Anders Behring Breivik on July 22nd 2011 to bomb government headquarters (killing eight people) in Oslo and thereafter shoot 69 young people at a Labour Party youth camp. It was contact with likeminded people online, and not face-to-face, that motivated Breivik to commit this horrible attack. The same was true for the perpetrator of the Al-Noor Islamic Centre shooting on August 10th 2019. Philip Manshaus said he was inspired by the Christchurch shooting in New Zealand in March 2019, and it was revealed that he had announced the shooting an hour earlier in extreme right-wing net-forums, where the forthcoming act was cheered by other activists (Burke, 2019). These cases prove that words spread online can motivate cruel acts.
Most of our interviewees do not support the use of violencerather they are waiting for others to take action. But their comments are obviously not innocent, and therefore it is not surprising that moderators deleted them. A strong argument for why the study of the self-understandings of those who write extreme comments online is important is that advanced by Conway (2016): 82) that we cannot develop effective counter measures to online radicalization if we do not know how it is occurring. She argues that research so far has been 'largely focused on analyzing digital content and not its producers'.
Indeed, there is a dearth of previous research on the topic of right-wing citizen activists online. We know that extreme violent attacks have been motivated by online hate speech, yet, we know little about how people who are responsible for such writing think. This is important to know, since unless we know what motivates their writing, it is hard to find the best measures to combat hate speech. The presence of hate speech is a major dilemma for the UN's declaration of human rights, article 19: 'Everyone has the right to freedom of opinion and expression; this right includes freedom to hold opinions without interference and to seek, receive and impart information and ideas through any media and regardless of frontiers'.
Much has been written about the dilemma between free speech and hate speech from a legal perspective (for an overview see Høy-Petersen & Fangen, 2018). But our study considers this dilemma with the self-understandings of censored commenters as a basis for better understanding the mechanisms that lie behind the writing of hateful and scornful comments online. Senders of hateful expressions are often mystified and demonized (Hoffman, 1996), and we aim to avoid this by analyzing how they themselves reflect around their commentary practice.
Norway is an interesting case, both because of the two recent terrorist attacks just mentioned, but also because a right-wing populist party has been part of the government since 2013 (Fangen & Vaage, 2018). 1 Our subjects favour even stricter measures than the current government, therefore they support smaller and more radical right-wing populist parties, not represented in parliament.

Far right populism, fake news and conspiracy theories
Although the studies of populism in general and right-wing populism in particular have increased massively during recent decades, research on the self-understanding of people with such views remains rare. Exceptions include studies of self-understandings among populist voters and non-voters (Kemmers, 2017) and of people belonging to a conspiracist milieu (Harambam & Aupers, 2015. Still lacking, however, is a study of people who belong to both of these categoriesthat is right-wing populists who believe in racialized conspiracy theories. Populism is a broad concept that includes both right-wing and left-wing populism, but the specific kind of populism we are interested in has variously been called far-right (Wodak, 2015), radical-right populism (Akkerman et al., 2016) or national populism (Eatwell & Goodwin, 2018). Blee and Creasap (2010: 270) define far-right populism as not only right-wing but to the right of rightwing, implying an idea of white supremacy and/or conspiracy theories. National populists, Eatwell and Goodwin (2018): ix), hold, "prioritize the culture and interests of the nation, and promise to give voice to a people who feel that they have been neglected, even held in contempt, by distant and often corrupt elites". It is important to see that there is a wide spectrumfrom more moderate rightwing populist parties such as the Norwegian Progress Party (Fangen & Vaage, 2018) to extreme right populist parties such as Jobbik and Golden Dawn (Halikiopoulou & Vasilopoulou, 2014).
Moreover, there is a difference between party politics (e.g. Akkerman et al., 2016) or political style (Moffitt, 2016) and the internal ideology of voters and non-voters (Kemmers, 2017). We are interested in the self-understandings of people who are not necessarily members of any group, but whose views resemble common definitions of far-right or nationalist populism. One element which is neglected in many studies of right-wing populism is that of white identity. Kaufmann's (2019) study of the white shift and Jardina's (2019) study of white identity politics are notable exceptions. Kaufmann identifies a fight response to increased immigration, which he defines broadly as including everything from voting behaviour to terrorist attacks. The fight response we are interested in is indulging in racialized hate speech online, which is important, as pointed out by Hughey and Daniels (2013): 334), as "not much research examines the new digital saga of racial" (…) "discourse in (…) online news formats".
As pointed out by Brown (2018): 298), "the instant nature of online communication encourages forms of cyberhate that are more spontaneous, and therefore, unconsidered". What is specific about online hate speech is the anonymity, which gives people the opportunity to engage in freer speech, but also often means that they say things they would never have uttered face-to-face. This is why online communication opens up a space for hateful comments and has led most major media companies to start regulating their commentaries. Even though we have studied people who post comments in the comment sections of mainstream newspapers, they themselves believe more in news posted in various alternative media platforms with strong anti-immigrant, anti-Muslim or anti-Semitic orientations (in another article, Ihlebaek & Riborg Holter, 2019, we analyze in detail their use of alternative media). An important feature of these alternative news sitesexamples include Breitbart and Stormfrontis that they build a community of like-minded individuals, although these individuals are not members of a party or defined group. And they typically present "false stories and conspiracy theories presented as news and which constitute racist expressions, incitement to hatred or Holocaust denial" (McGonagle, 2017: 204). In other words, some of the news that our subjects relate to falls under the umbrella of 'fake news', that is, "information that has been deliberately fabricated and disseminated with the intention to deceive and mislead others into believing falsehoods or doubting verifiable facts" (McGonagle, 2017: 203).
