Boundary Struggles

Arnfinn H. Midtbøen, Kari Steen-Johnsen and Kjersti Thorbjørnsrud (Eds.)

# **Boundary Struggles**

Contestations of Free Speech in the Norwegian Public Sphere

© Arnfinn H. Midtbøen, Kari Steen-Johnsen and Kjersti Thorbjørnsrud (Eds.), 2017

This work is protected under the provisions of the Norwegian Copyright Act (Act No. 2 of May 12, 1961, relating to Copyright in Literary, Scientific and Artistic Works) and published Open Access under the terms of a Creative Commons CC-BY 4.0 License (http://creativecommons.org/licenses/by/4.0/). This license allows third parties to freely copy and redistribute the material in any medium or format as well as remix, transform or build upon the material for any purpose, including commercial purposes, provided the work is properly attributed to the author(s), including a link to the license, and any changes that may have been made are thoroughly indicated. The attribution can be provided in any reasonable manner, however, in no way that suggests the author(s) or the publisher endorses the third party or the third party's use of the work.

Third parties are prohibited from applying legal terms or technological measures that restrict others from doing anything permitted under the terms of the license. Note that the license may not provide all of the permissions necessary for an intended reuse; other rights, for example publicity, privacy, or moral rights, may limit third party use of the material.

This book is made possible with support from Fritt Ord.

Typesetting: Datapage India (Pvt.) Ltd.

Cover Design: Cappelen Damm

Cover Illustration: Oda Sofie Granholt

Cappelen Damm Akademisk/NOASP

noasp@cappelendamm.no

# Contents


### **Chapter 2 Silenced by hate? Hate speech as a social**


### **Chapter 3 Willingness to discuss the publishing of religious cartoons. Spiral of silence in the private and public spheres ...........77**



### **Chapter 4 Perceptions of journalistic bias: Party preferences,**


### **Chapter 5 Editorial perspectives on the public debate**


### **Chapter 6 Debating freedom of expression in Norwegian**



### **Chapter 7 Boundaries of free speech in the political field........... 195**


### **Chapter 8 Ascribed representation: Ethnic and religious minorities in the mediated public sphere....................................... 229** *Marjan Nadim*


### **Chapter 9 Immigration critique: Moral boundaries,**



### **About the Authors............................................................................321**

### **Figures**


### **Tables**



# Preface and acknowledgments

The analyses in this book build on four years of work on a largescale project called *The Status of Freedom of Speech in Norway*. It was commissioned by the private, non-profit foundation Fritt Ord (literally: The Free Word), and included a broad, empirical study of the conditions for free speech in Norway. The project group consisted of a multi-disciplinary team of researchers – sociologists, media scholars, political scientists, and legal scholars. We have employed a wide range of methods: broad representative surveys as well as targeted surveys among ethnic minorities, journalists and editors, and artists; content analyses of the media coverage of key issues in the free speech debates; and qualitative studies of groups that are particularly relevant to a free speech perspective. In Norway, the subject of free speech had not previously been empirically studied in this manner or in this scope – and we posit, not in an international context either. In the present book, we draw on substantial parts of this rich data material in order to examine the social processes through which boundaries of free speech are maintained, challenged and changed in the Norwegian public sphere.

We have benefitted from the generous support and insightful comments of a number of people during this work. First of all, we wish to thank the Fritt Ord foundation for their financial support throughout these years of research, and especially to Bente Roalsvig for her valuable collaboration. Many thanks also to our editor at Cappelen Damm Akademisk, Dorte Østreng. We also want to thank the *Status of Freedom of Speech* project group and the authors contributing to this book for stimulating discussions and for the intense efforts made to finalize their chapters. For reading and critically discussing a previous version of the book as well as providing valuable comments to several of the chapters, we are grateful to our three anonymous reviewers, to Paul Sniderman, Mette Andersson, Stig Hjarvard, and to our dear colleagues at the Institute for Social Research. Many thanks to Frithjof Eide who did a great job in transcribing all the qualitative interviews on which chapters 7, 8 and 9 are based. Finally, we thank our project adviser, Synne Sætrang, for her continuous help and support in keeping the process on track, and Karin Kraglund, who made essential contributions to the crucial final phase of the book manuscript.

> *Arnfinn H. Midtbøen, Kari Steen-Johnsen and Kjersti Thorbjørnsrud. Oslo, March 15, 2017*

Chapter 1

# Boundary-making in the public sphere: Contestations of free speech

*Arnfinn H. Midtbøen, PhD, Senior Research Fellow, Institute for Social Research Kari Steen-Johnsen, PhD, Research Professor, Institute for Social Research Kjersti Thorbjørnsrud, PhD, Senior Research Fellow, Institute for Social Research*

# **Introduction**

Freedom of speech is a fundamental human right and considered a core value in liberal democracies. However, it is also one of our time's most contested issues, constantly claimed either to be too wide-ranging, allowing continuous repression of minority groups, or too limited – restricting dissent and democratic deliberation. In this book we depart from conventional approaches to free speech, which tend to focus on whether specific types of public utterances should be legally allowed or not. Instead, we study how the boundaries of free speech are contested and negotiated through social processes which silence certain groups and opinions while amplifying others.

Dramatic events in the past decade have demonstrated how free speech is deeply connected to global struggles over power and recognition. When the Danish newspaper *Jyllandsposten* published twelve caricatures of the prophet Muhammad in 2005, this led to heated debates and demonstrations in Europe as well as protests and the burning of Danish flags and embassies in the Middle East. The terror attack on the satire magazine *Charlie Hebdo* in Paris in 2015, resulting in the death of twelve people, led to renewed debate about the role of satirical cartoons in defining and pushing the frontiers of free speech in a global perspective.1 The *Charlie Hebdo* attacks also served as a forceful reminder that the exercise of free speech may be followed by acts of violence.

While debates over free speech are heavily marked by political and ideological cleavages on the global level, they take place within specific national contexts. In Norway the horror of July 22, 2011, in which 77 individuals were killed by an extreme right-wing terrorist, was perceived as an attack on the leading political party, the Labour Party, but also on multicultural society itself. The July 22, 2011 terror attacks led to intense debates over growing anti-Muslim and anti-immigration sentiments in Norway. A more responsible public debate, where people behaved decently and extreme views were cracked down

<sup>1</sup> We refer to both the Muhammed caricatures and the drawings published in *Charlie Hebdo* as 'cartoons'. As pointed out by Klausen (2009 pp. 6-7), these drawings are strictly speaking mostly caricatures – that is, 'wordless drawings that use exaggerated physiognomic features to make a statement about the fundamental nature of a person or thing'. However, we follow Klausen and use 'satirical cartoons' or simply 'cartoons' when referring to the publishing of such caricatures because the 'cartoon crisis' has been established as the main reference in both public and academic parlance.

on, was called for. However, over time a discussion about the consequences of constraints on free debate arose. The core of the argument was that perceived moral taboos surrounded debates on immigration and minorities to the extent that the voices of those concerned about the future of the nation state and their own social status were silenced. Moreover, a too strictly monitored mainstream public debate could potentially lead to increased support for populist right-wing movements capitalizing on this very concern. Even though a debate following July 22, 2011 has its particular reference points in Norwegian society, the rise of right-wing populism across Western democracies in recent decades has made pertinent the confrontation between an elite dominated public sphere and anti-immigration ideas in many countries, and concomitantly raises questions about the implications for free speech practices.

Rapidly changing media technologies and platforms are vital factors in the struggles over the norms, principles, and practices of free speech. Today, both traditional and social media permeate people's life, spread messages instantly across the world and democratize public debates. Digitalization has entailed both a democratization of the possibility to speak publicly, a fundamental change in the role of traditional media as implied in the change from 'gatekeeping' to 'gatewatching', and a blurring of the conditions for boundary-drawing related to free speech (Ash, 2016; Benkler, 2006; Bruns, 2005). The continuous debates about the protection of free speech, on the one hand, and the limits for acceptable utterances in debates over religion and immigration, on the other, are shaped by this new media context. Hence, the role of the media, both new and old, is a central topic in many chapters in this book.

While much of the literature on the boundaries of free speech has been in the form of normative discussions on its constitutional and legal limits (Maussen & Grillo, 2014, p. 176), there is now an emerging interest in studying the boundaries of free speech with perspectives from the broader social sciences. Arguing for a 'sociopolitical' approach to the study of the regulation of hate speech Maussen and Grillo, for example, advocate a conceptualization of hate speech as a social, cultural and political construct that depends on the context in which it is deployed (2014, p. 177). Moreover, they emphasize that speech, and how it is perceived and judged, is always embedded in power differentials, which has implications for how it can be analyzed.

The present book is in line with such a sociopolitical perspective, but places its emphasis on sociological processes and interactions on the elite, group and individual levels. Theoretically, we build on the concepts of boundaries and boundary-work. The study of symbolic and social boundaries has a long tradition in sociological and anthropological research (see Lamont & Molnár, 2002 for a review), but is less used in current scholarship on free speech. We believe that this field of research can benefit from employing this perspective because it allows us to study the social processes through which boundaries of free speech are drawn, maintained and changed. How are boundaries of free speech *defined* – explicitly or implicitly – by institutional elites? And how are these boundaries *perceived* by the mainstream public and from the margins of the public sphere?

These questions direct our attention to the fundamental dynamics of the public sphere: Public debates are shaped by social mechanisms which silence certain groups and opinions, while amplifying the voices of others. These mechanisms create boundaries that are not (primarily) defined through judicial paragraphs, but rather barriers made of different types of perceived pressure, self-censorship, exclusion and stigma. Sometimes the boundaries of free speech appear bright and clear-cut, based on a strong consensus regarding which opinions and groups are considered to be legitimate or illegitimate in the public sphere. However, they are more often blurred and ambiguous, leaving room both for explicit conflict over where the boundaries are or should be drawn, and for individual maneuvering in the public sphere based on assumptions about the subtle rules defining 'the game' of public participation. In a sociological perspective, we argue, the public sphere can be seen as a locus of 'boundary struggles': Constant debates over the boundaries of free speech shape the dynamics of public debates and gradually change which actors and opinions are granted a legitimate space in the public sphere.

Boundaries of free speech are shaped by a range of key actors and institutions. In this book, we look at how free speech is debated in Norwegian mainstream media, how it is conceived and experienced by young politicians, and how editors and journalists define the limits of the difficult immigration debate – encompassing questions of immigration policies, integration, and religious and ethnic diversity – perhaps the topic in which boundaries between legitimate and illegitimate opinions are most hotly discussed. However, boundaries are also set by and experienced through the everyday activity and interaction of ordinary people. Drawing on a survey among a representative sample of the Norwegian population as well as in-depth interviews with individuals at different margins in debates over free speech – ethnic and religious minorities and immigration critics – we demonstrate the value of a boundary perspective by showing how patterns of self-censorship may keep certain topics and opinions away from the public debate, and how groups at the margins may feel excluded from and stigmatized in mainstream society.

The book demonstrates that boundaries of free speech are 'real' in the sense that they shape individuals' propensity to speak their mind, but also that the way boundaries are perceived varies among actors with different social positions. Indeed, boundaries also vary across time and context: What were seen as illegitimate opinions in the field of immigration ten or twenty years ago, for example, may be seen as completely legitimate today. And what are viewed as boundaries of free speech in Sweden or the US may not be perceived as such in Norway. This book concentrates on the Norwegian context. However, the social and cultural processes analyzed are also part of a larger picture involving religious and political contestation on a global scale. We believe that both the empirical insights and the theoretical ideas presented here have relevance far beyond Norway, and may be employed in other contexts as well as in a comparative perspective. It is our hope that knowledge of how these processes work might contribute to the development of spheres of communication that are both sufficiently welcoming and open enough for people of all backgrounds to contribute and take part.

## **Conceptualizing free speech**

In its broadest sense, freedom of speech can be defined as the right to communicate in public unhindered by judicial, economic or social forces (Lipschultz, 2000). A long tradition, harking back to the classic works of Locke (2005) and Mill (1989), has defended the value of freedom of speech within a legal-normative framework (Greenawalt, 1989). On the one hand, freedom of speech is presented as a democratic good, by promoting truth, providing a check on the abuse of authority (especially government authority), and as a basic premise for deliberation and democracy. By affording people the opportunity to hear competing arguments, freedom of speech is thought to promote independent judgement, tolerance and individual autonomy. On the other hand, freedom of speech may also be argued for on the basis of the existence of inherent human qualities that warrant protection in themselves, such as rationality, autonomy, dignity and the right to self-determinacy (Waldron 2012). In liberal democracies the right to free speech is generally protected by national law, and also by international conventions such as the European Human Rights Convention2 and the UN Convention on Civil and Political Rights.3 A much-cited ruling by the European Court of Human Rights, states that freedom of expression is 'one of the essential foundations of [a democratic] society', and that the right applies also to information and ideas that 'offend, shock or disturb the State or any sector of the population.4

Still, freedom of speech is always subjected to limitations defined by alternative societal concerns or human rights. In this book we are particularly concerned with the boundaries to free speech that are drawn based on the concern for protecting groups and individuals from hate, prosecution and discrimination, what is often in legal terms defined as 'hate speech' (Wessel-Aas, Fladmoe, & Nadim, 2016, p. 19). There is also a set of national laws and international conventions that serve to define such a protection legally, as for example the EMC, article 17 and the UN Civil and Political Convention, article 20 (ibid.).

Crucial to the present book, however, is the perspective that such judicial frames provide necessary, but still not sufficient

<sup>2</sup> EMC, article 10, http://www.echr.coe.int/Documents/Convention\_ENG.pdf

<sup>3</sup> UN, International Covenant on Civil and Political Rights, article 19, http://www. ohchr.org/EN/ProfessionalInterest/Pages/CCPR.aspx

<sup>4</sup> *Handyside vs. Britain*, case no. 5493/72.

conditions for the exercise of free speech by citizens – or for agreement on its boundaries in social life. Hence, there is a need both to define a set of sociological perspectives for studying how freedom of speech is exercised and a need to garner empirical knowledge about the social processes which enable and restrain it. While there exists literature on polarization and debate within new and old public spheres (Brundidge, 2010; Stromer-Galley & Muhlbeger, 2009; Sunstein, 2003), growing literature on the occurrence of hate speech (Hawdon, Oksanen, & Räsänen, 2015; Nadim, Fladmoe, & Wessel-Aas, 2016), and studies that shed light on the public experiences of ethnic and religious minorities (Bangstad, 2015; Midtbøen, 2016; Midtbøen & Steen-Johnsen, 2016), each of these bodies of literature can only throw limited light on the particular social dynamics linked to the practice of freedom of speech in a given society. Our aim in this book is to provide a broader sociological lens on free speech and the dynamics of the public sphere.

In this book we use the terms 'freedom of speech' and 'free speech' interchangeably. However, as Terje Colbjørnsen shows in his chapter in this book (Ch. 6), freedom of speech may be seen as merely one aspect of the broader concept of freedom of expression, which includes not only written and oral speech, but, for example, the publishing of art and satirical cartoons. As most of the chapters deal with free speech issues, we have chosen to use this concept while pointing it out explicitly when we discuss freedom of expression in a wider sense. Importantly, all types of speech in any type of arena do not fall within the scope in this book. In line with Maussen and Grillo's approach (2014, p.175), we limit our study of boundary-drawing to speech that might be seen as *controversial*, which is expressed in *public forums* and which conveys *public matters*, whether they be political, religious or target values, norms or practices. Inspired by Bader (2014) we also tend to look at free speech as a matter of communication, implying an interactive perspective in which the speaker, the audience and the arena where a statement is uttered matter for how it is interpreted and acted upon socially (2014, p. 322). In this respect it is particularly important to distinguish between public, semi-public and private arenas. Even though our main focus is on public and semi-public spheres, we believe that the social processes that occur in these arenas interlink and mutually shape each other, and some of the chapters in the book are particularly concerned with these links (see especially chapters 3, 8 and 9).

The advantage of this project is to combine studies of how boundaries of free speech are established and maintained on an institutional level (media, politics) with studies of individual experiences of boundaries to free speech, both in the population at large and in very different groups at the margins (ethnic and religious minorities, immigration critics).5 The design allows us to examine the specific dynamics of boundary-making related to different groups and topics, but also to seek out more general social mechanisms through which free speech is being restricted. Conceptualizing the public sphere as a locus of 'boundary struggles' indicates that we view boundaries as shifting rather than permanent, and that they are objects of contestation; between institutions and individuals, between different groups in society, and as reflexive processes within individuals who are exposed to or challenge them.

Boundaries of free speech are drawn constantly, but appear particularly potent and contested in some areas and for certain groups. In this book we center on immigration, culture and religion – as key areas of public debate, and as lines of demarcation.

<sup>5</sup> Please consult the appendix for a detailed description of the survey methodology.

Since the Rushdie Affair, and, even more, since the Mohammed cartoon controversy in 2005/6, debates about free speech have come to be intertwined with discussions on religious freedom and concerns for the rights of specific ethnic, cultural or religious groups (Favret-Saada, 2015). In these debates, freedom of speech is, from one perspective, viewed as a problem for minority groups, as it allows the media to publish texts and drawings which may be considered offensive or blasphemous. From another perspective, freedom of speech is seen as not wide-ranging enough, since complaints about offensiveness and blasphemy may keep critical voices and satire from being published. As political scientist Erik Bleich has observed, all liberal democracies struggle with the dilemma of preserving the freedom of their citizens while simultaneously combating racism (Bleich, 2011). In this book, we analyze how media actors and politicians strike this balance, and how ordinary citizens – in the mainstream and at the margins – experience the opportunity to engage in matters important to them.

## **A sociological perspective on boundaries of free speech**

The literature on free speech originates from philosophy and is still dominated by legal-normative perspectives. Ever since Habermas ([1962] 1989), however, theories of the public sphere have formed an important element in the sociological tradition and in recent years a range of middle-level theories applicable for empirical studies of the dynamics of public debates have been developed. In this book, we build on this tradition by departing from the conventional focus on freedom of speech as a corollary to explicit legal and normative principles, and rather bring in sociological theories concerned with social norms, group identity and power in order to understand how the public sphere – and thereby principles of free speech – function in practice.

We argue that sociological perspectives may contribute to a more overarching model of the role of social and symbolic boundaries (Lamont & Molnár, 2002) for principles and practices of free speech. Our analysis is inspired by Jeffrey Alexander's (2006) theory of boundary-formations in the civil sphere, in which praised values in liberal democracies, such as individual freedom (including freedom of speech), dignity and autonomy, play a key part. In contrast to a legal-normative approach, Alexander argues that such values should be regarded not as static entities, but as the results of continuous boundary struggles. These higher values are, on the one hand, expressed in the founding documents of democratic societies, like laws, constitutions and bills of rights, and may thus seem given. On the other hand, the very same rights are historical, cultural and social in the sense that they are tied to a long chain of Western philosophy, religious thought, social movements and political struggle.

According to Alexander, civil societies are contradictory and fragmented. They are created by social actors at a particular time in a particular place. Arbitrary qualities (e.g. gender, race, nationality) become transformed into necessary qualifications for inclusion in the civil sphere. Processes of establishing or maintaining community, or solidarity, will always be characterized by struggle and contestation between interests, and by a tension between the particular and the universal. The contestation is, as we see in current debates over freedom of speech, not over the ideals as such, but over *who* or *what* can be defined as their antithesis, threatening the higher values of freedom through their uncivil, evil and contaminating force. The advantage of Alexander's perspective is that it links the communicative processes that take place in the public sphere to core societal processes, that involve cultural identifications and struggles over interest and power. In his view, the civil sphere is seen as an independent societal sphere, but still as deeply intertwined with and limited by the relationship with the public and the private spheres, and dependent on a set of societal institutions. This broader view has the potential to inform the core discussion on how boundaries to freedom of speech are drawn, in the light of more fundamental societal struggles.

In the overall theoretical framework of this book, Alexander's theories are supplemented by a specific sociological theory of boundaries and boundary making. According to Lamont and Molnár (2002), symbolic boundaries are 'conceptual distinctions made by social actors to categorize objects, people, practices, and even time and space', while social boundaries are 'objectified forms of social differences manifested in unequal access to and unequal distribution of resources (material and nonmaterial) and social opportunities' (Lamont & Molnár, 2002 p. 168). In public debates, symbolic boundaries are used by individuals and groups in struggles over what are conceived of as legitimate and illegitimate positions and standpoints. At the same time, these struggles over symbolic boundaries may have social consequences in the sense that they can take on a constraining character which excludes certain groups and points of view from public debates. Importantly, symbolic and social boundaries are created in different societal spheres and levels: between institutions and individuals, between different groups and organizations in society, and as reflexive processes within individuals who are exposed to or challenge them. This means that we need to study boundary contestations within these different contexts, and as seen from the perspective of a variety of groups with different degrees of formal and informal power.

Based on this basic perspective, we establish a threefold approach in this book. First, we believe that it is important to look into the institutions that hold the power to actually draw up and define the boundaries of free speech, explicitly or implicitly. Institutional elites, like politicians and editors, do have this type of influence, the first through legislation, and the latter as central gatekeepers of public debate (Gans, 1979). In liberal democracies, legislation and journalism alike are deeply committed to principles of pluralism and freedom of speech. Concomitantly, politicians and media professionals are engaged in boundary-work that is continuously challenged with regard to where the limits actually need to be drawn. Their mandate is, in different ways, to identify the types of speech acts that in form or content are deemed unacceptable or illegitimate, and hence should be excluded from the public sphere. The theoretical framework offered in this book enables an analysis which studies the intersection of explicit normative values of liberty with practices and decisions that take place in a social reality where an opinion climate and power relations are indeed strongly present.

Second, a boundary-making perspective on free speech means mapping out how different groups and sub-sets of groups experience boundaries or control mechanisms limiting their participation in the public sphere. We also focus on the dynamic character of such boundaries, asking how and why they change. In this, we are inspired by Richard Alba's distinction between bright and blurred boundaries (Alba, 2005). Although writing from the perspective of changing ethnic boundaries in Europe and the United States, Alba's distinction is useful for studying boundary-making in the public sphere more generally: When boundaries are bright, what is conceived of as illegitimate positions in, for example, debates over immigration, are hardly up for discussion. In contemporary contexts, however, boundaries between legitimate and illegitimate points of view are often not agreed upon. Rather, the boundaries are blurred, ambiguous and open to negotiation. Studying how this boundary-work plays out in public debates and how it gradually changes the demarcation line between insiders and outsiders, and legitimate and illegitimate points of view, is a task for sociological research.

Finally, boundaries may be drawn, not only by actors or institutions external to the individual, but by individuals themselves, as part of processes of inclusion or exclusion from the public sphere. Shared norms, beliefs and attitudes within a specific social culture create solidarity, integration, identity and belonging. Yet at the same time, these shared norms and attitudes can lead to the suppression of freedom of speech, because each person risks being excluded from the community if they challenge what is generally accepted. In order to shed light on such processes from the individual's point of view, several of the chapters in this book also draw on the theory of the spiral of silence, originally formulated by Noelle-Neumann (1974). This theory is based on the idea that individuals constantly relate to the opinions of others and that they adjust their behavior and their own opinions according to what is perceived to be the majority opinion (in a group or in a particular setting). Based on this theoretical perspective, a spiral of silence arises when individuals do not dare to deviate from the majority attitude and express unpopular opinions. The reason for this is fear of isolation, and a risk of being excluded from the communities one belongs to or feels part of. A crucial point in this theory is that individuals may misperceive the opinion climate, since they often do not have sufficient information about people's real opinions, and since some viewpoints are given more space than others in public debate.

The concepts of boundaries and of boundary struggles thus form the theoretical nexus of the present book. In our view these concepts add to our sociological understanding of how the conditions for free speech are formed in society, by looking at the institutional, group and individual levels. In this perspective, boundaries of free speech are the results of normative control and different actors' power to define where the boundaries are to be set, but always also as objects of cultural construction and struggle and thus continuously changing.

# **The Norwegian context**

This book offers a novel theoretical perspective on free speech and employs it on a rich collection of data, and we believe the insights offered are relevant in a range of contexts. Still, the analyses provided in the following chapters are based on empirical data from the specific Norwegian context which needs to be briefly outlined. We start by describing how free speech legislation in Norway has developed. Next, we describe the Norwegian history of immigration and the current composition of the immigrant population, and show how immigration has come to be a key political issue in public debates through the changing role of the right-wing Progress Party. Finally, we highlight important characteristics of the Norwegian media landscape, including the changing patterns of participation that we have witnessed over the past decade.

