ThesisPDF Available

Far Right, Right Here: Interconnections of discourse, platforms, and users in the digital mainstream

Authors:

Abstract and Figures

Background: This thesis explores the far right online beyond the study of political parties and extremist far-right sites and content. Specifically, it focuses on the proliferation of far-right discourse among ‘ordinary’ internet users in mainstream digital settings. In doing so, it aims to bring the study of far-right discourse and the enabling roles of digital platforms and influential users into dialogue. It does so by analysing what is communicated and how; where it is communicated and therein the roles of different socio-technical features associated with various online settings; and finally, by whom, focusing on particularly influential users. Methods: The thesis uses material from four different datasets of digital, user-generated content, collected at different times through different methods. These datasets have been analysed using mixed methods approaches wherein interpretative methods, primarily in the form of critical discourse analysis (CDA), have been combined with various data processing techniques, descriptive statistics, visualisations, and computational data analysis methods. Results: The thesis provides a number of findings in relation to far-right discourse, digital platforms, and online influence, respectively. In doing so it builds on the findings of previous research, illustrates unexpected and contradictory results in relation to what was previously known, and makes a number of interesting new discoveries. Overall, it begins to unravel the complex interconnectedness of far-right discourse, platforms, and influential users, and illustrates that to understand the far-right’s efforts online it is imperative to take several dimensions into account simultaneously. Conclusion: The thesis makes several contributions. First, the thesis makes a conceptual contribution by focusing on the interconnectedness of far-right efforts online. Second, it makes an empirical contribution by exploring the multifaceted grassroots or ‘non-party’ dimensions of far-right mobilisation, Finally, the thesis makes a methodological contribution through its mix of methods which illustrates how different aspects of the far right, over varying time periods, diversely sized and shaped datasets, and user constellations, can be approached to reveal broader overarching patterns as well as intricate details.
Content may be subject to copyright.
Department of Sociology
Umeå 2022
FAR RIGHT, RIGHT HERE
Interconnections of discourse,
platforms, and users in the
digital mainstream
Mathilda Åkerlund
This work is protected by the Swedish Copyright Legislation (Act 1960:729)
Dissertation for PhD
ISBN: 978-91-7855-724-0 (print)
ISBN: 978-91-7855-725-7 (pdf)
ISSN: 1104-2508
Cover design by Joakim Bergqvist
Electronic version available at: http://umu.diva-portal.org/
Printed by: Cityprint i Norr AB
Umeå, Sweden 2022
Table of Contents
Abstract ................................................................................... i
List of original papers ............................................................ ii
Acknowledgements ............................................................... iii
Introduction ............................................................................ 1
Background ............................................................................ 5
Delineating ‘the far right’ .................................................................... 5
Far-right party ideology and voters ........................................................ 5
Defining the digital far right .................................................................... 9
From the dark corners to the digital mainstream ............................ 10
Far-right discourse, platforms, and influential users ............ 14
Far-right discourse ........................................................................... 14
Discursive elements ................................................................................. 14
Efforts to normalise far-right discourse online ..................................... 16
Digital platforms ............................................................................... 19
The logics of platforms ............................................................................ 19
Far-right enablement through socio-technical features ....................... 21
Influence online ................................................................................ 23
The characteristics of online influence ................................................... 23
Far-right micro-celebrities and covert far-right influence ................... 25
Modelling far-right discourse, platforms, and influential users ...... 27
Case, data, methods, and ethical considerations .................. 30
Sweden as an empirical case ............................................................. 30
A brief overview of the contemporary Swedish far right ..................... 30
Situating ‘the Swedish case’ in the thesis ............................................... 33
Data and methodological approach .................................................. 35
Methodological benefits, compromises, and potential pitfalls ............. 36
The critical discursive perspective ......................................................... 39
Ethical considerations ...................................................................... 41
Paper summaries ................................................................. 45
Paper one .......................................................................................... 45
Paper two .......................................................................................... 47
Paper three ........................................................................................ 48
Paper four ......................................................................................... 50
Discussion ............................................................................. 52
Responding to the research questions ............................................. 52
The complexity of far-right discourse, platforms, and influence .... 57
Contributions and future research ....................................... 60
References ........................................................................... 63
Abstract
Background: This thesis explores the far right online beyond the study
of political parties and extremist far-right sites and content. Specifically,
it focuses on the proliferation of far-right discourse among ‘ordinary’
internet users in mainstream digital settings. In doing so, it aims to bring
the study of far-right discourse and the enabling roles of digital
platforms and influential users into dialogue. It does so by analysing
what is communicated and how; where it is communicated and therein
the roles of different socio-technical features associated with various
online settings; and finally, by whom, focusing on particularly influential
users.
Methods: The thesis uses material from four different datasets of
digital, user-generated content, collected at different times through
different methods. These datasets have been analysed using mixed-
methods approaches wherein interpretative methods, primarily in the
form of critical discourse analysis (CDA), have been combined with
various data processing techniques, descriptive statistics, visualisations,
and computational data analysis methods.
Results: The thesis provides a number of findings in relation to far-
right discourse, digital platforms, and online influence, respectively. In
doing so it builds on the findings of previous research, illustrates
unexpected and contradictory results in relation to what was previously
known, and makes a number of interesting new discoveries. Overall, it
begins to unravel the complex interconnectedness of far-right discourse,
platforms, and influential users, and illustrates that to understand the
far-right’s efforts online it is imperative to take several dimensions into
account simultaneously.
Conclusion: The thesis makes several contributions. First, the thesis
makes a conceptual contribution by focusing on the interconnectedness
of far-right efforts online. Second, it makes an empirical contribution by
exploring the multifaceted grassroots or ‘non-party’ dimensions of far-
right mobilisation, Finally, the thesis makes a methodological
contribution through its mix of methods which illustrates how different
aspects of the far right, over varying time periods, diversely sized and
shaped datasets, and user constellations, can be approached to reveal
broader overarching patterns as well as intricate details.
Page ii
ii
List of original papers
I. Åkerlund, M. (2021a). Dog whistling far-right code words: The
case of ‘culture enricher’ on the Swedish web. Information,
Communication & Society, 0(0), 1–18.
https://doi.org/10.1080/1369118X.2021.1889639
II. Merrill, S., & Åkerlund, M. (2018). Standing Up for Sweden? The
Racist Discourses, Architectures and Affordances of an Anti-
Immigration Facebook Group. Journal of Computer-Mediated
Communication, 23(6), 332353.
https://doi.org/10.1093/jcmc/zmy018 (co-authored equally)
III. Åkerlund, M. (2020). The importance of influential users in
(re)producing Swedish far-right discourse on Twitter. European
Journal of Communication, 35(6), 613628.
https://doi.org/10.1177/026732312094090
IV. Åkerlund, M. (2021b). Influence Without Metrics: Analyzing the
Impact of Far-Right Users in an Online Discussion Forum. Social
Media + Society, 7(2), 1-11.
https://doi.org/10.1177/20563051211008831
Page iii
iii
Acknowledgements
I am grateful to Simon and Sam for providing me a space with freedom
and independence to explore and learn on my own but where I have also
always felt supported, and prioritised. Thank you both. I would also like
to thank Mattias Ekman for your feedback and advise at my mid seminar,
and to Mattias Wahlström for pushing me in the final stages of writing.
Both your comments have greatly improved this thesis. Thanks also to
Maja, Moa and Jeff for being such great friends and co-workers. I would
also like to extend my appreciation to the rest of my colleagues at the
Sociology department and DIGSUM for your support. And to Joakim,
thank you for all your love and encouragement, and for the amazing work
with the cover design.
1
Introduction
As far-right parties and movements continue to gain political salience
throughout Europe and the rest of the world, beliefs which were
previously considered extremist are now progressively being normalised
as acceptable political ideas. Among other things, this includes efforts to
sow suspicion towards legacy media and state institutions, and to
undermine the rights of ethnic minorities, LGBTQI+ persons, and women
(Mudde, 2019). While the far right’s increasing political prominence
requires large-scale support among the general public, research shows
that it might not be people’s political attitudes that have suddenly changed
(Bohman, 2018; Bohman & Hjerm, 2016; Mudde, 2013). Instead, the
explanation could lie in the far-right’s ability to mobilise supporters
(Strömbäck & Theorin, 2018). This thesis seeks to deepen the
understanding of one underexplored dimension of far-right
mobilisationthat which takes place through everyday user-generated
content on popular social media platforms and online forums. Or in other
words, expressions of far-right discourse by users in the digital
mainstream.
Notably, the internet has proven important for the dissemination of far-
right discourse into mainstream public debate (Ekman, 2015; Karl, 2019;
Schwarzenegger & Wagner, 2018; A. Winter, 2019) but more than that,
the internet is often used actively and skilfully by the far right to help
increase their legitimacy and appeal to less radical audiences (Daniels,
2009; Gerstenfeld et al., 2003; A. Klein, 2012). It might be tempting to
dismiss the relevance of digital far-right settings, networks, discourses,
and users as marginal and merely virtual phenomena separate from
anything in the tangible ‘real world’. However, with contemporary society,
politics, and not to mention everyday life, characterised by an increasingly
complex media environment and interconnectedness of digital
technologies, such distinctions become futile (Chadwick, 2017; Couldry &
Hepp, 2017; Thompson, 2020). Aside from the fundamental impact the
internet has had on political communication and protest movements
more generally (Bennett & Segerberg, 2013; Castells, 2015; Earl &
Kimport, 2011), the prominence of digitally mediated communication has
also rendered the far right post-digital. That is, the online and offline
efforts by the far right are nowadays so closely intertwined that they
cannot anymore be neatly separated (Thurston, 2019; see also Simi &
Futrell, 2006).
2
While the focus of far-right discourse can of course vary depending on a
range of factorsincluding in what social, situational, or temporal
contexts it is expressedthe actual content of far-right discourse
communicated online differs little from that expressed in non-digital
settings. That is, far-right discourse online tends to incorporate the same
elements as far-right discourse offline (see e.g., Atton, 2006; Fielitz &
Thurston, 2019; Mondon & Winter, 2020; Mudde, 2019; A. Törnberg &
Wahlström, 2018). However, the ways in which these discourses can be
communicated and spread online provides the far right with never-before-
seen possibilities to popularise their political agenda. Therefore, to
understand and curb the threat to democratic values posed by the far right
more generally, it is crucial to comprehend how it functions online
(Albrecht et al., 2019; Davey et al., 2019; O. Klein & Muis, 2019; see also
Bowman-Grieve, 2009). This, however, is easier said than done.
Far-right discourse online is oftentimes deliberately covert, coded, and
downplayed, and can be expressed by anonymous users who might or
might not be affiliated with any specific parties or organisations, who
promote varying levels of far-right extremist values, and who skilfully
leverage different online settings for their own purposes. As such, it is not
surprising that research has historically often focused on aspects of the far
right online which are overt and easily outlined. That is, on the efforts of
political parties (Atton, 2006; Ben-David & Matamoros Fernández, 2016;
Ernst et al., 2017), and on explicitly extremist far-right sites and content
(e.g., Bowman-Grieve, 2009; De Koster & Houtman, 2008; Kleinberg et
al., 2020; P. Törnberg & Törnberg, 2021). However, these disregard that
far-right discourse online is also widely known to circulate in more
‘mainstream’ digital spaces. That is, settings which are not dedicated to
party-specific or extremist far-right ideas, and as such, are not places
which users can be expected to visit with intent to seek out far-right
discourse specifically. In these settings, content creation and interaction
can be unpredictable and diverse since its users can range from the most
committed far-right voters, members, leaders, and party-officials to far-
right extremists, non-organisational activists, right-wing conservatives,
political opponents, and those who might not know or have any interest
in the far right at all.
Consequently, in mainstream digital settingsunlike in those explicitly
dedicated to the far rightfar-right discourse risks becoming part of the
internet experience also for those who have not intentionally sought it out
(Daniels, 2009). These encounters can have severe consequences both
due to the gradual normalisation of far-right discourse more generally
online and offline but also in terms of the potential step-wise
3
radicalisation of individual users. While it is often assumed that insidious
content is just an inevitable consequence of hateful people being active
online, others instead argue that it is the way that many platforms
function which enable the creation, spread, and persistence of different
forms of repressive or hateful content (Munn, 2019). Consequently, it is
not enough to establish that the far right is now increasingly leveraging
mainstream digital settings and study the discursive strategies through
which this takes shape. The role of different platforms in enabling the
proliferation of far-right discourse must also be explored. And, while often
disregarded in research, far-right discourse does not appear organically
online nor do platforms create and circulate it only by themselves. Rather,
it is chiefly the result of deliberate user practices (Colley & Moore, 2020;
Kleinberg et al., 2020; Massanari, 2017; Zhou et al., 2019), and in
particular, some users tend to be exceedingly influential and matter more
than others when it comes to propagating far-right discourse (Berger &
Strathearn, 2013). Instead of merely exploring language use on a
collective level, research must also consider the role of particularly
influential users in disseminating far-right ideas and enabling their
potential normalisation.
This thesis aims to contribute understanding about far-right discourse
expressed by seemingly ‘ordinary’ users in the digital mainstream. To do
so it analyses what is communicated in these settings and how; where it
is communicated and therein the roles of different socio-technical
features in online settings; and finally, by whom, focusing on particularly
influential users. The research is guided by the following research
questions:
RQ1. How is far-right discourse expressed and normalised in different
online settings?
RQ2. How do online platforms enable the circulation and
amplification of far-right discourse?
RQ3. What is the role of influential users in propagating far-right
discourse and contributing to far-right community-building through
their uses of online platforms?
Taking a case study approach, this thesis explores four separate cases, all
in Swedish language settings online. While this thesis uses Sweden as an
overarching case to study the far right online, it seeks to provide insight
into the far right on a more general level, as is further elaborated in
Situating ‘the Swedish case’ in the thesis. The first paper maps the use of
coded far-right language over time and across online platforms; the
4
second explores far-right discourse and its enablement through the
technological features of an anti-immigration Facebook group; the third
and fourth papers respectively, investigate how far-right influence takes
shape on a conventional social media platform where influence is easily
showcased, and in an online discussion forum with few public measures
to indicate users social statuses on the site.
The rest of this thesis is structured as follows. The next section gives some
background to the far right and the overarching issues in focus in this
thesis. After this, section three, Far-right discourse, platforms, and
influential users describes previous research and the theoretical
framework of the thesis based on the three dimensions of digital
discourse, platforms, and influence in relation to the far right. Section
four, Case, data, methods, and ethical considerations, subsequently,
describes the use of Sweden as an overarching case in this thesis, a short
background of the Swedish far right, and the data, methods, and ethical
considerations taken throughout the project. Section five thereafter,
presents summaries of the four papers, then comes section six,
Discussion. Finally, the last section holds the conclusion, implications,
and suggestions for future research.
5
Background
Online aspects of far-right mobilisation have often seemed too difficult to
outline. Consequently, the overwhelming bulk of literature seeking to
define ‘the far right’ has focused on the characteristics of far-right ideology
as it pertains to far-right organisations, and primarily to far-right parties.
While the blurriness that is far-right mobilisation online is arguably very
different from the rigid delineations proposed to define far-right party-
political ideology, the lack of research on digital definitions nevertheless
prompts this thesis to proceed from these conventional views of the far
right. Accordingly, the coming subsection provides some background to
how the far right’ has previously been defined elsewhere, debates
surrounding these definitions, and some discussion on the supporters of
far-right parties. Finally, this subsection ends with a discussion on how
these definitions can be broadened to better fit this thesis and its concern
with digital dimensions of the far right. Thereafter, a second subsection
provides some background to the digital presence of the far right online
and the increasing prominence of far-right sympathisers and discourse in
the digital mainstream.
Delineating ‘the far right’
Far-right party ideology and voters
The far right has successfully put issues like national identity and
immigration on the public agenda (Art, 2011; Caiani et al., 2012; Mudde,
2019; Wodak, 2015). Around the world, and in Europe in particular, the
revival of the far right has been deemed as among “the most significant
political changes in democratic states during the past decades” (Rydgren,
2018a, p. 23). Yet, there is no general agreement among researchers of
how the contemporary far right should actually be defined (Carter, 2018;
Mudde, 2019). Far right, far-right populist, far-right extremist, right-
wing extremist, extreme right, extremist right-wing, radical right,
radical right-wing and populist radical-right are some of the terms used
interchangeably and synonymously to describe movements, ideas, and
organisations on the rightmost end of the political spectrum.