A problem with the concept of fake news is that, apart from the 'trolls' who deliberately post false information because they think it is fun (Shachaf & Hara, 2010), authors of what is defined as false newsand invariably their readerswould claim it is the truth, in contrast to the output of mainstream media. This is certainly true of our subjects. Moreover, fake news is a term used by some national leaders, such as Donald Trump, to discredit the mainstream press (McGonagle, 2017). This points to a link between fake news and conspiracy theories, since an important component in many conspiracy theories is that the press is allied with the establishment in publishing untruths.
A political conspiracy theory "is an explanation which postulates that an event is the (…) effect of activities that have been secretly planned and carried out by several actors" (Pelkmans & Machold, 2011: 68). Often, however, the term is used to imply a demarcation between rational, legitimate knowledge and illegitimate, irrational nonsense (Bjerg & Presskorn-Thygesen, 2017: 138). Barkun (2013) points out that the boundary between conspiracy theories and mainstream orthodoxy is unstable, however. What in one period is understood as a conspiracy theory might become part of commonly accepted knowledge in another. Furthermore, some conspiracies do in fact exist. Barkun (2013) underlines that there are a range of different conspiracy theories, and Harambam and Aupers (2017) stress that there are also many differences between adherents of conspiracy theories. We can only underscore this.
Moderators deleted the comments of our interviewees because they considered the comments to fall under the hate speech umbrella. Inherent in the hate speech concept is an assumption that the individual's mind is defined by hatred. Much has been written about the legitimacy of sanctioning hate speech by means of legislative measures 2 (Høy-Petersen &  and about the effect of hate speech on its victims (Fladmoe & Nadim, 2017). However, little is known about the senders of hate speechwho they are, or what motivates them (Erjavec & Kovacic, 2012;Jubany & Roiha, 2015: 46).

Methodology
Prior to data collection, we contacted and interviewed debate moderators in three major Norwegian newspapers. The moderators from Nettavisen and Dagbladet agreed to send us comments they had deleted from their online commentaries. The moderator from Dagbladet's Facebook page sent us comments before he deleted them, while we received the comments from Nettavisen from a server containing all deleted comments, from January to September 2016, in the aftermath of the 'refugee crisis' that brought a sharp increase in the number of asylum-seekers coming to Norway. We saved the altogether 4 000 comments we got access to in an Excel document. All comments included encrypted IDs.
We selected 300 comments that involved derogatory expressions, war rhetoric, and reference to conspiracy theories, and that labelled ethnic or cultural minority groups as threatening, criminal, perverse or dirty. Some comments used emotional arguments, others were threatening. Many comments linked criticism of immigration with anti-elitism, hinting at a connection between the media and politicians. To get a better understanding of the perceptions and reasons underlying the form and content of the comments, we examined them in their original form. This analysis is important for this paper, although due to space limits, we chose not to include citations of the comments themselves. The comments written by our interviewees are however available in an appendix. Not all these comments are equally extreme. However, according to the moderators, some are deleted because they are written by people who constantly derail the debates and chase away ordinary readers by their scornful words and hateful messages. In general, moderators deleted comments when they violated their own guidelines for a good and factual debate climate. These guidelines more or less follow the hate speech act (section 185 of the Norwegian Penal Code, 2019). The moderators' goal is to have a better debate, by controlling the small number of 'bad guys' who destroy the debate (Michaelsen, 2016).
The moderators contacted the individuals who had written the comments we had selected, asking if they would agree to be interviewed. 15 people (out of 135 people contacted by the moderators) agreed, and we contacted them by email, giving necessary information on ethical considerations and our project. We conducted 11 face-to-face interviews, three by telephone, and one by email. All the interviewees were men. This was not an intended result, but also not surprising, as previous research has shown that although women also participate in commentaries in general and hate speech in particular, there is a clear overrepresentation of men (e.g. Jubany & Roiha, 2015). The interviewer (one of the authors of this article) was a master's student at the time and in her 20 s, thus younger than all the interviewees. She managed to create an atmosphere of trust during the interviews. Interviews lasted three hours on average, as several respondents felt the need to "tell the whole story".
Semi-structured interviews give access to the interviewees' understanding of reality. By focusing on individuals, we seek to show how they justify their statements and reflect on them. Although the interviewer used an interview guide, she used it in a flexible way, thus allowing the conversation to naturally flow. During the first part of the interview, she listened, and focused on the subjects' understanding of reality and preferred themes. She wished to convey to the interviewees that she took them seriously. Later in the interview, she asked the subjects to elaborate about the comment(s) they had written. Many seemed to appreciate this opportunity to defend their views. The interviewer asked specifically about the interviewees' views on censorship, freedom of speech, immigration and the immigration debate, why they took part in online commentaries rather than using other forms of political expression (e.g. demonstrations, participation in NGOs or political parties etc.), and whether or not they were part of a political milieu.