# Free speech in Norway – historical development and current status

In the Norwegian Constitution from 1814, the right to the Freedom of Print was established by law, which implied a ban on

#### chapter 1

pre-censorship (NOU 1999 p. 27, 3.3). The paragraph included particular protection of political utterances, as it declared that any person might raise free criticism of the state and other objects. The formulation of the paragraph can be considered as liberal and progressive for its time, and reflected the need to build the foundations for a new nation and a new societal order (ibid.).

Although historically not the case, today all Norwegian citizens are formally granted the same rights – including the right to free speech. Freedom of speech is protected in the Norwegian Constitution (§ 100), in the European Convention on Human Rights (Article 10) and in the UN Convention on Civil and Political Rights (Wessel-Aas et al., 2016). When §100 was voted on in 2004 it was intended as a general strengthening of freedom of speech, which had been relatively unchanged in the Constitution since it was enacted in 1814. Most centrally, all forms of pre-censorship were abandoned,6 the role and responsibility of the state in ensuring the conditions for a 'positive freedom of speech' (Kenyon, 2014) through open and pluralist media was emphasized (NOU 1999 p. 27), and the abolition of the law against blasphemy was suggested. The latter change in the law was voted on at a later point, and first came into effect in 2015 after the attacks on *Charlie Hebdo* in Paris (Steen-Johnsen, Fladmoe, & Midtbøen, 2016). Norway also has a law against hate speech, revised in 2015. The Norwegian Penal Code, section 185, protects against hateful or discriminatory speech about persons or groups of persons because of their a) skin colour or national or ethnic origin, b) religion or life stance, c) homosexual orientation, or d) disability.7

<sup>6</sup> The only exception being censorship of films in the case of protecting children and youth, by the imposition of age limits.

<sup>7</sup> Translation cited from http://app.uio.no/ub/ujur/oversatte-lover/data/lov-19020522-010-eng.pdf

Surveys of attitudes towards free speech in the Norwegian population indicate general support for the principle of free speech, but with some reservations (Steen-Johnsen et al., 2016). Criticism of religion is far more accepted than criticism of religious and ethnic minorities. There is also a tendency to want strong social sanctions for racist speech or speech that is critical of ethnic minorities, especially when uttered in social media (Steen-Johnsen et al., 2016 37). In accordance with the so-called 'balance of harms' approach to free speech (Waldron, 2012), there is thus a general tendency to balance the value of freedom of speech against other values, such as protecting minority groups (Steen-Johnsen & Enjolras, 2016).

### The immigration context

Norway is often thought of as ethnically and religiously homogeneous. To some extent this is the case. Norway has historically been a country of emigration, mainly due to the substantial number of Norwegians emigrating to the United States in the 19th and early 20th centuries, and did not become a net immigration country until 1967, when large-scale immigration from outside Europe became a permanent phenomenon (Brochmann & Kjeldstadli, 2008). However, Norway has always had an indigenous Sami population, and for centuries Jews, Kvens, Forest Finns, Rom and Romani have been part of Norwegian society as relatively small ethnic minority groups immigrating to the country at different points in history. Several of these groups have been targets of aggressive state assimilation policies, and partly as a compensation for this treatment these five groups were granted status as national minorities in 1999 (Brochmann, 2002; Lund & Moen, 2010).

In the late 1960s and early 1970s, Norway received a substantial number of labour migrants from countries such as Pakistan, Turkey, Morocco and India, starting a process of ethnic and religious diversification that has continued ever since. Despite a moratorium on labour migration, introduced in 1975 and made a permanent policy measure in 1981, the immigrant population has steadily grown through humanitarian migration, family reunification and family establishment, and since the EU enlargements in 2004 and 2007 labour migration has again been the main source of immigration to Norway (Brochmann & Kjeldstadli, 2008). In January 2015, almost 16 percent of the Norwegian population was either immigrants or born in Norway with immigrant parents. The largest immigrant groups are currently from Poland, Sweden, Lithuania and Somalia. Among the children of immigrants born in Norway, individuals with Pakistani origin make up the largest group, followed by descendants of Somali and Iraqi immigrants (Egge-Hoveid & Sandnes, 2015). Today, immigration is an established reality in the Norwegian context. As in most other liberal democracies in the world, however, it is also a source of constant conflict in the political and public spheres.

### The political context

Norway is a small, stable and relatively consensual democracy, which ranks high when it comes to voter turnout and trust in institutions (Allern, Heidar, & Karlsen, 2016; Lijphart, 2013). Elections are based on proportional representation, within a multi-party system. Parliamentary democracy emerged in the late 19th century, and in the mid part of the 20th century Norway had one of the most stable party systems in the world, consisting of a conservative (the Conservatives), a social-liberal (the Liberal Party), a social democratic (the Labour Party), an agrarian (the Centre Party) and a Christian (the Christian People's) party (Allern et al., 2016, p. 36). From World War II to the end of the 1960s the social democratic Labor Party was the dominant party, and governed alone in an unbroken line until 1963. From this period on government has alternated mostly between minority governments in several combinations, either single-party Labour, centrist-right or centrist (2016 p. 38).

The Progress Party, founded in 1973 as a right-wing protest party against the growth of bureaucracy, state intervention and tax levels, has been important in shaping the Norwegian political debate about immigration (Hagelund, 2003). From a situation of marginal support in its early years, the Progress Party gradually gained importance when immigration came to be a contested political issue in the late 1980s, increased its support to above 20 percent after the turn of the millennium, and saw its electoral breakthrough in 2009 (Jupskås, 2015). The party experienced a decrease in the elections following the July 22, 2011 terror attacks (Allern et al., 2016 p. 38), which has been interpreted as a reaction to the relationship between the right-wing ideology of the perpetrator and the Progress Party (Bergh & Bjørklund, 2013). Nonetheless, after the Parliamentary election in 2013 the Progress Party entered the government in a minority coalition led by the Conservative Party.

The Progress Party is often seen as the Norwegian equivalent to right-wing parties in Europe, such as the Danish Popular Party, the Dutch Party for Freedom, and the National Front in France. Importantly, however, the Progress Party itself has rejected similarities to these other parties and has regularly turned down invitations for collaboration (Hagelund, 2003). This is probably an important explanation of why the party has been able to build political legitimacy in a country characterized by high affluence, high social and institutional trust, and comparatively low levels of anti-immigration attitudes. By denying labels and linkages related to fascism or right-wing extremism, the Progress Party has avoided connections to European right-wing parties while simultaneously emphasizing its opposition to immigration and multiculturalism (Jupskås, 2015). The ability to build political legitimacy and become part of the political establishment does not prevent the Progress Party and its political ideas from remaining a point of great contestation in Norwegian public debate. The party has continued to play a key role especially in debates about immigration and integration, domains which it has had political control of since entering the government with the Conservatives in 2013.

### The media context

Norway has a far-reaching freedom of information act (1970, revised 2009) to keep government agencies transparent and open to scrutiny. The objectives of Norwegian media policy have been to increase freedom of expression, improve access to information, and improve equality of access to information (NOU 1999 p. 27; NOU 2013 p. 4; Rolland, 2008). This relates directly to the requirement in the Norwegian constitution that state authorities must create conditions that facilitate an open and enlightened debate (§100).

The co-existence of commercial media, media with roots in civil society and political groups, and public service media, make the Norwegian media system representative of a Nordic democratic corporatist model (Hallin & Mancini, 2004). However, the Nordic media model has moved towards a commercial, liberal model in recent years, and this also applies to Norway (Allern & Blach-Ørsten, 2011; Nord, 2008; Syvertsen, Enli, Mjøs, & Moe, 2015). In line with global developments, Norwegian media have been deeply affected by the impact of digitalization and the challenges it poses to traditional business and organization models, and the industry is currently going through profound processes of adaptation and downscaling (Steen-Johnsen, Ihlebæk, & Enjolras, 2016). Digitalization has, for example, influenced how broadcasters work strategically to keep and move audiences between different platforms and services (Ihlebæk, Syvertsen, & Ytreberg, 2014). So far the impact on citizen news consumption, fragmentation and polarization remains unclear, but the public broadcasting still has a strong position and news reading remains at a high level, even if carried out on new platforms (Steen-Johnsen et al., 2016).

Contemporary Norwegian media are increasingly diverse, with high levels of social media use, high internet use per capita, and a decentralized media structure (Syvertsen et al., 2015). Comparatively, Norway is characterized by high access to technology. In 2015, 94 percent had used a PC at home during the past three months and 89 percent had used the internet to read or download newspapers and magazines. 97 percent of the population was connected to the internet at home, reaching almost 100 percent among people under 45. 8

Even though there are limited digital divides when it comes to access to digital technology, there is still a question as to whether divides exist in relation to use (Hargittai, 2010). Previous research has demonstrated that those who participate in public debates and political activities through social media are younger and less educated than those participating offline (Enjolras, Karlsen, Steen-Johnsen, & Wollebæk, 2013). Also, gender differences persist; women are less active than men in digital debates on most platforms, except on Facebook (ibid.). Further,

<sup>8</sup> Numbers collected from Statistics Norway: https://www.ssb.no/teknologi-og-innovasjon/statistikker/ikthus/aar/2015-10-01

a set of surveys carried out within *The Status of the Freedom of Speech* project showed that ethnic minorities were as active as the majority in expressing their opinions through social media, but unlike the majority population ethnic minorities are more likely to experience negative comments targeted at core identity features such as their ethnic, national or religious background (Midtbøen & Steen-Johnsen, 2016; Staksrud et al., 2014; see also Ch. 7 and 8 in this book).

Immigration has been debated intensely in Norwegian media in recent decades. The media coverage has been characterized as polarized and dominated by stereotyped representations of immigrants (Eide & Simonsen, 2007; Lindstad & Fjeldstad, 2005). More recent research suggests that whereas the coverage is still frenzied and dominated by single news stories rather than systemic analysis, the debate has become more inclusive: Stories with different perspectives, sources, debaters and to some extent reporters with a minority background have entered mainstream media, and are gradually challenging the majority dominance (Figenschou & Beyer, 2014; Midtbøen & Steen-Johnsen, 2016; Thorbjørnsrud & Ustad Figenschou, 2014). In other words, while immigration continues to be a major source of conflict, immigrants and their descendants are increasingly also active participants in public debates.

## **An overview of the book**

The book consists of ten chapters, including this introduction. The first three chapters paint a broad picture of how boundaries of free speech are drawn, based on population representative survey data. In the next two chapters we zoom in on the media, both from the production and the content side. The final three empirical chapters of the book delve more deeply into the processes where boundaries are drawn and enacted, through qualitative studies of the experiences of free speech among young politicians, ethnic and religious minorities and immigration critics, respectively. In the last chapter of the book, we develop a theoretical lens for studying free speech and the dynamics of the public sphere as social phenomena, as well as applying this theoretical framework on the empirical findings presented in this book. In what follows we present the chapters of the book in more detail.

Chapter 2, authored by Marjan Nadim and Audun Fladmoe, focuses on the extent and consequences of personal experiences of hate speech and other unpleasant comments in social media. As the authors note, hate speech brings discussions of freedom of speech and social boundaries to the fore, as any ban on hate speech is a limit to free speech, while hate speech simultaneously is a tool for creating and reinforcing boundaries and hierarchies between groups. Drawing on a large-scale survey among Norwegian adults, Nadim and Fladmoe find that people of immigrant backgrounds are more exposed to hate speech directed towards legally protected grounds, but that the majority population is as equally exposed as immigrants to other, more general unpleasant comments. However, hate speech directed towards legally protected grounds have more encompassing consequences for women and people of immigrant backgrounds, suggesting that hate speech may have negative democratic consequences by silencing certain groups.

In chapter 3, Audun Fladmoe and Kari Steen-Johnsen discuss the extent to which individuals self-censor when engaging in public debates. The theoretical backdrop of the chapter is Noelle-Neumann's (1974) theory of spirals of silence, which predicts that individuals will tend to adjust to what they perceive as dominant public opinion, and will be less willing to speak out if they perceive that they are part of a minority. Examining the case of religious cartoons, Steen-Johnsen and Fladmoe find that people with opinions incongruent with the dominant positions held by the general public, are less willing to discuss the publication of religious cartoons, but also that spiral of silence mechanisms seem to be stronger in private than in public arenas. This result indicates the significance of peer effects when engaging in controversial issues, and suggests that symbolic boundaries work to reinforce majority positions in both public and private discussions.

In chapter 4, Hallvard Moe, Kjersti Thorbjørnsrud and Audun Fladmoe explore how Norwegian citizens perceive the credibility of journalists to provide mediating information from the centres of power to the public, and to present issues in a fair way. The authors find that confidence in the impartiality of journalists in general is low, and that political party preferences and attitudes towards immigrants and immigration, along with general trust in the media, are important indicators of perceptions of journalistic bias. These findings represent a major challenge to the news media's position in a democratic society, and may be a signal of changing expectations on the part of the public, and a changing role for the media as channels between the public and the rulers.

The next chapter shifts focus from this challenge posed by the citizens to the editors' and journalists' point of view. These professions draw the lines of free speech every day, choosing which issues deserve attention and to whom to give a voice, as well as which issues, opinions or groups are not granted the privilege (or burden) of media attention. The chapter authors, Karoline Andrea Ihlebæk and Ingrid Thorseth, study how the editors of Norwegian news media deal with issues of hate speech and racism, and how they work to achieve balance and diversity in a polarized debate climate. The authors claim that even though social media have challenged the traditional media's position, opinion editors still represent an important type of gatekeeper who guards the legal and symbolic boundaries of public debate.

In chapter 6, Terje Colbjørnsen provides an analysis of newspaper debates on freedom of expression and freedom of the press in Norway over a twenty year period. Drawing on both quantitative and qualitative data, Colbjørnsen shows that the cartoon controversies in 2005/6 and 2015 stand out as 'critical moments' in the freedom of expression discourse. He also analyses the dominant arguments present in the debates, and discusses which types of argument are given validity. According to Colbjørnsen, historical-philosophical arguments appear as more legitimate in the media discourse than emotional arguments, which challenges the idea of a public sphere dominated by emotions and sentiments of 'offendedness'.

The next three chapters delve more deeply into subjective experiences and social processes related to free speech, based on in-depth interviews with individuals from three different groups. In chapter 7, Arnfinn H. Midtbøen centres the attention on youth politicians. Politicians have the power to influence which expressions are to be defined as legal and illegal, by virtue of their political involvement. However, they are also visible actors who face boundaries of free speech both as individuals and as representatives of particular organizations. Midtbøen explores how the leaders of Norway's political youth organizations experience being public figures in Norway, and how they deal with the 'cultures of expression' in different parties. He distinguishes between external and internal barriers to free speech, and discusses the implications of these barriers for the politicians' attitudes to free speech regulation and, more broadly, for the future of democracy.

Following a similar route, in chapter 8 Marjan Nadim explores the conditions for participating in public debate for individuals with a religious or ethnic minority background, and in particular how these individuals deal with their ascribed role as a representative of the group they are perceived to belong to. Drawing on in-depth interviews with people with an ethnic or religious minority background, Nadim shows how representation is a two-way street. On the one hand, ethnic and religious minorities are ascribed certain values and points of view by the media and mainstream society. However, being cast as representing specific communities in the public sphere also entails a question of 'internal legitimacy'; the people you are supposedly representing will to varying degrees accept you as their representative in public.

Importantly, those who are at the margins of society may be perceived in different ways. Often research in this area focuses on typical target groups, like ethnic, religious or sexual minorities. But in chapter 9, Kjersti Thorbjørnsrud explores experiences with public debates from the perspective of immigration critics; that is, people who defend more restrictive immigration policies and are concerned about the negative impact of immigration on society. Based on in-depth interviews with individuals who have influential positions in Norwegian public debate, as well as with people who are less visible and more active on social media, Thorbjørnsrud shows how immigration critics experience stigma, exclusion and marginalization in Norwegian society.

Although they are informed by theoretical concepts of symbolic boundaries, spirals of silence or competing philosophical ideas about justice, the above-mentioned chapters are first and foremost *empirical* in scope, showing how the boundaries of free speech in the Norwegian context are defined and maintained, but also experienced and challenged. In the final chapter, Bernard Enjolras aims at laying the foundation for a sociological perspective on free speech by placing the findings of the preceding chapters in relation to a broader theoretical framework, thus proposing a synthesis between sociology of the public sphere and sociology of social boundaries. This final chapter of the book outlines a conceptual framework, which enables us to recast the empirical findings presented in the previous chapters, and to interpret them in terms of the processes of symbolic boundary struggles in the public sphere.

### **References**


Feinberg, J. (1985). *Offence to others*. New York: Oxford University Press.

Figenschou, T. U., & Beyer, A. (2014). Elitene, minoritetene og mediene

– Definisjonsmakt i norsk innvandringsdebatt. *Tidsskrift for Samfunnsforskning*(01).

Greenawalt, K. (1989). Free Speech Justifications. *Columbia Law Review, 89*(1), 119–155. doi:10.2307/1122730


Hallin, D. C., & Mancini, P. (2004). *Comparing media systems : three models of media and politics*. Cambridge: Cambridge University Press.

Hargittai, E. (2010). Digital na(t)ives? Variation in internet skills and uses among members of the 'net generation'. *Sociological Inquiry, 80*(1), 92–113.

Hawdon, J., Oksanen, A., & Räsänen, P. (2015). Online extremism and online hate. Exposure among Adolescents and Young adults in Four Nations. *Nordicom Review, 37*(3-4), 29–37.

Ihlebæk, K. A., Syvertsen, T., & Ytreberg, E. (2014). Keeping Them and Moving Them: TV Scheduling in the Phase of Channel and Platform Proliferation. *Television & New Media, 15*(5), 470–486. doi:10.1177/1527476413479676

Jupskås, A. R. (2015). *Persistence of Populism. The Norwegian Progress Party, 1973–2009.* (PhD. Political Science), University of Oslo, Oslo.


Lijphart, A. (2013). *Patterns of Democracy. Government Forms and Performance in Thirty-Six Countries*. New Haven: Yale University Press.

Lindstad, M., & Fjeldstad, Ø. (2005). *Av utenlandsk opprinnelse: nye nordmenn i avisspaltene*. Kristiansand: IJ-forl.

Lipschultz, J. H. (2000). *Free Expression in the Age of the Internet*. Westview Press: Colorado.

Locke, J. (2005). *Two Treatises on Government: a Translation Into Modern English*. Manchester: Industrial Systems Research.


### Chapter 2

# Silenced by hate? Hate speech as a social boundary to free speech

*Audun Fladmoe, PhD, Senior Research Fellow, Institute for Social Research Marjan Nadim, PhD, Senior Research Fellow, Institute for Social Research*

Hate speech is central in discussions of the legal and social boundaries of freedom of speech. On the one hand, any ban on hate speech is a limitation of free speech. On the other hand, hate speech may in itself pose a social boundary on free speech through inciting fear and silencing individuals. Based on a large-scale survey among Norwegian adults, the chapter studies experiences of hate speech and other unpleasant comments in social media, and whether hate speech discourages people from voicing their opinions. The results suggest, first, that people of immigrant backgrounds are more exposed to hate speech directed towards legally protected grounds, but that the majority population are as equally exposed as immigrants to other more general unpleasant comments in social media. Second, the results suggest that more general unpleasant messages may have consequences similar to those of hate speech, in terms of willingness to voice opinions publicly. However, women and people of immigrant background are more affected by hate speech directed towards legally protected grounds than other groups. The chapter thus demonstrates how hate speech may have negative democratic consequences by silencing certain groups.

## **Introduction**

Free speech and the protection of minorities are not usually incompatible values; nevertheless they can come in conflict. Liberal democracies constantly engage in delimiting the legal boundaries between preserving freedom of speech and combating racism, harassment and discrimination (cf. Bleich, 2011). And hate speech – persecutory, hateful or degrading speech directed towards certain group attributes – is a core issue in discussions of the boundaries of free speech.

Legislating against hate speech and harassment means that some utterances are deemed unacceptable and unlawful. This can be problematic because it entails a constraint on freedom of speech and can potentially limit public discussions through a so called *chilling effect* – i.e. that individuals might be discouraged from engaging in legitimate political debate by threat of legal sanction (cf. Gelber & McNamara, 2015 p. 640). A further argument against hate speech regulation, and in favour of allowing such utterances, is that discriminatory and hateful speech is best met by counter-arguments in public debate. Thus, freedom of speech can in itself be seen as a tool to combat hate speech and discrimination through what can be called the 'cleansing function' of public debate (NOU, 1999 p. 10).

On the other hand, hate speech can have negative consequences for society and the targeted individuals. Allowing hate speech in public debate can contribute to making such rhetoric appear more legitimate and acceptable, paving the ground for even more hate speech. Furthermore, hate speech can in itself have a discouraging effect on the exercise of free speech. One purpose of hate speech is to incite fear in the groups targeted. Hate speech works to guard and reinforce boundaries and hierarchies between groups, and to remind those who are considered 'different' or 'other' of where they belong (cf. Perry, 2001). Experiences with, or fear of, hate speech can shape individuals' propensity to speak their mind, and make targeted individuals or groups more cautious in expressing their views and making themselves visible. A potential consequence of hate speech is that certain groups are silenced, thereby excluding particular voices and viewpoints from public debates. Thus, while legislation against hate speech poses a *legal boundary* on free speech, hate speech in itself can, in effect, also function to limit the individual's exercise of the right to free speech through instilling fear and causing withdrawal from public debate for those targeted. In this sense, hate speech can represent a *social boundary* for free speech.

The aim of this chapter is to study experiences with different forms of hate speech in social media, and whether such experiences discourage people from expressing opinions publicly. So far, discussions about freedom of speech and hate speech have largely been legal and normative, and there has been a remarkable lack of empirical contributions (Bleich, 2011). This chapter takes a sociological and empirical approach to hate speech, and speaks to the overall theoretical framework of this book by analyzing how hate speech can function as a *social boundary* for the individual expression of opinions, and how these boundaries may be different for different groups. The chapter draws on a large-scale population-based survey with more than 5000 respondents, carried out in June 2016 in Norway. The large number of respondents in the survey enables us to scrutinize variations among different subgroups of the population.

The rest of the chapter is structured as follows: We begin by discussing the concept of hate speech, before reviewing previous research on who are targeted by hate speech, and the potential consequences of hate speech for individuals, groups and society. Next, we present our data and empirical analyses, and finally we discuss the implications of our results.

### **What is hate speech?**

Hate speech is a contested term, and there is no shared definition of the concept (Gagliardone, Gal, Alves, & Martinez, 2015; Gelber & McNamara, 2016). Still, definitions of hate speech typically focus on two key features: the tone or style of the message, and what ground(s) the message is directed towards. Hate speech can be defined as persecutory, hateful, or degrading speech that is directed towards an individual or a group on the basis of certain (perceived) group attributes (Boeckmann & Turpin-Petrosino, 2002; Gagliardone et al., 2015 p. 10; Lawrence III, Matsuda, Delgado, & Crenshaw, 1993). Not all groups are included in the concept; it is usually reserved to cover hateful speech directed towards attributes associated with members of historically oppressed (minority) groups (cf. Lawrence III et al., 1993).

Hate speech reflects negative stereotypes, prejudice and stigma, and is based on perceptions of boundaries and hierarchies between groups. It builds on a rhetoric of exclusion, fear and contempt for individuals and groups that are deemed to be different, and can be seen as a way of 'doing difference' (cf. Perry, 2001). The purpose is to guard and highlight the boundaries between groups, and remind groups and individuals who are seen as 'other' of their rightful place in the social hierarchy (Nilsen, 2014; Perry, 2001). Thus, in understanding and

silenced by hate?

defining what hate speech is, it is central not only to look at the rhetoric and tone of the message, but also at what *grounds* the speech is directed towards.

Historically there has been a high degree of acceptance of racist expressions and discrimination, but after World War II, and in particular since the 1960s, the general trend is that European countries have brought hate speech under increasingly more stringent regulation; the USA remains one of the very few countries to resist the trend to ban hate speech (Bleich, 2011; Parekh, 2006). National and international legislation employ different definitions of hate speech (Gagliardone et al., 2015). The Norwegian Penal Code section 185 protects against hateful or discriminatory speech about persons or groups of persons because of their a) skin colour or national or ethnic origin, b) religion or life stance, c) homosexual orientation, or d) disability. Thus, in Norway, for an utterance to be defined as hate speech in judicial terms it must be directed towards one of these group-based identities (also referred to as *protected grounds*). This does not imply that hateful utterances directed towards members of others groups are necessarily lawful, but that these must rather be tried in relation to other laws, such as laws on threats, discrimination, defamation, etc. (see Wessel-Aas, Fladmoe, & Nadim, 2016).