This lack of consensus has been described as a ‘war of words’ (Mudde,
1996). In fact, when surveying the literature two decades ago, Mudde
(2000) identified no less than 26 definitions, and a total of 58 different
6
features incorporated into ‘right-wing extremism’; a number which is only
likely to have increased since. Nonetheless, there are some main
definitions that reoccur in scholarly literature. An often-used definition of
contemporary far-right party ideology is that it is nationalist, xenophobic,
welfare chauvinist, and authoritarian (Mudde, 2000). Mudde
determines, like many others (e.g., Bar-On, 2018; Eger & Valdez, 2019a,
2019b), that nationalism is at the very centre of far-right ideology in terms
of an aspiration for a homogenous and mono-cultural society within the
state. Anything that deviates from this homogeneity is a threat
(xenophobia). Consequently, welfare policies should put this homogenous
‘people’ first (welfare chauvinism), and the legal system should work to
protect the nation and its ‘people’ against perceived internal and external
threats and enemies (authoritarianism) (Mudde, 2000; see also
Hainsworth, 2000). Others also add some measure of criticism towards
democracy as a central feature of far-right ideology (e.g., Carter, 2018;
Crosset et al., 2019). Besides these characteristics, other definitions place
even more emphasis on the presence of perceived threats, including a
suspicion of federal authority, and beliefs in threats to personal freedom
and/or national sovereignty (Freilich et al., 2014). Others also highlight
the promotion of violence to defend or transform society to fit far-right
ideals (Blee & Creasap, 2010; Crosset et al., 2019).
It has also been argued that populism is an important feature of far-right
ideology (Rydgren, 2018a). Mainly, that it is ‘anti-establishment’ in the
sense that ‘the establishment’ or ‘elites’ are thought to stand in opposition
to ‘the people’ (Engesser et al., 2017; Rydgren, 2002; Rydgren & van der
Meiden, 2018; Wodak, 2015). Like other descriptions above, populist-
based definitions also often highlight features like nationalism,
xenophobia, and authoritarianism (Herman & Muldoon, 2019; Rydgren,
2018a). In fact, some argue that there tends to be little difference in
content between definitions including or excluding populism as a
characteristic of far-right ideology (Mudde, 2000).
Beside the diverging views among researchers regarding the constituting
parts of far-right party ideology, there is also something to be said about
the words making up the terms themselvesnamely the ‘extreme’,
‘radical’, and ‘far’ labels in relation to some sort of political ‘right’. The
leftright scale was previously primarily a division based in socio-
economic policies, wherein the left advocated for increased state
involvement and the right for market economy. More recently, it has
instead increasingly been expressed as a division between an
inegalitarian, authoritarian, and nationalist right wing, versus an
egalitarian, liberal, and internationalist left wing. (Mudde, 2019). With
7
this in mind and as the above definitions illustrate, the far right should
not be defined as rightist based only on the economic policies often
associated with the parliamentary political right, even though some far-
rightists are of course also rightist in an economic political sense. Rather,
the far right is ‘rightist’ based on its conservative socio-cultural views
(Jones, 2016; Rydgren, 2018b; Spanje, 2011), and ideas about inequality
as natural and necessary for the good of society (Bobbio, 1996).
Furthermore, while different descriptors are often used interchangeably,
the ‘extreme’, ‘radical’, or ‘far’ prefixes of rightism denote varying degrees
of intolerance. The ‘extreme’ label has often come to describe White
supremacy (Durham, 2000), and opposition to democracy (Fielitz &
Laloire, 2016; Rydgren, 2018a), while the ‘radical’ label instead tends to
describe more moderate ideas and organisationsthose which are “not
explicitly opposed to democracy but nonetheless hostile to the way
representative democracy functions in contemporary society” (Rydgren,
2018a, p. 24; see also Fielitz & Laloire, 2016). The ‘far’ label on the other
hand, is perhaps the most comprehensiveand therefore the label used
throughout this thesisas it is often described as encompassing “both
radical and extreme rightists as well as ultranationalists” (Durham, 2000,
pp. xiixiii; see also Fielitz & Laloire, 2016; Mudde, 2019).
However, these delineations are anything but clear-cut. The use of labels
like ‘extreme’, ‘radical’, and ‘far’ also indicate different degrees of
influence or marginalisation of rightism (Blee & Creasap, 2010). This is
problematic because focusing on the ‘radical’ and ‘extremist’ has helped
exaggerate the differences between the party political ‘mainstream’ and
‘the far right’, obscuring their potential closeness and interconnections
(Feischmidt & Hervik, 2015). In reality, there are often overlaps between
more or less extremist far-right ideas, organisations, and people
(Rydgren, 2018a). Moreover, the far right is far from homogenous, its
features can vary greatly over time and across national contexts (Art, 2011;
Hainsworth, 2000; Spanje, 2011; see also Eger & Valdez, 2019b, 2019a;
Schroeder, 2020), and might among other things differ in terms of
history, organisation, and leadership (Mudde, 2019).
Delineating the far right from the mainstream is further obfuscated by the
instability of these positions. Most recently, this has been due to the shifts
of far-right political parties closer toward the political centre (Cammaerts,
2018; Feischmidt & Hervik, 2015), and the political centre’s appropriation
of ideas and issues previously only associated with the far right (Fielitz &
Laloire, 2016; Mudde, 2019). Some argue that the increased voting for far-
right parties over time is in part due to this shift in positions, which better
8
fits with voters’ attitudes and demands (Boonen & Hooghe, 2014; Eger &
Valdez, 2015). Beyond this, there are a range of studies attempting to
pinpoint why people support far-right parties, not the least in a European
context (Arzheimer, 2017).
First, it is important to note that not all far-right supporters or voters are
extremists or fanatics (Arzheimer, 2017). They range from those holding
lifelong commitments to the far right, to those who have converted to far-
right beliefs, or whose allegiances my shift back and forth over time, and
finally, to those who perceive their far-right stance as matters “of
circumstances beyond their control (Mudde, 2019, p. 127).
Correspondingly, there are several reasons for why people support the far
right. Research on far-right voting specifically, has shown that some vote
for far-right parties in protest of mainstream parties, which they either do
not identify with or perceive are not addressing issues of importance
(Arzheimer, 2017). However, many voters of course also agree with far-
right ideology. For instance, people who hold welfare chauvinist (Eger &
Valdez, 2015) and certain nationalistic attitudes are more likely to vote for
a far-right party (Lubbers & Coenders, 2017). And most prominently,
many studies have shown that far-right voters tend to hold anti-
immigrant attitudes (Abou-Chadi et al., 2021; T. J. Allen & Goodman,
2021; Eger & Valdez, 2015; Spierings, 2021).
These attitudes depend on a range of factors. It is often argued that
working class people (Arzheimer & Berning, 2019; Rydgren & Tyrberg,
2020) and especially men (T. J. Allen & Goodman, 2021), as well as those
who experience a sense of economic inequality (Arzheimer, 2017;
Arzheimer & Berning, 2019; Engler & Weisstanner, 2021) are more likely
to vote far right. However, due to the increasing success of the far right,
these groups are broadening (Abou-Chadi et al., 2021). Despite
stereotypical ideas about far-right voters as white, male, young(ish),
moderately educated, and concerned about immigrants and
immigration”, most people who fit into this description do not vote for the
far right, while many who do vote for the far right are not necessarily
‘typical’ far-right voters (Mudde, 2019, p. 127).
The increasing popularity of the far right is the result of a complex
interconnectedness of attitudinal, social, economic, and political factors
across Europe and the rest of the world. However, while a great number
of studies have attempted to detail the subtle and intricate differences
between parties in the far-right ‘family’ (e.g. Mudde, 1996; Spanje, 2011),
and a great deal others have posed questions about the who’s and why’s
of far-right voting (Arzheimer, 2017), these research efforts are not
9
enough to paint a full picture of the far right today. As research points to
the importance of far-right mobilisation as crucial in its growing success,
the activities of far-right sympathisers online cannot be overlooked.
Defining the digital far right
With the potential meanings and understandings of various descriptors
and not to mention the ‘rightist’ term itself, and with the wide range of
characteristics that have been ascribed to the far right before, it is perhaps
not surprising that there is a lack of agreement regarding what actually
constitutes ‘the far right’. Online, the challenge to define far-right ideology
and categorise those supporting it becomes even greater.
There are a few primary reasons for why the far right is particularly
difficult to outline on the internet. Mainly, while some official party or
organisational accounts online might be easily identified, many individual
users have no official ties to the organisational far right or use anonymity
or pseudonymity to disguise these allegiances (Crosset et al., 2019; see
also Burris et al., 2000). There is also a disparate interconnectedness of
political parties, organisations, communities, networks, and individuals,
who operate in a range of different digital settings, but also globally, across
national contexts (Ben-David & Matamoros Fernández, 2016; Caiani &
Kröll, 2015; Chau & Xu, 2007; Gerstenfeld et al., 2003). With the plethora
of varying understandings and definitions among these users, groups, and
networks regarding what far right ideology is, it becomes exceedingly
difficult to delineate (Albrecht et al., 2019; Munn, 2019). Consequently,
discourse expressed by hardcore political party members, bots, or even
organisational channels themselves, might be very difficult to distinguish
from the voters, trolls, curious conservatives, or even political opponents.
Following the debates discussed throughout this section, ‘far-right
ideology’ is used in this thesis as an umbrella term (Schwarzenegger &
Wagner, 2018) to describe ideology which lies ‘right of the right’, that is
beyond the political right, in terms of socio-cultural values and ideas
about equality (Art, 2011; Bobbio, 1996; see also Castelli Gattinara & Pirro,
2019). ‘Far-right discourse’ describes the expression and reproduction of
far-right ideology through language use (see also van Dijk, 2006). Finally,
‘the far right’ refers to a diverse network of parties, organisations,
communities, and individuals, wherein it is in particular the expression of
far-right discourse by the latter two categoriescommunities and
individualswhich, to the extent that they are identifiable, are in focus
here. It should also be noted that while populism can often characterise
10
far-right discourse, consistently labelling the discourse and ideology
studied in this thesis as ‘populist’ is limiting. Especially online, where
there is no single, agreed-upon way of understanding ‘the far right’, such
a label provides unnecessary constraint.
While these definitions might be vague and diffuse, and while it can
certainly seem like a simplification to call the overtly extremist as well as
the slightly offensive by the same label, the point here is not to map and
categorise all individual users and their statements on a far-right
extremist scale. Unlike far-right parties, which have a party program, and
which present a coherent collective identity (see also Bennett & Segerberg,
2013), any individual user, group, network, or community might manifest
a range of different perspectives, viewpoints, and ideologies. Sometimes
these change over time or even contradict themselves from one sentence
to the next. A broad definition of the far right allows for exploring also the
sometimes inconsistent, seemingly less radical discourse, platforms, and
users which have often gone overlooked in studies adhering to strict
definitions.
From the dark corners to the digital mainstream
While the internet on the one hand, has proven important for
democratisationenabling the propagation of progressive ideas, voices,
and movements, in protest of issues like racism (Freelon et al., 2016; Kuo,
2018) and sexual harassment (Jackson et al., 2018; Lindgren, 2019;
Mendes et al., 2018)it is also a space in which reactionary political
expressions, like those of the far right, have come to thrive (Davey et al.,
2019; Futrell & Simi, 2004; Lumsden & Harmer, 2019; Siapera & Veikou,
2016).
The internet offers the far right (and others) space for elsewhere
stigmatised ideas, scalable, cheap, and efficient tools for organising, and
a perception of power through numbers as a result of broad, global reach
(Adams & Roscigno, 2005; Burris et al., 2000; Gerstenfeld et al., 2003;
Koehler, 2014). As such, the internet has often been highlighted for its
importance in enabling the connection of a heterogeneous network of far-
right sympathisers, effectively linking together and providing a sense of
community among people and organisations from various social and
geographical settings (Adams & Roscigno, 2005; Back, 2002; Perry, 2001;
Scrivens et al., 2020; Perry & Scrivens, 2016; Futrell & Simi, 2004; Burris
et al., 2000; Gerstenfeld et al., 2003).
11
While the far right’s presence on the web is as diverse as the users and
groups which constitute it, it can be roughly divided into two categories:
first, that which takes place on inward-facing or fringe, extremist sites
the ‘darker corners’ of the web (Chandrasekharan et al., 2017; Rogers,
2020)which have historically been very important for the far right’s
digital efforts; and second, the more recent and seemingly less extremist
forms of presence in the digital mainstream (e.g. Crosset et al., 2019;
Farkas et al., 2018a; Gaudette et al., 2020; Matamoros-Fernández, 2017),
which is in particular focus in this thesis.
The far right’s activity on fringe, extremist sites date back to the early days
of the commercial internet and have thus long constituted an important
site of study for understanding the far right online. As early as 1995,
Stormfront, often referred to as the first white Supremacist website, was
launched by former Ku Klux Klan member Donald Black (Kaplan et al.,
1998; Schafer, 2002). The site is still today considered “a breeding ground
for right-wing extremists worldwide” (Kleinberg et al., 2020, p. 2) not the
least due to its prolific usage by impending far-right terrorists (Southern
Poverty Law Center, 2014). This is likely, at least in part, due to the site’s
ability to foster meaningful interactions, social refuge, and a sense of
community among radical far-right sympathisers (Bowman-Grieve,
2009; De Koster & Houtman, 2008; A. Törnberg, 2021; P. Törnberg &
Törnberg, 2021).
For a long time, and still today, far-right blogs have also been important
for fringe far-right activity both internationally (Baele et al., 2020;
Cammaerts, 2007; Chau & Xu, 2007; Meleagrou-Hitchens & Brun, 2013)
and also specifically in Sweden and other Nordic countries (Horsti, 2017;
Pettersson & Sakki, 2017, 2020). The continuous significance of far-right
blogs is likely due to the fact that in a relatively inexpensive manner (Holt
et al., 2019), and with high levels of creative freedom and control,
individual far-right bloggers are able to easily discuss and circulate their
ideas within a wider ‘blogosphere’ of likeminded far-right sympathisers
(Horsti, 2017; Sakki & Pettersson, 2016).
More recently, the far right’s inward-facing activities online have involved
the establishment and use of new spaces and platforms, including a
number of ‘alternative’ social media sites like Gab, Telegram, and Parler
which in different ways have been found to enable the proliferation of hate
speech (Zannettou et al., 2018; Zhou et al., 2019) and legitimation of
violence (Munn, 2021), while providing safe, private spaces for elsewhere
banned (or ‘deplatformed’) users (Rogers, 2020; Urman & Katz, 2020).
Notably, the far right has also demonstrated an increasing need for their
12
own sources of news and information due to a surging disbelief in legacy
media, or ‘lügenpresse’ (lying/liar press) as far-right sympathisers have
often called it (Figenschou & Ihlebæk, 2019; Haller et al., 2019; Haller &
Holt, 2019). Consequently, the United States and Europe have during the
last decade seen the emergence of many alternative news sites
broadcasting the agendas of far-right parties and politicians (Haller et al.,
2019; Marwick & Lewis, 2017; Schroeder, 2019). Over time, these outlets’
appearances and marketing have grown increasingly professional,
consequently draping alternative news in a cloak of legitimacy despite
their rejection of the ethical and editorial standards adhered to by legacy
news media organisations (Figenschou & Ihlebæk, 2019; see also
Schroeder, 2019).
Fringe, inward-facing, extremist settings like the ones described above
attract those who hold the most radical far-right views. They are valuable
to them because they provide safe spaces to perpetuate their specific world
views and to nurture and strengthen their community in the absence of
criticism or contesting opinion. The ease with which to delineate these
users, networks, and sites, and the far right’s long tradition in these kinds
of settings has meant that inward-facing, extremist sites have to date
garnered much of the attention of research exploring the far right online.
However, more recently, far-right activity is also moving into the digital
mainstream.
In recent years, research has shown how far-right accounts have become
increasingly common on conventional social media sites like Facebook,
Twitter, and YouTube (Ben-David & Matamoros Fernández, 2016;
Crosset et al., 2019; Ekman, 2019; Matamoros-Fernández, 2017; Munger
& Phillips, 2019; Oboler, 2016; Wahlström et al., 2020), as well as popular
sites with highly eclectic user bases like Reddit, 4chan, and the Swedish
online forum Flashback (Colley & Moore, 2020; Gaudette et al., 2020;
Massanari, 2017; Phillips, 2019; Topinka, 2018; A. Törnberg & Törnberg,
2016a, 2016b).