We transcribed the interviews as verbatim as possible, and coded them through the qualitative data analysis program NVivo. We anonymized the total data material with all comments, so we know the identities of only the few individuals selected for interviews. All interviewees have been given fictive names. The Norwegian Data Security Service approved the project on July 1, 2016.
The analysis will proceed in two parts, first dealing with self-understandings and then with enemy images.

Self-understandings
In this section, we will present the self-understanding of our subjects. We followed an abductive strategy, where inductive reading and re-reading of the data has made us gradually change what theories and concepts we found most useful. The initial phase of data analysis revealed the uniform perception among our subjects that the censoring of their comment(s) was unfair. They felt that their comments contained descriptions that reflected reality, not normatively different from other 'fact-based' descriptions. This was something they repeatedly came back to without the interviewer explicitly asking. There was an obvious discrepancy between what posters and moderators perceived to be the meaning of a comment. Where the subjects see it as presenting facts or legitimate criticism, moderators have interpreted it as harassment of ethnic or religious groups.

Warners
Our interviewees do not aim to misinform; they claim to want to convey the truth. Similar to the conspiracy theorists studied by Harambam and Aupers (2017), our subjects have a self-understanding of being more able to take a critical stance than the majority of the population. When asked why they wrote their comments, they said that they wanted to inform the public about what reality is really like. Their goal was to show that a news article was wrong, and that journalists undermine and manipulate truth: I was stating an opinion that I believe to be true and there is evidence to support it. I was not indulging in racism, threats of violence or hate-speechsimply stating what I believe to be facts […]. My intention is to counter what I believe to be false information about Islam being spread in the media. Erik I think I'm not alone in my impression of the [lack of] truth in the media […]. I say to the media, you lie. The media can't tolerate this, and my comment gets deleted. Thomas We see here that the subjects relate not to folk wisdom or experiential knowledge, but use concepts derived from science, such as "evidence" and "facts". This is different from what has been coined 'post truth' in the sense that it still relates to a positivist conception of truth, only that society's mainstream sources of such truth are met with suspicion (cf. Ylä-Anttila, 2018). Birger wrote the following comment: "Hell's creeps. The new Norway. It goes to hell with our kids, and that's the AP's [the Labour Party's] fault". As seen below, he believes this was objectively true and not a generalization and therefore finds it incomprehensible that it was censored: As long as I do not understand why it was deleted, I perceive it as unfair. What is wrong? They should explain to me, "This is not correct," so I can prove that it is correct. Because I can post links documenting that it's correct.
Torstein, who had written a comment making an analogy between refugees and the German occupiers during World War II, the traitors now being the immigrant-friendly politicians allowing refugees into the country, had a similar conception of the truth.
The truth is made in the way that we only dare to talk around it. We don't dare to speak it outright, the way it really is. Torstein Several interviewees contend that the media delete "truths that may be uncomfortable to have to face", and also that "it's not the nastiest comments that get deleted, but the most uncomfortable ones". According to the news moderators we interviewed, comments are censored because they conform to the official definition of hate speech. The news commenters, however, perceived such censorship as a confirmation that power holders feel threatened because though they know what the commenters write is true. This is similar to what Barkun (2016): 1) writes about conspiracy theorists: "Rejection by authorities is for them a sign that a belief must be true." All our subjects claim to see things others do not. In this way, they attribute to themselves a unique position for promoting the truth. Our subjects see this notifying role as an individual responsibility because the media have failed to serve as the Fourth Estate. The role of 'society's watchdog' is therefore transferred to 'the common people': But our subjects also have a long-term goal, which they describe as "preventing war" and "securing the future of my children and grandchildren." Issuing warnings is presented as something they feel morally obliged to do. According to the interviewees, the warner role requires insight into the state of the nation and the ability to assess information and pick up details that others fail to notice. Torstein argues that most people are not conscious or critical enough to "understand" the news: Your generation is so hung up on a constructed truth that says we are so good. What is needed to escape this trap? Understand the lies served over Facebook, understand the lies being censored and what is "politically correct".
At the same time, this role is presented as 'the little man's struggle' against a totalitarian hegemony, through a kind of David versus Goliath typology. The commentary field becomes an arena where utterances can be used to throw light on veiled truths, trapping those who dominate the public debate. It is useful here to point to the spheres of consensus, legitimate controversy and deviance, as described by Hallin (1986: 117). By being censored from the comment sections of mainstream newspapers, our interviewees contribute to the sphere of deviance, "the realm of those political actors and views which journalists and political mainstream of society reject as unworthy of being heard". Hallin argues that journalism becomes a boundary-maintaining mechanism by exposing, condemning or excluding from the public agenda those who violate or challenge political consensus. The political purpose of our respondents is to undermine the legitimacy of political correctness. They have a self-image as 'warners'; they see what others don't. This special knowledge makes them into a kind of self-defined elite, differentiated from the ignorant, herd-like public (Barkun, 2016: 2). By publishing online, they want people to realize they are being tricked by mainstream media.