In popular debates hate speech is often understood in a broader sense than legal definitions, and the concept is used to refer to a wide spectre of phenomena, from online bullying and aggressive and intolerant statements in public, to racism and threats towards individuals (Gagliardone et al., 2015; Sunde, 2013 p. 42; Waldron, 2012 p. 34). For a sociological approach to hate speech that aims to understand and empirically study the phenomenon, it is fruitful to expand the understanding of hate speech from its strict legal understanding. First, it is methodologically challenging to restrict empirical studies of hate speech to a legal definition. Second, and more substantially, the distinction between criminal and lawful speech is not clearcut, and expressions that are not covered by the legal definition can also have negative consequences for individuals and society at large.

The definition of hate speech can be extended from the legal version regarding both of the key features in the definition. First, a broader understanding of hate speech can include other grounds than those protected by law. The grounds that are offered legal protection against hate speech are a reflection of historical struggles for group recognition, but they do not necessarily mirror who is most exposed to hate speech or the consequences such speech has for different groups. The creation of group boundaries and hierarchies is an ongoing process, and there are ongoing debates about whether other groups should be included in definitions of hate speech (Jenness, 2003; Maher, McCulloch, & Mason, 2015; McPhail, 2002). In the popular understanding of hate speech, the term is often not restricted to speech directed towards group attributes at all. Second, a legal understanding of hate speech necessarily needs defined criteria distinguishing expressions that are sufficiently harmful in their tone and style to be considered unlawful from those that are not. Such criteria can be difficult to adhere to in empirical investigations. Third, Section 185 in the Norwegian Penal Code limits its definition of hate speech to speech that is expressed in public or in the presence of others. Empirically, it is also relevant to include direct messages to individuals. We wish to emphasize that while we are arguing for employing a broader understanding of hate speech in empirical research, this is not in itself an argument for expanding the regulation or legal definition of hate speech.

In this chapter, we reserve the term hate speech for hateful expressions that are directed towards potentially vulnerable groups. In the empirical analyses we will measure hate speech using different definitions of the phenomenon, distinguishing between a 'protected grounds' definition that is restricted to hateful speech directed at the grounds protected by the Norwegian Penal Code, and an 'expanded definition' also including other characteristics related to people's identities. We also measure experiences with hateful messages directed towards other types of grounds, further from the conventional understanding of hate speech.

### **Targets of hate speech**

As mentioned above, hate speech is understood as persecutory, hateful and degrading speech directed towards historically oppressed groups or individuals, based on their (perceived) group attributes (cf. Lawrence III et al., 1993). Hate speech can be understood as an expression of prejudice, stereotypes and perceptions of differences and hierarchies between groups (cf. Chakraborti & Garland, 2015; Perry, 2001). Thus, the targets of hate speech are first and foremost members of minority groups. But also more general (majority) group attributes, such as gender, may be targeted.

There has been little empirical research that specifically examines experiences and the prevalence of hate speech. One of the few studies that provides information about which group attributes hate speech is directed towards, is Hawdon and colleagues' (2015) international comparison of experiences with hate speech among young adults in the USA, the UK, Germany and Finland. They asked survey respondents whether they had witnessed hate speech online, and if so, what grounds the hate speech was directed towards. Hatred towards ethnicity and sexual minorities were the most common forms of hate speech observed in all four countries. Ethnicity accounted for between 48 percent (Germany) and 67 percent (Finland) of the hate speech observed in the four countries. Religion was also high on the list in all the countries. Hatebase, a database that gathers instances of hate speech globally, similarly finds that ethnicity and nationality are the most common targets for hate speech, and indicates that there has been a clear increase in hate speech based on religion and class background (Hatebase, 2016). Hawdon and colleagues' comparative study further found large national differences regarding hate speech directed towards gender. Gender was a much more common ground for the observed hate speech in the UK than in the other three countries included in the study (Hawdon et al., 2015 p. 34).

Based on existing research and the insight that hate speech reflects prejudice, stereotypes and group hierarchies in society at large, our expectation is that people with immigrant backgrounds will be especially at risk for receiving hate speech. Thus, our first hypothesis is:

H1: People of immigrant background are to a larger extent exposed to hate speech than other individuals.

However, there is a difference between the groups of individuals having the most experiences receiving hate speech, and which grounds the hate speech is directed towards. For instance, studies of individuals' experiences with online harassment and with receiving unpleasant and degrading comments indicate that group differences in exposure to these phenomena are not necessarily large, but that the *content* of the comments received by different groups varies considerably (Midtbøen & Steen-Johnsen, 2016; Pew Research Center, 2014).

### **Hate speech as a silencing mechanism**

Hate speech is found to have a range of consequences for individuals, such as fear and other emotional symptoms, lowered selfesteem, loss of dignity, and withdrawal from the public –both physically and in terms of participation in public debate (Boeckmann & Liew, 2002; Boeckmann & Turpin-Petrosino, 2002; Eggebø, Sloan, & Aarbakke, 2016; Gelber & McNamara, 2016; Herek, Cogan, & Gillis, 2002; Leets, 2002; Midtbøen & Steen-Johnsen, 2016; Pew Research Center, 2014). All instances of hate speech will of course not have these consequences, but the empirical studies demonstrate that hate speech *can* produce such outcomes.

In this chapter we examine one possible consequence of receiving hate speech, namely discouraging people from voicing their opinions publicly. One puprose of hate speech is to incite fear in the groups targeted, and to remind those who are considered 'different' or 'other' of where they belong (cf. Perry, 2001). If hate speech works to silence its targets, it can be seen to pose a social boundary on free speech. Furthermore, if certain groups are more likely than others to refrain from voicing opinions publicly due to experiences with hate speech, hate speech is potentially a democratic problem. A precondition for an enlightened democratic debate is that all group-based interests are represented in public discourse (cf. Phillips, 2009).

The idea that hate speech is more harmful than other types of negative and unpleasant expressions, is part of the rationale for passing legislation against this specific type of speech. Also some researchers hold that hate speech can have more adverse consequences than other types of negative speech (Boeckmann & Liew, 2002; Herek et al., 2002). Boeckmann and Liew (2002) find that hate speech produces stronger emotional responses in the recipients than do other forms of degrading speech. Based on a study of sexual minorities, Herek and colleagues (2002) argue that even less severe expressions of hostility against minorities can be experienced as traumatic because minorities are very aware of the violence and injustice members of their group have been subject to. The argument is that since hate speech triggers the awareness of belonging to a vulnerable group, it incites more fear than other types of negative speech.

Furthermore, because hate speech is not only directed towards individuals, but is also – intentionally or not – targeted against groups, it can have consequences beyond the individuals targeted. Because the content of hate speech is based on certain group attributes of an individual, publicly expressed utterances also send a signal to other individuals with similar attributes (Bell, 1998; Kunst, Sam, & Ulleberg, 2013; Perry, 2014). For members of a minority group, perceptions of other members' experiences – and consequently knowledge about the risk of being subject to the same oneself – can incite fear, even if they themselves have no personal experiences with hate speech (Gelber & McNamara, 2016 p. 327; Perry, 2001).

Does receiving hate speech discourage people from publicly expressing their opinions? A Norwegian study found that, compared to the majority population, ethnic minorities are substantially more prone to become cautious about expressing their opinions after experiencing harassment. While 19 percent of the majority population reported that receiving unpleasant or degrading comments has caused them to be more cautious in expressing their opinions, 36 percent of respondents with immigrant backgrounds reported the same (Midtbøen & Steen-Johnsen, 2016). However, this study did not examine the significance of the *content* of the messages, i.e. whether hate speech directed towards legally protected grounds have more adverse consequences compared to messages directed towards other grounds.

Based on a review of previous research on the consequences of hate speech, our second hypothesis is:

H2: Hate speech directed towards legally protected grounds (i.e. skin colour, nationality, ethnicity, religion or life stance, homosexual orientation, or disability) has more adverse consequences, in terms of discouraging the expression of opinions publicly, than other types of negative speech.

### **Data and method**

We rely on a web-based survey, carried out in June 2016 as part of the project *Social Media in the Public Sphere* (SMIPS). The sample consists of 5054 respondents, drawn from TNS Gallup's access panel (response rate: 44.6 %). Members of this panel are recruited by means of random sampling through the National Register, and no self-recruitment is allowed.

A limitation of using survey data to study the prevalence of hate speech is that we rely on subjective assessments. Different respondents may understand what constitutes a 'hateful message' differently. In effect, the empirical results must be interpreted precisely as subjective assessments of hate speech. In the following we describe the variables used in the analyses.

### Dependent variable 1: Personal experience with hate speech

In order to assess personal experiences with hate speech, respondents were first asked if they themselves had received hate speech via social media, and second towards what grounds these messages were most often directed. In the survey, 'hate speech' (*hatefulle ytringer)* was defined as statements that are 'degrading, threatening, harassing, or stigmatizing', but the question did not specify any particular grounds. The term hate speech (*hatefulle ytringer*) does not function as a synonym for racist or discriminatory speech in the Norwegian context, as it more commonly does in the American context. Rather it is predominantly understood as containing very negative expressions, without necessarily being related to an individual's group attributes. Thus, the first question captures what respondents themselves perceive as hate speech in general terms, allowing for hateful utterances beyond the legal definition.

In order to be able to distinguish between different definitions of hate speech, the second question asked what the received hateful statements were most often directed towards. It was possible to select one or more attributes from a list of 13. The list included grounds protected by the Norwegian Penal Code (skin colour, nationality, ethnicity, religion, sexual orientation, and disability), in addition to other potentially relevant attributes such as gender, content of one's argument, political views, etc. In effect we can distinguish between hate speech directed towards grounds protected by the Norwegian Penal Code on the one hand, and other types of unpleasant messages perceived as hate speech, on the other.

## Dependent variable 2: Reluctance to express opinions

In order to assess reluctance to express personal opinions publicly, we rely on a follow-up question of whether experiences with receiving hate speech have caused the respondents to be more cautious in public debates: 'After experiencing hate speech, have you become more reluctant to express opinions publicly?'

### Independent variables

We include the same set of independent variables across different analyses: gender (female=1), age, education (university/ college=1), political left-right orientation (1-11), and propensity to share personal opinions on the internet (1: Never – 4: Often). Additionally, initial inspections of the data suggested that political ideology is not linearly related to receiving hate speech, but rather that the relationship is curvilinear – that people on the far left and far right are most likely to have received what they perceive as hate speech. In order to capture this relationship we include squared transformations of the left-right scale. Finally, we include a dummy variable distinguishing between the majority population and respondents with immigrant backgrounds (=1). Following the definition employed by Statistics Norway (see for instance Egge-Hoveid & Sandnes, 2015), this variable includes both people born abroad and people born in Norway of two foreign-born parents. Unfortunately, we have limited information about the country of origin of respondents with immigrant backgrounds. Descriptive statistics for the independent variables are summarized in Table 2.1.

As shown in Table 2.1 women, young people, and low education levels are somewhat underrepresented in the sample. We therefore employ sampling weights in all analyses.

## **Results**

We present our results in two steps. First, we estimate the number of people who have experienced hate speech, how the number varies according to different definitions of the phenomenon, and how the estimates vary among different subgroups. Second, we explore how different types of hate speech may discourage people from expressing opinions publicly.


**Table 2.1.** Independent variables. Descriptive statistics.

*Source: SMIPS (2016).*

### Experiences with hate speech

Table 2.2 displays the number of respondents who reported that they had experienced what they perceive as hate speech via social media, and what these messages were usually directed towards. The table distinguishes between respondents with immigrant and non-immigrant backgrounds.

The table shows that 7.2 percent of the full sample reported having received what they perceived as hate speech. However, the results suggest that the content of most of these messages falls outside conventional definitions of hate speech. Most of the reported messages are directed towards the content of one's argument, political standpoint and personality. Fewer respondents mentioned any of the legally protected grounds (ethnicity, nationality, skin colour, religion, disability, and sexual orientation). Each of these characteristics is mentioned by less than 1 percent of the total population. The fact that the majority of messages reported are not directed towards protected grounds, shows that the popular comprehension of the concept of 'hate


**Table 2.2.** Has received what was perceived as hate speech via social media – and what these messages were directed towards. Percent.

*Source*: *SMIPS (2016). Light blue shading: significant difference (p<0.05) between respondents with immigrant and non-immigrant backgrounds.*

NOTE: Weighted according to gender, age, and education.

speech' is broader than the legal definition (cf. Gagliardone et al., 2015; Sunde, 2013 p. 42; Waldron, 2012 p. 34). Thus, to fully capture how people experience hate speech (at least in the Norwegian context), it is necessary to employ a rather broad definition of the phenomenon.

With regard to respondents' immigrant background, the table shows that immigrants (10.7 percent) more often than non-immigrants (7.0 percent) report having experienced what they perceive as hate speech. Furthermore, the grounds the received hate speech is directed towards differs. More immigrants than non-immigrants report hate speech directed towards gender, nationality, religion, skin colour, and ethnicity.

In order to distinguish between different forms of hate speech and other unpleasant messages, we categorized the experiences according to three different definitions of the phenomenon.1 *Protected grounds* include experiences with hate speech directed towards grounds that are covered by Section 185 of the Norwegian Penal Code, i.e. ethnicity, nationality, skin colour, religion, disability, and sexual orientation. The second group (*expanded definition*) includes the same attributes as the first definition, but adds those who had experienced what they perceived as hate speech directed towards gender, personality, and appearance, which are all characteristics related to people's identities. Finally, in a third 'rest category' (*Other*) we grouped messages directed exclusively towards the content of the argument, political standpoint, occupation, education, other, and 'don't know'. This category thus includes comments that are further from the conventional understanding of hate speech. Table 2.3 sums up the share of non-immigrants and immigrants who reported having received what they perceived as hate speech, grouped by the three definitions.

The table shows that in total about 2 percent of the population have received what they perceive as hate speech directed towards protected grounds. When expanding the definition to include other characteristics related to personal identities, personality, gender and appearance (*Expanded definition*), the share

<sup>1</sup> This differentiation between different types of unpleasant expressions is based on which *grounds* or content the expressions are directed towards. We do not have information about the tone or style of the messages.


**Table 2.3.** Has received what was perceived as hate speech via social media. Different definitions. Percent.

*Source: SMIPS (2016). Light blue shading: significant difference (p<0.05) between respondents with immigrant and non-immigrant backgrounds.*

NOTE: 'Protected grounds' include religion, ethnicity, skin colour, nationality, sexual orientation, and disability. 'Expanded definition' includes in addition personality, gender, and appearance. 'Other' includes the content of the argument, political standpoint, occupation, education, other, and don't know. Weighted according to gender, age, and education.

reporting having received hate speech increases to 4.4 percent of the full sample. Finally, 2.8 percent of the full sample reported having received what they perceived as hate speech, but only directed towards other characteristics, such as the content of one's argument and political standpoint.

We see that a relatively large share of the reported experiences with hate speech fall outside a conventional (legal or academic) understanding of hate speech, as they do not refer to speech directed towards group attributes in any sense. Thus, a substantial share of what the respondents report as hate speech should rather be understood as more general unpleasant experiences with online harassment. This underlines the ambiguity of the concept of hate speech in the general public, and illustrates how subjective perceptions of hate speech are broader than the legal definition.

We hypothesised (H1) that people of immigrant background would be more exposed to hate speech than other individuals. With one exception, the results in Table 2.3 give initial support to this hypothesis. As would be expected, more immigrants (7 percent) than non-immigrants (1.6 percent) have received what they perceive as hate speech directed towards protected grounds. The difference between these two groups is reduced when expanding the definition to also include, gender, appearance and personality (7.8 vs 4.2 percent), but the difference is still statistically significant. However, considering unpleasant messages directed towards other attributes, the data suggests no difference between immigrants (2.9 percent) and non-immigrants (2.8 percent).

Descriptive statistics thus gave initial support to H1. The question is whether or not this relationship holds when controlling for other relevant factors. In order to test this we estimated two logistic regression models for each definition. Model (1) controls for gender, age, immigrant background, education, political ideology and political ideology squared. Model (2) adds propensity to share personal opinions on the internet. The dependent variable is 'has received [what respondents perceive as] hate speech' (1='yes', 0='no'). Odds ratios from the regression models are summarized in Table 2.4. Ratios above '1' indicate a positive relationship, while ratios below '1' indicate a negative relationship.

Controlling for a host of other factors, we see that the hypothesized relationship between experiences with hate speech and immigrant background is less clear-cut. Narrowing the definition of hate speech to *protected grounds*, there is a clear tendency that respondents with immigrant backgrounds are more exposed to hate speech compared to the majority population. Even controlling for propensity to share personal opinions on the internet (model 2), respondents with an immigrant background are almost four times (odds ratio=3.8) as likely as non-immigrants to have received hate speech directed


**Table 2.4.** Has experienced [what respondents perceive as] hate speech via social media. Logistic regression. Odds.

*Source: SMIPS (2016). Sig:* †≤*0.1* \*≤*0.05 \*\*≤0.01 \*\*\*≤0.001.*

NOTE: 'Protected grounds' include religion, ethnicity, skin colour, nationality, sexual orientation, and disability. 'Expanded definition' includes in addition personality, gender, and appearance. 'Other' includes the content of the argument, political standpoint, occupation, education, other, and don't know. Weighted according to gender, age, and education.

towards protected grounds. This is not surprising considering that several of these grounds – nationality, ethnicity, skin colour – are more relevant to the immigrant population than to the majority population.

If we expand the definition of hate speech to also include personality, gender and appearance, the difference between respondents with immigrant backgrounds and non-immigrant background is reduced. The odds coefficient is still positive,

#### chapter 2

indicating that immigrants are also somewhat more exposed to hate speech according to the expanded definition. But when we introduce propensity to share personal opinions on the internet (model 2), the difference is no longer statistically significant. This might be due to the fact that few respondents have experienced what they perceive as hate speech. Nevertheless, based on this survey we must conclude that the majority population and respondents with immigrant backgrounds are equally exposed to this expanded definition of hate speech.

Finally, considering unpleasant messages directed towards other attributes, the data suggests that immigrants are *less* exposed to such messages (odds coefficient is below 1). Again, however, the difference is not statistically significant.

In other words, if we only consider hate speech directed towards legally protected grounds the first hypothesis (H1) is clearly supported. However, if we instead employ wider definitions of hate speech, that are perhaps closer to the popular understanding of the concept, H1 is not supported.

The results in Table 2.4 also reveal some other noteworthy findings. Men are more likely than women to have experienced what they perceive as hate speech directed towards protected grounds, and also to have experienced unpleasant messages directed towards other attributes. However, the gender difference is insignificant when messages targeted at gender, appearance and personality are included (expanded definition). The reason is obvious: Women are more likely than men to have received what they perceive as hate speech directed towards gender as an attribute (see Table 2.2). Furthermore, young people are more likely than older people to have received hate speech, but level of education is more or less irrelevant. People who place themselves on the far ends of the political left–right scale are more exposed to hate speech compared to political moderates. This relationship does however disappear when we introduce propensity to share personal opinions on the internet (model 2). A probable explanation is that radicals, on either side of the political spectrum, are, on average, more politically active than moderates, leading them to engage in more heated discussions in social media. In general, propensity to share personal opinions is a very strong predictor for the likelihood of receiving hate speech. The odds of receiving hate speech directed towards protected grounds increases by about 3 for each increase on the four-point scale of propensity. Thus, the more active you are in debates on the internet, the more exposed you become to hate speech.

To sum up this first empirical part, about 7 percent of the sample reported having received what they perceived as hate speech through social media. These utterances were most often directed towards characteristics other than those covered by Section 185 of the Norwegian Penal Code, but rather directed towards the content of the argument, political standpoint and personality. About 2 percent have received hate speech directed towards legally protected grounds. When we expanded the definition to also include gender, appearance, and personality, 4-5 percent of the sample reported having received hate speech. Finally, we saw that people with immigrant backgrounds are much more exposed to hate speech directed towards legally protected grounds than non-immigrants, but non-immigrants are equally exposed to (what they perceive as) hate speech and unpleasant messages directed towards other attributes.

### Discouragement from expressing opinions publicly

Next, we look at one possible consequence of receiving hate speech, namely discouragement from expressing opinions publicly. As in the previous section we distinguish between three definitions of hate speech: protected grounds, expanded definition, and other. The following analyses are based only on those respondents who had experienced hate speech, and consequently the number of observations is limited and the results must be treated with caution.

Table 2.5 shows the answer distribution on the question of whether respondents who had received hate speech would be more cautious to express their opinions in public.

We hypothesized (H2) that hate speech directed towards legally protected grounds has more adverse consequences, in terms of discouragement from expressing opinions publicly, than other types of negative speech not directed towards minority group characteristics. This hypothesis is *not* supported by the results in Table 2.5. The results rather suggest that on the aggregated level the consequences are the same regardless of what kind of hate speech you measure: Across definitions more than one fourth of the respondents answered that they will indeed be more cautious when expressing their opinions in public. About two thirds said they would not be more cautious,


**Table 2.5.** Discouragement from expressing opinions publicly after experiencing hateful messages via social media. Percent.

*Source: SMIPS (2016).*

NOTE: 'Protected grounds' include religion, ethnicity, skin colour, nationality, sexual orientation, and disability. 'Expanded definition' includes in addition personality, gender, and appearance. 'Other' includes the content of the argument, political standpoint, occupation, education, other, and don't know. Weighted according to gender, age, and education.

whereas the rest answered that they do not know. Differences across definitions of hate speech are not statistically significant. As such, these findings suggest that any experience with what one perceives as hate speech may lead to a retreat from public debates, and that hate speech directed towards protected grounds may not have more negative democratic consequences than other similarly unpleasant messages.

Aggregations may however hide important group variations. One can argue that hate speech and other unpleasant messages only have *democratic* consequences if particular groups are more likely than others to be silenced. We therefore end the empirical investigation by exploring variations in willingness to express opinions among different groups of respondents. As in the previous section, Table 2.6 summarizes results from two logistic regression models for each definition. Model (1) controls for gender, age, immigrant background, education, and political ideology (and political ideology squared), while model (2) adds propensity to share personal opinions on the internet. The dependent variable is 'experience with hate speech will limit willingness to express opinions' (1='yes', 0='No/Don't know').

The results in Table 2.6 suggest that H2 may hold for some segments of the population, most notably women. Across definitions, the regression models clearly suggest that women are more likely than men to state that experiences with hate speech have lead them to be more cautious in expressing personal opinions. However, the magnitude of the gender difference varies: women who have received hate speech directed towards protected grounds are about 5 times more likely than men to state that they will be more cautious. One could have expected that women experiencing hate speech directed towards gender, appearance, or personality, *in addition* to the protected grounds, would be even more affected. This is however not the case. By expanding

#### chapter 2


**Table 2.6.** Discouragement from expressing opinions publicly after experiencing hateful messages via social media. Logistic regressions. Odds ratio.

*Source: SMIPS (2016). Sig:* †≤*0.1* \*≤*0.05 \*\*≤0.01 \*\*\*≤0.001.*

NOTE: 'Protected grounds' include religion, ethnicity, skin colour, nationality, sexual orientation, and disability. 'Expanded definition' includes in addition personality, gender, and appearance. 'Other' includes the content of the argument, political standpoint, occupation, education, other, and don't know. Weighted according to gender, age, and education.

the definition to also include these three attributes, the odds coefficient is reduced. The gender difference is however still sizeable: women are about 3 times more likely than men to state that they will be more cautious after having received hate speech according to the expanded definition. Finally, the gender difference is reduced even more when regressing hate speech directed towards other characteristics that are further from characteristics related to personal identities, and remains significant only at the 0.1 level. In other words, relative to men, the consequences for women seem to be strongest when receiving hate speech directed towards legally protected grounds (religion, ethnicity, skin colour, nationality, sexual orientation, and disability). Thus, for women, it does seem to matter what type of hate speech they receive.

Respondents' immigrant backgrounds are not statistically significant related to reluctance to express personal opinions publicly. Considering hate speech directed towards protected grounds, the size of the odds coefficients are substantial (odds=3), suggesting that immigrants are more affected by these utterances. However, due to few respondents the differences are not statistically significant.2 Combining these findings with other recent studies (Midtbøen & Steen-Johnsen, 2016; Nadim, Fladmoe, & Wessel-Aas, 2016), we do however see clear indications that people of immigrant background in Norway are more affected by hate speech directed towards protected grounds than the majority population.