The penetration of far-right discourse into such mainstream online
settings, from a societal perspective, means that it is more readily
available online and that users who are not affiliated with the far right risk
encountering and potentially becoming affected by far-right ideology
(Daniels, 2009). Due to the widespread use of mainstream digital
platforms, the users expressing far-right discourse in these settings also
range greatly in levels of extremism. Consequently, those who might come
across it could be any given search engine, social media, or forum user.
This could have severe consequences as mainstream digital settings have
13
been found to work as stepping-stones for users towards increasingly
more radical far-right ideas and contexts (Munn, 2019, 2020; Ribeiro et
al., 2020). From a far-right perspective, the move from far-right enclaves
into mainstream online spaces provides a far greater discursive reach and
opportunities to access entirely new audiences compared to far-right
content which circulates on dedicated extremist sites among far-right
sympathisers (Hawley, 2017). However, these mainstream settings come
with the collapse of many different social contexts and audiences to
navigate (Marwick & boyd, 2011) and therein, entirely different rules
when it comes to interaction and the expression of far-right discourse.
Including for instance what can be seen as shared political knowledge and
common sense, and in regard to what is thought of as socially acceptable
behaviour. To deal with these conditions and ultimately to leverage them
successfully, far-right users have to act in ways that can be considered
appropriate according to the rules and regulations of these platforms
themselves, while still managing to appeal to both old and potentially new
sympathisers.
14
Far-right discourse, platforms, and
influential users
The following section discusses some of the previous research and central
theoretical underpinnings which have been important in shaping the
empirical research carried out in this thesis. The chapter gives an overview
of far-right discourse, digital platforms, and user influence online and
thereafter ends with a theoretical discussion on the need to go beyond
individual explorations of these dimensions to gain a more comprehensive
understanding of far-right efforts in the digital mainstream.
Far-right discourse
Where the previous section discussed the central characteristics of far-
right ideology, this section instead seeks to identify the expression and
reproduction of this ideology through language use. It focuses first on
what meanings these discourses contain, that is on the elements of far-
right discourse. Second, it discusses how far-right discourse can be
expressed online, through discursive strategies.
Discursive elements
Across national contexts, research has shown how the nationalist
dimension of far-right ideology tends to revolve around the discursive
construction of an usversus a them(Hameleers & Schmuck, 2017;
Wodak, 2009). Here, an exclusionary defined in-group, supposedly
constituted by a homogenous native ‘people’ (Hameleers et al., 2019;
Krämer, 2017) stands in opposition to various perceived ‘threats’ posed by
external as well as internal out-groups (see also Sakki & Pettersson, 2016).
The xenophobic dimension of far-right ideology is manifested in
particular through the delineation of external threats. This refers to those
who are considered not to be part of the imagined, native people. This has
since long meant racism and violence towards Jews (Lööw, 2016; Quinley
& Glock, 1979) and Black people (Fredrickson, 2015; J. Wilson, 1976), and
more recently, also towards Muslims. While the perceived threat of Islam
and the ‘Orient’ can be traced all the way back to medieval times (Said,
1978), as a consequence of 9/11, Islamophobia has surged throughout the
United States and Europe in the last decades (C. Allen, 2010; Sheridan,
2006) to become one of the central elements of far-right discourse (e.g.
15
Awan, 2014, 2016; Awan & Zempi, 2017; Ekman, 2015; Farkas et al.,
2018a; Vieten, 2016).
Research on far-right expressions of Islamophobia show depictions of
Islam as an anti-democratic and non-progressive opposite to ‘the West’.
With this, Muslims are portrayed as incompatible with Western culture,
values, traditions, and heritage (Cammaerts, 2018; Feldman & Jackson,
2014; Mondon & Winter, 2017; Oboler, 2016). Muslim men in particular,
are thought to also pose a threat to native people, especially its women,
due to their believed inherently violent and overly sexual nature (Horsti,
2017; Sakki & Pettersson, 2016).
These discursive representations are closely connected to anti-
establishment, authoritarian and welfare-chauvinist dimensions of far-
right ideology, through depictions of the perceived internal threat posed
by legacy news media, progressive movements and people, and
mainstream political partieswhat the far right calls ‘the (politically
correct) establishment’ (Atton, 2006; Dignam & Rohlinger, 2019;
Feischmidt & Hervik, 2015; Keskinen, 2018; Korolczuk, 2020; Nilsson,
2021). According to previous research, the far right often argues in a
populist fashion that the so-called ‘establishment’ betrays the nation and
its legitimate people by favouring the needs and rights of immigrants over
those of its native population. Supposedly, this is done by lying and
covering up for immigrants by supressing the truth about the social and
economic consequences of immigration, and by threatening native
people’s freedom of speech and opportunity to protest these conditions
(Krämer, 2017; Merrill, 2020).
Finally, previous research has illustrated how nationalism is expressed by
far-right in-groups through a longing to restore a lost agehowever
loosely definedof native glory, which is discursively reconstructed as an
idealised time before immigration (Engesser et al., 2017; Duyvendak,
2011; see also Elgenius & Rydgren, 2017, 2019). In Sweden, these nostalgic
constructions have been found to invoke a mishmash of idealised periods
in Swedish history, spanning from the perceived pinnacle of traditional
gender roles, social cohesion, and welfare during the mid-20th century
‘folkhem’ (people’s home) all the way back to the masculine, conquering
days of the Viking age (Merrill, 2020; Schroeder, 2019). Oftentimes, these
depictions portray Swedishness as something homogenous, masculine,
and importantly, white (Horsti, 2017).
The close interwovenness of whiteness and Swedishness originate in
longstanding ideas about ‘the Swede’ as in every way racially superior, not
16
just compared to non-white persons, but as a purer and whiter ‘people’
than any other (Hübinette & Lundström, 2014; Kjellman, 2013, 2014).
These ideas, however readily discarded as a dark and distant chapter in
Swedish history, were perpetuated through scientific research carried out
by the Swedish State Institute for Race Biology from the 1920’s up until a
not-so-distant past (Björkman & Widmalm, 2010; Ericsson, 2021).
In Sweden, like elsewhere, racism and whiteness tend to be downplayed
as a thing of the past, while in reality, it has instead only taken new and
often ‘colour-blind’ forms (Bonilla-Silva, 2015; Hughey & Daniels, 2013),
not the least online (Brooks, 2020; Kettrey & Laster, 2014). Consequently,
while the elements or actual content of far-right discourse remains the
same whether it is expressed online or offline, the internet means different
conditions for how it can be communicated and spread.
Efforts to normalise far-right discourse online
The development of the internet and especially the emergence of the
interactive, multi-modal, user-centred, and platform-based ‘Web 2.0’
(O’Reilly, 2009), has meant changing conditions for “who gets to say
something to how many” (Jensen & Helles, 2011, p. 520). Online, people
have gone from being passive audiences of one-to-many mass media
broadcasts, to being producers and consumers of content of their own
choosing (Bennett, 2012; Castells, 2009; Jenkins, 2006). And where
discursive control over public debate before the emergence of the internet
largely resided with electoral politics and legacy news media (van Dijk,
1993a), the internet has expanded the possibilities for ordinary people to
disseminate political discourse inexpensively, easily, quickly, and widely.
In turn, this expansion of opportunity to be heard can have considerable
societal consequences. This is because language use, while being socially
conditioned, at the same time contributes to maintaining or challenging
existing hegemony. As such, increasing opportunities for people
(including for far-right sympathisers) to be heard can reinforce or
challenge power relations, impact our knowledge, values, and identities,
and by extension even affect societal structures (see also Fairclough,
2003, 2010; Wodak, 2015). So, the internet provides space for individual
users to create their own counter-discourse and challenge established
political orders, ideas, and meaning making (Cammaerts, 2009; Jackson
& Foucault Welles, 2016). Thereby, the internet grants (also) the far right
possibilities to provide alternative narratives to mainstream political
discourse.
17
Among the most important efforts by the far right online as well as offline
are those which help normalise or ‘mainstream’ far-right ideology (e.g.,
Feischmidt & Hervik, 2015; Karl, 2019; Mondon & Winter, 2020; A.
Winter, 2019). These more or less active efforts include deploying
discourse otherwise commonly associated with mainstream politics
(Cammaerts, 2018; Feischmidt & Hervik, 2015), for instance attempting
to leverage progressive politics’ protection of freedom of speech to justify
and disseminate far-right discourse (Castelli Gattinara, 2017; Gerstenfeld
et al., 2003; Topinka, 2018). But perhaps most importantly, the far right
in Europe and elsewhere has also in recent years successfully rebranded
their ideas into more palatable, less overtly ‘far-rightist’ packaging and
with this, incorporated strategies to conceal far-right discourse in
different ways, making it more easily able to ‘hide in plain sight(Kallis,
2013; May & Feldman, 2019; Topinka, 2018). While such strategies were
themselves not invented online, the internet has meant new possibilities
for how they can take shape and be disseminated (Daniels, 2018).
The most obvious efforts to normalise far-right discourse in the digital
mainstream have been those of the ‘alt-right’ (short for ‘alternative right’),
primarily in the United States. This predominantly digital, mostly
anonymous, seemingly leaderless, and loosely organised far-right
movement has gained prominence through its affiliates’ tech savviness
and knowledge of internet culture. By using humour and trolling as
strategies to spread their political agenda, the alt-right appears from the
outside as a youthful, carefree, mischievous, and in some sense ‘fun’
movement, not to be taken too seriously. Consequently, they might
sometimes be thought to not pose any real danger to mainstream society
(Hartzell, 2018; Hawley, 2017; Tuters & Hagen, 2020; A. Winter, 2019).
Humour in the form of sarcasm, satire, and irony is used by the far right
online with the consequence that it to some extent can disguise far-right
discourse as unproblematic and fun (May & Feldman, 2019; A. Winter,
2019). Previous research shows that over time, such humorous
expressions can work to desensitise those who are exposed (Marwick &
Lewis, 2017; Munn, 2019), while creating distance and plausible
deniability for users spreading such content as they can always claim they
were only joking (Tuters, 2019). Also outside the American alt-right
culture, in European digital settings, humour has been identified as an
increasingly important strategy for the far right (Hakoköngäs et al., 2020;
Hervik, 2019; Malmqvist, 2015; Nilsson, 2021).
Another way in which previous research has shown efforts to conceal far-
right discourse in the digital mainstream is through the use of code words
18
(Davidson et al., 2017; Magu et al., 2017; A. Schmidt & Wiegand, 2017; see
also Lööw, 2016). Previous research for instance shows that users might
substitute letters or words to decrease the searchability of far-right
content (Gröndahl et al., 2018; Magu et al., 2017) but also that they
employ ‘dog whistling’ practices so as their far-right claims appear
inaudible to the general public, while simultaneously being able to
resonate with far-right sympathisers (Bhat & Klein, 2020; Haney López,
2014; Kien, 2019).
Sometimes, far-right efforts to deceive have been found to stretch as far
as forging antagonistic content on social media in order to spark conflict
(Farkas et al., 2018a, 2018b). Users and groups have also been shown to
use information or news from legitimate sources in a selective, piecemeal
fashion to substantiate skewed, far-right claims (Haller & Holt, 2019).
It should be noted that these are not efforts taking place solely in the
digital mainstream. In line with the far right’s general clean-up efforts, the
repackaging of far-right discourse online has also been seen in attempts
by far-right sites to conceal and convey less blatant versions of far-right
discourse (De Koster & Houtman, 2008; Gerstenfeld et al., 2003;
Meddaugh & Kay, 2009). More specifically, previous research has
illustrated how far-right sites online ‘cloak’ their true political intentions,
disguise authorships and use seemingly non-radical, deceiving domain
names in order for visitors to mistake these websites as legitimate and
non-extremist (Daniels, 2009). Relatedly, others have shown how far-
right sites borrow aesthetics and content from credible sources, like legacy
news media, to appear more trustworthy (A. Klein, 2012; see also
Gerstenfeld et al., 2003). This can make far-right content appear palatable
and therein might enable far-right sites to attract (unsuspecting) visitors.
In the digital mainstream, subtlety in the expression of far-right
discourse, for individual users, can be a means of avoiding human and
automatic content moderation and the potential repercussions that such
detection might bring (Gröndahl et al., 2018; Hughey & Daniels, 2013;
Magu et al., 2017). In turn, the covertness of such far-right discourse
makes it hard to detect and police (Meddaugh & Kay, 2009), often with
the consequence that it is able to persist in public, online spaces. Here, it
is easily accessible to a range of different audiences to whom covert far-
right discourse might appear as mere ‘jokes’ or ‘common sense’, whilst it
in reality could be laden with hateful meaning (Bhat & Klein, 2020;
Hawley, 2017; May & Feldman, 2019).
19
Because understanding and dealing with the subtle practices of far-right
language use online requires in-depth knowledge (Hughey & Daniels,
2013) it seem as though it has too often been considered difficult to study.
Nevertheless, the difficulty in detecting and grasping such covert efforts
make them all the more important for the far right’s advancement in
mainstream settings. Therein, they are crucial to understand also for
efforts to halt its expansion (Albrecht et al., 2019; May & Feldman, 2019;
Schwarzenegger & Wagner, 2018).
Digital platforms
To contextualise the discussions in the previous sections about how the
far right is able to leverage the internet to their advantage, this section
provides some explicit discussion regarding digital platforms. First, it
discusses how platforms work, that is, the general logics of platforms, and
second how these logics can enable far-right efforts.
The logics of platforms
The internet has indeed expanded the possibilities for regular people to
create and spread political discourse, bypassing traditional editorial
standards as well as gatekeepers (Thompson, 2020). This, however, does
not mean that the internet is without mediation. Despite social media
companies’ best efforts to promote themselves as neutral ‘platforms’ for
users to engage on (Gillespie, 2010, 2018), digital settings are never value-
free
1
. On the contrary, (the people in charge of) digital platforms are
ultimately responsible for shaping how users can act and interact, and
how content can be created and spread (Gillespie, 2015; Klinger &
Svensson, 2015; Matamoros-Fernández, 2017).
Famously, van Dijck and Poell (2013, p. 3) highlight four primary socio-
technical featuresprogrammability, connectivity, popularity, and
dataficationwhich are central to understanding discourse, platforms,
and users in many online settings, and on social media especially. By
programmability, they refer to how platforms and user practices are
mutually constitutive. That is, different technological features will limit
some user practices while enabling others. At the same time as users’
1
This is not to say that technology in itself is either ‘good’ or ‘bad’ (boyd, 2008; Kranzberg, 1995), nor
that technologically deterministic explanations suffice to comprehend user practices online. As
argued by Ito (2008, p. 4) technology is not “an external force that impacts society and culture. Rather,
technologies are embodiments of social and cultural structures that in turn get taken up in new ways
by existing social groups and cultural categories”.
20
interactions with, as well as understandings and expectations of,
technological features contribute to shaping technology itself (Bucher &
Helmond, 2018; Nagy & Neff, 2015; see also boyd, 2008). The connective
feature of platforms is about how, instead of connecting people who are
geographically proximate, online settings enable customised connections
between individuals or groups based on other criteria. This creates
networks of users which are sometimes global in character and stretching
beyond nation states (Castells, 2012) or which are structured based on
personal relevance (J.-H. Schmidt, 2014) or interest in particular issues
(Bruns & Burgess, 2015; Rambukkana, 2015). However, the concept also
refers to forms of connectivity which have at times been criticised among
scholars. For instance, the one between users and advertisers (Fuchs,
2010; Fuchs & Sevignani, 2013), and algorithmically, between users and
various recommendation systems (Munn, 2020).
Platforms also have various mechanisms for quantifying popularity and
promoting certain content and users over others. These can be based in
for instance ‘follower’, ‘like’, or ‘share’ functionsso-called ‘vanity
metrics’ because of how they are often used to show off social status in
online settings (Rogers, 2018). Depending on how a platform’s ranking
system works, different content, groups, and users will be favoured. In
turn, some will receive a disproportionate amount of exposure.
Finally, through digital technology, all these elements are based in
‘datafication’ (van Dijck & Poell, 2013). First conceptualised in 2013 by
Mayer-Schönberger and Cukier, to datafy something is “to put it in a
quantified format so it can be tabulated and analyzed” (Mayer-
Schönberger & Cukier, 2013, p. 78). In many online settings, technological
features and users’ deployment of these, their popularity, and
connectivity, are all datafiedaccumulating into what in the last decade
has come to be known as ‘big data’ (Boellstorff, 2013; boyd & Crawford,
2012).