Male protectors
The warner role was rooted in the subjects' identity as males. Men are described as having a dual role as both risk-taking warriors and protectors. This sense of masculinity has been called toxic masculinity, where traditional norms of masculinity include expectations that boys and men must be active, aggressive, tough, daring, and dominant (Haider, 2016). Similar views are well-known in far right and radical-right circles (e.g. Fangen, 2003;Mulinari & Neergaard, 2014). Similarly, our subjects described men's ability to take risks as a uniquely male characteristic and as a reason why men had to take responsibility: Men have a different approach to risk than women. We like to drive fast. Women tend to choose the safe and good and be kind, but that won't work in the long run. They just push the problems in front of them […] For us men, it's not like that. I have to look ahead, anticipate what will come. Ulrik The assertion that "men like to drive fast" is presented as one reason why online commentary fields are male-dominated. Women are depicted as reserved and passivebut also incompetent or indeed dangerous. Feminists especially are portrayed as the main actors behind multiculturalism and the "fall of the nation". Our subjects draw parallels between a threatened national identity and a threatened masculine identity: We see how defenceless countries become when there is a very feminized male population. How little warriors we are. We don't recognize this until it comes. Look at the defence ministers in Europe now, almost only women. They have no military background […]. They don't have that instinct […]. I'm worried, because they don't realize that hell is approaching. Ulrik Women are thus unable to defend the nation. As Tor says, "it is most often those [left-wing feminists] who stand on the barricades to protect Muslims." Feminists are also held responsible for the censorship of immigration-critical opinions: for example, Magnus says "feminists protect those who treat women worst of all" and that there are "feminists who will censor everything they do not like". Ulrik believes that immigration is governed by a "socialist feminist mafia". Our subjects hold that gender equality has made men weak and feminine, which weakens the nation's ability to protect itself. Implicit is the idea that men are the protectors of the nationwhereas women, by their very nature, will always be either a threat or a weakness for its security and sovereignty.
As a consequence of a "feminized male population," as Ulrik describes it, many become paralyzed and passive. Norway no longer has "real men" who can take matters into their own hands. Magnus explains: If anything happens in the neighbourhood, because earlier you had adult men who would roll up their sleeves and get to work. If there was a big water leak, for instance, they'd fix it. Today, you see adult men in weird clothes and rings in their ears. Who's going to roll up their sleeves and fix the problem that's unfolding now? No one. Those people, they're gone. Magnus Immigration is for Magnus the water leak 'flooding' the country. When he and others post critical comments, they "roll up their sleeves" and, at the same time, fulfill their subjective expectation of masculinity. This gender perspective is an integral part of their understanding, serving to motivate and justify their activism in the commentary fields.

Victims, players or warriors
Parallel with the view of themselves as male protectors is a feeling of being stigmatized and held down. This feeling has been described by several researchers such as Jardina (2019) on white identity politics and Eatwell and Goodwin (2018): 113) on national populism. It is the feeling that their identity is under threat and that their views are pilloried and censored (literally). Our subjects K. Fangen and C.R. Holter Poetics xxx (xxxx) xxxx clearly felt misunderstood in this sense, and several of them referred to the "racism card" as being played against them: You are labelled Nazi or right-wing. I do not feel this is correct […] Labelling people in a negative sense is a ruling technique. They want to just kill a debate before it gets started. Sondre According to our subjects, when they are labelled racists or extremists, their arguments lose value and their role in the debate is undermined. However, they also see this as a victory: their opponents have reached the point where they have no more arguments and have to "switch to the unpleasant". This, they hold, is because the opponents realize they are right, but are afraid to acknowledge the "truth". This has a self-reinforcing effect: the more they are censored and called racists, the more convinced they become that their version is the only truth.
The interviewees feel that banning certain words is an attack on freedom of speech, leading to self-censorship and obedience to the system. However, they do not deny using stronger expressions to make the message clear. Sverre says he uses capital letters to highlight specific words, while Tor feels that employing stronger language makes it easier to get noticed: You often have to simplify, to get the message throughbecause if you say it too nicely, they wonder "what are you talking about?" Therefore, it must be sharp, in big letters, war rhetoric.
The concept of "hate speech" 3 is interpreted as a tool used to criminalize and gag their opinions. Several respondents feel provoked by being labelled hateful. The following statement illustrates how the concept of hate speech is turned into a victim discourse: It's labelled hate speech. But is it? Is there an angry person talking? [No, this is] someone who has been held down, one who has been called racist and stigmatized. There are very many provocative statements on the other side too. But I try to write in a subdued way. Rather sarcasm than anger online. And use intelligence to get the message through […] To me, hate is a worn-out term. Ulrik Nevertheless, not all respondents deny that hatred might be a driving force behind their comments. For Erik, hatred is a completely legitimate feeling "in the face of evil": Saying that you hate Muslims is not a hate crime, it is expressing an opinion […]. The new laws have made the police into 'thought police', not what they should be: preventers of crime. Any law that makes having an opinion illegal is wrong. I think Islam is evil, I think all Muslims should be deported. I think immigration should be severely restricted in order to preserve our nation's culture and ethnic/racial identity.