Summing up this final empirical section, we have seen that on the aggregate level the consequence of hate speech in terms of discouragement from expressing opinions publicly seems to be similar irrespective of what grounds the hate speech is directed towards. However, there are indications that women and people of immigrant background are more likely than men and the majority population to be affected by hate speech directed towards protected grounds. It is however important to treat the results in this final section with caution, as the number of

<sup>2</sup> Of the 73 respondents who had been exposed to hate speech directed towards protected grounds, 14 had immigrant backgrounds. 7 (50 percent) of these said they would be more cautious. 18 (31 percent) of the 41 majority respondents gave a similar answer.

respondents is limited. More research is still needed in order to understand the consequences of experiencing different forms of hate speech and unpleasant comments.

# **Discussion and conclusion**

The empirical analyses in this chapter were motivated by two research questions: (1) Which groups are most exposed to hate speech?, and (2) Are people who have experienced hate speech directed towards legally protected grounds more reluctant to express opinions publicly, compared to people who have experienced other types of negative comments? Survey data from Norway suggested that people with immigrant backgrounds are more exposed to hate speech directed towards grounds that are protected by the Norwegian General Civil Penal Code (skin colour or national or ethnic origin, religion or life stance, homosexual orientation, and disability), but that non-immigrants are equally exposed to hateful messages directed at other grounds, such as gender, appearance, political viewpoints, etc. It is especially people who often share personal opinions on the internet who are vulnerable to hateful and other unpleasant messages.

Several scholars have argued that hate speech directed at protected grounds have more severe consequences compared to other forms of negative speech (e.g. Boeckmann & Liew, 2002; Herek et al., 2002). One explanation for this is that since hate speech triggers the awareness of belonging to a vulnerable group it incites more fear than other types of negative speech (Herek et al., 2002). We tested this claim on one possible consequence, namely discouragement from expressing opinions in the public. We found that a substantial share of those who had received hate speech were indeed discouraged from expressing opinions. However, contrary to what we expected, on the aggregated level the analyses showed that reluctance to express opinions publicly is more or less on the same numeric level irrespective of what grounds the hate speech is directed towards. This suggests that negative or derogatory speech may function as a social boundary for free speech irrespective of content – as long as people subjectively perceive messages as hateful. Disaggregating the general public, we did however see that women and people of immigrant background seem to be more strongly affected by hate speech directed towards protected grounds, than by other types of negative comments (see also Midtbøen & Steen-Johnsen, 2016; Nadim et al., 2016). A possible interpretation of this finding is that women and immigrants, more than men and the majority population, see themselves as belonging to vulnerable groups, and that they therefore react more negatively to messages directed towards group-based identity characteristics. This suggests that hate speech, more than other types of negative and derogatory speech, can represent a democratic problem in that it might silence *specific* groups and discourage them from voicing their opinions publicly. These findings are based on a relatively small number of respondents, and a task for future research should be to examine more carefully whether hate speech is distinct from other types of speech for minority groups.

Is hate speech in social media an extensive phenomenon in Norway? In the survey analyzed in this chapter, 7 percent said they had experienced what they perceived as hate speech, and 2 percent had experienced hate speech directed towards protected grounds. These are small numbers, and, as such, one may view hate speech as a marginal phenomenon. Such an interpretation is, however, problematic. First, the legal definition of hate speech aims at protecting vulnerable *minorities*. Minorities obviously make up a limited share of the total population, and by analyzing a national representative sample the number of respondents with any type of minority background will be limited. Indeed, if we only look at respondents with immigrant backgrounds about 7 percent reported that they had experienced hate speech directed towards protected grounds, and the regression models also suggested that – all else equal – this group was almost four times as likely as non-immigrants to have experienced such speech. A second objection is that hate speech can have consequences not only for those who receive messages directly, but also for those who observe the messages (Bell, 1998; Kunst et al., 2013; Perry, 2014). A comparative study of young adults in the US, UK, Germany, and Finland, found that between 30 (Germany) and 50 percent (USA) had during the *past three months* witnessed 'writings or speech online, which inappropriately attacked certain groups of people or individuals' (Hawdon et al., 2015). Thus, although few people have direct experience with receiving hate speech, it appears to be relatively common among young adults to have witnessed it. Observing hate speech can also incite fear among individuals who are not directly targeted, because it highlights the risk of being subjected to it (Gelber & McNamara, 2016 p. 327; Perry, 2001).

Hate speech brings the question of boundaries of freedom of speech to the fore. This chapter has illustrated how hate speech and other unpleasant messages can represent social boundaries to the exercise of free speech. Empirical evidence suggests that a substantial number of individuals who receive hateful messages, become reluctant to express opinions publicly. One purpose of hate speech is to incite fear in the groups targeted, and fear can be an effective silencing mechanism. Hate speech as a response to an individual's public expression of opinions, is an attack on the legitimacy of that person's position as an equal member in public debate (see also Enjolras, ch. 10). If certain groups are systematically silenced, hate speech ultimately has democratic consequences. Legal regulation of hate speech does, however, also represent (potential) boundaries on freedom of speech. Where to draw the line between freedom of speech and protection against hate speech is a delicate balance, and it is ultimately a political and normative question.

## **References**


Nilsen, A. B. (2014). *Hatprat*. Oslo: Cappelen Damm akademisk.

NOU. (1999). *'Ytringsfrihed bør finde Sted' - Forslag til ny Grunnlov § 100*. (NOU 1999: 27). Oslo: Justis- og politidepartementet Retrieved from https://www.regjeringen.no/no/dokumenter/nou-1999-27/id142119/.

Parekh, B. (2006). Hate speech. *Public policy research, 12*(4), 213–223.


Pew Research Center. (2014). Online Harassment.

Phillips, A. (2009). *Multiculturalism without Culture*. Princeton: Princeton University Press.


## Chapter 3

# Willingness to discuss the publishing of religious cartoons. Spiral of silence in the private and public spheres

*Audun Fladmoe, PhD, Senior Research Fellow, Institute for Social Research Kari Steen-Johnsen, PhD, Research Professor, Institute for Social Research*

The publishing of religious cartoons has spurred crucial debates about freedom of speech in Western societies. Cartoon debates represent contestations where symbolic boundaries are drawn towards what are 'acceptable' and 'unacceptable' positions in public debates. According to Elisabeth Noelle-Neumann's spiral of silence theory, individuals who perceive that their opinions are incongruent with the dominant opinion climate are more likely than others to remain silent in public debates. Based on survey data, the empirical analysis explores people's willingness to discuss whether news media should publish potentially offensive religious cartoons. Two dimensions are explored: different arenas for discussion (public, semi-public and private) and different climates of opinion (general public and peers). The results suggest, first, that people with personal opinions perceived to be incongruent with the dominant positions held by the general public are less willing to discuss the publication of religious cartoons. Second, the results suggest that spiral of silence mechanisms are stronger in private than in public arenas, i.e. that people are especially wary of both the general opinion climate and the opinions of their peers when discussing the publication of cartoons among friends, family and workmates. One implication of the findings is that symbolic boundaries work to reinforce majority positions in both public and private discussions.

## **Introduction**

Particularly in the past two decades, the publishing of religious cartoons has spurred crucial debates about freedom of speech in Western societies. The Mohammed Cartoon Crisis in 2006, in important ways, set the stage for the debate about the role of religion in modern societies, and how respect for religious identities and feelings should be weighed against the principle of freedom of speech. The terrorist attacks on the satirical magazine *Charlie Hebdo* in Paris and on the local cultural centre Krudttønden in Copenhagen in January 2015 set aflame renewed debates on the current threats to this freedom. Different as they were, both of these events were characterized by the use of violence to protest against cartoon publication. In the case of the Mohammed cartoons, their publication had global repercussions, entailing the torching of consulates and embassies in the Middle East, and violent protests in a range of countries in Asia and Africa. As many as 241 people are estimated to have died in connection with demonstrations during the spring of 2006 (Klausen, 2009, p. 107).

The publishing of religious cartoons has thus been marked by highly dramatic events, which have brought to the fore the conflict between the principle of freedom of speech and questions of blasphemy and intercultural tolerance. In Norway, public debate on freedom of speech reached its peak during the cartoon crisis and the *Charlie Hebdo* attacks (see Colbjørnsen, chapter 6). These debates were characterized by high temperature and strong disagreement. Crudely put, advocates of unlimited free speech state that no limits should be enforced on the publication of strongly provocative cartoons. More restrictive voices, on the other hand, argue that freedom of speech is only one of many bricks in a liberal democracy, the protection of minorities and their religious beliefs being another. Hence freedom of speech must be balanced against its possible negative consequences, especially in terms of reinforcing social cleavages and hurting particular minority groups. As pointed out by Erich Bleich (2011) both positions tend to accuse the other side of creating a situation where crucial democratic principles are undermined.

In relation to this book's discussion of the boundaries of freedom of speech, the publishing of religious cartoons is relevant in a double sense. First, exploring the attitudes of the population towards publishing religious cartoons is an indicator of where people draw the boundaries for freedom of speech vis-à-vis the protection of religious feelings and even blasphemy. Second, cartoon debates represent types of high temperature contestations where one might assume that symbolic boundaries are drawn between what are acceptable and unacceptable positions, either among individuals or for individuals themselves. It is this latter social process that will be the main focus of this chapter.

As we have defined symbolic boundaries in this book, they are linked to contestations between groups and individuals in society, based on disagreement over values, ideas or principles. In this chapter we turn our attention to another type of social process, i.e. the process whereby individuals adjust their expression of opinions to what they perceive to be the normatively sanctioned opinion in their social environment, exercising what may be termed self-censorship. Boundaries, in this context, could thus be seen as self-imposed limitations on expression, rather than as limitations drawn by others. The core of the matter, however, is that such processes of potential self-censorship are closely linked to public debates and opinion.

Based on survey data we ask whether and to what extent people are willing to take part in discussions about the publishing of religious cartoons, if they believe their personal opinions are incongruent with the current opinion climate. The survey was carried out in August 2015, at a point in time when the *Charlie Hebdo* attacks and the right to publish religious cartoons had been fiercely debated for several months (Colbjørnsen, chapter 6). Through our data we can then study mechanisms of selfcensorship and the willingness to express one's opinion in the particular case of discussions on publishing religious cartoons. Moreover, the timing of the survey enables us to reflect upon the question of whether spiral of silence mechanisms did occur in the Norwegian context, since it taps into the question of selfcensorship at the end of a long period of public discussion.

Noelle-Neumann's (1974) theory of spirals of silence forms the theoretical backbone of our analysis. Spiral of silence theory states that individuals will tend to adjust to what they perceive to be the dominant public opinion, and to be less willing to speak out if they perceive that they are part of a minority. We have a particular interest in what has been termed 'peer effects', i.e. that people may primarily fear and react to the danger of isolation from peer groups such as family and friends, rather than to isolation from the wider public (Neuwirth & Frederick, 2004). In order to analyze which opinions play a role when people evaluate whether to express their opinion or not, we differentiate between public and private opinion climates and the willingness to speak in public, semi-public and private arenas. A Norwegian study from 2013 showed a general tendency among respondents to be less selfrestrictive in the private than in the public setting when faced with various types of risk (Steen-Johnsen & Enjolras, 2016). However, the question asked in that study was more abstract, given that respondents were not presented with a specific case. More importantly, based on the present study we are able to disentangle the effects of various opinion climates and the arenas in which the potential utterance takes place.

# **Spiral of silence and the opinion climate: our approach**

The term 'opinion climate' refers to how individuals *perceive* aggregated public opinion. According to Paul Lazarsfeld (1972), the opinion climate of an issue is closer to the more permanent and subconscious 'value system' of a society compared to the more fleeting everyday reactions expressed by citizens in surveys. Opinion climates have been described as heavily loaded with social and normative meaning, guiding acceptable attitudes and behavior in a social group (Noelle-Neumann, 1974; Shamir & Shamir, 2000). In other words, people's willingness to express opinions may be affected by their perceptions of the opinion climate. A person who senses that her opinions run contrary to the majority may be less willing to express these opinions compared to a person with mainstream opinions.

In her famous study of the effects of opinion climates Noelle-Neumann (1974) argued that, due to fear of isolation, people with diverging opinions would gradually be less willing to express opinions, leading to a 'spiral of silence' where only dominant opinions remain. Based on empirical evidence from West Germany on several different political issues, Noelle-Neumann found that people on the 'losing side' were consistently less willing to discuss controversial issues compared to people on the 'winning side' (ibid.). Spiral of silence theory does not assume that people know what the opinion climate of a given issue is in reality; rather what counts is how people *perceive* the opinion climate.

The core tenet of the spiral of silence theory is that willingness to express opinions is influenced by perceived support for those opinions, which is also the topic of this chapter. In the aftermath of Noelle-Neumann's publication however, several empirical studies have found very small spiral of silence effects, leading some scholars to argue that '…the literature provides little support for this [spiral of silence] notion' (Glynn, Hayes, & Shanahan, 1997). Noelle-Neumann herself has responded to the criticism that her theory lacks strong empirical support, by stressing the importance of studying value laden issues: '…a situation that involves *real* struggle for public opinion…' (Noelle-Neumann & Peterson, 2004 p. 352). It is a mistake to believe that the spiral of silence theory applies to all situations. This claim is supported in a study by Bodor (2012), which, among other things, stresses the importance of *timing*. Bodor found that during the 2004 US presidential campaign the spiral of silence mechanism was vulnerable to weekly shifts in opinion climates. In a period during the campaign when George W. Bush's chances of reelection seemed to erode, his supporters suddenly became less willing to discuss politics in the workplace. When the opinion climate again shifted in Bush's favor, his supporters became more willing to discuss politics.

In addition to the criticisms that have been raised about the generalizability of the spiral of silence mechanism on the macro level, concerns have also been raised about the psychological mechanism at the core of the spiral of silence theory, i.e. the fear of isolation (Moy, Domke & Stamm, 2001). For example, Pollis and Cammalleri (1968) pointed out that people's tendency to conform may quite easily be broken if they receive support from just one relevant other, or if they enter the setting together with a friend. Another question concerns the impact of the general, abstract public as compared to the impact of peers and relevant reference groups. Several studies have shown that opinions held by family and friends are more important when deciding whether to speak out (Glynn & Park, 1997; Krassa, 1988; Moy, Domke & Stamm, 2001).

Based on the assumption that people may fear isolation from their reference groups more than isolation from more remote groups or from society at large, one hypothesis states that feeling out of sync with the opinion of family and friends will impinge on the willingness to speak in any context, also public ones. One point that might underpin this line of reasoning is that the boundaries between public and private speech are becoming increasingly blurred, as what is posted publicly in social media might reach a very diverse crowd of friends, family, colleagues and faint acquaintances. The private and the public are hence becoming increasingly interwoven (Mutz & Silver, 2014, p. 77). Alternatively, one might hypothesize that the public opinion climate and opinions among peers ('private opinion climate') are experienced as distinct, and that the willingness to speak in a particular arena depends on the corresponding opinion climate.

In our analysis, we take account of the criticisms raised in relation to Noelle-Neumann's original theory and the new digitalized context for voicing one's opinion, and explore two different mechanisms for self-censorship in the religious cartoon debate in Norway. The first is related to the impact of different *types of opinion climates*, from the private (family, friends and the workplace), to the more public (people who are on social media and comment fields in online newspapers, and whom you don't necessarily know personally), and to the wider public of the edited mass media. Are people more concerned with the opinions of their peers than with the opinion of the general public? Hence, although 'opinion climate' is typically associated with public opinion, in our analysis we distinguish between 'private opinion climates' and 'public opinion climates'.

The second mechanism is related to *the arena* in which the expression of an opinion might take place, i.e. to the question of whether people act differently in public as opposed to the semipublic or private spheres. This is the question of whether congruence with different groups plays a different role in different contexts. More concretely – if people feel aligned with their family and friends on the issue of cartoon publishing, will this make them more willing to discuss the issue, not only in the family context, but also in social media and in edited media contexts? We use the term 'semi-public arenas' to indicate willingness to discuss on social media sites and in online comment sections of newspapers. These arenas are semi-public in the sense that they may include a mix of known and unknown people; they are interactive and not as formal as the edited public sphere.

Taken together these two entries then enable us to examine more closely the question of boundaries that people draw for their own speech in contested issues such as the religious cartoon debates, and what social forces influence the drawing of these boundaries. In a broader sense the issue of boundary making that we are discussing here is of vital importance to the exercise of free speech. It is both a debate on substance and a meta-debate on principles, laying foundations for the functioning of the public sphere.

In a broader view it is also worth pointing out that most people do not and never will, discuss religious cartoons in public arenas. In this light, it makes particular sense to distinguish between different arenas for discussion, and to be concerned with the private arena as well, since discussions in the private arena may be of great importance to opinion formation. We would also like to emphasize that opinions about contested issues are not necessarily formed prior to discussion, but may rather result from it. Still the theory of the spiral of silence implies that the sense that people have about the relationship between the dominant view in the environment surrounding an issue and their own, may play a role in their willingness to discuss the issue at all.

### **The case: Publishing of religious cartoons**

Ever since the 'Cartoon Crisis' in 2005-2006, ignited by the publishing of cartoons depicting the Prophet Muhammad, in the Danish newspaper *Jyllands-Posten*, cartoons have been at the center of public debates on free speech (Colbjørnsen, 2016, this book). It can be argued that debates over boundaries for cartoon publishing thus crystallize contemporary, global debates on free speech, which makes this a relevant case to explore, both substantively and theoretically.

As described above, the debate has in the main elucidated two main positions: a position opposing most restrictions on free speech, and a more restrictive stance arguing for caution against insulting (religious) minorities (Favret-Saada, 2015). Boosted by the attacks on the French satirical newspaper *Charlie Hebdo* in January 2015, the unlimited free speech stance arguably became the most dominant position in the Norwegian debate, as it was expressed through op-eds and commentaries in the main newspapers. This was also a moment when many opinion leaders, such as pundits and political commentators, used the occasion to revisit the 2006 debate in Norway on the Mohammed cartoons and to criticize those who had taken a more restrictive position at that point. Given the violence of the *Charlie Hebdo* attacks, it seemed easy to pinpoint the more restrictive position as a failed attempt to reconcile principles, and to argue that if one gives in on the principle of free speech, society is laid open to the oppressing forces of those who are willing to use violence.

While one side of the debate seemed to be clearly dominant in the elite debate, as reflected in op-eds and commentaries, an interesting question is whether these positions were reflected among ordinary people. Before turning to the empirical analysis of people's willingness to discuss the publication of religious cartoons, we briefly present the real distribution of opinions in the population (the 'public opinion climate') concerning whether cartoons insulting religion should be published or not (Figure 3.1). In our August 2015 survey among the general population and journalists (described in chapter 4) respondents were confronted with the question of whether or not media should publish potentially offensive religious cartoons1 . In this context, we conceive of the journalists as part of an opinion elite, and are interested in seeing whether there are differences between their opinions and those of the broader population.

<sup>1</sup> See the note attached to Figure 1 for the precise question wording.

**Source:** *Freedom of speech survey (2015).*

**NOTE:** Question wording: 'Which of the following statements is most in accordance with your own opinion?' Don't know answers are excluded. In the surveys the samples were randomly divided into six groups, of which five groups received additional contextual information before the question: In the last few years there has been some discussion about the publishing of religious cartoons in the media. Given that a religious cartoon can be perceived as [1. offensive; 2. offensive among Christians; 3. offensive among Muslims; 4. offensive among Jews; 5. offensive among Christians, Muslims or Jews]... Population data weighted according to age, gender and education.

It was possible to give an unconditional 'yes' (media should publish) or unconditional 'no' (media should not publish) answer, in addition to a conditional answer (media should be restrictive, if the cartoon can be perceived as offensive). In the survey, respondents were randomly assigned to six groups, in which four of the groups got specific information about *who*  would potentially feel offended (Christians, Muslims, Jews, and all three together). Figure 3.1 displays mean scores across groups,2 among respondents who had an opinion on the matter (don't know excluded).

Figure 3.1 may be interpreted in different ways. On the one hand it may be read as illustrating a relatively one-sided opinion climate, especially among journalists, given that so few state that media should not publish religious cartoons. 45 percent of the population and 63 percent of the journalists held the opinion that media *should* publish potentially offensive religious cartoons (unconditional yes). Only 15 percent of the population and 3 percent of the journalists answered that media should not publish potentially offensive religious cartoons (unconditional no). On the other hand, the population is divided almost in two between the unconditional and conditional yes categories, which indicates that there exist two sides of the question with almost equal strength. Even though one might argue that the main positions in the Norwegian debate were an absolutist yes and an absolutist no to publication, the opinion climate rather seems to have been divided mainly between the unconditional yes and the more conditional position. A relatively large minority of the journalists also favored the conditional 'yes' option, but the general public is more equally divided between the unconditional and the conditional standpoints. This is an important finding, which illustrates the fact that *perceived* opinion climates may differ from what is the actual distribution of opinions in a population. When observing the opinions expressed most strongly in the public sphere through the mass media during the 2015 cartoon debates in Norway, one can hypothesize that people leaning towards the middle position

<sup>2</sup> Variations between the groups were mainly between unconditional and conditional yes. The highest share of conditional 'yes' was given by the Muslim group. There were no significant differences in unconditional 'no' answers. See (Steen-Johnsen, Fladmoe, & Midtbøen, 2016) for detailed analysis.

might have felt that they were part of a minority, even though they were not.

In other words, based on the spiral of silence theory it is reasonable to expect that most people viewed the public opinion climate on this specific issue as leaning predominantly towards publishing religious cartoons unconditionally. In the public debate, expressing more conditional views could easily be viewed as anti-liberal and as 'attacks on the principle of free speech'. For example, when the leader of the social democratic Labour Party, Jonas Gahr Støre formulated his position as a defense of free speech, but also as more conditional on the right to exercise blasphemy, this was described as surprising and shocking, given that 12 people had been killed3 . Our data, collected in August 2015, are thus well-suited to explore the willingness to speak or to remain silent on a value-laden and rather one-sided political issue. This provides a 'best case' for a study of the impact of spiral of silence mechanisms in debates about free speech.

# **Data and variables**

We use data from the population survey on freedom of speech, carried out in August 2015 (see Online Appendix).

## Dependent variables

The dependent variable is the willingness to participate in discussions about the publishing of religious cartoons in the media. We followed the proposed method of Glynn et al. (1997) and asked respondents about their willingness to express opinions in different scenarios: 'Imagine a discussion in the near future in one of

<sup>3</sup> http://www.dn.no/meninger/kommentarer/2015/01/08/2200/Terroraksjonen-i-Paris/de-fornrmedes-forsvarer

chapter 3

the following arenas about the publishing of religious cartoons. How willing or unwilling would you be to participate in such a debate?' It should be noted that these questions followed the questions on whether media should publish (different types of) religious cartoons or not (as illustrated in Figure 3.1). As such, the respondents were primed on recent debates regarding this issue. A total of seven different arenas were listed (closest family and friends, at work, in social media, in comment sections of online newspapers, in debate sections in newspapers, on radio, and on TV), and answers were given on a four-point scale ('Very unwilling' to 'Very willing'). We re-coded 'Don't know' answers in a neutral position (3), meaning that the variables had five values4 . The mean value for each arena is illustrated in Figure 3.2.

The figure suggests that, among closest family and friends and at work, a majority of the respondents are willing to take part in discussions about the publishing of religious cartoons. Only a minority of the respondents, however, are willing to take part in discussions in the five other public or semi-public arenas.

All seven items are significantly correlated, but the size of the correlation coefficients varies extensively (0.21-0.95). As suggested by Figure 3.2 the pattern is that the two items 'Among closest family and friends' and 'At work' correlate strongly with each other, but weakly with the six other items. A principal factor analysis confirms this pattern, by distinguishing between three factors (Table 3.1).

The factor analysis suggests that all six public and semi-public arenas could be collapsed into one single variable (Factor 1), but since we are interested in examining the willingness to speak in

<sup>4</sup> Omitting 'Don't know'answers on the dependent and the main independent variable from the analyses yield basically the same results (available upon request), but reduces the sample by 782 respondents.

**Figure 3.2.** Willingness to take part in discussions about the publishing of religious cartoons. Mean score and 95 % confidence intervals.

### **Source:** *Freedom of speech survey (2015).*

**NOTE:** Question wording: 'Imagine a discussion in the near future in one of the following arenas about the publishing of religious cartoons. How willing or unwilling would you be to participate in such a debate?' 1=unwilling, 3=Don't know, 5=Willing. Weighted according to age, gender and education.

a digitalized public sphere (Mutz & Silver, 2016), we keep the distinction between public and semi-public arenas. Even though most social media platforms and comment fields are by default public arenas, most people do not necessarily perceive them as public to the same extent as newspapers, radio and television. Hence, we constructed three dependent variables. *Willingness to discuss – public arenas* includes 'in debate sections in newspapers', 'on radio', and 'on TV', *Willingness to discuss – semi-public arenas* includes the two arenas 'in social media' and 'in the comment sections of online newspapers'. Finally, *Willingness to discuss – private arenas* includes the arenas 'Among closest family and friends' and 'At work'. Descriptive statistics for these three variables are displayed in Table 3.2.