While datafication provides unique opportunities for researchers to gain
insight into settings that are becoming increasingly central to everyday
life, politics, and news media, the opportunities of large amounts of
aggregated data have also been leveraged for repressive purposes like
surveillance (Brayne, 2017) and political persuasion campaigns (Tufekci,
2014). Some, in fact, argue that the ways in which many platforms
function are particularly efficient in enabling the creation, spread and
persistence of harmful content, Previous research shows that the far right
has been exceedingly skilful in leveraging these opportunities (Colley &
Moore, 2020; Gaudette et al., 2020; Massanari, 2017; Munn, 2019, 2020).
21
Far-right enablement through socio-technical features
While it might seem strange that many digital platforms would enable the
far right through their designs, it should be noted that issues like sexism
and racism have since long been built into digital technology (Eglash,
2007; Noble, 2018). Moreover, it is important to recognise that most
digital platforms are not freely provided communication tools for the
general public, but for-profit organisations that base their businesses
around ad revenue (Fuchs & Sevignani, 2013; Gillespie, 2010). And
oftentimes, a desire for profit trumps the social responsibilities these
organisations could be expected to meet. This is problematic as it has been
found that polarising and antagonistic content stimulates the greatest
engagement, and thus, provides the most ad revenue for platforms (Ben-
David & Matamoros Fernández, 2016; Munn, 2020).
Relatedly, forms of algorithmic sorting and recommendations, while often
opaque to users, shapes much of the online experience. For instance, while
search results on search engines do not appear in an order which displays
the most ‘important’ content first, users can sometimes perceive it as
though it is. Therein, search engines can provide credibility to far-right
content, allowing users to search for and find far-right sites, like-minded
others and organisations, and access skewed information that confirms
their far-right beliefs (Daniels, 2009, 2018; A. Klein, 2012). On social
media sites like YouTube, Facebook, and Reddit, recommendation
systems and content feeds have been found to support and emphasise
incendiary content (Gaudette et al., 2020; Massanari, 2017; Munn, 2020),
and on YouTube especially, recommendations have been found to steer
users towards increasingly radical far-right content (Munn, 2019; Ribeiro
et al., 2020; see Ledwich & Zaitsev, 2020 for opposing view).
Online sites tend to state in their rules of conduct that they forbid hate
speech, but many have often historically taken a hands-off approach in
actually dealing with these issues. This is usually a means to not pick
sidesas has often been the case with conventional social media
platforms (Gillespie, 2018), or as a freedom of speech actas it has often
been argued by those running eclectic sites and forums (Colley & Moore,
2020; Costello & Hawdon, 2018). Often, digital platforms have
transferred the responsibility to report rule-violating content onto users
themselves (Ben-David & Matamoros Fernández, 2016; Oboler, 2008),
but after ample criticism in recent years, many online platforms have
increased their efforts both when it comes to automated and human
content moderation (Munn, 2020). Sometimes, resulting in the banning
22
of far-right users and content (Gaudette et al., 2020; Lima et al., 2018;
Rogers, 2020; Urman & Katz, 2020). While some research has shown that
bans lead to decreases in concentrations of far-right content in the current
setting (Chandrasekharan et al., 2017), others argue that banning merely
relegates it for other settings to deal with (Hughey & Daniels, 2013;
Urman & Katz, 2020).
Moreover, while most platforms forbid the use of hate speech as a self-
protective measure, there tends to be little regulation or consequence for
those who just pass it on (Ben-David & Matamoros Fernández, 2016). The
fact that many reaction, sharing, and content organising features online
are so easily accessible and quick to use means far-right discourse can be
easily shared with little effort or reflection (Munn, 2020). In turn,
platforms’ opportunities to allow users to share and link content, both
within and across sites could facilitate a ‘network of hate’like the one in
the far-right blogosphere (Burris et al., 2000; Caiani & Wagemann,
2009)of easily accessible channels, sites, content, and users (Ben-David
& Matamoros Fernández, 2016; Ekman, 2014; A. Klein, 2012; Sakki &
Pettersson, 2016). It also means that the categorising function that
hashtags can have on some sites helps users organise and spotlight far-
right discussions which would otherwise go unnoticed (Burgess &
Matamoros-Fernández, 2016; Crosset et al., 2019; A. Wilson, 2018).
Lastly, possibilities for users to ‘follow’ or ‘subscribe’ to certain users or
channels is yet another way in which the amplification and circulation of
far-right discourse can take shape with few repercussions online. Notably,
the emergence of far-right influencers on for instance YouTube could not
have been made possible without the continuous exposure and status
granted by the ‘follower’ or ‘subscription’ logic and vanity metrics (e.g.
Lewis, 2018; Maly, 2020; van der Vegt et al., 2020).
Finally, the enablement of the far right has often also been connected to
users’ possibilities for anonymity and pseudonymity online. Anonymity
and pseudonymity can provide a feeling for individuals of not being
constrained to certain frames of being (boyd, 2015; Hogan, 2013), but
have also proven important for enabling the often-times socially
stigmatised activities of the far right (Awan, 2014; Costello & Hawdon,
2018; Crosset et al., 2019; Evolvi, 2017; Urman & Katz, 2020; A. Winter,
2019). For instance, by enabling users to act and talk more radically than
they might, had they been held personally accountable (Koehler, 2014; see
also Suler, 2004). While previous research has shown how anonymity can
blur the distinction between individuals in an online setting, making their
far-right claims appear more unified (de Zeeuw & Tuters, 2020),
23
pseudonymity has been found to enable users’ creation of distinct far-
right personas (Crosset et al., 2019).
Different digital and social settings might impact which technological
features are important for the proliferation of far-right discourse and how
far-right enablement takes shape. Explorations of different user
constellations and cultures, on different platforms (including those
beyond YouTube which remain underexplored), with different types and
scopes of metrics, degrees of anonymity and pseudonymity, and over
different time periods, will contribute to a deepened understanding of far-
right efforts online. Especially, there is need for research to clearly identify
which socio-technical features are distinctively biased in favour of the far
right online.
Influence online
This section discusses how influence can be understood online, in the
context of the far right specifically. First, it seeks to identify what online
influence is by exploring some of its main logics and characteristics.
Thereafter, it explores how far-right influence in particular, has been
understood and studied in digital contexts, and how they can contribute
to far-right community-building.
The characteristics of online influence
The internet has historically often been thought of as a leaderless utopia
of equal opportunity for people of different ages, genders, sexualities, and
ethnicities (Curran, 2012b; Nakamura, 2002). Not the least, these ideas
have been perpetuated through the widespread use of theoretical concepts
like ‘crowdsourcing’ (Howe, 2006), ‘participatory culture’ (Jenkins,
2006), ‘collective intelligence’ (Levy, 1997), ‘peer production’ (Benkler,
2006), and ‘the wisdom of crowds’ (Surowiecki, 2005), which all express
the collective and non-hierarchical ideals of online participation. And
while it has even been argued by some that online settings may function
without leadership, through horizontally organised networks of users
(Bennett & Segerberg, 2013; Castells, 2012), others argue that some users
will come to be more influential than others, even in seemingly democratic
settings online (Bakshy et al., 2011; Jacobson et al., 2019; ONeil, 2009).
The way that this influence functions online, however, has often been
conceptualised as different from the control over public discourse
exercised by traditional (offline) ‘elites’ (Mills, 1956; Marty, 1974; Scott,
24
2008) or ‘agenda setters’ (van Dijk, 1993b; McCombs & Shaw, 1972)
through their official positions of status and power in for instance
parliamentary politics or legacy news media. In fact, many argue that
offline influence does not directly translate into influence online at all
because it is not necessarily those who have the most influence offline who
become influential online, and vice versa (Huang & Yeo, 2018; S. Winter
et al., 2020). However, this needs some unpacking because what actually
constitutes influence online varies in the extant literature with definitions
ranging from users own self-assessments (e.g. Hwang, 2015; Park, 2013;
Weeks et al., 2017; S. Winter et al., 2020), an ability to elicit quantifiable
reactions from others (Berger & Strathearn, 2013; Bigonha et al., 2012;
Chorley et al., 2015; Romero et al., 2011), to positions of centrality in
digital social networks (Bakardjieva et al., 2018; Della Ratta & Valeriani,
2012; Diani, 2003; Weimann, 1994). In sum, and as noted by Räbiger and
Spiliopoulou (2015), there is no agreed-upon definition of ‘online
influence’.
Still, there are some general characteristics that many agree are important
to distinguish influential users from others online. Participation in digital
settings tends to follow certain logics. Generally, it follows a power-law
distribution wherein some users or sites come to be more highly
connected than others and these then tend to increase their
connectedness more easily (Barabási, 2003; Barabási & Albert, 1999;
Chau & Xu, 2007; Zhou et al., 2007; see also Merton, 1968). There is
relatedly also often a skewedness when it comes to content production in
digital settings where a very small share of users produces the majority of
content while most users participate very little or not at all (Nielsen,
2006). Moreover, the technological features of digital platforms, as
previously noted, will favour some content over others and therein grant
some more exposure and opportunities to influence (van Dijck & Poell,
2013).
Influential users online therein tend to be understood as those who have
grown increasingly important in the eyes of other users due to the content
they create and share. A reliance of continuous support by other users to
stay relevant means that influence online has been conceptualised as
‘performative’ (Bakardjieva et al., 2018) and ‘crowdsourced’ (Papacharissi
& Oliveira, 2012). Correspondingly, this type of influence has an informal
character similar to the pre-digital, long-standing notion of ‘opinion
leaders’ who function as mediators of information from legacy media to
less active sections of the public through direct personal relationships
(Lazarsfeld et al., 1944).
25
In light of these discussions, influence is defined in this thesis in terms of
two things. First, it is argued here that those whose content reaches many,
are in a unique position to exude influence on a large scale. Thus,
influence requires some measure of reach in terms of content, for instance
through vanity measures or network centrality. Second, influence is
defined here as a form of power connected to an ability to shape the
opinions of others with one’s ideas (van Dijk, 2008).
Far-right micro-celebrities and covert far-right influence
Leadership has long been important for the expansion of the far right. This
concerns the form of intellectual leadership provided by key philosophical
thinkersthose who have developed the literary works on which the more
sophisticated parts of the far right relies (Sedgwick, 2019). But also, the
charismatic, rhetorically skilled, and approachable leaders who have
made far-right ideology seem relatable and propelled far-right issues into
the mainstream political debate (Mudde, 2019).
Despite this, research into the far right online has often tended to overlook
the role of influential actors and has instead focused on collectives of
users. As a consequence, far-right discourse online has often been
described as though it is shaped by a leaderless mass of far-right
sympathisers who are equally responsible for its dissemination. While it
might seem obvious, it is important to note that this is not the case, nor
are platforms themselves solely responsible for the creation and
circulation of hate, independently of its users.
On the contrary, some far-right users have learnt to game digital platforms
and are skilled in producing content which gets plenty of engagement
(Laaksonen et al., 2020; Munn, 2020) and visibility with the right
audiences (A. Wilson, 2018). Some users are able to leverage content
moderation systems so as to not get banned or break community
regulations (Ben-David & Matamoros Fernández, 2016; Gaudette et al.,
2020; see also Meddaugh & Kay, 2009) and organise in constellations and
groups through which they can create and circulate far-right discourse
(Merrill, 2020; Topinka, 2018). Despite this, it was not until recently that
research started recognising the role of such influential users in the
context of the far right online. Most notably, research efforts have begun
exploring the emergence of so-called ‘micro-celebrities’ or ‘influencers’
(Abidin, 2018; Senft, 2013) in the context of the far right, particularly on
YouTube.
26
Research exploring these influencers has shown how they leverage the
popularity feature (van Dijck & Poell, 2013) of digital platforms to increase
their reputation and exposure. This has been shown to involve a range of
strategies concerning ‘self-branding’, in terms of conveying a hip online
persona (Lewis, 2018), and by staging content for it to appear relatable,
intimate, and personal. These strategies, while providing the influencers
with an appearance of authenticity and transparency (Lewis, 2018, 2020),
are nevertheless based in political purposes (Maly, 2020). Far-right
influencers often dedicate their content to political commentary and try
to build trust with their fans byunlike traditional media outletsbeing
approachable and endorsing feedback and conversation (Laaksonen et al.,
2020; Lewis, 2018, 2020). Beside efforts of self-branding, and sustaining
the interest and involvement of fans, far-right influencers’ strategies also
involve leveraging the technological features of platforms themselves to
increase their reach (Lewis, 2018; Maly, 2020).
Far-right influencers are not a homogenous group, instead they have been
found to promote a range of different levels of far-right extremism,
ranging from those promoting white Nationalist ideas to those promoting
conservative political values. However, on YouTube and elsewhere, these
boundaries become blurred as far-right influencers are often highly inter-
connected, forming a network of ‘alternative’ political micro-celebrities;
the so-called ‘Alternative Influence Network’ (Lewis, 2018; see also
Munger & Phillips, 2020). This network positions far-right influencers as
‘bridges’ between varying levels of extremist far-right discourse,
platforms, and users, and ‘the mainstream’, making it easy for audiences
to find and explore far-right content (Albrecht et al., 2019; Laaksonen et
al., 2020; A. Winter, 2019; see also Barkun, 2017). It also lets far-right
influencers leverage each other’s fame and social networks to increase
their own, and to cross-promote their content through collaborations and
guest appearances (Laaksonen et al., 2020; Maly, 2020).
In the context of language use which can sometimes be seen as socially
stigmatised and potentially even illegallike that of far-right discourse
it is however not necessarily the goal of all users to become famous (see
also Crosset et al., 2019; Futrell et al., 2018). Consequently, there might
be some ‘ordinary’ users who are deeply political in their use of platforms
and in their interactions with others, but who operate under the radar due
to the potential consequences their exposed far-right sympathies could
have. These users, nevertheless, could be important in shaping far-right
discourse and community culture in any given online setting. Potentially,
some users are able to spread far-right discourse widely but without any
public accountability, and there are some initial findings showing that the
27
universal logics of online participation (unsurprisingly) also apply to the
far right, beyond celebrity influencers.
In the context of the far right online, research has shown how some users
are exceedingly active (Kleinberg et al., 2020; Scrivens, 2017), some
content and users are garner disproportionate amounts of attention
(Berger & Strathearn, 2013; Zhou et al., 2019), and that some far-right
sites are especially well-connected through links (Chau & Xu, 2007). This
means that some are more deeply responsible for the proliferation of far-
right discourse online and therein are also particularly important in
shaping the far-right community. However, little is still known about how
these users operate. The few studies exploring such users suggest that they
might be senior and more extremist in their far-right language use (Berger
& Strathearn, 2013; Kleinberg et al., 2020; Scrivens, 2017). Still, most of
this research has explored extremist sites exclusively, and far-right
influence in mainstream settings might function differently. Not the least
due to the different audiences and social rules at play there, as previously
mentioned.
Gaining and sustaining influence in any online setting requires knowledge
of that specific social context, its users, and its culture (Bernstein et al.,
2011; Maly, 2020; Nissenbaum & Shifman, 2017). Previous research on
far-right interaction online for instance points towards the importance of
using the appropriate jargon and fluency in style of communicating
(Bernstein et al., 2011; A. Törnberg, 2021; van der Vegt et al., 2020). Far-
right communities online have been seen to impact individual users’ ways
of communicating (P. Törnberg & Törnberg, 2021) but also more
profoundly affect their norms and values (Bowman-Grieve, 2009). Thus,
research which focuses on the actions of influential users could expand
understandings of how far-right communities’ function and, importantly,
how they may contribute both to the normalisation of far-right discourse
and to the radicalisation of individual users (Colley & Moore, 2020).
Modelling far-right discourse, platforms, and
influential users
Research exploring the far right online has often focused either on far-
right discourse or platforms or influential users. However, such efforts can
only reveal fragments of far-right endeavours online, instead, as noted by
Daniels (2018, p. 62), there is need to “think about several things at once”
to understand the far right online. So, in order to better grasp the constant
28
advancement and technological adaptiveness by the far right online, and
the way that covert far-right discourse, digital settings, and influential
users enable it, this section seeks to highlight the complex
interconnectedness of these three dimensions. Following the theoretical
perspectives and previous research presented thus far, the roles of
technology and of sociality are seen as deeply intertwined and co-
constitutive, both generally in user-generated content online but also in
the context of the far right specifically.