[…] Should I be imprisoned or fined for holding these opinions? Should I lose my job? Am I evil? I do not incite to violence and I wish no harm to anyone, except for violent criminals. Erik Our interviewees differed in the way they presented the kind of emotions they experienced whilst writing the comment(s). Some said they responded impulsively to news articles. They believed that emotional reactions lead to the "truth" becoming more apparent. Because they are emotionally engaged in the debate, they may overstep the mark in expressing their feelings verbally: If you get involved, you can become very provoked and angry, and then you have a blow-out online. Thomas While such emotional comments may appear hateful and offensive, that does not necessarily mean that these posters are the most radical.
We found that the comments that were written with provocation as a method differed substantively from those written in anger. The former were quieter in tone; they were probably censored because of their content, rather than their choice of words. Posters with 'rough' comments might perhaps be expected to represent more radical attitudesbut that was not the case. Of our subjects, those who had written 'quieter' comments were the ones who gave the most 'extreme' statements. They showed considerable control over the statements they had written, and expressed conspiratorial worldviews that placed them further to the right.

When I post provocative comments, I […] trick someone into responding. […] I use words that trigger socialists […] They respond to certain words, like when the Holocaust is mentioned […]
. Then, when they start attacking, they are into things that I know a lot about […]. So I can confront them with the lack of logic behind their philosophy. Magnus Our interviewees describe a sense of being treated as worse than "outsiders", of losing "their community". They alternate between war rhetoric and patron rhetoric. One is either for or against, friend or enemy. Some subjects say that, although they personally do not support violence, they see violence as an inevitable and partly necessary consequence, because other alternatives that can ensure peaceful change have proven ineffective. Both Birger and Torstein distance themselves from violence, but see it as necessary to protect family or country: I'm totally against violence, absolutely. But if I see that my family is attacked, then I'll use whatever means I have to protect them. I will not let myself be slaughtered just to say "no, I'm not going to be violent". Birger I'm willing to take up arms to defend my country. I hope this won't be necessary, because I'm afraid of bullets and everything like that, and I have no interest in hurting people. Torstein Our subjects legitimize their implicit support of the use of violence by claiming it is "them" who attack "us". Torstein speaks of "grabbing weapons" if "the rights of the Norwegian people are violated, especially the rights of women" and "when sharia is introduced in some areas".
Even though Torstein argues that there is a long way from words to actions, he also states that most of the victims of the July 22, 2011, terrorist actions were "traitors". Despite Torstein's widespread use of war rhetoric, he recalls how he was later frightened at his own reaction to the events of that day, an experience that opened his eyes to how dangerous his attitudes can be: I really shouldn't tell you this, but I will. You must get me right, this is very important. On July 22, when it was said on the news that a bomb had exploded in Oslo in the government headquarters where Stoltenberg ruled, I cheered for a brief second […]. Then there came the thought, "Hey, wait a minute, the top politicians are not at work today, it's in the middle of summer. There are only innocent people who have been blown into the air." I must emphasize this was before the shooting at Utøya [the Labour Party youth camp]. And I was very scared afterwards, over my own reaction […] I thought I'd have to work with myself and my political hatred.
Torstein and other interviewees distance themselves from the actions of the perpetrator, Behring Breivik, but agree with parts of his manifesto. According to Magnus, "Breivik had some good ideas, but he did the worst thing, namely to create sympathy for the other side." The potential for violence in the views of these subjects should thus not be underestimated. For example, Tor says that "all forms of violence are really stupid," but he believes that it is acceptable to remove politicians by force if they do not serve the people. He refers to "his plan" without elaborating further: I have a plan and … well, that's another matter. There are people who are willing to kill for political change in Norway.
Although most subjects were not so explicit about violence, there is a latent potential for violence in many of their expressions: It's a warning. There are dangerous things happening now. I'm not really a doomsday person, but I see the danger in this. Sverre The system that is supposed to protect you is against youwouldn't then the natural reaction be for people to say "OK, then we'll beat up that bloody policeman or that damn judge, we'll chop the head off whoever did this thing"? Magnus We see here that the different self-understandingswarners, protectors or warriorsare not separate categories, but rather different aspects of the same kind of self-understanding, the sense of seeing what others don't, taking the role of protecting both the nation and its women, being misunderstood and held down, and even fighting back.

Enemy images
Hostility to others is constitutive of the self-image of these online actors. Our subjects present two parallel enemy images, which resemble contemporary definitions of right-wing populism (Moffitt, 2016: 43), namely the elite (or related signifiers such as the system or the establishment) and particular Others, such as refugees or minority groups. Importantly, these two enemy images are understood as mutually dependent on one another. The power-holders are the core of the problem, however, fostering a new enemy through their politics.