All three indexes were constructed by taking saved factor scores and standardizing on a 0-1 scale, where a higher value equals more willingness to discuss. As none of the three indexes are normally distributed, we ran additional analyses with normalized versions of the three (natural logarithm of public and


**Table 3.1.** Principal factor analysis. Varimax rotation (n=1984).

*Source: Freedom of speech survey (2015).* 

#### **Table 3.2** Constructed willingness to discuss variables. Descriptive statistics


*Source: Freedom of speech survey (2015).* 

semi-public arenas and the exponentiation of private arenas). Compared to the 'raw' variables results did not change substantially when regressing normalized variables. For the ease of interpretation, we therefore present results with the original variables in main text and normalized variables in the online Appendix.

### Independent variables

The main independent variable is perceptions of the opinion climate surrounding the publishing of religious cartoons in the media. We followed the strategy used in several studies (e.g. Liu & Fahmy, 2011; Moy, Domke, & Stamm, 2001; Perry & Gonzenbach, 2000) and assessed perceived opinion climate by means of a question distinguishing between opinion congruence in different contexts. More specifically, respondents were asked to estimate how many people they believed shared their opinion on the issue (1) among close family and friends, (2) people living in the municipality, and (3) among people living in Norway in general.5 Answers were given on a five-point scale from 'Almost no one' to 'Almost everyone'. 'Don't know' answers were re-coded in the middle/neutral category ('about half ') (see footnote 4).

Some former studies have summed up these different levels of opinion climates and created one single index (e.g. Liu & Fahmy, 2011; Moy et al., 2001; Perry & Gonzenbach, 2000). We are, however, interested in variations between the private and the public sphere, and therefore expand this approach by studying variations between the private and the public opinion climates. Thus based on these items we constructed a variable consisting

<sup>5</sup> The survey also included the item 'People living in your neighborhood', but this is left out of this analysis.

### chapter 3

of four categories: (1) Most people agree, (2) Most people disagree/neutral, (3) Family/friends agree, larger public disagree/ neutral, and (4) Family/friends disagree/neutral, larger public agree. 'Agree' is the sum of the values 4 and 5 on each scale, while 'disagree/neutral' is the sum of the values 1 thru 3. 'Larger public' consists of the two items 'People living in your municipality' and 'People living in Norway in general'. We have labelled these four categories of perceived opinion climates (1) 'Fully supported', (2),' Unsupported', (3) 'Peer supported' and (4) 'Publicly supported'.

Table 3.3 summarizes the distribution of perceptions of the opinion climate. More than 40 percent of the respondents believe that most people in both the private and public opinion climates agree with them on the issue, while 32 percent believe most people disagree or that the opinion climate is divided halfand-half. 24 percent believe their opinions are shared among people in their private opinion climate, but not among people in the public opinion climate, while only 1 percent believe their opinions to be congruent with the public opinion climate but not with the private opinion climate.


**Table 3.3.** Perceptions of the opinion climate on the publication of religious cartoons. Typology.

*Source: Freedom of speech survey (2015).* 

In the analysis we also include a set of control variables. It is reasonable to think that willingness to take part in discussions is related to awareness of the public debate on the publishing of religious cartoons, and also whether one has felt offended by religious cartoons. We included two items in the survey measuring these factors: 'How closely would you say you have followed the recent years' debate on the publishing of religious cartoons with religious and political content?' (followed very or somewhat closely coded as '1') and 'Have you yourself felt offended by religious cartoons published in the media?' (Yes=1).

We also include controls for the usual suspects: gender (female=1), age, education (higher education=1) and immigrant background. The latter variable is important to include in the analysis because the subsample of immigrants in the survey is not statistically representative of the total immigrant population in Norway. Descriptive statistics for the control variables are summarized in Table 3.4.


**Table 3.4.** Control variables. Descriptive statistics.

*Source: Freedom of speech survey (2015).* 

## **Who plays a role when deciding whether to discuss the publishing of religious cartoons?**

In order to tap into how the Norwegian opinion climate on religious cartoons was perceived in the autumn of 2015 we begin by mapping out the opinion climate on the publishing of religious cartoons, by perceptions of opinion congruency. Then we go on to the main analyses by exploring the bivariate relationship between willingness to take part in discussions and perceptions of the opinion climate. Finally, we estimate the net effect of perceptions of the opinion climate, controlled for other relevant factors.

Figure 3.3 displays the opinion climate on the publishing of religious cartoons that was presented earlier in the chapter, contingent on perceptions of the opinion climate. Based on what we found in Figure 3.1, we might expect that those in favor of unconditional publication of cartoons would be more likely to think that others agree with them. The figure confirms this, and thus suggests that opinion congruency is clearly related to position in the debate. The majority of respondents that gave an unconditional 'yes' response to the publication of cartoons believed that their opinion was congruent with both family/friends and the general public (Fully supported) or with the general public only (Publicly supported). Conversely, the majority of respondents that gave an unconditional 'no' response to the publication of cartoons believed their opinions to be *in*congruent with the general public. As shown in Figure 3.1, in the population the conditional 'yes' response was almost as widespread as the unconditional 'yes'. However, only 27 percent in this group feel fully supported, which indicates that the impression they get from public debate is that the unconditional position is the dominant one. Concomitantly it is interesting to see that this

**Figur 3.3.** The opinion climate on the publishing of religious cartoons, August 2015, by perceptions of the opinion climate. Percent.

**Source:** *Freedom of speech survey (2015).*

**NOTE:** Question wording: 'Which of the following statements is most in accordance with your own opinion?' Don't know answers are excluded. In the surveys the samples were randomly divided into six groups, of which five groups received additional contextual information before the question: 'In the last few years there has been some discussion about the publishing of religious cartoons in the media. Given that a religious cartoon can be perceived as [1. offensive; 2. offensive among Christians; 3. offensive among Muslims; 4. offensive among Jews; 5. offensive among Christians, Muslims or Jews]...' Population data weighted according to age, gender and education.

group feels more peer supported than the unconditional group, which probably indicates that a conditional yes is more often expressed in everyday conversations. These patterns underscore the importance of distinguishing between *actual* and *perceived* opinion climates, of which the latter is central to the spiral of silence theory.

Figure 3.4 displays willingness (mean score) to take part in discussions in public, semi-public, and private arenas, by

**Figure 3.4** Willingness to take part in discussions about the publishing of religious cartoons, by perceptions of the opinion climate. **Source:** *Freedom of speech survey (2015).*

**NOTE:** 0=unwilling, 1=Willing. Weighted according to age, gender and education.

*perceptions* of the opinion climate. The figure suggests that those who believe that their personal opinions are congruent with the opinion climate are more likely to be willing to take part in discussions, compared to people who believe that their opinions are incongruent with the opinion climate. In other words, there is a difference between the unsupported on the one hand, and the fully or publicly supported on the other. This already supports the original spirals of silence thesis, that feeling part of a majority increases the willingness to speak out (Noelle-Neumann, 1974).

At the same time, the figure also shows that willingness to discuss in different arenas depends on how different opinion climates are perceived. In public and semi-public arenas, willingness to discuss the publication of religious cartoons is on the same level, irrespective of whether respondents perceive their opinions to be incongruent with the general public or supported by their peers. In other words, when it comes to discussing religious cartoons on radio, in the newspapers or in social media, it doesn't matter much whether you think that your friends and family agree with you. What matters is whether you feel that you are in line with the dominant view of the public.

A different picture emerges when it comes to willingness to discuss in private arenas. The peer supported, who feel that their family and friends agree, while the public disagree are more willing to speak in the private sphere than those who believe that most people disagree or are neutral (unsupported). This suggests that when discussing the publication of religious cartoons with family, friends or colleagues people are less preoccupied with any incongruence with the public opinion climate. In private arenas what matters is if personal opinions are congruent with the perceived opinions of peers.

Conversely, although the number of respondents is limited, the figure also suggests that the opposite picture may be true. The publicly supported – people who believe their opinions are congruent with the public opinion climate but incongruent with the private opinion climate – are more willing to discuss in the public sphere compared to the two other groups of people who feel unsupported on the whole or who perceive that they are supported by family and friends only.

In sum, these findings underline the importance of taking context into consideration, and looking at the relationship between specific opinion climates and the arenas where a potential discussion might take place. While our findings support the notion that spiral of silence processes do take place, both chapter 3

in the private and the public spheres, the analyses also show the occurrence of a peer effect. In conversations about religious cartoons in the private sphere, it is not public opinion that counts the most in determining whether to speak out, but rather the opinions in that same group of people. However, this peer effect does not stretch into the semi-public or public arenas, such as social media or the newspapers. In these contexts, public opinion is what matters. Finally, one fundamental, and not surprising insight that might be garnered from this analysis, is that most people are much more willing to discuss controversial issues in private than in semi-public or public arenas, notwithstanding the perceived congruence with the relevant opinion climate.

## How unequivocal is the spiral of silence mechanism?

In what follows we examine whether other demographic or personality factors have an impact on the willingness to speak out in the case of the cartoon debates, and whether the relationship between perceptions of opinion climates and willingness to speak still holds when controlled for such factors. Table 3.5 summarizes results from a set of regressions of each of the three dependent variables. For each variable two models were estimated. Model (1) includes the main explanatory variable – perceptions of the opinion climate (most people disagree as reference), and socio-demographics. Model (2) introduces variables measuring attentiveness to the debate about the publishing of religious cartoons, and whether respondents have themselves felt offended by religious cartoons.

First of all, across all models the analyses confirm that those who believe their opinions to be congruent with both the private


**Table 3.5.** Willingness to take part in discussions about the publishing of religious cartoons, by perceptions of the opinion climate. OLSregressions. Unstandardized coefficients.


NOTE: Dependent variable is willingness to discuss (0=unwilling – 1=willing).

chapter 3

102

and the public opinion climates (Fully supported) are more willing to take part in discussions about the publishing of religious cartoons. Even though the size of the coefficients is reduced by introducing other relevant variables, they remain significant. The difference between the unsupported and the fully supported is most sizable when it comes to willingness to discuss in private arenas. In model 2 in the regression of private arenas the predicted difference between these groups, adjusted for several background characteristics, is 9.2 percent (0.092 on a 0-1 scale). In semi-public arenas the corresponding predicted difference is 5.2 percent, while it is 3.7 percent in public arenas. As such, these findings suggest that spiral of silence effects are stronger in private than in public arenas.

In line with what was suggested in Figure 3.4, those in the second group – Peer supported - are not more willing to discuss religious cartoons in public arenas compared to those who believe their opinions are incongruent with everyone (Unsupported). However, the peer supported respondents are about as equally willing as the first group – Fully supported – to discuss religious cartoons in private arenas. Controlling for a range of other variables, the coefficient remains highly significant, underlining the importance of distinguishing between different arenas.

Finally, we see a tendency that the third group – Publicly supported –are somewhat more willing to discuss in the public sphere. The coefficients for this group are as sizeable as the Fully supported, however they fail to meet statistical significance when controlling for other factors. This is probably due to the low number of respondents in this group (n=25).

Considering other variables, across all models people who have paid attention to the debate about religious cartoons are more willing to take part in discussions compared to those who have not paid attention. This variable is probably a proxy, both chapter 3

for interest in this particular debate and for political interest in general.

Those who have themselves felt offended by religious cartoons are more willing to take part in public discussions, but not in private discussions. Women are less likely to discuss in all arenas, older people and lower educated are less willing to discuss in the private sphere, while lower educated are *more* willing to discuss in the public sphere. While the gendered dimension of willingness to speak in public has been demonstrated in previous studies (Steen-Johnsen & Enjolras, 2016), the differentiated effect of education related to the public and semi-public vs the private spheres is perhaps more surprising. This particular finding does however correspond with a study on participation in social and political debate in Norway. Enjolras et al. (2013) found that the highly educated were overrepresented in offline debates, but not in online (semi-public) debates. As they put it, the contribution of internet debate was to alleviate education based differences in participation in the public sphere (2013 p. 76). Our findings with regard to the semi-public vs private arenas may point to a similar mechanism.

## **Discussion and conclusion**

Since the 'Mohammad Cartoon Crisis' of 2006, Norway and other countries have witnessed intense and value-laden debates on the boundaries between free speech, protection of religious minorities and blasphemy. These debates have contributed to constructing and reinforcing moral boundaries between majority and minority groups, but also between 'acceptable' and 'unacceptable' opinions in debates on free speech. By applying the theory of spirals of silence we have explored public opinion in the case of publishing religious cartoons, and showed that people's willingness to discuss this particular issue in different arenas depends on how they perceive different opinion climates. Those who perceived their opinions to be congruent with the general public opinion climate, as well as with the opinions of their family and friends, were more willing to speak out than other groups.

Moreover, we identified a more specific mechanism connected to different reference groups: Those who felt that family members and friends tended to agree with them had a stronger willingness to speak out in private arenas. In other words, we do find a peer effect in our analyses in the sense that feeling the support of peers has an influence on the willingness to speak out when among them. However, being peer supported does not enhance willingness to speak in social media, to debate in comment fields or to write an op-ed piece in a newspaper, i.e. it does not impact debate activity in what we have termed semi-public and public arenas. Rather, speaking privately and publicly stand out as two separate types of activities, that require different types of considerations. To gain a better understanding of these mechanisms, more studies are required.

As pointed out, the spiral of silence mechanism has been difficult to find and replicate in studies following Noelle-Neumanns's original contribution. This study has also showed that the strength of the spiral of silence mechanism relating to the willingness to discuss in public arenas is limited, with the predicted difference between the groups of unsupported and fully supported being 3.7 percent. Thus this particular part of our study supports previous findings of a significant, but weak spiral of silence effect (Glynn et al., 1997). One obvious explanation for this is that many people would not be willing to discuss any value-laden issue in public, irrespective of how they view the opinion climate. There are other barriers to participation in chapter 3

public debates beyond perceptions of opinion climates. The threshold for participating in discussions in private arenas is lower, and we did indeed see a much stronger spiral of silence effect here (9.2 percent predicted difference).

Theoretically, the results in this chapter therefore suggest that spiral of silence mechanisms are relevant, but that such mechanisms may play a greater role in close social relations on the micro-level than in public debates. This is also underlined by the finding that spiral of silence mechanisms were somewhat stronger in semi-public than in public arenas. Mutz and Silver (2016) have suggested that digital public spheres may feel more like private than public arenas, given that activities here are linked to a network of friends and followers. This would lead to the assumption that the opinions of peers would be of stronger importance here than in the public arena consisting of newspapers and other edited media. Although the differences are small, we do see indications of this in our analysis. This finding speaks to the question of whether social media may contribute to creating meaningful opposition to elite opinions and to what is published in the mass media. The evidence presented here does not suggest that semi-public arenas are particularly apt in breaking spirals of silence.

The observed spiral of silence mechanisms illustrate the stickiness of symbolic boundaries constitutive of the moral order (see Enjolras, chapter 10). People are conscious of what speech is 'acceptable' in different arenas - and what is not, and they adjust to what they perceive as the dominant opinion. It is a reasonable assumption that the spiral of silence mechanism may to some extent have minimized the amount of utterances in favor of the minority position in the cartoon debate that took place during the spring of 2015. As we saw in the empirical section, those believing their opinions to be incongruent with the opinion climate were more likely to take a restrictive standpoint in the discussion on the publishing of religious cartoons. Hence, although we do not have time-series data, one might infer that after years of debating religious cartoons, intensified by the *Charlie Hebdo* events, a spiral of silence mechanism occurred in Norway where the restrictive position was increasingly less heard. Thus, one implication of the findings is that symbolic boundaries work to reinforce majority positions, both in private and public arenas, through processes of self-censorship.

Cartoon debates are interesting in the sense that they crystallize some of the core debates on free speech in Western societies, and when they have erupted in the past two decades have served to confront some fundamental positions on this question (see Colbjørnsen, chapter 6). If heated cartoon debates lead to the occurrence of spiral of silence mechanisms, this might serve to draw boundaries for free speech that are not in concordance with the views of a small or large minority. As a result, these minority positions may not be sufficiently debated in the public sphere.

### **References**


Steen-Johnsen, K., Fladmoe, A., & Midtbøen, A. H. (2016). *Ytringsfrihetens grenser. Sosiale normer og politisk toleranse*. Retrieved from Oslo:

### Chapter 4

# Perceptions of journalistic bias: Party preferences, media trust and attitudes towards immigration

*Hallvard Moe, Professor, University of Bergen Kjersti Thorbjørnsrud, PhD, Senior Research Fellow, Institute for Social Research Audun Fladmoe, PhD, Senior Research Fellow, Institute for Social Research*

Journalists are entrusted with a key role in mediating information from the centres of power to the public, and to present issues in a fair way. This chapter takes the perspective of the citizens and explores the perceived credibility of journalists to fulfil this important role in relation to the issue of freedom of speech. We combine qualitative and quantitative analysis of survey data from 2013 and 2015 to discuss perceptions of journalistic bias among the Norwegian public. Our interest lies in the gaps between these perceptions and the professional normative ideal of journalism. We find that confidence in the impartiality of journalists is low among the audience. Party preferences and attitudes towards immigrants and immigration, along with general trust in the media, are important indicators of perceptions of journalistic bias concerning source selection and the ability to separate personal views from professional practices. With this as a backdrop, we analyze how respondents, in their own words, define the groups they believe are ignored by the news media. In the qualitative part of the analysis, we describe how such perceptions play out among different voter groups.

## **Introduction**

For the vast majority of the public, freedom of speech is mundane. Only a small minority among us participate in mainstream public debates, even counting social media. Still, we all fundamentally enjoy the right to freedom of speech in our everyday lives. For one thing, the freedom to make one's voice heard is not restricted to mainstream public arenas such as newspapers or other mass media, it also pertains to face to face communication at work, in class or at the mall. But freedom of speech also has another dimension which is crucial for all members of the public, including those who do not themselves regularly participate in mainstream mediated debate. This dimension is often referred to as freedom of information: The right to seek, impart, and receive information of different kinds relevant to us as members of a public. It is a dimension of freedom of speech we all meet, at least indirectly, every time we watch the news, search the web or go to a movie.

The experience of the right to freedom of speech among the public rests on citizens' assessment of the legislative boundaries drawn by public authorities concerning access to documents and proceedings. But the experience also depends on less clearcut boundaries made by producers, editors and journalists (see Ihlebæk and Thorseth, ch 5). Through their selection, representation and interpretation of relevant information, journalists create and redefine boundaries in the public sphere, allowing the public to engage in democratic deliberation and make informed choices. The legitimacy of this role builds on a particular set of claims about professional journalism: It must be relatively independent from partisan struggle, group interests and personal motivation; journalists should speak truth to power and work for the common good (McNair, 1998; Waisbord, 2013). We are interested in the gaps between the professional normative ideal of journalism, and the citizens' perception of journalistic bias. Do the public trust journalists to keep personal political views apart from their professional practice, and balance sources of different political leanings? And how can we understand divisions between different groups in society on these issues?

In conjunction with the British referendum to leave the EU and the 2016 US election, commentators and pundits have described politics with buzzwords such as 'post-truth' or 'postfactual', to a large extent fuelled by social media campaigns, partisan media coverage of populist political movements and candidates, and allegations of 'fake news'. Public trust in the media is reportedly at an all-time low (e.g. Gallup 2016). Such a description is, importantly, based on predominantly Anglo-American sources. Our aim in this chapter is to explore the perceived credibility of journalists to act as important safeguards for freedom of speech in a Norwegian context: a Nordic welfare state with little political polarization, high levels of institutional trust, a strong public service broadcaster, public press subsidies, and high levels of news consumption. This makes for a critical case study in how the public express confidence in the impartiality and professional independence of journalism.

The empirical analyses are based on quantitative and qualitative data from population surveys on freedom of speech, carried out in 2013 and 2015 (see chapter 1). Through the analysis, we seek to map patterns of difference among groups of citizens. To get a more nuanced understanding of where, exactly, citizens think journalists are drawing the wrong boundaries, we use open-ended data where respondents were asked to name groups they felt did not get to speak in Norwegian news media.

In what follows, we first lay out the theoretical underpinnings of our analysis – the issue of media credibility for a well-functioning culture of freedom of speech. Next, we describe the data and method, before we present the results and discuss consequences for an understanding of how those who do not actively participate in public debate assess the workings of freedom of speech in Norway.

## **Freedom of information and perceptions of journalistic bias as a dimension of freedom of speech**

The UN's Universal Declaration of Human Rights' article 19 defines freedom of expression to include the freedom 'to seek, receive and impart information and ideas through any media and regardless of frontiers'. Freedom of information concerns the receiving end of speech. It covers not only information held by public authorities, but also, in a wider sense, news and current affairs as well as access to a varied menu of political, moral and aesthetic ideas and cultural experiences. Freedom of information is laid down for instance by the US Supreme Court (e.g. Gripsrud 2002), and in legislation in several nations. In Norway, the Constitution's paragraph on freedom of speech includes a so-called infrastructural requirement that obliges the state to facilitate open and enlightened public debate. This requirement encompasses the right to freedom of information.

While freedom of information covers the possibility for any citizen to access government information, intermediaries are helpful in enabling us to enjoy the fruits of such a right. Ideally, someone should spend time scrutinizing authorities to prevent corruption and misuse of power, and someone should collect, edit and present relevant information to members of the public in order for them to make decisions on whom to elect as representatives. In modern societies, journalistic media have been entrusted with this task (e.g. Keane, 1991).

To facilitate open and enlightened public debate, the media are supposed to perform in a certain way. They should allow for a plurality of voices and the representation of diverse interests. The public needs to be informed about important events, and issues should be covered from different angles. Lastly, the media is expected to undertake self-scrutiny of how they fulfil their role, and keep a distance from centres of power. As such, the choices made by journalists about what to publish, whom to use as sources, and how to present their views are central to the actual experience of freedom of information for the general public.

Trust in the media, specifically, is regarded as key to any notion of a working democracy. In the words of Stephen Coleman 'Unless we can trust the news media to deliver common knowledge, the idea of the public – a collective entity possessing shared concerns – starts to fall apart' (Coleman 2012, 36; also Livio and Cohen 2016). The legitimacy of the media to select, frame and forward different types of information and messages is related to general trust in the ethics and moral foundation of journalism. Journalistic norms of impartiality, neutrality, factuality and integrity are there to justify the unique position of mass media as the interpreters, arbiters and gatekeepers of information to the public (Alexander 2006). The vital role of professional journalism within a vibrant and diverse public sphere depends on the belief in a set of norms that lifts journalism above subjectivism, personal self-interest and political motivation (Waisbord, 2013).

Facing these ideals, research within news sociology and media studies (e.g. Gans, 1980; Schudson, 2003; Tuchman, 1972) has for decades maintained that journalistic representation and selection, like every other type of interpretation of social reality, necessarily construct a picture of the world rather than mirror it. Different schools of research have focused on the impact of economic interests, political connections as well as the particular moulding force of a media logic on journalistic representations (e.g Altheide, 2004; Berkowitz, 2009).

The great worry expressed in recent public debate by researchers and media experts over an increasing popular scepticism towards the established news media, however, indicates that these types of studies, even if they qualified the principles for new production, never fuelled a wholesale discarding of the ideals of professional journalism. Established news media, run by media professionals, are conventionally recognized as vital contributors to an open democratic debate. Wide reaching freedom of information and press laws express the general acceptance of the right of journalists to access information and powerful sources (Ackerman & Sandoval-Ballesteros, 2006). The question is if this recognition and trust in the authority of journalistic texts is diminishing.

The context for our analysis is the increased fragmentation and polarization in media use in general and news consumption specifically (e.g. Prior 2007, Ksiazek et al 2010). Coupled with a dramatic downturn in the funding of traditional journalistic institutions, especially those based in print news (Nielsen & Kuhn, 2014), the result has been seen as a crisis in trust in journalism. Moreover, the fragmentation and polarization in media use is, in the public debate, related to the rise of protest movements and populist parties where distrust in elites and appeals to lay people are common ground (Aalberg & De Vreese, 2017). Even if these movements may belong to different parts of the ideological spectrum, the rise of movements combining distrust in elites with immigration critique and the exclusion of outgroups have come to dominate the public agenda (Sheets, Bos, & Boomgaarden, 2015). A seemingly increasing scepticism towards mainstream media coverage is related to a more general distrust in political, economic and cultural elites.