So, while language use shapes the social context in which it is expressed,
it is seen here as also being shaped by available opportunities and
structures for its expression (Fairclough, 2003, 2010). This means that
language is deployed in relation to the social conventions, traditions, and
norms in a given context. At the same time, language use can challenge or
uphold these social conventions, traditions, and norms as well as people’s
relations to each other in a given social context (Fairclough, 2001).
Correspondingly, in the context of the far right online specifically, this
means that users have to relate to pre-existing discursive elements and
strategies. Simultaneously, through their own use of far-right discourse,
users have opportunities to reinforce or alter these conventions and to
impact the social context in which it takes place.
Furthermore, while platforms set the rules for what can be expressed, how
and by whom, as well as possibilities for social interaction, it is not
adequate to resort to technologically deterministic explanations to
understand the role of platforms in enabling the proliferation of discourse.
Users themselves also have agency to steer how information flows and
choose who they interact with, why, and in parts also how (Bucher &
Helmond, 2018; Klinger & Svensson, 2015; van Dijck & Poell, 2013);
actions which might not necessarily be in accordance with how the
platform’s designers intended. This is also true for platforms’ roles in the
propagation of far-right discourse specifically (Munn, 2020, 2021). In the
end, the spread of far-right discourse is most often the result of user-
generated content, and while far-right users must conform to the
technological prerequisites of online settings, they might also leverage
them for their own purposes, and to strengthen far-right community
culture and shape its customs.
This theoretical perspective which interconnects far-right discourse,
platforms, and users is synthesised in Figure 1. The conventions of far-
right discourse seen as ‘discursive conditions’ on the top-left, set certain
limits for how influential users can express themselves while still
appealing and making sense to other far-right sympathisers, that is, to the
29
given far-right community. As such, this dimension is more about
discursive strategy than discursive elements since these as previously
discussed have proven consistent over time and context. At the same time,
influential users, as seen in the bottom-centre are especially skilled in
their use of far-right discourse and through theirdiscursive competence
they are able to convincingly display their far-right ideas and make others
support them. Platforms, on the bottom-right create technological
conditions’ which provide certain possibilities to spread far-right
discourse. Finally, as seen on the top-right, influential users possess a
particular technological competence’ which allows them to leverage
platforms for their own purposes.
Figure 1. Interconnections of far-right discourse, platforms, and influential users
While there is no direct link between platforms and discourse in Figure 1
the figure should be read as suggesting that these dimensions are
indirectly interconnected through the mediation of influential users. More
concretely, far-right discourse impacts digital platforms through the
discursive actions of influential users while the technological features of
digital platforms simultaneously affect far-right discourse via the
considerations taken by influential users. Furthermore, it can be noted
that while this model focuses solely on influential users as the actors and
far-right discourse as a particular type of language use, variations to the
model could potentially also illustrate the interconnectedness between
other user types, discourses, and digital settings.
30
Case, data, methods, and ethical
considerations
Sweden as an empirical case
Sweden presents an interesting case through which to study the far right.
While previously considered a ‘political exception’ in 21st century
Europefor a long time having had no electorally successful far-right
party (Dahlström & Esaiasson, 2013; Elgenius & Rydgren, 2017)Sweden
has in recent years experienced a spike in far-right political sentiment.
Consequently, Sweden’s third largest party in parliament today is far-
rightist. Moreover, as previously touched upon, Sweden has a problematic
background in regard to whiteness and eugenics (Berggren, 2014;
Björkman & Widmalm, 2010; Hübinette & Lundström, 2014). And, while
perhaps less well-known, also a long history of colonialism (Naum &
Nordin, 2013), anti-Semitism (Berggren, 2014), and fascist and national
socialist ideology (Lööw, 2004).
A brief overview of the contemporary Swedish far right
In 1924, the National Socialist Freedom League was formed and therein
became Sweden’s first Nazi party. However, this and other related
organisations and coalitions were ripe with internal conflict, and while
some were able to gain a bit of popularity in the period leading up to
WWII, their traction remained limited both during and after the war
(Lööw, 2004). Then, in 1956, the Nordic National Party was formed and
has since proven important for bridging those early 20th century Swedish
Nazi groupings with contemporary far-right movements (Lööw, 2004).
During subsequent decades, far-right groups and organisations would
come and go, and it was not until the 1980s and 90s that the Swedish far
right became progressively more organised and visible. During this period
and into the early 2000’s, a number of far-right groups and organisations
were formed (see Figure 2), and alongside these, a growing cultural scene
of white power music, publications, events, and street actions (Lööw,
2015; Wåg, 2010).
31
Figure 2. Timeline of the contemporary far right in Sweden
During this time, in 1991, a freshly formed far-right party called New
Democracy briefly gained electoral seats in the Swedish parliament
(Widfeldt, 2008), and extremist, neo-Nazi activist organisations like
White Arian Resistance and the Swedish Resistance Movement paved the
way for organised and radical far-right action (Lööw, 2015). It was also
during this period, in 1988, that the far-right party the Sweden Democrats
(SD), was formed as the result of a merger of two far-right organisations
(Rydgren, 2006).
While SD remained insignificant during the 1990s, the party slowly
started gaining traction in the 2000s (Widfeldt, 2008) and in the 2010
general election Sweden ceased being a political exception in Europe as
SD entered into parliament. Since then, the party has seen continuous
electoral success. Alongside this, and likely as a result of this success, there
has been an overall increased political focus on immigration-related
issues in Swedish electoral politics. Concurrently, after a peak in
immigration reception in 2015, restrictive immigration policies have also
followed (Schroeder, 2019). Finally, in Sweden’s last general election, held
in September of 2018, SD received nearly 18 percent of the votes and
became Sweden’s third largest political party in parliament; a position
they are most likely going to maintain in the 2022 election, according to
all major polls.
Rydgren and van der Meiden (2018) point to four primary reasons for the
eventual electoral success of SD. First, a decrease in class voting allowed
SD access to working class voters; second, an increased political salience
of immigration played in favour of SD’s already large focus on anti-
immigration issues; third, beliefs that mainstream parties differed little
from each other allowed SD to portray itself as a distinct alternative; and
2000
1990
2010
1980
Vitt Ariskt
Motstånd
1991
Sverigedemokraterna
1988
2020
Sverigepartiet
1988
Soldiers of Odin
2015
Fria Nationalister
2008
Nordiska Förbundet
2004
Nordiska
Nationalsocialister
2009
Nationaldemokratiska
partiet
1983
Nationell Ungdom
1997
Nationaldemokraterna
2001
Nationalsocialistisk
front
1994
Svenska
Motståndsrörelsen
1997
Metapedia
2006
Motpol
2006
Nordisk.nu
2007
Ny Demokrati
(enters into parliament)
1991
Info-14
2000 Nordfront
2009
Fria Tider
2010
Nya Tider
2017
Nyheter Idag
2014
Nordisk
Alternativhöger
2017
Realisten
2009
Nationell.nu
2003
Samhällsnytt
2017
Avpixlat
2011
Flashback
2000 Dispatch
International
2012
Samtiden
2014
Politiskt
Inkorrekt
2008 Petterssons
blogg
2010
Motgift
2018
A timeline of the contemporary
Swedish far-right
2010
Online
Offline
32
finally, like many other similar far-right parties and movements in Europe
and elsewhere, SD has actively tried to distance itself from its openly racist
and neo-Nazi origins in more recent efforts to clean up its image. The
appearance of a more acceptable and less overtly far-rightist party
program has helped attract voters also outside of a far-right fringe.
Beyond this, SD has also been exceptionally powerful on social media,
where they dominate in terms of reach and user engagement not only
compared to other Swedish political parties but also compared to other
Swedish social media presences more generally (Medieakademin, 2021).
With the emergence of the commercial internet in the late 1990s, far-right
activity would in Sweden like elsewhere come to take place online to an
increasing extent (Lööw, 2015). At this point in time, many far-right
organisations started websites. Simultaneously, various far-right chats,
blogs, and electronic newsletters started appearing on the Swedish web.
In total, around forty different dedicated far-right websites were active in
Sweden at the end of the last millennium (Taloyan, 1999). Oftentimes,
these were explicitly connected to different far-right organisations.
In the first decade of the 2000s, several significant digital efforts by the
Swedish far right were beginning to blur the boundaries between
organisational and individual expressions of far-right discourse in
Sweden. In particular the umbrella organisation Nordiska Förbundet,
founded in 2004, was responsible for launching several of these efforts in
the mid 2000s. Including the internationally successful Wikipedia-like,
far-right encyclopaedia Metapedia (Ahlin & Ranstrorp, 2020; Arnstad,
2015); the blog portal Motpol, which synthesised a number of far-right
blogs (Lundquist, 2010); and last but not least, the widely popular
discussion forum Nordisk.nu. The now inactive Nordisk.nu forum was
able to span over different parts of the Swedish far-right spectrum, while
actively distancing itself from overt far-right symbols in favour of a
youthful and seemingly non-extremist image (Ekman, 2013; Wåg, 2010;
see also Lagerlöf, 2011). Since it was able to engage individuals also
without previous affiliations to the Swedish far-right scene, it can be
thought of as among the first public and successfully ‘cloaked’ (see also
Daniels, 2009) far-right sites on the Swedish web.
In the last decade, a number of far-right news sites, like Fria Tider,
Nordfront, Nyheter Idag and Samhällsnytt (previously Avpixlat and
Politiskt Inkorrekt) have started popping up on the Swedish digital far-
right scene (Kaati et al., 2017). And with the emergence of social media in
the latter half of the 2000s, far-right organisations like Nationell
Ungdom, Nordisk Ungdom, and Soldiers of Odin also became active on
33
conventional social media platforms (Ekman, 2013, 2014, 2018),
Alongside these developments, far-right accounts, groups, and
discussions have continued to emerge outside and across organisational
borders on sites like Flashback (Blomberg & Stier, 2019; Malmqvist, 2015;
A. Törnberg & Törnberg, 2016a), Swedish speaking parts of Twitter
(Urniaz, 2016), and in Facebook groups (Ekman, 2019; Merrill, 2020;
Wahlström et al., 2020).
The Swedish far right has also played an important role in the formation
of the alt-right movement. Co-founder of the previously mentioned
Nordiska Förbundet and long-time key figure in the Swedish far right,
Daniel Fribergwho also co-founded one of the most important
publishing houses for far-right literature globally, Arktosis one of the
founding fathers of the alt-right. Along with for instance American neo-
Nazi profile Richard Spencer, he formed the Alt-Right Corporation and
the website AltRight.com in 2017, from which the alt-right movement
would come to take shape (Ahlin & Ranstrorp, 2020; Teitelbaum, 2019).
Finally, alongside these developments of national Swedish far-right
efforts, there are also external forces reframing the image of Sweden.
While Sweden has historically been perceived internationally in terms of
progressiveness, equality, modernity, and political neutrality (Andersson
& Hilson, 2009), more recent framings by the transnational far right
portray a Sweden in decline (Thorleifsson, 2019; Titley, 2019).
Situating ‘the Swedish case’ in the thesis
While Sweden as a country has often portrayed itself as politically neutral
and progressive, there is nevertheless an abundance of historical and
contemporary examples of public far-right efforts that range from neo-
Nazi organisations and eugenics research, to movements, organisations
and individuals seeking to limit the rights and freedoms of women, ethnic
minorities and LGBTQI+ persons. With this, the Swedish far-right case is
in many ways different from other countries. It is for instance distinct in
its history regarding its relationship to whiteness and the closely
intertwined concept of ‘Swedishness’; its colonialism of and eugenic
experiments on Sámi people; its national position of political neutrality;
and its relative progressiveness and equality (Andersson & Hilson, 2009;
Björkman & Widmalm, 2010; Hübinette & Lundström, 2011; Körber,
2019). While Swedish-speaking users and settings, and sometimes
Sweden-specific discussions and problems explored in this thesis will
inevitably provide insight into the particularities of Swedish society and
34
the Swedish far right, it is argued here that the knowledge it can provide
goes beyond Sweden alone.
Mainly, Sweden has followed the development of many Western countries
in terms of increasing influence and importance of the parliamentary far
right. Like elsewhere, this development cannot be explained by any great
changes in Swedes’ attitudes (Bohman, 2018; Bohman & Hjerm, 2016).
Moreover, Swedes are generally skilled English speakers, which allow
them to find information and partake in international far-right
conversations and contexts that might then influence their ways of
expressing far-right discourse also in exclusively Swedish-speaking
settings. Furthermore, the far right is deeply transnational in a range of
ways: in terms of networks and connections of far-right users and sites
(Atton, 2006; Burris et al., 2000; Caiani & Kröll, 2015; Chau & Xu, 2007);
when it comes to internet-specific, digitally born phenomena like the
creation and circulation of memes (Askanius, 2021; Kien, 2019; Wagner
& Schwarzenegger, 2020); in terms of more traditional cultural
expressions such as the white power music scene (Lööw, 2016); and when
it comes to the discursive elements themselvesthe so called ‘content’ of
far-right discourse. Finally, research has also identified overlaps in far-
right discourse transnationally, for instance in practices of ‘othering’
(Doerr, 2017; Sakki & Pettersson, 2016), and in circulation of far-right
tropes (Horsti, 2017; Titley, 2019).
Accordingly, this thesis uses Sweden as its empirical case to study the far
right, but it seeks to provide insight into the far right online on a more
general level
2
. The use of Sweden as a case is carried out in a universalising
manner, where Sweden is considered an example among others of far-
right internet use. This is in contrast to using Sweden as a case to bring
forth the particularities of the Swedish far right (cf. Tilly, 1984). The
decision is based in an assumption that the technological functions of
online platforms and principles of influence are not country-specific
entities. But mainly, this decision is based in the previous research
presented above, which has identified considerable overlaps within the far
right both internationally and transnationally.
2
This is not a claim for universal generalisability. These claims are limited to places in which the far
right operates in a similar manner as in Sweden, implying for instance that whiteness is the norm,
and that social media platforms are used in a comparable way and for similar purposes by the far
right.
35
Data and methodological approach
This thesis uses material from four different datasets, collected at
different times through different methods. The digital data used in this
thesis have been collected from APIs, manual Google searches, and
scraped from HTML code. The datasets range in size from two thousand
up to seven million texts of varying types and characteristics, with
sampling periods spanning from around three months up to 21 years.
These datasets have been analysed primarily using mixed-methods
approaches. Interpretative analysis, primarily in the form of critical
discourse analysis (CDA), has been combined in this thesis along with
various data processing techniques, descriptive statistics and
visualisations created through Python code. Moreover, interpretative
analysis has been used in combination with three main computational
data analysis methods, namely, topic modelling, sentiment analysis, and
network analysis. Table 1 provides an overview of the datasets and
methods used in the four papers.
In this spirit of combining CDA with quantitative approaches, the thesis
takes some inspiration from the field of corpus-assisted discourse studies
or CADS, in which large-scale, textual dataset are studied by combining
methods of corpus linguisticssuch as identifying word-frequencies,
concordances and collocationswith discourse analysis (Baker &
McEnery, 2005; Mulderrig, 2011; Partington et al., 2013). However, I
approach the datasets described above following Lindgren (2020), who
Table 1. Data collection and analysis methods
Paper one
Paper two
Paper three
Paper four
platform(s)
multiple
Facebook
Twitter
Flashback
sample criteria
all instances of
culture enrich*
all texts posted
snowball sampling
via 23 hashtags
the entire Politics
forum category
sampled texts
2,336 texts
entries
59,489 posts and
1,754,425
comments
74,336 tweets
7,488,468 posts
posted by/on
287 websites
73,928 users
6,809 users
113,550 users
sample
timespan
Nov 1999 to
May 2020
Feb to Sept 2017
Sept to Nov 2018
May 2000 to Dec
2019
computational
method
-
topic modelling
sentiment analysis
network analysis
interpretative
method
CDA
CDA
close readings
CDA
36
advocates the need for ‘methodological bricolage’ wherein the research
problem should be what guides the use of methods, and that these
methods should then be used with flexibility and creativity for them to suit
the research task at hand. As such, he argues that “[f]inding good
solutions rather than adhering to rules should be the end goal of any
analytical strategy” (Lindgren, 2020, p. 26). Accordingly, through a
combined approach that encompasses both computational and
interpretative elements, this thesis seeks to make a methodological
contribution with regard to how the far right can be better studied in
online settings, through innovative applications and combinations of
methods.