Politicians and the media
Our subjects' vision of society corresponds fully with Mudde's (2007): 23) definition of right-wing populism: 'the pure people' and 'the corrupt elite' as two homogenous and antagonistic groups, and the insistence that politics should be an expression of the general will of the people. They display a deep distrust of Norwegian democracy and it main societal institutions, starting with where power is located: the politicians. This distrust has developed over time. Some interviewees state that they began to feel distrust during the 1980s, when "the Labour Party failed the working class." They were once faithful Labour voters, but found that things changed drastically when "academics" took over the party. Eatwell and Goodwin (2018) similarly describe the disappointment right-wing populists feel with liberal left-wing politicians' concern with diversity and anti-discrimination and their equivocation over issues such as how to improve conditions for all workers and tackle inequality. Further, our subjects refer to politicians as "power abusers", irrational and selfish: they will "do anything" to accrue more power: The urge of power. […]. It's really more addictive than […] drugs. It does something to their motives. Magnus Our interviewees see ruling politicians as weak leaders who fail their own people, and are thus unfit to govern. In particular, they feel that the politicians do not take the consequences of immigration seriously. Some think that politicians are deliberately engineering immigration; some think that politicians are just stupid, and others that they are cowards. Some interviewees hold that politicians have a "hidden agenda," which implies a silent and strategic Muslim takeover. In line with the Eurabia conspiracy theory (Ye'Or, 2005), Torstein describes this as "fifth column" for building "a new state hidden within the state".
Although politicians appear to be the foremost antagonists, they have important accomplices, such as the media. Our respondents find that the media direct what one may and may not think. For example, media reporting of the refugee crisis is designed to make people feel "sorry for the refugees". The distrust of commercial media among our respondents starts with their interpretation of media angles as manipulation or veiling reality: I don't have faith in anything that comes from the media as regards immigration […] There is very little credibility.
[…] These people come to Norway and will not deliver. See what our prisons are filled withnot Norwegians. Jarle Our respondents think the media withholds information about perpetrators' ethnic backgrounds. Further, they are convinced that the political left has a direct impact on the media. They differ, however, as to whether this influence is accidental or intentional. Magnus believes that socialists and journalists have different "biological dispositions" that affect how they respond to the outside world, and therefore they do not have the same perceptions of danger as the average member of the public. Other subjects think there is strategic cooperation between politicians and the media. They believe that politicians or party delegates are strategically placed in leadership positions in Norwegian media companies and organizations, where they deliberately "plan, hide, and manipulate." Others don't go as far as this, merely pointing to "a pre-arranged game" where politicians control the media and "regime critics" are silenced, and journalists have a "clear political agenda" aimed at manipulating readers' political views. But the outcome is the samethe Norwegian people are subject to "ideological propaganda", aimed at making the left side appear the "good guys" and the right side "evil".

Religious or ethnic enemy images
The representation of people with immigrant backgrounds appears to run parallel to the enemy image of the power elite, together constituting a two-sided "evil". The relationship between these two enemies is aptly described by Moffitt (2016): 146): "it is 'the elite' who are the helm of all that is threatening, or the Other whom 'the elite' has 'let in' or been too lax towards." Our interviewees direct their focus on specific religious minoritieson Muslims especially, but a minority identified Jews as the most dangerous threat.
When describing migrants and Muslims (these two terms were often used interchangeably), our subjects typically used metaphors like "fortune hunters," "cowards" or "luxury refugees". At the centre is the idea of "economic" or "illegal" refugees, who are "wreckers" of the welfare state. For example, when asked about the use of the word "parasite" to characterize Muslims, Ulrik replied: I have used that word a few times, and it is on the border of what's legitimate, I know. Yet, when people consciously and demonstrably undermine the Norwegian welfare system, it is a very reasonable term.
Several interviewees believe it is acceptable to use derogatory expressions in referring to certain groups if these groups violate liberal ideals. Furthermore, they distinguish between legitimate and illegitimate refugees, to support their claim that contemporary refugees should be deported. For example, Magnus contrasts the refugee crisis in 2015 with established narratives about Norwegian emigration to America in the 1800s and early 1900s. These "real refugees" are presented as true workers, in contrast to today's refugees who are given whatever they want and remain ungrateful.
Moreover, they justify their comments by reference to "the people" and "national citizens". Antagonists are presented using war rhetoric with strong symbolism, like "barricades," "invasion" and "occupation," playing on collective reference frames from World War II, when Norway was occupied by the Nazis. A word commonly used is "traitor" -a direct reference to Norway's prosecution of traitors after World War IIas a collective term referring to crimes against the independence and security of the state. Here our subjects are relating to a pre-established narrative, which they then redefine to use against people they see as a threat to Norwegian sovereignty today: I feel that there are a lot of politicians today who ought to be lined up at Akershus Fortress, instead of those politicians who were executed after World War II. There are many today who have a lot to answer for. Sondre The 'traitor' metaphor supports the idea that there are actors in society who actively facilitate a Muslim takeover. Holding a conspiracy-oriented understanding of reality, our interviewees believe that left-wing politicians with a multiculturalist agenda want to destabilize the country and promote "cultural eradication".