There is extensive empirical literature on trust in the media and perceived media credibility.1 A part of this literature aims to develop comprehensive measurements to grasp how citizens assess the workings of news media in general. For instance, Kohring and Matthes (2007) define media credibility as depending on four dimensions: trust in the selectivity of topics, trust in the selectivity of facts, trust in the accuracy of depictions, and trust in journalistic assessment (Kohring and Matthes 2007, 240). More recent work has zoomed in on one single aspect to look for changes in the credibility of journalism following digitalization. Karlsson et al (2014) probed into whether new opportunities to increase transparency in the journalistic process and product in online media improved users' assessment. They

<sup>1</sup> As Kohring and Matthes (2007) show, contributions to the field have used terms such as 'trust' and 'credibility' interchangeably. This lack of conceptual uniformity is also illustrated in the specific wording of questions in relevant surveys. For instance, a recent Gallup survey in the US asked for respondents' 'trust and confidence' (Gallup 2016). In Norway, surveys have used different terms to capture the same idea, including 'tiltro' and 'tillit' (e.g. NMD 2016).

chapter 4

found almost no effect on source and message credibility. Another line of research focuses on how involvement in an issue matters for trust. Matthes and Beyer (2015), in a recent contribution that also includes data on Norway, measure the effect of cognitive and affective involvement, and find right-wing political ideology to be related to less media trust in Norway (but not in the US and France). A recent related study based on Israeli data finds trust in journalists to be positively predicted by a leftwing political orientation, and also sees a general decline in media trust (Livio and Cohen 2016; also Tsfati and Ariely 2014 for a large-scale comparative analysis).

These latter studies point to interesting country-specific differences, related, among other factors, to the media systems and their characteristics. Importantly, though, to probe the effect of involvement, such studies use specific issues or events as starting points for their analysis. This is also the case for some of the more comprehensive attempts at measuring media credibility (e.g. Kohring and Matthes 2007). Yet other efforts have been made to track differences in trust across different media outlets. For Norway, Sjøvaag and Ytre-Arne (forthcoming) find support for the assumption that people express more trust in quality newspapers than tabloid ones.

Our interest is more specific, but at the same time more general. We want to concentrate on one central aspect of perceived media credibility, and relate that not to a single, current issue, but to a general impression of the Norwegian media system's journalistic practices. We are interested in what citizens think is lacking in media coverage, in the sense of which voices get ignored. We are also interested in whether or not citizens think the selection of sources and journalists' practices more generally are coloured by their personal political leanings. And we are interested in explaining differences among citizens on these issues.

## **Norway: A 'critical case' media system, with a partisan history**

Norway is often described as a Nordic model of society (e.g. Hilson 2008) with welfare policies aimed not only at creating a security net for those who fall, but at utilizing public support schemes to advance equal opportunities. Politically, it is a multiparty system with comparatively little polarization. Economically, the country experienced rapid growth from the 1970s due to a booming oil industry. The media system can be seen as an integral part of this welfare state. Comparatively, the Nordic region including Norway, stands out because of (1) a long history of universally available communication systems, emphasizing these as public goods; (2) the early development of and stillstrong commitment to institutionalized editorial freedom of the press; (3) an extensive cultural policy for the media, and; (4) a tradition for consensual policy-making between key stakeholders (Syvertsen et al 2014). The Nordic countries tend to rank high on indexes of new media technology use, and also perform well on lists measuring editorial freedom and related concepts (Syvertsen et al 2014). In 2017, Norway still has a publicly funded public service broadcaster enjoying high user numbers, as well as a diverse press structure both locally and nationally, subsidized by VAT exemption and some direct press subsidies.

Taken together, these factors create an image of Norway as a 'critical case' for studying media credibility and perceptions of journalistic bias held against professional norms. One key historical development does, however, need to be highlighted: As argued by Hallin and Mancini (2004) in their much-quoted work on media systems, countries similar to Norway are characterized by strong institutionalized professionalism in the media and strong state intervention through positive regulatory measures, but also by a shift away from political pluralism towards a neutral commercial press. For roughly 100 years, from when the parliamentary system was formed in the late 19th century, up until the last decades of the 20th century, the Norwegian press was a so-called party press. Political parties owned, staffed and directed newspapers (Høyer 2005, Syvertsen et al 2014 p. 53ff). As a result, each newspaper represented one political view or ideology. From the 1970s, the dominance of the party press diminished slowly (e.g. Allern and Blach-Ørsten 2011). By the end of the 1990s, a public report found only one newspaper declaring party attachment in its preamble (Syvertsen et al 2014 p. 54).

Yet, while the party press as an organizational rule is long gone, replaced by a commercially owned and professionally run press, studies have found content to follow political partisanship (e.g. Allern 2007; Allern and Blach-Ørsten 2011). Such lines of partisanship have laid the basis for assumptions of bias in the Norwegian media. This argument has in particular been forwarded by the Progress Party. The former leader of the party for many years made a point of referring to the public service broadcaster as 'the Labour party's national broadcaster'. Originally based in an anti-tax liberalist movement, the Progress Party has steadily transformed into a moderately populist party with growing voter appeal in the last decades of the 20th century, promoting protests against established political elites and for restrictive immigration policies. Until recently, the Progress Party was the only political party who actively profiled itself through immigration policies, and for voters critical to established levels of immigration, the party became the likely choice (Aardal & Berg, 2015).

Following the 2011 attack by a sole extreme right terrorist in Oslo and on Utøya, the extent to which mainstream media shut out deviant voices and thereby fueled isolated echo chambers, has also been a topic of public debate (Ihlebæk & Løvlie, 2013; Ustad Figenschou & Thorbjørnsrud, 2016; see also Ihlebæk and Thorseth, Ch. 5). This debate, along with claims of mainstream media being biased against the political right, especially on the issue of immigration, persisted in the following years, especially as the Progress Party took office along with the Conservative party in a minority government following the 2013 election.

Existing data on how Norwegians assess the media and journalists provide a mixed bag of results. In 2016, a survey found merely 16% of respondents expressing quite low or very low confidence in journalists, while 47% claimed to have quite or very high confidence (NMD 2016). The same survey – which in different forms has been repeated since 2007 – reports a slight increase in respondents who express less or no confidence in the media in general (up from 16% in 2007, to 20% in 2016), and a corresponding slight drop in those having some or high confidence in the media in general (down from 83% in 2007, to 79% in 2016). Our own survey data from 2013 show that with regard to confidence in the media in general, a majority think that the media provide important information (56%) and that they offer a diverse debate (53%). However, considerably fewer believe that the media are able to cover an issue from different angles (38%). The degree of confidence is related to party preference; people who identify with the left have higher confidence than those on the right. A substantial minority of the respondents doubt the media's ability to critically evaluate their own role (37%) (Staksrud et al, 2014). On this basis, we find that Norway provides an interesting case for studying in detail the perceived bias of journalism.

# **Data and method: combining quantitative and qualitative analyses**

In our analysis, we use both quantitative and qualitative data from the two population surveys on freedom of speech, carried out in 2013 and 2015. In the quantitative analysis, we rely mainly on the 2013 survey, but also add two questions from the 2015 survey. About half of the respondents in 2013 also answered the survey in 2015. In the qualitative analysis, we explore an openended question from the 2015 survey.

## Quantitative data

The dependent variable in the quantitative analysis – *biased journalists* – is an index consisting of mean scores on two survey items, meant to describe key aspects of journalistic boundarymaking: to what extent journalists (1) '… favour sources with similar opinions to themselves', and (2) '…allow their personal political views to affect them' (Pearson *r*=0.675). Answers were given on a 1 (disagree) – 5 (agree) scale. Don't know answers were recorded into the value '3'.

The main independent variables in the quantitative analysis are *party choice*, *attitudes towards immigration and immigrants*, and *trust in the media*. *Party choice* is based on a question on which party respondents voted for in the previous national (Storting) election. In order to measure *attitudes towards immigration and immigrants* we constructed an index consisting of three survey items, of which (1) and (2) were included in the 2015 survey and (3) was included in the 2013 survey: (1) 'Most immigrants enrich cultural life in Norway', (2) 'Immigration is a serious threat to our national distinctiveness', and (3) 'We have enough immigrants and asylum seekers in this country'. Answers were given on a 1 (disagree) – 5 (agree) scale. Don't know answers were recorded into the value '3'. The three items were mean-centred before creating an index consisting of mean values. On the two 2015 items there are about 50 percent missing values, meaning that about half of the respondents only have values from the third item. However, among those responding to all three items the alpha reliability score was .835, suggesting a strong correlation between responses given in 2013 and 2015.

*Trust in the media* relies on a single survey item: 'How much trust do you have in the following institutions (…) media'. Answers were given on a 1 (no trust) – 5 (great trust) scale. Don't know answers were recorded into the value '3'.

Additionally, in the models we control for gender (women=1), age, education (higher education=1), and immigrant background. About 25 percent of the sample consists of respondents with immigrant background, but these are not representative of the immigrant population in Norway. This variable is therefore only included in order to 'control out' the effect of this group.

### Qualitative data

In the qualitative analysis, we explore an open-ended question from the 2015 survey about which groups in society respondents believe are underrepresented in news stories (The specific question wording was 'Do you think that the voices of certain groups in society are underrepresented in Norwegian news stories?'). A total of 414 respondents (21% of the full sample) gave a more or less comprehensive response. The purpose of our analysis of this material is to gain a better understanding of the results from the quantitative analysis, namely why some people are more inclined than others to believe journalists are biased. The qualitative analysis is based on a close reading of the material, where central patterns and categories were detected. These categories were then analyzed with regard to party preference.

# **Results: Perceptions of journalistic bias**

We present our results in three steps, of which the first two are based on the quantitative data and the third step focuses on the qualitative data. First, we take a look at the answer distribution of the two survey items included in our dependent variable. Second, we estimate the net impact of our independent variables on the *biased journalists* index. Third, we explore replies to the open-ended question on which groups in society respondents believe are underrepresented in terms of voices in Norwegian news media.

### Quantitative data

Figure 4.1 shows the distribution of answers to respondents' perception of the two survey items: the extent to which journalists are thought to prioritize opinions that correspond with

**Figure 4.1.** Perceptions of biased journalists. Percent.

their own views; and the extent to which respondents express discontent in journalists' separation of their own personal political opinions from the news stories they report, and the ways they go about it.

Figure 4.1 displays generally low levels of trust in journalists on both issues: A clear majority believe that sources will be favoured when they are in line with the opinions held by journalists (65% to some or a great extent), and even more clearly, the majority express a lack of trust in the journalists' ability to keep personal views distinct from their professional practices (72% to some or a great extent). It is however important to note that few respondents choose the most extreme option ('To a great extent'), rather most respondents choose the more moderate option ('To some extent'). Thus, the overall picture is not totally one-sided.

The next step in the analysis focuses on identifying characteristics of the groups that voice the most discontent. Results from three regression models are summarized in Table 4.1. The dependent variable is perceptions of journalistic bias (1=unbiased – 5=biased). Model 1 estimates the impact of party choice, model 2 adds attitudes towards immigrants and immigration, whereas model 3 adds trust in the media. Control variables are included in all three models.

Beginning with model 1 we see a tendency that the farther one goes to the right on the political spectrum, the more likely respondents are to believe that journalists are biased. Compared to Labour voters (reference category), those who voted for the centric Christian Democrats or the Liberal Party are located 0.2 points higher on the five point bias scale. The corresponding coefficients for voters of the right-wing Conservative Party and the Progress Party are 0.3 and almost 0.5. Thus, Progress Party voters are a half scale point more likely than Labour Party voters to believe that journalists are biased.


#### **Table 4.1.** Perceptions of journalistic bias. OLS regressions.

† p ≤ 0.1, \* p ≤ 0.05, \*\* p ≤ 0.01, \*\*\* p ≤ 0.001.

NOTE: Immigrant background is not representative for the immigrant population in Norway.

Moving on to model 2 we see a clear tendency that negative attitudes towards immigrants and immigration is correlated with perceptions of journalistic bias. For each increase on the five point attitude scale, the bias scale increases by 0.115. In other words, those expressing the most negative attitudes towards immigrants place themselves more than 0.5 points higher on the bias scale compared to those expressing the most positive attitudes towards immigrants.

Adding this variable does not change the party coefficients for the Liberal Party or the Christian Democrats, and only reduces the coefficient for the Conservative Party by about 0.05 points. This suggests that attitudes towards immigrants are not particularly relevant for these voters' perceptions of journalistic bias. The coefficient for the Progress Party is however reduced by as much as 0.145 points, suggesting that much of the distrust in journalists remaining unbiased among Progress Party voters is explained by their negative attitudes towards immigration and immigrants.

Finally, adding trust in the media to the equation (model 3), we see that this variable is negatively correlated with perceptions of journalistic bias. This means that the more people trust the media in general, the less likely they are to believe that journalists are biased. The difference between the most and least trustful is more than a scale point on the bias scale. Adding the trust in media variable, the coefficients for the Liberal Party and the Conservative Party remain more or less unchanged, whereas the coefficient for the Christian Democrats and the Progress Party is (further) reduced. This suggests that distrust in the media is part of the explanation why these two groups of voters believe journalists are biased. In fact, in model 3 the coefficient for the Progress Party is almost down to the level of the other mentioned parties, indicating that negative attitudes towards immigrants and distrust in the media are two important reasons why Progress Party voters have a stronger tendency than other voters to believe that journalists are biased.

In the next step of the analysis we dig deeper into this finding by exploring answers to the open-ended question about which groups in society respondents believe are underrepresented as speakers in the news media.

## Qualitative data

In the open-ended question about media bias, the respondents were asked to state any groups (if any) they felt were neglected or underrepresented in Norwegian news media. Given that the question only invites respondents to think about failures, the input does not shed light on the positive aspects of people's perceptions. Instead, the material provides a rich basis for exploring the issue of (dis)content with the selections or boundary-making undertaken by journalists. Our interest lies in exploring dissimilarities among groups of citizens. Given the results from the quantitative data showing differences among voter groups, and that these differences for some voter groups are related to attitudes towards immigration and media trust, we here focus on identifying different factors brought up in the responses. We identify, illustrate and discuss factors of discontent related to a general criticism of media tabloidization, ignorance of vulnerable groups, as well as a centre vs periphery criticism. Having described these types of media critique, we then concentrate on the significant discontent with the representation of the issue of immigration, and how this is associated with a perceived bias against the political right.

Many of the respondents who answered the question expressed, in their own words, not only what type of actors they believed received too little media attention, but also which groups or interests disproportionally dominated the news media. These answers give an idea of the level of frustration or even cynicism in parts of the public. Some briefly mentioned a group or two they felt needed more attention – ranging from religious minorities or disadvantaged groups, to age groups, specific interest groups, and even people with certain hobbies. Other respondents based their discontent on scepticism not only to the news media, but to the established elites and the democratic system as a whole.

An initial factor of discontent could be described as a general critique of media tabloidization and sensationalism. This was voiced through comments on the lack of space given to the moderate majority, the 'Average Joe' in the media, who do not hold extreme views and are unwilling or uninterested in making outrageous claims to instigate controversy. In an illustrative reply, this is expressed in the following way: 'Those who do not seek attention, who are not extreme, who do not seek confrontation'2 . Some respondents even demonstrate an understanding of the inescapable fact of confrontation in the media, but still miss those who represent the status quo:

It's in the media's 'nature' to let the extreme speak up the most. Debates where most agree quickly turn boring! But I think the big majority feel things are pretty ok as they are, and this view is expressed too seldom in debates.

We find this aspect of media critique among voters from several parties, ranging from the far-left Red party through Labour and the Conservative party to the right-wing Progress Party, and also among those who did not state party preference. A related way of expressing similar opinions was found in comments on the lack of reasoning, or the lack of knowledgeable speakers on specific issues. Again, this can be interpreted as a discontent

<sup>2</sup> All quotes are translated from Norwegian by the authors.

chapter 4

with the selection of high profile, vocal sources, at the cost of balanced, rational and reasoned input in the coverage of a given issue.

A second prominent line of critique points in the direction of vulnerable groups. We find respondents including such groups in general, but also singling out specific ones. In one instructive example, a Red party voter states: 'Weak groups such as those receiving disability benefits, those receiving social security benefits, disabled persons and the mentally ill.' For some respondents, such opinions are linked to patients in health care and their next of kin, or to recipients of specific welfare schemes. We do not know to what extent such replies are linked to the respondents' own experiences, but we do find various disadvantaged groups being mentioned by voters across the political spectrum, from the radical Red, to the political centre and to the Progress Party.

The third factor has to do with centre vs periphery, where respondents pointed to the lack of speakers from anywhere outside the most central areas of the capital city of Oslo. Antagonism between urbanised centres of power and rural areas is well known in most societies. Norway has a dispersed population, traditionally strong primary industries, and a long tradition of upholding rural development through diverse regional policies. In addition, the oil industry along with newer export industries such as salmon farming, are located off the Western coast. As such, the centre vs periphery statements should not be surprising.

For some, the criticism is of the mainstream media's Oslocentricity. For others, the discontent is linked to a portion of the public, for instance: 'the normal part of the population who live in the countryside without public-transport facilities, and have a [household] gross income below €50 000.' In effect, this is more a critique of the media's failure to broaden their range of sources, than it is a general criticism of the centralization of power in the capital city. We find this sort of critique across the political spectrum.

While the factors and groups mentioned so far address a perceived general tendency of source selection undertaken by journalists, one specific issue does stand out in the replies: Immigration. Immigration, then, constitutes the fourth factor of discontent. Comments on immigration do not all land on one side of the debate. Rather, it is a vehicle to express opposite forms of discontent. For some, the problem with mediated public debate in Norway is the lack of nuanced representation of immigrants. One example quote would be: 'The average Muslim', he/she who does not have extreme ideals and wishes to live a good life in peace'. On the opposite side, other respondents point to the lack of voices critical to immigration: 'Persons who are opposed to immigration. They are often/always accused of being racists'.

These two examples illustrate how the respondents are divided on the issue in accordance with party preference: By and large, those pointing to the lack of immigrant voices, or to the misrepresentation of specific immigrant groups, are found among leftist party voters. Those who express a lack of voices critical towards immigration are scattered on the political spectrum from the political centre to those who do not vote. In accordance with the findings from the quantitative analysis above, this concern however, is conspicuously expressed most often and with most frustration by Progress Party voters.

In the replies addressing the immigration debate, respondents hint at a political spectrum, and the media's bias towards the left. Beyond the specific issue of immigration, other respondents relate their criticism to the left-right spectrum of politics chapter 4

explicitly. This is a fifth dimension of discontent found in the open-ended answers. These respondents relate to the parties on the right, and they feel that the right is getting too little attention. For voters from the Christian Democrats, such failures can be linked to religion too: 'Religious and political minorities. Political minorities on the right are censored much more strongly than those on the left '.

For other respondents, the issue triggers a more targeted critique of the political left. In the following illustration, the left is evoked by the colours of the previous left-centre coalition government and the right is tagged with the then-in-office Conservative-Progress Party government:

'The Norwegian media are to a large extent leftist – in other words there is a lot of positive writing/talking about the redgreen side of Norwegian society, and at the same time a very clear focus on negative stuff on the blue-blue side. I observe this difference almost on a daily basis.'

In other replies, the same line of thinking is expressed more precisely as a criticism of the Labour Party: 'I would say, to put it bluntly, all people who do not vote Labour speak out all too seldom in Norwegian news media'. Whereas the former quote was attributed to a Conservative voter, the latter was written by a voter from the Progress Party. In our analysis, the subtle difference between them is illustrative: Progress Party voters offer a harsher discontent with the media, and link the problem to the Labour party, whereas Conservative voters and those in the political centre, offer more general critiques without mentioning a specific group or party.

This fifth factor of discontent with journalistic bias also triggers a more fundamental scepticism. Some of the Progress Party voters express distrust in the openness of the democratic system and the public sphere in general. This is coupled with a principled idea of the role of journalists: the media should be impartial, and journalists should be 'apolitical'. 'Due to the fact that the press is by and large owned and run by the Labour press and their opinions, they should act as apolitical media and also give a voice to others who have different opinions from themselves'. Such a statement is not accurate concerning current media ownership in Norway. Still, it expresses, first, a perception of Labour bias reminiscent of the party press era, and, second, it clearly states a basic expectation of journalists' balanced and transparent professionalism.

Importantly, we find no criticism in the material of Norwegian media blocking the voice of the political left.

To sum up, we see here traces of five different forms of discontent: (1) general criticism of media tabloidization and sensationalism, (2) a lack of attention given to speakers from vulnerable groups, (3) a centre vs periphery criticism, (4) the contested issue of immigration and, finally, (5) the perceived bias against the political right.

### **Discussion and conclusion**

This chapter has looked at the issue of freedom of speech and boundary-making from the perspective of the citizens. With a starting point in a widespread perceived scepticism towards the media, coupled with a crisis for the business models journalism has traditionally relied on, the aim of our analysis was to explore how citizens evaluate the independence and impartiality of journalists. The case of Norway, we argued, is particularly interesting. It is a 'critical case' in the sense that Norway has high levels of institutional trust and a media system described as professional and with a high audience reach.

Our results show the distrust in these aspects of journalists' boundary-making to be quite high, and to go beyond any specific subgroup of the population. The results indicate that a large majority of the general public in Norway have little faith in the conventional journalistic claim that the news profession represents a type of disinterested or third party interpretation of current affairs.

This type of scepticism stands out as strong compared with levels of confidence in the media as a whole. Compared to the trust in other institutions, people's trust in the media is low, almost half of the respondents answering that they have no or little confidence in the media. When asked more specific questions related to diversity, the results are more positive. More than half of the respondents answer positively when they are asked if they believe that people with different opinions are allowed to speak in Norwegian media (Staksrud et al, 2014). This apparent difference in trust towards the more roughly defined 'media' versus the journalistic profession, and thereby the more narrowly defined news media, can be interpreted in different ways. In general, though, the finding might be bad news for those who wish to distinguish the traditional mainstream news media as particularly trustworthy with regard to impartiality and integrity.

Through quantitative analysis, we have pointed to different variables that help us to understand why the level of discontent is so high, and what kind of explanations might lie behind different groups. The results show that party preferences, attitudes towards immigration and trust in media are related to the perception of journalistic bias. We also identify differences among political groups, e.g. that negative attitudes towards immigrants and distrust in the media are more important reasons for Progress Party voters to believe that journalists are biased, compared to people with other party preferences. In our explorative analysis of the open-ended question about those who do not get to speak up in the Norwegian news media, we identified five different factors, and showed how two of those – the immigration issue and the perceived bias against the political right – seem to follow the same kind of pattern found in the quantitative material.

We do not know if the comments on this issue are related to the media covering immigration differently than other issues, or are due to the political (and humanitarian) urgency of the issue. But the salience of this issue indicates that immigration engages people particularly. Furthermore, their concern over issues related to immigration seems to be related to their evaluation of media performance, as indicated in previous studies.

Our analysis confirms the impression of a widespread discontent with the way journalists handle their own political views in their professional practices: The majority of the Norwegian public have little trust in the journalists' ability to draw legitimate borders for freedom of speech when it comes to these particular issues, and this distrust is linked to political preferences and to the pertinent societal issue of immigration. In spite of the old declaration of the 'death of the party press', and almost five decades with media outlets defined as professional and nonpartisan, the audience is not convinced. Rather, a qualified guess is that people are aware of the documented leaning of journalists towards the liberal left (NMD 2016), and that they to some extent support the popular contention from right wing voices about a media bias that reflects this political position.

Such findings could be interpreted as a blow to the journalistic news media as an institution in and for the democratic public sphere. Professional journalism in Norway does not seem to be a particularly exceptional case in this way. Rather, the Norwegian media system looks to be facing similar challenges as do legacy media in other liberal countries: Low confidence in the press follows partisan cleavages and a deeper level of distrust in the democratic quality of the public sphere. Thus far, the impact of

### chapter 4

these trends seems less grave in Norway compared to states with other media systems and political cultures. At the time of writing, the turmoil surrounding US President Donald Trump and the rise of populist parties in several European countries has pushed the issue of journalistic bias to the top of the agenda. This invites replication of our study to track developments over time, and to measure the effect of such general trends in specific societal settings. If the premise is that freedom of information and open democratic deliberation is intimately linked to widely trusted sources of information, some worry is warranted. Tendencies of polarization, group thinking and scepticism need to be taken seriously into account – and to be researched in different societies.