Methodological benefits, compromises, and potential pitfalls
The availability of massive volumes of user-generated content and
metadata online enable large-scale digital research. Moving beyond the
‘either or’ view of quantitative and qualitative approacheswhen
exploring the internet especiallycan enable insights into issues and
situations which would otherwise not be possible (Lindgren, 2020).
Quantitative or computational methods help harness the possibilities of
voluminous and rich datasets by providing overview of large-scale
patterns in ways that qualitative methods cannot. However, because of the
disorderliness and sheer volume of such datasets, there is a risk of
identifying patterns which are not really there (boyd & Crawford, 2012) or
assuming the data will “speak for themselves” (Lindgren, 2020, p. 13).
Qualitative or interpretative methods on the other hand can provide
detailed insight into the messy, informal, and subtle nature of online
communication. With the potential risk nevertheless, of cherry-picking
especially interesting pieces of data for analysis (see Baker & Levon, 2015).
By combining computational and interpretative approaches, the strengths
of each methodological type can be harnessed, while their respective
drawbacks can be moderated.
However, these opportunities also come with a number of challenges.
Notably, the creation and storage of user-generated data online takes
shape independently of research efforts and data are often generated
before the research even begins kerlund, 2021a). When collecting
‘naturally occurring’ user-generated datathat which was not created for
the specific research purposes of this thesisthere were no opportunities
to get clarifications, follow-up questions or explanations from users about
the feelings and thoughts behind their content creation or interaction with
others. While this has at times been frustrating, it has on the other hand
37
also been beneficial in the sense that I have not interfered or projected any
preconceptions or biases on the creation of the data. Relatedly, this thesis
has sometimes focused on extended time periods, going many years back.
This has required spending lots of time getting acquainted with the
settings and users who had been active there and interpreting the meaning
of discussions which had sometimes taken place many years ago.
Consequently, these approaches have been immensely time consuming.
Moreover, the structures and regulations of different platforms affect
what data can be collected and with what precision. While in an ideal
scenario, the research questions would fully guide what data material was
collected and how, this has rarely been the reality in the empirical work of
this thesis. Instead, as I have noted elsewhere, digital data collection
procedures often require a measure of flexibility and adaptability for
instance in terms of adjusting research questions, but also potentially to
forego parts of the intended materials or analyses kerlund, 2021a). For
instance, even though they are an important part of far-right
communication, images and memes have been entirely excluded from
analysis in this thesis for the simple reason that there is still a lack of
reliable methods for studying these on a large-scale and automated
manner.
Furthermore, the thesis has only used data which were openly accessible
that which has been posted in open groups and on sites which do not
require any special access beyond ordinary user accounts. And
importantly it has only accessed users and content which have not been
deleted. This can be thought of as a drawback as there are of course plenty
of other spaces online which do not conform to these criteria. On the other
hand, it serves the purposes of this thesis to explore the far right where it
takes shape in accessible spaces, among ordinary users.
Relatedly, while influence is thought of in this thesis as a form of power
connected to an ability to shape the opinions of others with one’s ideas
(van Dijk, 2008), the data collected has not allowed for a first-hand
exploration of people’s thoughts, such as if they were in fact actually
affected and influenced by the content posted by certain users. Instead,
influence has been operationalised in terms of reach and resonance
among other users in the ways identified as most reasonable for each of
the explored sites and social contexts. These definitions have also been
made in consideration of firstly, that (far-right) users’ ways of acting can
depend on a given social setting and thus what constitutes influence might
change between platforms; and secondly, that the technological functions
of different platforms provide varying possibilities to define influencean
38
issue which has also complicated any use of a universal measure of
influence.
Similarly, the technological features of different sites have also
complicated efforts to compare the role of different functions and
platforms in the proliferation of far-right discourse. In addition, the
methodological combinations used in this thesis are at times
unconventional and untested, especially in the context of these particular
settings and topics. As such, much time and energy has been spent on
exploration through trial-and-error, testing, evaluating, and reworking
data collection, data processing, and analysis procedures.
This has partly also been due to the fact that unstructured pieces of user-
generated content, metadata, and user information comes in varying
scope and quality, oftentimes containing anomalies and errors (Pink et al.,
2018). Data can be difficult to capture when sampling because user-
generated content can be unpredictable. Especially, as far-right discourse
can be subtle and intentionally covert, it easily gets overlooked by research
efforts (e.g., Ben-David & Matamoros Fernández, 2016; Colley & Moore,
2020; Gröndahl et al., 2018; Magu et al., 2017). This became especially
obvious where users discussed far-right issues ironically and for instance,
where the use of a method designed to determine emotions in texts coded
irony as positively charged content. Similar difficulties were encountered
in qualitative identifications of far-right code words and whether these
were used in literal or coded ways.
Another issue that I have found difficult to address is the notion of the
‘black box’ when working with automated data collection and data
analysis methods. That is, it can be difficult knowing what data is collected
and what is not, and how reliable, and reproducible computational results
might be. This has been a reoccurring issue and topic of discussion during
the work with this thesis, especially when I have encountered unexpected
or contradictory findings. For instance, when measuring the levels of
emotions between influential and non-influential users on Twitter there
was a clear discrepancy between my findings and what I was expecting
based on previous research. It was not until after a number of reiterations
and close-readings of the material that I could confidently argue that my
results actually did diverge from previous research. Similarly, when
exploring the most influential users on Flashback in terms of betweenness
centrality, I was very surprised to see that some of the most central users
were not in fact the most influential in terms of resonance among other
users. Accordingly, the analytical strategy had to be reworked.
39
To mitigate the limits of my knowledge in data science, machine learning,
programming, and computational methods, I have throughout this work
balanced these efforts with in-depth investigations, close-readings, and
trial-and-error approaches. In fact, the interpretations and qualitative
evaluations of the coherencies and reasonability of these approaches have
been fundamental for this work. It is important to recognise that human
judgement is at the core of data-driven research even though typical big
data methods and approaches are often thought of as especially objective
and accurate (boyd & Crawford, 2012). In the empirical work of this thesis,
human decision-making has been central to all stages of the research,
from deciding on which data to collect, the methods by which to do it, how
to analyse the material, and in choosing the best procedure for gaining the
richest quantitative as well as qualitative insights from the material.
The critical discursive perspective
The use of computational methods in this thesis provides valuable insights
and perspectives to the papers. Nevertheless, it is the critical discursive
perspective that is the primary focus throughout the thesis and also what
guides its ontological and epistemological starting points.
Central to understanding discourse is the idea of ideology, and in this
thesis, far-right ideology in particular. Ideologies are shared systems of
ideas or beliefs within social communities, which organise their social
identities as groups, their goals, norms, cultural values, and activities.
These socially shared belief systems are in turn perpetuated through
language use (van Dijk, 2006).
CDA posits that language use and society are mutually constitutive.
Accordingly, language use is socially conditioned but will at the same time
have social effects which contribute to maintaining or challenging existing
hegemony (Fairclough, 2003, 2010). These struggles over the power to
define ideological meaning are discursive, taking place through language
use (van Dijk, 2009b; Wodak, 2009). The perspective also recognises
discourse as historically situated. Therefore, to understand discourse, its
social context must be considered (Fairclough et al., 2011). Accordingly,
CDA should be concerned with scrutinising ideology, the construction and
reproduction of power relations through language use, and the relation
between discourse and wider social and political structures.
CDA is not however, applied in this thesis in a syllable-by-syllable
linguistics type manner, as it is used in CADS (Baker et al., 2008) and as
40
it was first developed in CDA (Blommaert, 2005), nor does the thesis
adhere strictly to one critical discourse theorist. With this said, however,
the thesis takes inspiration from the methodologies of three prominent
critical discourse analysists. Namely, Ruth Wodak, Norman Fairclough,
and Teun van Dijk, which all advocate their own specific approaches to
CDA. Wodak adheres to a critical theoretical methodology known as the
discourse-historical approach, or DHA (Reisigl & Wodak, 2009).
Fairclough takes a dialectical-relational approach to CDA based in critical
realism, wherein discourse is seen as dialectically related to other social
elements (Fairclough, 2008). Finally, van Dijk’s approach places
emphasis on socio-cognitive dimensionsthe mental as well as socially
shared representationsof discourse (van Dijk, 2009a). Despite these
differences in focus, they also share many of the same foundational views
on language, power, and ideology, and have been used in combination in
this thesis due to their particular strengths. More specifically, Wodak and
van Dijk have both spent great effort analysing far-right and racist
discourse, but their applications of CDA, however, provide limited tools
with which to actually conduct analysis. Fairclough on the other hand,
while oftentimes focusing his studies on neo-liberal discourse in news
media, has an approachable and easy-to-use analytical framework.
As such, and in line with the flexible methodological approach outlined
above, this thesis sees CDA as a toolkit for qualitative interpretation.
Wodak, Fairclough and van Dijk view CDA in a similar manner, noting
that “there is no set procedure for doing discourse analysis; people
approach it in different ways according to the specific nature of the
project, as well as their own views of discourse” (Fairclough, 1992, p. 225).
They argue against strictly adhering to one narrow school of thought or
theorist (van Dijk, 2001), instead stressing that CDA should be viewed as
a “problem-oriented interdisciplinary research movement” (Fairclough et
al., 2011, p. 357; see also van Dijk, 1993b; Wodak & Meyer, 2009).
Practically, in this thesis, the critical discursive perspective helps expose
the contestation of dominant, mainstream ideas, values, and ideology by
the far right in potentially subtle and covert attempts to redefine what is
politically as well as socially acceptable. It provides a perspective for
understanding the particular conditions of far-right discourse online, as
well as platforms’ powers to shape how and what can be communicated,
and the hierarchies within the far right which renders some users more
influential than others.
More than this, the critical discursive perspective also constitutes an
explicit political stance (Wodak & Meyer, 2001). CDA is an inherently
41
activist approach drawing from theoretical perspectives aiming to critique
and fundamentally change society (Lazar, 2005; Wodak & Meyer, 2009).
As such, the critical discursive perspective is, as noted by van Dijk (1993b,
p. 253) “unabashedly normative” this means, as van Dijk continues, that
“any critique by definition presupposes an applied ethics”. More
concretely, the far right, far-right ideology and discourse pose threats to
democratic society and have real-life consequences for immigrants,
women, and other minority groups. A critical discourse analysist cannot
take a neutral approach to these issues but must instead make their
perspectives and positions transparent (see also Wodak & Meyer, 2009).
While some discredit any adoptions of political positions in research, such
apparent ‘neutrality’ or ‘objectivity’ is in itself an explicit political stance
one that denies, ignores, or justifies issues like inequality, dominance,
discrimination, racism, and sexism (van Dijk, 1993b). It is the firm belief
held in this thesis that research cannot view the far right as a political
constellation like any other, and its interconnected ideology and discourse
as any example of political ideas and language use. Such an approach
ignores the discriminatory, exclusionary, and ultimately harmful views
and actions perpetuated by the far right. However, there is a limit to this
political stance that I am not willing to cross. Moving further along the
activist spectrum might encourage ‘doxing’ or exposing particular far-
right actors. This thesis seeks to address issues and ultimately explore
opportunities for societal change (see also van Dijk, 1993b), through
analytical insights based in theory, analysis, and against the background
of previous research, but not to engage directly in activist action. This is
motivated by the ethical concerns outlined in the following section.
Ethical considerations
The project adheres to the EU General Data Protection Regulation
(GDPR), and the complementary rules on data protection exemptions for
research in Sweden. As this research deals with political opinion and thus
handles what is considered ‘sensitive personal data’, the project within
which this thesis is written has undergone ethical vetting from the
Regional Ethics Review Board
3
. The processing of personal data has also
been reported to the Umeå University Data Protection Officer. Collected
data has been processed through the open-source coding platform
Jupyter and stored on the research group’s own encrypted server.
3
The process for ethical vetting has since changed and is now carried out by
Etikprövningsmyndigheten
42
With the large amounts of user-generated data available online come a
number of ethical considerations. User-generated content online is easily
traceable both manually through platform search functions, as well as
through various third-party software. Hence, there is need to exercise
caution in order to protect user privacy. While social media and other
easily accessible data online can to some extent be considered public, it
must be recognised that users might not be fully aware or comfortable
with their information being used outside of the specific social media
settings in which it was created (boyd & Crawford, 2012; boyd & Marwick,
2011). As such, care has been taken throughout this project not to disclose
users’ personal information. Most often, data have been presented at an
aggregated and abstracted level to protect anonymity. Statistics
concerning users’ accounts and platform usage patterns are presented
using centrality measures so that no individuals nor their social ties can
be identified. In all instances where users have been quoted, these have
been translated from Swedish and slightly altered so as to complicate
tracing them back to their sources (Markham, 2012).
While this can be considered problematic especially in linguistically
oriented discourse analysis, user privacy cannot be compromised in
favour of disclosing quotes verbatim. The choice to slightly alter quotes
has been important not the least because several of the articles have
explored smaller groups of users in depth. Even though these datasets
have sometimes contained public figures, which according to Williams,
Burnap, and Sloan (2017) require less consideration in terms of
anonymity, disclosing their identities could still have repercussions for the
individuals behind these user accounts. Especially, considering the
sometimes discriminatory or even illegal nature of language use explored
in these papers.
Still, the research carried out in this thesis has not meant recruiting
respondents as you would for survey or interview-based studies, and no
consent has been acquired from social media accounts explored in this
project. The large amounts of data collected is one reason for this.
However, it should also be noted here that as issues of researcher safety
are gaining increased attention in discussions on internet research ethics
(franzke et al., 2020; Massanari, 2018; Rambukkana, 2019), the potential
risks I could expose myself to as a researcher of these far-right settings,
discourses, and users have been an even greater driving force in deciding
not to acquire users consent. Furthermore, GDPR
4
states that informing
4
Article 14 of Regulation 2016/679 on the protection of natural persons with regard to the processing
of personal data and on the free movement of such data (EU General Data Protection Regulation),
2016 O.J. (L 119) (EU)
43
the data subject is not required when “the provision of such information
proves impossible or would involve a disproportionate effort”. When
consent is not feasible, GDPR
5
states that processing of personal data is
lawful as long as it is “carried out in the public interest”. This project
addresses several dimensions that are important for a socially sustainable
society, including political participation, democracy, and civil society, as
also acknowledged by the funding of the project by the Swedish Research
Council and the approval of the research approach by the Regional Ethics
Review Board.
It is also important to emphasise that users might have a variety of
motives for participation online. All users explored in the papers of this
thesis do not necessarily associate themselves with the far right. Some
perhaps participate unaware of a given far-right context, while some
might participate consciously as a form of protest or counteraction.
Furthermore, because a few users are much more active than others, not
all of its contributors are equally active in co-constructing far-right
discourses in the studied settings, and others still may represent ‘fake’
accounts.
While it could seem obvious that any data sample drawn from the internet
does not represent the general population (boyd & Crawford, 2012), this
should still be addressed explicitly. Even though most Swedes use the
internet daily (91%), what people do online vary greatly between different
age groups, income and education levels, and gender. This becomes
evident when looking at the specific platforms explored in this thesis. Only
74%, 32% and 24% of Swedish internet users use Facebook, Flashback and
Twitter, respectively. And the differences in usages between men and
women are ten percentage points on all three platforms. Students use all
three platforms more than those who work or who are retired, and all
three platforms are used more by people who live in cities than by those
who live in the countryside (The Swedish Internet Foundation, 2019). As
such, it becomes evident that what is explored in this thesis is not the
opinions, ideas, networks, or relationships of a representative sample of
the Swedish population. Relatedly, while far-right sympathisers are
generally avid internet users, it should also be noted here that the Swedish
far right online explored in this thesis might not be representative of the
far-right view held by Swedish people in general, or by all far-right
internet users for that matter.
5
Recital 35 of Regulation 2016/679 on the protection of natural persons with regard to the processing
of personal data and on the free movement of such data (EU General Data Protection Regulation),
2016 O.J. (L 119) (EU)
44
Finally, the point of this thesis is not to give attention to particular
accounts, groups, or ideas as a means to promote the far right. However,
it must still be acknowledged that this research might unintentionally
contribute to normalising and spreading far-right discourse by providing
knowledge about what behaviours make users influential, and how certain
ways of framing far-right discourse make it harder to detect and police. At
the same time, making these issues explicit could potentially also make
far-right efforts online easier to counteract.