The use of war metaphors such as "traitor" reflects a general idea that the country will be, or already is, in some form of war. Media commentary fields provide expressions of this, and indications of what is to come: Because there's already a war. But it's still not at the physical level […]. It is a kind of war that takes place online, which many of us participate in. Ulrik You are in a war. You respond in order to defeat the enemy […]. You need to know the weaknesses of the opponents. You must know that if you act in such-and-such a way, they will respond. Tor Our subjects showed an 'orientalist' reality orientation, by representing Muslims as primitive and barbaric (Said, 1978). They used adjectives such as "medieval," "backward," and "undemocratic", and described Islam as an underdeveloped culture. They criminalize Muslims as dangerous and different from "us" through a rhetoric intended to create anger and fear: Their patriarchal family life, where father and the sons decide on the girls […]. "Here all the women shall be slaves, all the daughters shall be married and used, and as long as they live at home, they must be slaves to the sons and the father of the house." I think that's completely shocking. Sondre Some subjects tried to influence the interviewer by appealing to her being young and female. For example, Torstein said, "But it's your children I'm thinking aboutand especially the girls." Similarly, Sverre asked, "How would you react if your friend or mother was raped because she showed her legs on a summer's day?" Our subjects hold that Muslims lie to hide their true loyalties and intentions. For example, Ulrik states that Muslims have "zero loyalty to Norway and Norwegians and what we stand for" and Tor maintains that a Muslim firefighter in a burning house will save only those who share his faith. Our interviewees believe that Muslims can never be trusted, even those who "claim to be moderate". The idea is that Muslims are gaining power through a slow process whereby they infiltrate European countries from the inside. Ulrik describes this process as "boiling the frog": When the consequences of immigration are being hidden, people don't notice anything, because they don't feel it directly, right?
[…] It's like putting a frog in a pot of cold water. Then you turn on the heat, and it gradually gets warmer. The frog doesn't notice until the temperature creeps closer to boiling point. "Ooh, it's a bit uncomfortable here!" Then it tries to jump out, but it's paralyzed. It's too late, and it dies. This is what's happening here. Ulrik This statement illustrates the conviction that people are the victims of a subtle or hidden attack. This is identical to Ye'Or's (2005) Eurabia term, dhimmitude: European countries gradually surrender control of their territory to the Muslims. The concept of Islamization meets all the requirements of a conspiracy theory: our elites plot secret conspiracies and aim at nothing less than domination and territorial takeover: They enter political parties […]. They're already established in the Norwegian Labour Party. Take Hadja [Tajik], for example […]. They will use democracy to destroy democracy. That's what they intend to do. Jan Most of our subjects focused on Muslims, but Magnus and Tor highlighted the Jews as the greatest threat. They differed from the others by quoting from certain literature to substantiate their points, and used a biologically deterministic understanding. Drawing on established narratives about eugenics and 'racial hygiene' (Reisigl & Wodak, 2001), they argued that differences in physique, mentality andnot leastintelligence, are genetically based.
The fact that Jews have lived in Norway for several generations is irrelevant, since their genetic origins will always determine their sense of belonging and loyalty. According to Magnus, one cannot feel Norwegian and Jewish at the same time: Because they are primarily Jews, they do not feel Norwegian. Being a Jew is not just religion […]. The consequence of living as Jews for centuries has also done something with their genes […]. They will always work to promote Jews, regardless of where they are.
Magnus draws on an anti-Semitic conspiracy theory based on the (now discredited) Protocols of the Elders of Zion, also used by the Nazis. He claims that Jews "are conspiring and trying to undermine state power", and that they represent "a danger to Norway". Both Tor and Magnus speak of a secret conspiracy to take control of the state apparatus to promote Jewish interests. Although there are few Jews in Norway today, they are seen as a kind of invisible enemy, infiltrating power positions from the inside. Their chief aim is to weaken Europe by introducing multiculturalism and opening up for immigration. Unlike our other respondents, Magnus sees Jews as more dangerous than Muslims: The way they work in society poses a much greater threat. We accept them as our own because they look like us. But they have a perverted ideology that causes society to be corrupted.
However, the majority of the interviewees were more concerned about Islamization, resulting in dystopian future perspectives. They used strong expressions to describe this alleged development such as "civil war" "bloodbath" and "revolution": It goes straight to civil war […]. It will be like that as long as the left side doesn't turn around and get a different mind-set. Thomas I know there's going to be civil war throughout Europe. The whole of Europe will be one big Bosnia. Ulrik I'm not exaggerating. In 30 years, we will have full civil war in this country. Torstein Our subjects created a consistent threat scenario, based on the logic: "If (x) does not change, then (y) will happen." The conflict itself is expected to take place in 30-50 years. It will be a consequence of the ongoing all-encompassing Islamization of Europe, including limitations emplaced on national sovereignty, with politicians exercising speech dictatorship, and enforcing the criminalization of opponents. Muslims will take over political power from within and enforce special requirements as universal laws.
It is the bloodbath that follows from immigration, the consequence of forcing people to be politically correct […] I look at the massive invasion of strangers, like bacteria eating rotten flesh in an infected wound, and harming our society. That's what's going on. Magnus Some describe the coming conflict in terms of a calculated plan on the part of specific actors. Groups of people are "sitting on the fence," waiting for the right moment. Tor believes the "brown right" is waiting for a signal to start the war. Several interviewees hold that Muslims are just waiting, and then will control the Europeans with an iron hand; that Muslims pretending to be kind or moderate will show their "true selves" when the time comes.
The conflict will be triggered by a single event that will create chain reactions with more serious consequences. The uprising will be caused by the gradual accumulation of negative pressures, until "the glass overflows". Countries like Sweden, Germany and France which already have considerable immigrationwill fall first, and very violently. Norway will follow but more slowly: There is a certain limit, and then there will be a damn revolution […] I fear a bloodbath when it comes. Things are going to be very violent, I think […] It's going to happen in many places in Europe. It's going to happen here too, but it will take a little longer. Birger In Central Europe, it's going to drain rivers of blood. Magnus 4 Our subjects argue that one should not underestimate this vision of future conflict. They believe others who think the same are not visible in society. But when the changes become noticeable, many will come together and eventually form a social movement.