Still, we should be careful not to overstate the ramifications of our findings. Our study is based on data from two surveys, but does merely tap into the topic of trust in the media through one specific angle. Further work should extend the analysis of indicators of perceptions of journalistic bias by zooming in on specific media outlets, preferably linking media use to perceptions of trust.

Fundamentally, one could argue that a critical stance towards the media is key for citizens wanting to take part in the democratic process. The design of our study rests on the assumption that news in general should be impartial and adhere to established professional journalistic norms. The findings of widespread criticism could, then, also be seen as a signal of changing expectations, and a changing role for the media as channels between the public and the rulers.

## **References**

Aalberg, T., & De Vreese, C. (2017). Introduction. Comprehending Populist Communication in Europe. In T. Aalberg, F. Esser, C. Reinemann, & J. Stromback (Eds.), *Populist Political Communication in Europe* (pp. 3–12). Taylor & Francis: Routledge.

Aardal, B., & Berg, J. (2015). *Valg og velgere*. Oslo: Cappelen Dam.

Ackerman, J. M., & Sandoval-Ballesteros, I. E. (2006). Global Explosion of Freedom of Information Laws. *The Administrative Law Review* 58, 85–123.

Alexander, J. C. (2006). *The Civil Sphere.* Oxford: OUP.


*Status for ytringsfriheten i Norge. Resultater fra befolkningsundersøkelsen 2014.* Oslo: Fritt Ord, ISF, IMK, FAFO.


### Chapter 5

# Editorial perspectives on the public debate on immigration

*Karoline Andrea Ihlebæk, Department of media and communication University of Oslo Ingrid Endresen Thorseth, Department of media and communication University of Oslo*

A key responsibility of the news media is to facilitate the public debate; thus newspaper editors have traditionally wielded a good deal of power in regard to the selection and rejection of texts for publication. However, the dominant position of the editorial driven news media as an arena for public discourse has been altered dramatically as a result of the growing number of arenas where ordinary citizens can express their opinions. As a result, the role of the editor as a gatekeeper has also been transformed, and new forms of editorial mechanisms have come into use. In this chapter, we investigate editorial perspectives on boundaries by using debates on immigration as a point of departure. The study is based on qualitative interviews with opinion editors in national and regional newspapers carried out in 2014 and 2016. We focus in particular on how editors talk about the management of the reader section of the newspaper, including letters to the editor, comment sections, and social media. We argue that immigration as a topic represents an interesting entry point for studying editorial attitudes because, as societies become increasingly diverse and multicultural, such debates have become more prevalent and demanding. These discourses are often described as heated, emotional and polarised, and they are frequently accompanied by discussions about the normative boundaries in the public sphere.

### **Introduction**

In recent years, the role of the editor as a gatekeeper has been transformed, although editors in the news media still represent a form of authority grounded in the professional codes of journalism. This form of authority, we argue in this paper, is manifested in the idea that we still need qualified gatekeepers to filter information and help us to 'make sense of the world' (Barber, 2004, p. 44). However, as recent public debates have indicated, trust in how the media execute this form of authority has regularly been questioned (see also chapter 4), and the news media have to defend their legitimacy by pointing out the professional ethics and norms that guide their work.

Thus, this chapter will investigate editorial perspectives on the public debate on immigration. The focus is on how editors discuss the management of the reader section of the newspaper, including letters to the editor, comment sections, and social media. As mentioned above, the topic of immigration represents an interesting entry point for studying editorial attitudes because, as societies become increasingly diverse and multicultural, such debates have become more prevalent and demanding (Balch & Balabanova, 2011; Brochmann, Hagelund, Borevi, Jønsson, & Petersen, 2012; Horsti, 2008). Furthermore, such discourses are often described as heated, emotional and polarised and they are frequently accompanied by discussions about the normative boundaries in the public sphere (Bangstad, 2013; Eide & Nikunen, 2011; Figenschou & Beyer, 2014; Midtbøen & Steen-Johnsen, 2016; Thorbjørnsrud & Figenschou, 2016).

This study uses two rounds of interviews with newspaper editors responsible for managing the debate section of the newspaper. The first round of interviews was conducted in 2014 and focused on general issues related to freedom of speech and the administration of public debates on multiple platforms. The second part of the study was conducted in 2016 in the aftermath of the migrant and refugee crisis that escalated during the autumn of 2015. This dramatic situation gained comprehensive media coverage and produced heated public debate about how this situation should best be handled. While migration and related topics like integration, religion, Islam and terrorism are regularly discussed in the media, the situation in 2015 can be described as a peak moment when attention to such topics was particularly prominent in newsrooms around the world (Askanius, Linné, Berry, Garcia-Blanco, & Kerry, 2015).

### **Theoretical perspectives on gatekeeping**

The news media have traditionally played an important part as an arena for public debate. This perceived obligation has been part of what has been described as a 'social contract' between democracy and journalism, indicating a certain dependence. The news media should serve the public by providing reliable information, act as a public watchdog and function as an arena for public debate, while, at the same time, journalism requires a democratic system that secures such principles as freedom of information, freedom of expression and freedom of the press, which are necessary to fulfil these tasks (Christians, Glasser, McQuail, Nordenstreng, & White, 2009; McQuail, 1992; Steel, 2012; Strömbäck, 2005). This interdependence has, in the Norwegian context, been met by political measures like press subsidies, tax reduction and strong public service media, which are meant to ensure a well-functioning and diverse national media sector (Hallin & Mancini, 2004; Syvertsen, Enli, Mjøs, & Moe, 2014).

How the media fulfil their democratic responsibilities has been of pivotal concern to media scholars. In the following we will in particular explore perspectives on the media's gatekeeping power when facilitating the public debate, and how their concept of boundaries in the debate is influenced by professional norms and ethics. Studies of boundaries are about analyzing power, in other words how boundaries are drawn, sustained, negotiated, contested and changed by different stakeholders (Abbott, 1995; Carlsson & Lewis, 2015; Gieryn, 1999; Lamont & Molnár, 2002; Lewis, 2012). An important dimension in how boundaries in the public debate are drawn by journalists and editors is through their role as gatekeepers. On the most general level, gatekeeping studies have been concerned with information control and how and why some kinds of information become news (Shoemaker & Vos, 2009; Shoemaker, 1991). Since the iconic gatekeeping study carried out by White (1950), in which he observed how an editor he called 'Mr Gates' filtered information for publication, several studies have explored individual as well as organisational and institutional mechanisms at play in the news-making process (Gans, 1980; Reese & Ballinger, 2001; Shoemaker & Vos, 2009; Shoemaker, 1991; Tuchman, 1978). While earlier gatekeeping studies have focused predominantly on journalistic processes, the gatekeeping metaphor is also relevant when studying how the news media orchestrate and facilitate the public debate (Bruns, 2008; Ihlebæk, 2014; Ihlebæk & Krumsvik, 2015). The editors responsible for such content have the power to select and reject texts based on particular criteria. Studies have, for instance, explored how professional norms like relevance, quality, novelty and originality are commonly used to explain why some letters to the editor get published and some don't (Kleis-Nielsen, 2010; Wahl-Jorgensen, 2001; Wahl-Jørgensen, 2002). In addition, editors can take a more active role and order articles and commentaries from people they want to partake in the debate.

In more novel forms of debate formats like comment sections, blogs and social media sites, new forms of editorial control mechanisms have been developed with the aim of maximizing the participation from the audience, but minimizing uncivil behaviour (Carpentier, 2001; Carpentier, 2009, 2011; Goode, 2009; Holt & Karlsson, 2011; Hujanen, 2016; Ihlebæk & Krumsvik, 2015; Larsson, 2011; Lewis, 2012; Mitchelstein, 2011; Ruiz et al., 2011; Singer, 2006; Singer, Paulussen, & Hermida, 2011). Editors have, for instance, experimented with different forms of design, levels of identification, rules for participation, limiting access through closing the comment fields at night or by restricting the type of newspaper articles users can comment on, as well as by using moderators who can delete inappropriate comments and throw people out (Ihlebæk, 2014).

Economic, ethical and legal factors come into play when editors manage the debate and decide on appropriate control mechanisms. First, the news media are commercial enterprises and as Doyle (2013) points out, most decisions in the media industry are influenced by economic factors either explicitly or implicitly. In the context of managing the public debate, the available resources (allocated staff, technical solutions etc.) at any given time is of course of importance (Ihlebæk, 2014). Furthermore, the need to attract readers and advertisers might impact editorial decision-making concerning what kind of content that is published or not, and in the online environment news organisations gather extensive knowledge about what kind of content that is successful in generating clicks, shares and likes. In relation to the management of comment sections, studies have indicated that news organisations have been reluctant to impose stricter control in such formats partly because they feared it would lead to less activity and traffic (Ihlebæk, 2014). However, the motivation to enhance participation in quantitative terms, had to be balanced against the need to take into account the effect certain forms of control, or lack of control, had on the quality of the content in these kinds of formats, which leads us over to the second dimension. The news media are democratic institutions guided by particular normative ideals. The idea that journalism serves the public, constitutes an important part of journalistic ideology and allows journalists and editors to legitimize their own position and authority as gatekeepers (Deuze, 2005; McQuail, 1992). Furthermore, media professionals defend their autonomy based on the adherence to professional norms and ethics that guide their work, and that separate them from other publishers and debate arenas. Such norms could be independence, balance, neutrality, factuality and accuracy (Alexander, 2006; Alexander, 2016; Carlson, 2015; Singer, 2015).

In the Norwegian context, the notion of editorial responsibilities can be found in *The Rights and Duties of the Editor*, and ethical guidelines are outlined in *The Code of Ethics of the Norwegian Press.* These standards have been developed by the industry as a self-regulatory measure, which is central to the Northern European Democratic-Corporatist media system (Hallin & Manchini, 2004). The *Code of Ethics* is not a legally binding document, but it is supported by all editorial-driven news organizations in Norway. If a newspaper breaks the *Code*  *of Ethics* a complaint can be made to the *Norwegian Press Complaints Commission* and if found guilty the editor needs to publish a public apology. The self-regulatory system, in other words, represents a particular framework for the established editorially driven media (both online and offline) that differs from their non-editorial driven competitors. In the guidelines, the social responsibility of the press is stated as advice on how to conduct their work. In relation to their role as an arena for public debate, the code of ethics emphasizes the need for diversity which is stated in point 2.1: 'The press has important functions in that it carries information, debates and critical comments on current affairs. The press is particularly responsible for allowing different views to be expressed' (point 1.2). Furthermore, norms like factuality, fairness, respect and truthfulness are emphasised. The *Code of Ethics* also states that those who are subjected to strong accusations or attacks shall have the opportunity to reply (points 4.14 and 4.15).

A third factor that is relevant in regards to how editors manage the debate is based on legal boundaries (Amos, Harrison, & Woods, 2012; Bing, 2008; Steel, 2012; Wessel-Aas, 2013). As in most countries there are limits to free speech in Norway (see chapter 2). According to *The General Civil Penal Code*, editorsin-chief can be held legally responsible for what is published in their newspaper. While illegal forms of expressions are easily avoided in texts that have been through an editorial pre-editing process, post-moderated forms represent a different risk. Bing (2008) argues that the law implies that the publication must ensure that they have good routines and control mechanisms for detecting potential violations. While questions concerning the editorial responsibility of news organisations' online services remain largely untested in the Norwegian legal system, a point of reference is the case against the Estonian newspaper *Delfi* in the European Court of Human Rights, which held the newspaper responsible for anonymous and defamatory statements in the comment section.

## **Method**

The study rests on twenty-two qualitative interviews with editors who are responsible for the debate.i Qualitative interviews are useful, in the words of Brinkmann and Kvale (2015) when the goal is to 'understand the world from the subjects' point of view, to unfold the meaning of their experiences, to uncover their lived world prior to scientific explanations' (p. 3). In the context of this article where the aim is to gain insight into editorial perspectives on the management of the public debate, it was necessary and helpful to apply this method and to get as much information as possible on the topic of concern.

The informants can be categorised as 'elite informants'. There is no clear definition of what constitutes an elite, but the term is often used to define informants who, through their profession or role, exercise some kind of power or authority that non-elites do not have (Harvey, 2011; Hertz & Imber, 1995). In our study the informants have been chosen because they represent a powerful position in guarding the public debate, even though many editors don't necessarily see themselves this way, particularly as the public debate expands onto a multitude of platforms. A general challenge with elite interviews is that the informants might want to control the interview, and that they use standardized phrases to

i Newspapers included in the study: 2014: Aftenposten, Bergens Tidende, Dagbladet, Dagsavisen, Drammens Tidende, Klassekampen, Nationen, Nordlys, Stavanger Aftenblad, Vårt Land og VG. 2016: Aftenposten, Agenda, Bergens Tidende, Dagbladet, Dagens Næringsliv, Drammens Tidende, Klassekampen, Nordlys, Stavanger Aftenblad og VG.

portray their work or organisation in a favourable way. While this is hard to avoid, we have tried to counter this by probing for specific examples and dilemmas that they experience in their work.

The newspapers in the sample are predominantly national and regional. Twelve of the interviews were carried out in 2014 and then ten more in 2016. The first round of interviews focused on general issues relating to freedom of speech and the administration of controversial debates on multiple platforms. The study identified immigration as one of many challenging topics to manage. The second part of the study focused specifically on the immigration debates and explored issues like polarisation, boundaries and the inclusion and exclusion of voices.

All the interviews have been transcribed and analyzed using the analytic software *NViVO*. The software helps to systematically categorize the data through coding the material by particular 'nodes'. The nodes were inspired by the research question and developed by thorough reading of all the interviews. In the end six main nodes were used: debate culture, debate platforms, immigration, racism/hate speech, editorial practices, challenges/dilemmas. The nodes overlapped and the same text could be coded several times. In the analytical process, we could then read texts relating to one of the nodes across all interviews.

It is necessary to point out that we do not analyze change over time, even though changes in editorial perspectives might occur; Rather the interviews complement each other with insights on how editors talk about the management of public debates and how they practice boundary work. Finally, using qualitative interviews as the main method limits the possibility of saying anything about how actual *practices* develop. Instead this paper contributes some insight into how editors *talk* about the debate, based on the professional ethics of journalism.

## **Managing the debate in a fragmented public sphere**

To be able to understand editorial perspectives on the immigration debate, it is first necessary to gain more insight into the kind of portfolio editors are responsible for, and how they talk a about their own role in a fragmented public sphere.

As referred to in the theoretical section, editors are responsible for many kinds of debates in the newspaper domain, and most newspaper have an active online strategy apart from one newspaper in the sample that on the contrary focused mainly on the printed paper. In terms of editorial management, a general distinction can be made between formats subject to pre or post forms of control. The first category involves formats like letters to the editor, commentaries and op-ed-articles that are commissioned by the editor, or submitted by a person or organisation, and either accepted or rejected by the editor based on particular criteria. The latter category consists of post-moderated forums, like comment sections in the newspaper or on Facebook, online forums or blogs where anybody can participate without going through the scrutiny of the editor.

These pre and post formats fulfil different functions in the newspaper and are guided by different principles of editorial management. Pre-edited formats are meant to set the agenda, and the editors have much higher expectations as to the quality of such texts than for the post-moderated ones. Pre-edited debate formats are by some editors referred to as the 'elite debate' or 'top division' that the readers can then react to. It is important, according to the editors, that the authors manage to present their points clearly, with civilized language and ideally that they have something original or newsworthy to say. Furthermore, it is important that the pre-edited pages represent a diverse range of opinions. It is common that editors from national newspapers actively commission op-ed articles to achieve such goals, while editors in regional newspapers more often select from the texts that are sent in to the newspaper. Some editors claim that up to fifty per cent of published op-ed articles have been commissioned.

Comment sections and online forums are, on the other hand, meant to facilitate spontaneous reactions and debates that resemble everyday talk between ordinary people. Editors consequently accept a different form of language and type of argumentation in such formats than in pre-edited debates. Most newspapers have developed some kind of ground rules for participation in the comment fields that participants have to follow. Such rules often promote civility and respect, as well as prohibiting racism, hate speech or defamation. Participants who break the rules can be thrown out by the moderators. While many editors claim that they do not experience this as a major problem, comment sections are nevertheless regularly criticised for being uncivil and sometimes also racist or xenophobic. However, editors defend (with a certain ambivalence) their existence based on ideals like democratic participation and inclusion. Recently though, many newspaper have, or are thinking about, shutting down the comment section of the newspaper altogether and moving such services to Facebook. Editors in the study argue that they have to be on Facebook because it is where their audience are. However, they are also aware of that when they outsource the comment fields to Facebook, they at the same time become more dependent on the algorithmic power of the platform and the technical solutions it provides for moderating practices. This is a double-edged sword as one editor bluntly put it:

On Facebook, it is not possible to turn off the comment section, and it is not possible to delete the entire comment section unless you delete the whole post. And if you delete the whole post you get punished by Facebook's algorithms... Facebook gives us limited opportunities to be the kind of editor we would like to be.

This leads us to the final point. In a fragmented and hybrid media landscape where public debates take place everywhere all the time, editors have become more dependent on utilizing social media services to be able to defend and maintain their role as agenda-setters. Many editors in the study explain how they use Twitter and particularly Facebook not only to invite audiences to comment, but to attract new voices and to pick up on interesting debates. It is a major advantage for editors, then, to gain the skills and knowledge necessary to exploit such platforms, as this editor explains:

I personally have between 1500 and 2000 friends. I use Facebook professionally, so I monitor debates and public voices through my personal Facebook profile… and we have been trained to search effectively so that if there is a story trending in a particular geographic region we can search for people who have been posting something about it, in that specific area.

In relation to the immigration debate, the editors who emphasise the importance of an active social media presence, have to be constantly aware of particular people who are active in such debates. Editors can then try to include them in their network, or invite them to write something for the newspaper. However, this is sometimes a complicated task, as this editor explains:

On Facebook, you don't always know where the most interesting debates are. Is it on the page of Kjetil Rolness2ii? Is it somewhere else? Are you friends with that person? Have some of the participants blocked you? Can you read everything that is there?

The growing dependence on Facebook, in other words, creates some extraordinary opportunities for editors to include new

ii Rolness is a commentator who writes for the tabloid newspaper Dagbladet. He also has a very active Facebook-page where immigration is frequently debated.

and original voices, while at the same time the increased importance of global intermediaries creates some dilemmas. Facebook, for instance, operates with its own rules for participation including the right to delete posts that contain certain forms of nudity, hate speech or violence. While these kinds of norms to some degree coincide with the normative foundations outlined in the rules for participation in the comment sections of the news media, the differences can also sometimes lead to conflicts in terms of what is understood to be legitimate boundaries. Such disagreements became apparent when Facebook deleted the Pulitzer Prize winning photograph of the 'Napalm girl'. This move was highly criticised by newspaper editors who viewed it as an unacceptable form of censorship and it led to international discussions about the consequences of changing editorial power and responsibility.

# **Diversity and deviance**

It is an overall goal for editors working in the news media to facilitate and to stimulate a diverse and fact-based debate on immigration, mirroring the professional norms outlined in the *Code of Ethics*. Editors also generally defend a liberal stance on free speech, stressing the importance of arguments being confronted with counter arguments, and advocating the need for controversial and original opinions in the public conversation in the news media. This view is by some informants defended in terms of the 'pressure cooker' discourse, claiming that it is better to let 'deviant' voices participate in the wider public debate (particularly in the comment sections), rather than pushing them into smaller forums where their viewpoints are not challenged, and which consequently could work as echo chambers. A competing perspective, which editors also take into account, acknowledges the potentially negative effects chapter 5

of allowing certain viewpoints or forms of argumentation into the general public debate. While such considerations can of course be difficult to navigate between, they are a natural part of the professional role of an editor.

What is meant by 'deviance' in the immigration debate is not given, but rather a part of the ongoing struggle concerning what is evaluated as acceptable or unacceptable at a given time and in a given context by different people. Hate speech or other kinds of illegal forms of utterance are of course not allowed in any format in the news media (see chapter 2). Furthermore, editors do not generally accept crude generalisations of minority groups, particular not in the pre-edited formats, and they are wary of insulting or offensive language. At the same time, boundaries are often not clear-cut and editors sometimes face difficult choices as to whether or not they should publish a controversial text, and that might push the boundaries for what is understood to be legitimate to express in public. Of course, such viewpoints might be raised elsewhere, on blogs, online forums or alternative media sites that operate with their own boundaries for what is deviant or not. Editors in the news media have to balance between pushing and protecting the boundaries in the public debate to stay relevant - and can consequently be criticised for both strategies.

One example provided by an editor in a national newspaper illustrates this kind of dilemma. The blogger, 'Fjordman', became well-known to the wider public after the terrorist, Anders Behring Breivik, claimed he was inspired by Fjordman writing. Fjordman later wrote a book sponsored by the Free Speech Foundation to present his side of the story, and both the decision to finance the book as well as the book itself were highly criticized by commentators in the media. In the

152

aftermath of the book's release, the editor in question published a text by the blogger, but was unsure about whether or not it should have been:

Sometimes I think it is difficult to draw the boundaries. For instance, when Fjordman published a book, it was obvious that we had to write about it, and when he received a lot of public criticism he had to be given the chance to reply. But one of the texts he wrote and that I chose to publish at the time, I would perhaps normally not have published. I let him use his form of rhetoric to show what he actually stands for, but it was on the borders of being racist.

The question concerning whether or not to publish a text that might be considered deviant, as alluded to in the quote, then, is considered in relation to the particular context. The evaluation in this case was based on the normative ideal that the blogger should be able to respond and defend himself to public criticism as well as stimulating a public debate about his viewpoints. The editor explains that the risk of publishing such texts is that the boundaries for what is defined as legitimate might be stretched or even normalised at one side of the spectrum. In other debates about immigration, the editor explained, it is important that different perspectives are presented and that there is some kind of balance between opposite viewpoints over time. However, the concern is that in a polarized debate a kind of 'false balance' might develop if the more extreme views, in this context representing the radical right-wing side of the political spectrum, is allowed to represent one of the poles in the debate, while more moderate liberal viewpoints represent the opposite pole. Of course, what constitute 'radical' or 'extreme' is again a matter of negotiated boundaries.

Another point is that it is sometimes difficult to draw the line between acceptable forms of scepticism and criticism in the

### chapter 5

immigration debate and unacceptable forms of racism, because such boundaries are often blurry, as illustrated by this editor:

I sometimes find it hard to define when something can be considered racism. What is racism? I do not think it is easy to define. That is why I call it 'muddy'. Because there is something there, but you cannot put your finger on it and call it pure racism either.

This editor further points out that it might be the totality of the opinions raised in a particular debate, across platforms - and particularly non-editorially driven ones - that might in sum turn out to be racist or interpreted as racist. Of course, editorial gatekeeping takes place within a fragmented public sphere, and editors can only influence and control the debate that takes place within their own newspaper and their own comment sections. Furthermore, some editors also point out that terms like 'racism' or 'racist' are challenging because they have been used in the debate not only to pinpoint actual xenophobia or hate speech, but also to silence opponents. People defending strict immigration and integration policies have claimed that it has been difficult to raise such perspectives in the public debate, and that critical viewpoints against immigration often have been denounced by a left-wing, politically correct cultural elite for being uncivil or racist (for a more thorough discussion, see chapter 9). 'Cultural elites' in the debate, the argument goes, is motivated by the need to be perceived as 'good people' and consequently they base their arguments on humanitarian grounds rather than political and economic challenges and solutions (Brox, 1991; Brox, Skirbekk, & Lindbekk, 2003). This kind of reasoning was again part of the public discourse in 2015 when a large number of refugees and migrants fled to Europe. Sylvi Listhaug, the Norwegian Minister of Immigration and Integration, argued that those who promote restrictive and realistic immigration policies are labelled as 'awful people' and that those who defend a liberal position are viewed as 'morally good' (NRK, 03.11.2015). The minister also characterized the discourse on immigration as 'the tyranny of the good', referring to how liberal voices belonging to the left on the political scale have dominated the immigration debate in Norway.

Several editors acknowledge that it used to be difficult to present critical arguments against immigration, particularly in the elite debate, due to the potential stigma of being labelled a racist. However, many editors claim that there has been a significant shift in such dynamics after the refugee crisis in 2015, as expressed in this quote:

During my time in the newspaper, there has been a distinct change. I wouldn't say it was a taboo, but people were met with a lot of condemnation if they defended critical perspectives on immigration. Now it is almost the other way around, you risk getting get ridiculed if you present a liberal point of view. So, I think the hegemony in the debate has shifted.