45
Paper summaries
This section provides an overview of the main results from the empirical
work carried out in this thesis. As previously noted, these papers are the
result of four case studies from as many datasets. Through different
methodological approaches and focuses, these take on the issue of far-
right discourse, platforms, and influential users. The papers are first
summarised separately below. Thereafter follows an empirical discussion
which seeks to respond to the three research questions posed in the
introduction of the thesis.
Paper one
Dog whistling far-right code words: The case of ‘culture
enricher' on the Swedish web.
Published in Information, Communication & Society, 2021.
The first paper kerlund, 2021b) explored how far-right, coded discourse
gets mainstreamed online, both over time and across online platforms
using the once Neo-Nazi expression ‘culture enricher’ (kulturberikare) as
a case study. Data were collected through manual Google searches of the
Swedish term for culture enrich* (kulturberik*). Search results were
localised, manually read, dated, and downloaded, along with associated
URL’s from whenever they were first recorded by Google until May 2,
2020. This resulted in a sample of 2,336 instances of text, posted between
November 1999 and May 2020, on 287 different websites. A detailed CDA
was carried out on the entire dataset to explore how the culture enricher
expression was articulated and where, and if, how and to what extent the
expression was being used in mainstream settings. Specifically, the CDA
involved analysis of wording and word co-occurrences on the textual level;
identifying discourses, styles, genres, and discursive strategies as means
of representing meaning and identities from particular perspectives;
uncovering the underlying assumptions of common ground in the texts;
and intertextuality (Fairclough, 1992, 2003).
The analysis showed that the culture enricher expression was used in
relation to well-known far-right discourses criticising ‘the establishment’,
othering and criminalising immigrants, and concurrently portraying
Sweden and supposed ‘real’ Swedes as being under threat and in
opposition to immigrants and their needs. It illustrated a practice of co-
articulating coded expressions which facilitated a broader vocabulary of
46
coded language. Furthermore, the analysis identified a use of culture
enricher as a (non-neutral) replacement for ‘immigrant’ without
acknowledgement of its neo-Nazi background. Specifically, the analysis
showed how the use of the expression enabled two things which the words
‘immigrant’ and ‘immigration’ do not. First, it worked like a ‘dog whistle’
by enabling far-right sympathisers to self-identify with an imagined white,
Swedish ‘people’ while simultaneously showing opposition to a
generalised ‘establishment’ and its priorities. Second, as the expression
took new forms online, and was co-articulated with other hateful, coded
expressions, it was used to pin racist stereotypes to Muslim men in subtle
ways.
The analysis furthermore identified that culture enricher was mainly
circulated in the large Swedish online forum Flashback, and in Swedish
far-right blogs. On Flashback, the expression was widespread throughout
a variety of different forum parts, even those unrelated to issues of
immigration and several users also claimed to have learnt it through the
platform. Thus, signalling Flashback’s importance as a gateway site to
more radical far-right discourse, also exposing those not interested to
such language use. Because overt far-right discourse is generally not
socially acceptable, promotion of these ideas was often expressed with
subtlety. Far-right blogs in particular tried (sometimes successfully) to
borrow credibility from legacy news media and actively downplayed overt
far-right expressions to increase the spread of their ideas. The analysis
also identified an interconnectedness of blogs and content, which could in
turn enable the exposure and spread of far-right discourse.
Finally, the analysis illustrated two ways in which culture enricher made
its way into the mainstream: through counter-discourse and resistance to
far-right ideology, and through appropriation of far-right ideas. The
analysis illustrated how when coded far-right expressions were used in
mainstream settings, uninitiated individuals risked encountering them
even though they had not actively sought them out. It also showed how
the expression peaked in mainstream settings in 2008 and 2009 with
legacy news media reports calling out Sweden Democrat politicians for
using culture enricher and other hateful expressions online. Potentially,
the attention from mainstream media prompted caution, from which even
subtler variations of the expression not sampled in the current study (e.g.,
enrichers) and use of other coded words might have come to take its place.
47
Paper two
Standing Up for Sweden? The Racist Discourses, Architectures
and Affordances of an Anti-Immigration Facebook Group.
Published in Journal of Computer-Mediated Communication, 2018
(co-authored equally
6
with Samuel Merrill).
The second paper (Merrill & Åkerlund, 2018) explored the large Swedish
Facebook group Stand up for Sweden. With the group itself claiming to
only discuss immigration politics but following massive criticism of hate
speech faced by the group, the paper aimed to establish how racism
contributed to the group’s discussions of immigration, and how Facebook
as a platform allowed it to do so. All posts and comments from when the
group was founded in February 2017 up until the time of data collection
in September 2017 were collected using the open-access Python tool
fb_scrape_public (Freelon, 2015/2020). In total, it explored a sample of
59,489 posts and 1,754,425 associated comments which were then
analysed using a combination of topic modelling and CDA. Topic
modellinga form of automated thematic analysiswas used to gain an
initial overview of the overarching topics discussed in the Facebook group.
It was also deployed as an entry point and sampling method for a
subsequent CDA. Through the topic model, a number of distinct themes
were identified as being discussed in the group and ten themes in total,
which in different ways related specifically to immigration were explored
further. This analysis showed that users expressed well-known far-right
tropes, nostalgic longings to an imagined masculinised and white past
while fearing a dystopic future for Sweden. Specifically, seven of the ten
themes reflected more concrete discussions on economy, criminality,
deportation, international immigration, accommodation, employment,
and party politics, while the remaining three concerned more general
debates about freedom of speech, definitions of racism, and truth.
The CDA took its primary starting point in the five questions posed by
Wodak and Reisigl (2001) to identify racist discourse. This analysis found
that users often identified with an imagined, homogenous group of ‘real’
(white) Swedes, who saw themselves as in opposition to immigrants and
‘the establishment’. Immigrants were generally thought of in negative
termsas benefit cheats and criminals, placing economic strain on
Sweden’s social welfare, housing, and job market. And ‘the establishment’
6
The division of authorship on the paper: Intellectual input 50/50; Data collection and analysis 60/40
Write up and reviews 40/60 between me and Samuel Merrill, respectively.
48
was presented as responsible for Sweden’s alleged decline and as traitors
in their favouring of immigrants over ‘real’ Swedes. To justify these
categorisations, the group’s users selectively presented news articles and
pseudo-statistics to prove their skewed points, they engaged in conspiracy
theories and compared Sweden’s immigration politics to that of countries
they believed had dealt with the issue successfully (more restrictively).
The CDA also showed that users positioned themselves in specific ways
when it came to the issue of immigration; both as truth-tellersspeaking
up about the declining state of Sweden, as victims of reverse racism, and
as realistswho want what is best for both Swedes’ and ‘immigrants’. The
analysis also found instances in which users tried to deny accusations of
racism or redefine the concept altogether.
The findings suggested, furthermore, that Facebook’s technological
features allowed both covert and overt forms of racism to accumulate.
Specifically, the analysis identified that weaknesses in Facebook’s
reporting system allowed racist content to persist even when moderators
acted against offending individuals. The analysis of technological features
showed furthermore how the visibility and reach that Facebook could
afford content contributed to the circulation and normalisation of racist
discourse. Finally, it also highlighted how content was magnified by users
who took advantage of how easy it was to duplicate, like, comment, and
spread others’ content, links, and images. This helped the proliferation of
racist content both within but also beyond the Facebook group.
Paper three
The importance of influential users in (re)producing Swedish
far-right discourse on Twitter.
Published in European Journal of Communication, 2020.
The third paper (Åkerlund, 2020) analysed the differences between
influential and non-influential far-right users when it came to language
and platform usage patterns on Twitter. Furthermore, it analysed how
these social media usage patterns contributed to (re)producing far-right
discourse in this setting. To explore this, tweets were collected via
Twitter’s API using a series of hashtags as search terms. With data
collection starting just before the Swedish 2018 general election, and with
speculation of the Sweden Democrats imminent success there, a snowball
sampling of twenty-three ‘far-right hashtags’, starting with #SDwas used
to collect a dataset totalling 74,336 tweets, posted by 6,809 users between
September and November 2018. Influential users were identified in this
49
paper by determining which users had been retweeted the most by the
largest number of individual users. Sentiment analysisan automated
method to identify emotions in textwas used to analyse differences in
language use between the influential and non-influential far-right Twitter
users. Through this method, the median level of emotion for each user’s
tweets were compared between influential and non-influential users. The
model also provided scores for each individual post for further close
readings to provide context to the sentiment scores.
The analysis showed firstly that influential users were a diverse group
including people holding official positions within far-right organisations,
private individuals, as well as anonymous accounts. In common were that
they all seemed to support far-right organisations such as SD but also
smaller organisations like Alternative for Sweden, Citizens’ Coalition, or
far-right ideology more generally. These users had profiled themselves as
entirely focusing on far-right politics, both in terms of the content they
shared but also in terms of their profile presentations, thus making them
easy to find for other far-right sympathisers. Overall, the analysis showed
high numbers of mutual followings within the dataset, suggesting a
seemingly stable constellation of accounts who had personal connections
with one another. The analysis also found that influential users had more
followers and friends and were more active in posting and liking than
other users. Influential users also produced more original content than
others, suggesting that influential users were more opinionated, keen, or
confident about sharing original content while others preferred to pass on
ready-made ideas. Considering that influential users rarely mentioned
others, it could be argued that they were not so much engaging in
conversation themselves, as they were defining the conversation for
others to engage in.
Surprisingly, influential users tended to be much more neutral than other
users in terms of the levels of sentiment in their posted content. This is
not to say however, that they were not expressing far-right ideas. The
appearance of neutrality seemed to be used as a way for influential users
to spread far-right discourse without having to be responsible for or
accept the consequences of such content themselves. The analysis showed
that many of these influential users maintained their influential positions
throughout the sample period. Potentially because they were using
Twitter actively in such a way that they did not get reported, they were
able to tweet content which had a clear far-right tone, but which did not
explicitly express any negative sentiment. The analysis showed instead
how it was those commenting in response to influential users’ neutral
posts who expressed overt hateexplicitly stating what the influential
50
users were simply implying. Finally, influential users were found to enable
engagement and (re)production of far-right discourse through the
openness of their neutral tweets, which allowed for an inclusive and vague
‘us’, allowing a wider range of users to potentially identify with the far
right online.
Paper four
Influence Without Metrics: Analyzing the Impact of Far-Right
Users in an Online Discussion Forum.
Published in Social Media + Society, 2021.
The fourth paper kerlund, 2021c) analysed the processes through which
subtle forms of influence take shape in eclectic online forums with few
vanity metrics. The paper explored who becomes influential, their
strategies for appealing to the given community, and others’ support of
them in this setting. Using the Swedish online forum Flashback as a case
study, the entire forum category Politics with all the threads, posts, and
users in its twenty subcategories was collected using the Python-based
scraping tool Flashbackscraper (Kullenberg, 2018). In total, the dataset
contained 124,175 threads and 7,488,468 posts, created by 113,550 users,
and spanning from May 2000 when the forum was first launched, until
January 2019. Network analysis was used as a starting point for
determining influence within the Flashback forum. Specifically,
betweenness centralitya measure of how often nodes are on the shortest
paths (the geodesics) between other nodeswas used to determine
influence. A small selection of users with particularly high betweenness
centralities, determined in terms of quote activity, were explored further,
and used as the basis for interpretative analysis. This interpretative
analysis involved CDA, which specifically focused on how other texts and
voices were incorporated (intertextuality), on ideas about common
perspective or knowledge (assumptions), how actors self-identified, and
how they positioned others in the texts (styles and representation of social
actors) (Fairclough, 2003).
The analysis showed that those users who were deemed influential in
terms of holding strategic network positions and consistently producing
content that resonated with others, throughout the sample period
reiterated well-established far-right ideas, including anti-establishment,
anti-feminist, and anti-immigrant discourse. This was expressed with the
help of pseudo-rational arguments, decontextualised use of external
references, and through an extensive use of code words and jargon. In
51
turn, these practices illustrated their embeddedness in forum as well as in
far-right culture. While it has been known that Flashback hosts more or
less extremist far-right users and content, this article shows that these
sentiments are neither fringe nor obscure, but instead seemingly widely
supported and influential in the studied parts of the forum.
The analysis showed that despite few visible markers, other users had
learned to know and recognise influential users and their content as
reliable and worth supporting. The absence of vanity metrics meant that
the importance of users on Flashback instead became based on reputation
on the site. Instead of using built-in functions, some engaged in manual
pinning or saving, ‘liking’, and ‘sharing’ of influential users’ content
presented via their replies as means to promote and legitimise these users.
However, with self-centred motives often in play online, this was
discussed as responding users’ ways to potentially increase their own
statuses within the forum.
The analysis also found that some users acquired the best positions in
terms of network centrality not because they were influential, but rather
because of how other users disagreed with them and their content. These
widely disliked users all advocated progressive gender and immigration
ideas. While these users were not censored by the forum itself, those who
did not conform to the more or less informal rules were nonetheless
reprimanded by the community. But because of an absence of public
metrics, these users, whom others were in a sense trying to silence through
their hateful responses, were in fact being magnified.
Finally, the role of Flashback in enabling the success of far-right users and
ideas should not be understated. Flashback users cannot delete their own
accounts, posts, or threads. With this, the likelihood that far-right content
persists in the forum is high. Furthermore, the analysis shows that the
forum’s regulations promoting anonymity and freedom of speech become
a veil which disguises far-right content and users, and results in
difficulties in detecting and scrutinising far-right influence. With this, and
its ease of use, accessibility, and widespread adoption, hateful ideas
expressed in the forum can easily reach outside of fringe far-right circles.
So, providing a limited number of vanity metrics is convenient for
Flashback and other forums like it because it obscures what types of
content and users are most highly appreciated.
52
Discussion
This section begins with an empirical discussion which covers how the
four case studies address the three research questions posed in the
introduction of this thesis. Thereafter follows a discussion on the
interrelations of far-right discourse, and the enabling roles of platforms,
and influential users, and how this complexity can be better understood.
Responding to the research questions
The four case studies have in different ways, using a variation of methods
and different data materials addressed the issues of far-right discourse,
and the enabling roles of platforms, and influential users. It has sought to
do so by answering three interrelated research questions which are
discussed separately below.
First, in regard to how far-right discourse is expressed and normalised in
different online settings (cf. RQ1), the studies carried out support previous
research showing that the elements of far-right discourse in online
settings are neither new nor particularly different compared to far-right
discourse expressed outside of the internet. While varying in focus, the
papers illustrated how Swedish far-right discourse online revolves around
whiteness and exclusionary understandings of Swedishness. However,
rather than users adhering strictly to any narrowly delineated ideology or
specific party-political affiliations, the case studies showed how joint
opposition toward progressive politics was often what brought far-right
sympathisers together online kerlund, 2020, 2021c). Far-right
discourse was also a way for users to identify with an in-group as well as a
lens through which to understand those perceived as ‘others’ (Merrill &
Åkerlund, 2018).
The case studies, furthermore, illustrated how users appropriated a range
of strategies which helped to normalise far-right language use in the
contexts studied. Åkerlund (2020) for instance, showed the effectiveness
of an appearance of openness and neutrality in tweets, and Merrill and
Åkerlund (2018) highlighted the use of a range of strategies to persuade,
justify, and mitigate racism on Facebook. Not only can strategies such as
those identified in the case studies generate lots of ‘noise’, but they might
also increase probabilities that unsuspecting users will encounter at least
one argument or strategy by which they will be convinced of the supposed
reasonability of far-right discourse.
53
Of course, not all individual posts or comments explored in these papers
perpetuated far-right discourse. Instead, the papers highlighted the
importance of recognising the overarching patterns of language use when
studying how far-right discourse is expressed and normalised online.
Notably, Åkerlund (2021b) identified how far-right jargon interconnects
overt and covert, and newly formed as well as repurposed far-right
expressions, which ultimately enable their exposure. The case studies also
showed how seemingly non-extremist far-right content sometimes
prompted more explicit and overtly hateful responses and comments by
others kerlund, 2020, 2021b; Merrill & Åkerlund, 2018).