Conclusion: the battle for truth
Our study has made an important insight into how online news commenters on the far right see their own commentary practice. This is particularly important in a time when, not only in Norway, but also in many other Western countries, there have been several violent attacks perpetrated by actors who explicitly state that their main inspiration stemmed from online communication platforms. Communication on the web, including the commentaries of mainstream newspapers, has become so extreme, that it is hard to distinguish what is meant seriously and what just constitutes a heat of the moment reaction. We argue that it is important to make further investigations into the self-understandings of people who post extreme comments online.
The self-understanding of our interviewees is not just the self-understanding of a selective group of people, but rather resembles a world-view that has similarities both to national populism in general (Eatwell & Goodwin, 2018), and more recent phenomena like the intellectual dark web (Weiss, 2018). Particularly the view of themselves as heroic supporters of the "real truth" as opposed to the false truth that dominates the public sphere is found among adherents of national populism in general, or in different variants of white identity politics (Jardina, 2019;Kaufmann, 2019). Typical is the view that political correctness is shot through with a socialist ideology that sees immigration as enriching society and Islam as a peaceful religion. People like themselves, who criticize immigration and Islam, are branded as racists and extremists. They experience being excluded from the public sphere of expression such as when their comments are deleted from news sites.
The presence "of overt racism" and hate speech "in online comments has prompted news sites to change their practices" (Hughey & Daniels, 2013: 343). This also holds for the two media companies that initially published the comments used for this article; they have both since closed down their commentary fields (Dagbladet, 2016;Nettavisen, 2019), the reason being that there was such a flow of hateful comments, and it was too extensive a task to moderate them. As rightly pointed out by Hughey and Daniels (2013): 343, such a solution, where the newspapers are disallowing commenting, may provide a false sense of a 'post-racial' or 'colour-blind nation', as many "conflate the absence of racism with the moderation of racism". The problem is that hateful comments instead becomes a 'backstage discourse', and enters into more extreme net forums such as chan channels. Also, moderation hinders the ability to interfacing hate speech in a meaningful way. By banning the problem, "frame the offenders as rogue commenters, and then bury their heads in the sand, they hope not to encounter more racism". Therefore, Hughes and Daniels argues, moderation does not address racism (or hate speech), but rather hides and passes the problem along to other virtual venues where racism will find expression. They also point to the fact that current racist discourse is packed into culture war rhetoric and therefore goes under the radar of moderators.
In the interviews our informants, we find a strong sense that the moderation of their comments was unfair. Importantly, not all of their comments were equally extreme. However, they had all experienced being censored many times, and it might be that they were known to the moderators of often writing unacceptable comments. The problem with the censorship of their comments is that the very censorship confirms their world view of being stigmatized, being held down and therefore a sense of being "unfairly victimized white person living under the onus of political correctness where one can't tell it like it is" (Hughey & Daniels, 2013: 342). Such a perception might trigger their extremeness rather than diminish it, and it is a fair chance that they next will join more extreme and closed platforms such as hateful Facebook groups or so-called chan channels. Therefore, even though it is obviously better for victims of hate speech to be spared of reading hateful comments in online newspapers, the problem of hate speech is not solved once and for all with such a solution.
On the other hand, it is not easy to challenge the problem of hate speech by going into discussion with the commenters. Since the subjects think theythrough their alternative sourceshave access to the "real truth", there seems to be little room for meaningful correction. They brand comments that challenge their counter-knowledge as "lies". They deem their own perspectives, and the sources they refer to, as more appropriate and valuable than more mainstream ones.
The latent violent potential of some of the comments means that society needs to take them seriously. It is, however, hard to know whether the commenters' reference to civil war is a ploy aimed at the gallery, or whether they really believe this will happen. Our respondents presented such ideas with conviction and seriousness, and their feelings appeared genuine. How then would they stand in such a conflict, not least regarding the use of violence? In general, they position themselves as outside the conflict, as outsiders looking in. Several say that they "really hope it does not happen," and they are afraid of the day it comes. They usually refer to violence in the passive, without wanting to exercise violence themselves. Interestingly, the outsider sensibility and positioning is similar to how proponents of the so-called intellectual dark web (Weiss, 2018) have described themselves. Echoing our subjects, they hold that they are locked out of mainstream media.
The subjects report different motivations for their commentary practice, ranging from play to provocation, from a call to tell the truth to an urge to warn about the consequences of immigration in order to avert a bloodbath. It is sometimes unclear what constitutes fun and play and what is a serious belief in a truth that most people do not realize. Also, it seems that the same subjects have different motivations simultaneously. Society therefore needs to adopt an equal multi-faceted strategy to deal with such comments. Fact checking, respectful dialogue (even with extreme commenters) and moderation of the most extreme comments seem the best solutions we have, but more research into the effect of different strategies of moderating is needed in order to find the best solution.
Funding Norwegian Research Council, Norway, grant number 144255.