Beyond this, the studies demonstrate how far-right language use
constantly adapts, develops and morphs to avoid exposure and to uphold
its availability in different online settings kerlund, 2021b). For example,
each individual, seemingly neutral tweet worked over time to propagate
far-right discourse widely, yet discreetly, while avoiding content
moderation on Twitter (Åkerlund, 2020). In some cases, the papers even
identified concrete requests by users to make others tone down the
explicitness of far-right discourse so as to not be “playing into the hands
of censorship” (Merrill & Åkerlund, 2018, p. 347) and to “make content
more easily sharable, and to not reflect badly on certain political
organisations” kerlund, 2021b, p. 11). Importantly, these findings
highlight the significance of not overlooking the seemingly banal in
explorations of far-right discourse online. While appearing less severe,
such expressions might nevertheless provoke exceedingly extremist far-
right behaviour. Or, as Åkerlund (2021b) saw, might also lead to
normalisation of far-right discourse through the unknowing adoption of
such language use by those who are unaware of its far-right nature.
Second, the thesis asked how online platforms enable the circulation and
amplification of far-right discourse (cf. RQ2). The papers showed how the
proliferation of far-right discourse was enabled by platform accessibility.
Notably, Flashback lets any internet user access and read much of its
content without having a user account. And while Facebook and Twitter
do require user accounts for accessing all content on these sites, many
groups, including the one explored by Merrill and Åkerlund (2018), are
‘open’ on Facebook, and on Twitter users can easily find plenty of topic-
specific content neatly organised via hashtags. As such, the ease with
which users can access far-right discourse either through search engines
or on the sites themselves without befriending, following, or subscribing
to specific users or groups was seen to afford far-right discourse broad
reach. At the same time, different functions to connect with others more
permanently through follower-friend relations, group memberships, and
54
homepage hyperlinks also contributed to the persistence of these user
relationships and thus, to the proliferation of far-right discourse
kerlund, 2020, 2021b; Merrill & Åkerlund, 2018).
The papers furthermore identified how several vanity metrics helped
amplify both content and users, and that the ‘like’ and ‘share’ functions on
Twitter and Facebook magnified and worked to circulate far-right
discourse for broad reach (Åkerlund, 2020; Merrill & Åkerlund, 2018).
Interestingly, in the absence of visible metrics, progressive content
instead became excessively amplified, as was identified on Flashback
kerlund, 2021c). This is not to say that just removing all vanity metrics
will promote progressive ideas on any platform. On Flashback specifically,
the lack of metrics made it difficult to detect the prominence of far-right
users and discourse on the site, obscuring that far-right ideology was
widely spread, popular, and influential. Likewise, the anonymity provided
by some of these settings made it easy for users to contribute content and
interact with others without the risk of social or legal repercussions.
While these functions were identified as important in the proliferation of
far-right discourse in the case studies, these socio-technical features are
likely to benefit many different types of ideas and movements also beyond
the far right. For instance, opportunities for anonymity have been found
to promote benign forms of self-disclosure (Andalibi et al., 2018; Clark-
Gordon et al., 2019) and emotional support (Pezaro et al., 2018).
Furthermore, accessibility, opportunities to link, befriend, follow, like,
shareand organise via hashtags has for instance been identified as
important technological features also for progressive movements like
#BlackLivesMatter (Anderson & Hitlin, 2016; Cox, 2017; Ince et al., 2017;
Jackson et al., 2020; Mundt et al., 2018).
Beyond accessibility and functions to amplify content and users, the case
studies identified how the analysed sites, threads, and groups often
worked to bridge between a range of different users and content, and as
such sometimes functioned as forms of gateways between ‘the far right’
and ‘the mainstream’. Prominently, Flashback was identified as a
particularly important gateway as it constitutes a place in which users
picked up on far-right jargon, and which exposed users to far-right
discourse even though they had not actively sought out such content
themselves kerlund, 2021b, 2021c). Moreover, groups like Stand up for
Sweden (Merrill & Åkerlund, 2018)which at one point had around
170,000 members, or the equivalence of more than 1,6 percent of
Sweden’s populationcould, due to its size alone, be thought to
potentially bridge a range of different levels of far-right discourse and
55
users. Simultaneously, these spaces worked to bring together otherwise
disconnected far-right sympathisers and therein functioned as breeding
grounds for far-right community kerlund, 2020, 2021b, 2021c; Merrill
& Åkerlund, 2018).
While there are of course no ‘far-right affordances’, some functions
nevertheless seem especially important for enabling content which might
not so easily have been accepted in physical public spaces. So, even though
the idea of gateways between varying levels of radical content and users is
not an exclusively far-right phenomenon, it can nonetheless be though to
hold particular relevance to movements, users, and discourse of
comparably harmful nature. Those whose best opportunity to recruit new
sympathisers is through a gradually increasing exposure to extremist
content and radicalisation (e.g., Munn, 2019).
Relatedly, policies for content removal were shown to contribute to the
build-up of residual far-right discourse. In line with previous research, the
case studies illustrated how much responsibility was placed on users
themselves to report and thus to counteract rule-violating content (Merrill
& Åkerlund, 2018), and showed how discriminatory posts were sometimes
permitted by platforms under the guise of freedom of speech kerlund,
2021c). Moreover, Facebook policies allowed for the removal of only the
previous week’s worth of content from suspended users (Merrill &
Åkerlund, 2018) and on Flashback, any deletion of content or accounts is
restricted to the will of moderators and administrators kerlund, 2021c).
This absence of moderating tools enabled far-right content to endure on
the sites studied, and likely contributed to it being seen, read, and
potentially absorbed by far too many users. The persistence of far-right
content was also made obvious by the long periods of time that Google
provided continuous availability to many racist texts via its search engine,
dating back as far as to the 1990’s kerlund, 2021b).
While longitudinal access to various content is a foundational premise for
the internet at large, the permittance of discriminatory and sometimes
even criminal content on easily accessible sites is something that
unequivocally benefits the far-right, whose content would likely not be
tolerated for long in a non-digital public space. The lack of responsibility
taken by different platforms to remove insidious content is central to how
the far right is able to leverage other functions to their benefit too. Even
those which might not themselves be inherently biased towards the far
right.
56
Finally, the thesis sought to understand the role of influential users in
propagating far-right discourse and contributing to a form of far-right
community online through their uses of online platforms (cf. RQ3). In line
with previous research, the case studies showed the importance of a select
few influential users in the contexts studied. While this took different
forms throughout the case studies, it could clearly be seen in terms of
some users being exceedingly active compared to others when it came to
contributing content and therein defining and setting the agenda and tone
for far-right discussions in the given settings kerlund, 2021b; Merrill &
Åkerlund, 2018; Åkerlund, 2020). Moreover, influence was present in
terms of the high reputations and statuses that only a minority of users
enjoyed kerlund, 2020, 2021c).
These positions, however, were not stable but instead required users to
continuously gain and sustain the interest and involvement of others.
While exposure was enabled by vanity metrics on Twitter (Åkerlund,
2020), on Flashback, these processes were much subtler and required
familiarity and tenure by users. With this, the thesis also highlighted the
importance of acknowledging not only influential users themselves but
also the need to analyse those supporting and co-creating influence.
Through the mixed methods approach, with particular focus on
qualitative interpretation, the thesis highlights that influence cannot be
based entirely on statistics, especially as it pertains to covert far-right
influence.
Relatedly, the case studies also explored the role of influential users in far-
right community-building. The papers illustrated that influential far-right
users seemed to use platforms more strategically than others, for instance
in their avoidance of moderation on Twitter (Åkerlund, 2020). They were
also particularly knowledgeable of community culture as seen among
other things through their use of jargon kerlund, 2021c). More than
this, the papers provided insight into the communities more broadly. For
instance, Åkerlund (2021c) showed that those who were the most
influential on Flashback exclusively supported far-right ideology. With
this, the paper highlighted the prominence and popularity of these ideas
in the forum category at large.
Finally, it was shown how influential users could serve as important
bridges between different types of users kerlund, 2021c), that they
assisted in making far-right discourse approachable and easily shareable
and provided an easy entry-point into far-right community (Åkerlund,
2020).
57
The complexity of far-right discourse, platforms,
and influence
The papers of this thesis have tried to highlight how the different
dimensions of far-right efforts online are interconnected and how they
cannot be neatly separated. To briefly summarise these interconnections,
paper 1kerlund, 2021b) illustrated that very few sites were responsible
for the creation and circulation of most uses of culture enricher on the
Swedish web. Users actively engaged with and cooperatively developed
the expression over time, morphing it into more racist forms. While at the
same time, through more or less deliberate attempts to deceive and cloak
the expression, finding ways to increase its spread. With the indirect
assistance of Google Search and other sites, this content was able to
persist over long periods of time, despite its sometimes explicitly racist
nature. Similarly, paper 2 (Merrill & Åkerlund, 2018) showed how most
content was published by very few users to a vast audience in the Stand
up for Sweden Facebook group. These users perpetuated far-right
discourse through a range of discursive strategies, but also through more
explicit efforts to game Facebook’s moderating system. Facebook in turn,
failed to efficiently deal with the racist discourse that had accumulated
through these efforts. Relatedly, paper 3 kerlund, 2020) found that a
small concentration of Twitter users posting far-right content enjoyed
excessive retweets by large numbers of others. With this, these users were
propelled into attention by Twitter’s vanity metrics. These and other users
posted about the same far-right issues, but influential users did so in a way
that rendered far-right discourse inclusive and widely appealing. At the
same time, their way of posting made them difficult to deal with by
Twitter’s moderating system. Finally, paper 4 kerlund, 2021c) showed
how a small group of users garnered the most positive attention by others
on Flashback. These users consistently advocated far-right ideas, which
were in turn promoted and spread through other users’ engagement with
platform functions. Correspondingly, Flashback’s protection of freedom
of speech, anonymity, and lack of public metrics served to disguise far-
right content and users. Below, these complex processes are made explicit
through the illustrations of Figure 3.
While it seems slightly reductive to try to fit all empirical research of this
thesis into a single model which can only ever be a crude simplification of
the findings of this research, it can still provide a graphical representation
of some of the thesis’ main takeaways. Accordingly, below is an
elaboration of the model presented earlier, in Figure 1. The reworked
58
model in Figure 3 attempts to illustrate how the dimensions of Figure 1
have actually taken shape in the empirical work carried out in the case
studies.
Figure 3. Reworking of Figure 1 in accordance with the main findings of the thesis
Starting from the top-left, the conventions of far-right discourse
(described as ‘discursive conditions’ in Figure 1) constitute the ways
influential users must package far-right discourse in order to garner the
support of other far-right sympathisers. In this thesis this dimension has
tended to revolve around a need to be broadly appealing in order to
engage heterogenous and disparate constellations of far-right
sympathisers, and to create joint opposition to those considered by the
community to be members of different out-groups.
On the bottom right, influential users responded to the discursive
conditions through their particular skills in leveraging far-right discourse
(described as ‘discursive competence’ in Figure 1). Influential users
successfully appealed broadly to far-right sympathisers and contributed
to the creation of a sense of opposition to out-groups through their use of
jargon and slang, displaying their skill and belonging in far-right settings.
They also presented far-right discourse with exclusionary openness by
making it appear reasonable, open, and approachable also for non-
extremist far-right sympathisers while at the same time directing
themselves exclusively towards ‘real’ and white Swedes.
The bottom-right part of Figure 3 shows the technological prerequisites
(described as ‘technological conditions’ Figure 1) which were especially
important for enabling far-right discourse. The case studies illustrated
how some settings provided a gateway access that helped bridge between
different levels of far-right extremism, and thus by extension also enabling
its spread, potential user radicalisation and general normalisation.
Furthermore, whether through a lack of reporting and moderating tools,
desire, or priority to remove far-right discourse, residual content made
far-right discourse easily available and accessible to users over time
59
online. This in turn, also helped influential users reach others broadly and
continuously.
Finally, On the top-centre, influential users possess particular skills
(described as ‘technological competence’ in Figure 1) which allows them
to leverage digital platforms. Influential users displayed this knowledge in
the case studies through strategic subtle language use which made their
content elude moderation and reporting systems and allowed its continual
availability. This in turn, was also what made their approach to far-right
discourse seem widely appealing and approachable, as illustrated on the
bottom-left of Figure 3. Furthermore, influential users were able to
leverage support as exposure. By engaging others, influential users were
propelled into visibility on different platforms, in turn, granting them
better opportunities to display their far-right knowledge, appeal to other
users, and unite them against perceived out-groups.
Despite the many limitations of a model like the one above, it begins to
illustrate the complex interconnectedness of far-right discourse,
platforms, and influential users.
60
Contributions and future research
In conclusion, this thesis aimed to contribute understanding to the
complexity of far-right efforts in Swedish digital mainstream settings
through its expression by ‘ordinary’ users. Through its case studies, this
thesis has strengthened the findings of previous research, illustrated
unexpected and contradictory results in relation to what was previously
known, and made interesting new discoveries. With this, it has begun
unravelling the complex interconnectedness of far-right discourse,
platforms, and influential users.
Given the fundamental impact the internet has had on society, including
the formation of public opinion, knowledge production, and information
sharing (Benkler, 2006; Curran, 2012a; Earl & Kimport, 2011; Ito, 2008),
the matter of understanding how the far right operates online is as noted
by Hughey and Daniels (2013, p. 342) “more than academic”. The
exploration of the far right carried out in this thesis provides an important
new perspective on this widely engaging and concerning topic. The thesis
makes the argument that to gain a more comprehensive understanding of
the workings of the far right online, we must move outside of the easily
delineated, not just treating the propagation of far-right discourse as a
collective phenomenon, incorporating analyses of how different platforms
are leveraged by the far right and notably by individual far-right users,
and explore how the structures of platforms themselves contribute to the
far right’s success online. While these issues are far from fully explored
through this thesis alone, the work undertaken here nevertheless makes
several contributions to their understanding.
First, this thesis makes a conceptual contribution by taking several
dimensions of the far right online into account at once. This, as the thesis
has attempted to show, offers deeper insight into the workings of the non-
party sector of the far right online. Instead of exploring only expressions
of far-right discourse, the cases have been able to show differences
between influential and non-influential users (Åkerlund, 2020) and how
other users can co-construct their positions of influence kerlund,
2021c). By taking platforms into account, the case studies explored have
shown how far-right discourse is enabled by the socio-technical features
of different online settings kerlund, 2020, 2021c; Merrill & Åkerlund,
2018). At the same time, had these studies focused solely on platforms or
influential users, and not taken discourse specifically into account, they
would have overlooked the centrality and importance of far-right
discourse on a site like Flashback kerlund, 2021c), and missed the
61
subtle development of far-right language use over time on the Swedish
web kerlund, 2021b). It is the combination of these three dimensions
that help enrich our insights into far-right efforts online.
The thesis also makes an empirical contribution. Most obviously, the
overwhelming bulk of research on the far right remains focused on
electoral politics and parties (Rydgren, 2018a), while the multifaceted
grassroot or ‘non-party’ dimensions of far-right mobilisation have
received much less attention (Castelli Gattinara & Pirro, 2019; O. Klein &
Muis, 2019; Veugelers & Menard, 2018). More than that, these papers
have explored and illustrated the significance of understanding the less
easily delineated aspects of the grassroot far-right online. Not the least
those which are not outspokenly far-right extremist. With this, this thesis
contributes knowledge about what remains an underexplored dimension
of far-right mobilisation. Furthermore, with the exception of a few
researchers and studies (Ekman, 2013, 2014, 2015, 2018, 2019; Merrill,
2020; A. Törnberg & Törnberg, 2016a, 2016b; A. Törnberg & Wahlström,
2018; Urniaz, 2016; Wahlström et al., 2020; Wahlström & Törnberg,
2019), little is still known about the Swedish far right online. While this
thesis seeks to provide insight into the far right beyond Sweden, it
nevertheless makes an empirical contribution to this context by
illuminating different aspects of far-right discourse, platforms, and
influential users in a Swedish language setting specifically.
Lastly, the thesis also makes a methodological contribution. While not
brought forth in the thesis as a focus in itself, the covert and blurry aspects
of the far right explored in this thesis have throughout the empirical work
posed methodological challenges which have required flexibility and
adaptability. The mix of methods used throughout the case studies have
illustrated how different aspects of the far right, over varying time periods,
diversely sized and shaped datasets, and user constellations, can be
approached to recognise the broader overarching patterns as well as the
intricate details. With these innovative applications of methods, the thesis
illustrates the benefits of going beyond the quantitative/qualitative divide
in order to better comprehend the far right’s efforts and activities online.
This thesis constitutes only a part of the academic work that needs to be
carried out to understand the contemporary far-right. Future research
efforts should delve deeper into the different dimensions of the far right’s
presence and activities online, in Sweden and elsewhere. This requires, for
instance, exploring the intersections between the far right and other anti-
democratic movements, across sites and national boarders; analyses of
<