Why such hegemonic shifts occur, is of course difficult to explain. Editors in the study point to how the political climate changed in the aftermath of the migrant and refugee crisis. As chaos occurred at the national border in the north of Norway as well as other places in Europe, the situation seemed out of control and the rhetoric changed across the political spectrum. The focus moved from the humanistic frame of helping refugees, to the shortcomings of the political system, as well as concerns about how high levels of immigration would impact society economically, socially and culturally. In particular politicians from the Progress Party that historically has defended strict immigration policies, including the Norwegian Minister of

### chapter 5

Immigration and Integration, used strong words to describe the situation and the challenges the migrant and refugee crisis represented to the Norwegian society.

Finally, many editors believe that the debate about immigration has become more diverse than it used to be. They point to the fact that new voices have joined the public conversation, and that this partly is a result of effective editorial strategies. Several editors in the study argue that they work hard to find moderate and original perspectives that can illuminate and enlighten the debate in different ways. Stimulating a fact-based debate is viewed as an important part of their professional responsibility in the fragmented media landscape, and particularly so because many editors believe the overall debate is highly polarized. Many editors argue that they work strategically to include certain groups who are less represented in the debate, like young people, women and minority voices, and that they believe this has contributed to the debite climate in a positive way, as this editor points out:

With a few exceptions, we have managed to establish a relatively reasonable, not stigmatizing kind of debate in a field that is incredibly emotional. It has been one of my strategies and probably for others too, to include voices from different minority groups … This is something I feel has contributed positively to the discussion climate and to the way we talk about these things in Norway. We have many immigrants that I feel have strengthened the quality and made the public debate more concrete. There is a decrease in words such as "politically correct" and other meta-descriptions. We have managed to get closer on peoples lives, and younger minority voices have particularly contributed in a positive way. I believe this is a good and a very smart editorial strategy.

Recent studies exploring this from the perspectives of people with a multicultural background, support this claim: many feel that editors in the news media are eager to get them to take part in the public debate (Midtbøen & Steen-Johnsen, 2016). In other words, access to the public debate is not necessarily an issue for minorities with a multicultural background. However, other challenges like 'ethnification' and ascribed identities might be prevalent (Bangstad, 2013; Eide, 2011), as well other kinds of repercussions (see also chapter 8).

## **Pushing and protecting boundaries**

In this chapter, we have explored editorial perspectives on the public debate, with particular focus on immigration discourses. We have focused on how editors talk about their own role as editors in a fragmented public sphere and how they interpret their responsibilities as gatekeepers. We have argued that gatekeeping is most often *not* about guarding the boundaries of freedom of speech in the legal sense of the word. Hate speech, defamation and threats do occur, most often in comment sections, but this is not regarded as a huge problem by a majority of the editors, indicating that such forms of problematic expressions to a large degree take place outside the editorial driven media. Rather, editorial gatekeeping includes a varied set of practices, guided by the ethical norms embedded in professional journalism. Such norms are grounded in ideals like pursuing diversity, balance, originality, factuality and quality. The adherence to such ethical norms is an important part of how editors legitimize their position and authority in a fragmented public. Secondly, editors defend a liberal take on free speech, however, boundaries are of course drawn. Editors explain that it can sometimes be challenging to determine if an opinion crosses the line for what is understood as legitimate, because boundaries are blurry, dynamic and dependent on contextual factors. If editors opt to publish a text that they feel is

### chapter 5

particularly controversial or deviant, they might follow up with a competing or supplementary standpoint the next day. Such practices illustrate how editors both push and protect the boundaries in the public debate at the same time. Furthermore, editors legitimise their position by highlighting their responsibility to stimulate diversity. Many editors, particularly from national publications, claim they actively seek out interesting and new voices in social media, and that they have been particularly concerned about getting people with a multicultural background to participate.

Finally, editors work in a rapidly changing media landscape and their responsibilities have in a relatively short time expanded from predominantly administrating the opinion pages of the newspapers, to managing a variety of services. Competition and collaboration with global intermediaries like Facebook have become even more imperative in the last couple of years. An important aspect of editors' work in the near future is how successful they will be in utilizing the social media logic where new forms of inclusion and exclusion mechanisms are at play. Novel forms of curation and navigation practices are necessary to detect interesting debates and to invite new voices to write for the newspaper. Such skills and knowledge are essential to protect their role as gatekeepers and agenda-setters in a fragmented public sphere.

### **References**

Abbott, A. (1995). Things of Boundaries. *Social Research, 63*(1), 857–882. Alexander, J. C. (2006). *The Civil Sphere*. Oxford: Oxford University Press. Alexander, J. C. (2016). Journalism, democratic culture, and creative

reconstruction. In J. C. Alexander, E. B. Breese, & M. Luengo (Eds.), *The Crisis of Journalism Reconsidered. Democratic Culture, Professional Codes, Digital Future* (pp. 1–31). Cambridge: Cambridge University Press.


Balch, A., & Balabanova, E. (2011). A system in chaos? Knowledge and sense-making on immigration policy in public debates. *Media, Culture and Society, 33*(6), 885–904. doi:10.1177/0163443711411007

Bangstad, S. (2013). Inclusion and exclusion in the mediated public sphere: the case of Norway and its Muslims. *Social Anthropology, 21*(3), 356–370. doi:10.1111/1469-8676.12034


*European Journal of Communication, 16*(2), 209–232. doi:10.1177/ 0267323101016002004


*Journalism & Mass Communication Quarterly, 78*(4), 641–658. doi:10.1177/107769900107800402

Ruiz, C., Domingo, D., Mico, J. L., Diaz-Noci, J., Meso, K., & Masip, P. (2011). Public Sphere 2.0? The Democratic Qualities of Citizen Debates in Online Newspapers. *International Journal of Press-Politics, 16*(4), 463–487. doi:10.1177/1940161211415849

Shoemaker, P., & Vos, T. P. (2009). *Gatekeeping Theory*. New York: Routledge.

Shoemaker, P. J. (1991). *Gatekeeping* (Vol. 3). Newbury Park, Calif: Sage.


Wahl-Jørgensen, K. (2002). The Construction of the public in letters to the editor. Deliberative democracy and the idiom of insanity. *Journalism Studies, 3*(1), 69–81.

Wessel-Aas, J. (2013). *Jus og sosiale medier*. Oslo: Kommuneforlaget.

White, D. M. (1950). The 'gate keeper': A case study in the selection of news. *Journalism Quarterly, 27*, 383–390.

### Chapter 6

# Debating freedom of expression in Norwegian media: Critical moments, positions and arguments

*Terje Colbjørnsen, PhD, Researcher, Department of media and communication, University of Oslo*

While the legal and fundamental protection of freedom of expression enjoys strong support in liberal democratic societies, there exists simultaneously a lively debate on whether and how to restrict utterances that are deemed hateful, hurtful or simply not beneficial to the common good. Departing from quantitative and qualitative data sets of Norwegian media debates, this chapter provides a longitudinal analysis of media coverage of the issue of freedom of expression, and of the positions and types of arguments that can be outlined in selected cases of the freedom of expression debate. The findings suggest that media coverage takes the form of short intense bursts of attention followed by a return to relative normalcy, and that the cartoon controversies in 2006 and 2015 stand out as 'critical moments' in the freedom of expression discourse. Further, the study finds two opposing positions in the debates, the absolutists and the consequentialists. The cartoon controversies are seen as examples of muddling the concepts of freedom of expression and freedom of the press. Findings suggest that historical-philosophical arguments are more legitimate in the media discourse than emotional arguments, challenging the idea of a public sphere dominated by emotions and sentiments of 'offendedness'.

## **Introduction**

In Norway, as in other liberal democracies, freedom of expression is a fundamental right, anchored in § 100 of the Norwegian Constitution and in Article 10 of the European Convention on Human Rights. That does not put the issue to rest. Freedom of expression is not only contested in the courts, but also in the public sphere, where legal, normative, moral, ethical and aesthetic boundaries are negotiated in debates over how to understand the right to free expression.

In this chapter I look in detail at the freedom of expression debate in Norway, examining firstly general trends in the media coverage, and secondly the positions and arguments characterizing these debates. The analysis is based on quantitative and qualitative data sets of newspaper debates on 'freedom of expression'1 in the period 1993-2015. The main emphasis is on 2005- 2015, covering the period from the Mohammad cartoons in

<sup>1</sup> The central search terms in Norwegian were «ytringsfrihet», which roughly covers both freedom of expression and freedom of speech, and «pressefrihet» (freedom of the press). Secondly, the search also covered neologisms such as «ytringsansvar» (responsibility when expressing opinions), «ytringsrett» (the right to express opinions) and «ytringsplikt» (the obligation to express opinions).

2005 to the Paris terror attacks in late 2015 and allowing for a closer examination of the freedom of expression debates raised in relation to 'cartoon disputes' in 2006, 2010 and 2015.

An apparent feature of the freedom of expression debates is that they tend to increase on particular occasions, in connection with specific non-planned events or concrete utterances. Religion, migration and the rights of minorities in particular have been catalysts for many of the recent debates. Liberal democracies struggle to balance individual freedoms against protection from racism, incitement to violence and other speech acts deemed to be harmful (Bleich, 2011; Peters, 2005). In recent years, the Mohammad cartoons controversy in 2005/06, and the brutal attacks on the Charlie Hebdo offices in Paris and the Krudttønden cultural centre in Copenhagen, both in early 2015, in particular have sparked debate, analysis and commentary concerning the issue of freedom of expression. In relation to events such as these, the boundaries of freedom of speech are discussed, negotiated and contested. As Jytte Klausen (2009) has noted in her study of the Mohammad cartoon crisis: '[E]veryone regarded the cartoons as an opportunity to draw a line in the sand, albeit for different reasons' (2009 p. 3). Wessler, Rinke & Löb (2016) in their study of the Charlie Hebdo case find an opportunity in such a crisis 'to symbolically draw inclusive boundaries in defense of central values' (2016 p. 323). In other words, boundary-drawing (Abbott, 1995; Lamont & Molnár, 2002) comes forth as a characteristic feature of the disputes, consistent with the overall argument of this anthology.

While specific events can provoke debate, there is also a recursive dynamic to free speech debates. Freedom of expression debates bring out the interplay between the specific and the principled: While the events that trigger debates are singular and based in different circumstances and contexts, 'freedom of expression' remains a common, recurring and dominant frame. As such, 'Many of the most important cases [of freedom of speech] are not about substantive issues but about the principle of free speech itself ' (Durham Peters, 2008 p. 275).

Freedom of expression debates in the public sphere, then, not only contain issue-specific arguments, but take the form of *meta-discourses*, contestations over principles of argumentation and the limits of free speech. A starting point for this study is to examine how people argue for freedom of expression and its boundaries in the media, whether it be by referencing law and philosophy, by laying claim to a sense of victimhood or by other standards of justification. In other words, what part did emotions and emotional arguments play in the public exchange over the cartoon controversies? All in all, the chapter seeks to contribute empirical data and background to an understanding of freedom of expression debates. Two over-arching research questions have guided the study:


I will first present central concepts and theories, before moving on to describe the methodological approach and then discuss the findings of the two studies that are integrated here.

# **Critical moments and framing**

This chapter traces general trends in freedom of expression debates and the arguments and positions that can be outlined in specific debates in the past two decades. To explain the tendency for some news stories to be conceptualized as cases of freedom of expression, I draw on theories of 'critical moments' and 'framing'. The concept of 'critical moment' helps to identify the landmark events that are crucial to the freedom of expression discourse. 'Framing' as a concept is used to explain the importance of labelling certain events as cases of freedom of expression.

## Critical moments

The study draws on Luc Boltanski's (2011) concept of critique and more specifically the notion of 'critical moments' (Boltanski & Thévenot, 1999). While freedom of expression debates are recurring phenomena over the entire period examined here, the public debate is in general connected to specific moments when media attention increases sharply. These can be conceptualized as 'critical moments', understood as unusual and particular moments that play an important part in social life, as Boltanski & Thévenot (1999) argue.

The notion of a 'critical moment' is part of the sociology of critique (or pragmatic sociology of critique), outlined in Boltanski's *On Critique. A Sociology of Emancipation* (2011). A central aspect of this theory of social action is to acknowledge the capacity of actors to reflect on their actions and environments (Boltanski, 2011). This can be observed in examples of self-reflexivity and meta-discursive practices (i.e. debating how we debate freedom of expression). To understand how social disputes and contestations take place, and how agreements can be made, Boltanski and his colleagues within the pragmatic school of critique have developed theories and concepts of social 'justifications'. For the arguments of one actor or party in a dispute to be accepted as legitimate, these need to be consistent with a certain *régime d'action.* Within such regimes exist 'orders chapter 6

of worth' (*économies de la grandeur*); in short, common standards of how to resolve issues. According to Boltanski & Thévenot, disagreements are harder to settle when the situation at hand is unclear and fuzzy because of mismatched orders of worth ('*situations troubles'*) (1999 p. 374). An example can be regimes based on justification confronted by regimes of violence or emotions. Geographic and cultural distance makes it harder still to resolve disagreements (Boltanski, 1999).

Two (related) aspects of the concept of critical moments are particularly pertinent to the issue here. First, the authors draw on the double meaning of 'critical' as both something of decisive importance (*a critical matter*) and as a quality of judgement, interpretation, or analysis (*criticism*). The critical moment is thus of the utmost importance while simultaneously open to debate. Second, while specific events such as the publication of the cartoons seem to generate public debate, the 'moment' in question may just as well be the reaction to an event or an utterance. More so than the notion of a 'critical event' (Das, 1995; Andersson, Jacobsen, Rogstad, & Vestel, 2012), a 'critical moment' leaves room for the critical potential of reactions and meta-arguments.

### Framing

Besides contributing to a debate, historical and contemporary, freedom of expression can be seen to constitute a 'frame', a way of highlighting a certain interpretation of an event or utterance. Framing, according to Entman's widely cited definition, means 'to select some aspects of a perceived reality and make them more salient in a communicating text' (Entman, 1993 p. 52). Further, analogous to how the concept of 'terrorism' (Entman, 2003a) is used to describe some violent acts but not others, the phrase 'freedom of expression' has implications. Frames, as emphasized by Entman (1993) have consequences in terms of what we understand as the problem, how we can interpret and evaluate the situation, and what treatments are required to solve it. Thus, when social actors make the claim that something is a matter of freedom of expression they are also saying that it is 1) a matter of principle; 2) of importance; 3) that the issue goes beyond the single case or the single story; and 4) that it forms a connection to a longer historical discussion of free speech. The framing literature contains multiple possibilities for how to understand and apply the concept (Matthes, 2009). Vreese (2005) distinguishes between generic and issue-specific frames. In our case freedom of expression can be understood as a generic frame, while the cartoon controversy exemplifies the issuespecific frame.2

# **Data material and method**

The quantitative part of the study was conducted mainly in October 2015 – January 2016. A search for the period January 1st 1993 – December 31st 2015 was conducted on 'ytrings\* OR pressefri\*' (approx. 'freedom of expression' OR 'freedom of the press') in the media archive Atekst.3 The search was narrowed down to newspaper items only and finally to items from the five selected news outlets that had been indexed in Atekst throughout the entire sampling period, resulting in a data set of 22,428 news items.4

<sup>2</sup> Recent applications of framing theory on the cases discussed here, include Jørndrup (2016) on the Krudttønden attack and Walter et al. (2016) on the Charlie Hebdo attack.

<sup>3</sup> Atekst, also known as Retriever, is a Norwegian media database owned by Retriever Norge AS. The archive provides searching and access to newspapers, online news, websites and magazines dating back to 1945.

<sup>4</sup> For details on search criteria and methods, see Colbjørnsen (2016).

The qualitative part of the study was conducted in January 2016 – June 2016. News items were sampled from 2005-2015, allowing for a wider sample of source material.5 Taking the findings of the quantitative approach as a starting point, media coverage was found to spike at several different times, with weeks in 2006, 2010 and 2015 standing out. To avoid paying undue attention to specific disruptive events and to garner more of the argumentative afterthought, three sample weeks occuring three weeks following each peak were selected. This resulted in a total of 194 articles from week 8, 2006, week 7, 2010, and week 9, 2015. Argumentative genres (editorials, opeds, and letters to the editor) accounted for 140 of these items. The news items were coded in Nvivo, following a multi-step coding process, partly inspired by the grounded theory approach (Glaser & Strauss, 1967; Thornberg & Charmaz, 2014). A first step involved searching the text and tracing frequent keywords, then moving step-wise towards a more systematic categorization. Each item was assigned one of the six main argumentative categories, and quotes within each item were marked as characteristic of a certain argument or justification standard.

To determine what kind of argument was dominant in each article, I looked for cues as to *how* the author chose to make freedom of expression relevant to the matter at hand. In this process I would look for answers to questions such as whether legal principles were invoked or if statistics and research played a part; if reference was made to historical parallels or to religious or philosophical traditions; or whether the author explicitly appealed to emotions.

<sup>5</sup> Sources include: Two national dailies (Dagbladet; VG), two regional dailies (Aftenposten; Stavanger Aftenblad), two local news outlets (Brønnøysunds avis; iTromsø) and two national niche outlets (Morgenbladet; Dagens Næringsliv).

# **The peaks and valleys of freedom of expression coverage**

This section provides an overview of general trends in the coverage of freedom of expression in Norwegian newspapers, looking at the time period 1993-2015. 6 The starting point of 1993 is motivated by a wish to include the attempted assassination of Salman Rushdie's Norwegian publisher William Nygaard in October 1993, a wellknown freedom of expression case in Norway (cf. Bangstad, 2014). Rushdie's novel *The Satanic Verses* (published 1988) sparked demonstrations and the infamous *fatwa* against the author and other people involved with publishing the book. While the Nygaard crime remains unsolved and the shooter unknown, the case has been highlighted as an attack on freedom of expression.

Based on a hypothesis that coverage coincides with certain events or moments, search parameters were set to weekly intervals.7 Figure 6.1 shows coverage patterns in two parts, for the first (93-04) and the second (05-15) parts of the period respectively.

A pattern emerges in Figure 6.1 of clearly defined peaks and valleys, indicating how coverage spikes in quite short bursts of attention, before returning, within a week or two, to a level of normalcy, as indicated by the average level. There is, however, a marked difference between the earlier and the later parts of the period.

The period 1993-2004 contains several smaller spikes, but the overall tendency is towards minor deviations from the average

<sup>6</sup> The sources were a small sample, only five media outlets (Aftenposten, Bergens Tidende, Nordlys, NTBtekst and VG), due to the lack of sources indexed in Atekst going back to 1993. While the sample is only partially representative of the Norwegian news media system, analyses of wider samples confirmed the overall coverage patterns as represented here. Colbjørnsen (2016) provides further explanations of the methodology and its limitations.

<sup>7</sup> This is as close in time as the Atekst software allows.

**Figure 6.1.** Coverage of freedom of expression 1993-2004 and 2005-2015, weekly intervals8.

level (standard deviation 7.3). Building on the notion from Boltanski & Thévenot, we can understand an especially prominent peak as a 'critical moment'. As there exists no agreedupon quantitative measure for a critical moment, we will have to be content here with a rough approach, identifying spikes that are evidently most pronounced. A look at the media articles behind the minor peaks for the 93-04 period, indicates how they correspond to certain events and news items:<sup>8</sup>

— The minor peak in week 41 of 1993 corresponds to the attempted assassination of publisher William Nygaard, October 11, 1993.

<sup>8</sup> Please note that while the actual number of news items is comparatively low (< 250), the crucial aspect here is the overall trend and to what extent the number of articles per week deviates from the average level.


As Figure 6.1 reveals, the pattern of peaks and valleys is a lot more pronounced for the 05-15 interval. The analysis indicates an increase in media coverage over the years, but even more conspicuous are the deviations from the average, far more pronounced than for 93-04. Relevant indicators such as standard deviation, relative standard deviation and mean deviation lend support to this interpretation of the graphic representation: For 1993-2004 the standard deviation is 7.3 vs. 19.9 for 2005-2015 (relative standard deviation is 52.5 % vs. 83.6 %; mean deviation is 10.2 vs. 5.3).

More precisely, what we may refer to as the 'critical moments' are clustered in three main periods of events and reactions. Unsurprisingly, the caricature conflicts in 2006 and in 2015 stand out:


Apart from the clearly defined peaks described above, media attention is fairly evenly distributed. In between the peaks and the 'critical moments' are periods of comparatively little attention to freedom of expression. As Boland (2007) has stressed, 'the 'critical moment' is transient, and will pass, by the re-establishment of order' (2007 p. 125).

News sociology and framing theory suggest that news and critical events are constructed (cf. Molotch & Lester, 1974; Entman, 2003b, 2003a). That implies a selection process in which some events or occurrences that could possibly have been framed as instances of freedom of expression were in fact not. To compare the critical freedom of expression moments with other significant events in the period, I selected two prominent and widely debated issues that have both been linked with the freedom of expression discourse. The first is the terror attack by Anders Behring Breivik on government buildings in Oslo and the Labour Party summer camp at Utøya July 22, 2011 (see also Midtbøen, Ch. 7). The second case is the terror attack in the Stade de France, the Bataclan concert venue and various shops and cafés in Paris on November 13, 2015. July 22nd and the Paris November 2015 attacks are marked out on the timeline in Figure 6.2 below. In addition, I have included the initial publication of the Mohammad cartoons in the Danish newspaper Jyllands-Posten on September 30, 2005.

Figure 6.2 gives graphic clues for comparing the cartoon controversies' ability to activate the freedom of expression frame with other events. Coverage of freedom of expression spikes in relation to the caricature conflicts, while the frame appears to be less relevant for the interpretation of the July 22nd and November 2015 terror attacks. Even though a full explanation for this requires a

**Figure 6.2.** Timeline and coverage of freedom of expression events 2005-2015.

separate study, we may point to theories of competing frames (Chong & Druckman, 2007; Nijkrake, Gosselt, & Gutteling, 2015) and how one frame can trump another and become dominant. Figure 6.2 suggests that freedom of expression was not a dominant frame for either the July 22nd or the Paris attacks. In fact, the week of July 22, 2011 stands forth as a defined low point in this material. Similarly, there is reason to believe that competing interpretative frames were more readily available than freedom of expression for the November 2015 Paris attacks.

Additionally, Figure 6.2 shows that the spike in media attention in the Mohammad cartoons case was lagging months after the first publication. The Mohammad cartoon crisis was indeed a 'long and messy event' (Hervik, Eide, & Kunelius, 2008), involving multiple actors across the globe with conflicting and often hidden agendas, meetings and protests, misunderstandings and attempts at reconciliation (cf. Klausen, 2009; Sniderman, Petersen, Slothuus, & Stubager, 2014). The time-lag lends support to the contention that the 'critical moment' can form in relation to *reactions* to an event rather than to the event itself.

### **Positions and arguments in the freedom of expression debates**

'Critical moments' do not appear out of nothing, but are part of a longer build-up: 'Critical moments' arise when the critical capacity of social actors is triggered by a feeling that *enough is enough*. As a dispute emerges, the first characteristic is that 'persons involved are subjected to an imperative of justification' (Boltanski & Thévenot, 1999 p. 360). Aspects of these disputes and justifications are the subject of this section.

The analysis is based on Norwegian newspaper items in the period 2005-2015, using a selection of 140 articles within argumentative newspaper genres – editorials, op-eds and letters to the editor. The section first looks at overall themes and topics that arose in the three sampled weeks. Further, significant positions in the debates are identified and analyzed, specifically looking at polarization and the muddling of concepts. Finally, I identify arguments and standards of justification, looking in some detail at the role of emotional arguments in the freedom of expression debates.

## Themes and topics

A quick look at the dominant topics in the sampled periods indicates how debates over cartoons and caricatures were prevalent:

In week 8 of 2006, the Mohammad cartoons and the following uproar was the dominant theme. The debate was more event-driven than that of the following periods. In addition to the cartoons, the sentencing of David Irving, British Professor of History, to prison for Holocaust denial in Austria, garnered attention.

In week 7 of 2010, the dominant theme was a drawing of the prophet Mohammad as a pig republished as a facsimile in the newspaper Dagbladet. It provoked demonstrations and debate in Norway. A controversial public appeal made by a young Norwegian Islamist, Mohyeldeen Mohammad, spurred the debate. In addition, the issue of the freedom of speech of a sacked university professor was widely debated in this week.

In week 9 of 2015, the most recent sample period, the Charlie Hebdo and Krudttønden attacks are the most prominent issues, both explicitly framed in relation to freedom of expression. While the attacks took place outside of Norway, they were both linked to the national political, religious, cultural and social contexts.

Although this summary of the predominant topics provides an overview, it also begs for a more detailed analysis of how the debates took place and what arguments were raised.

## Polarization in the freedom of expression debates

Looking more specifically at how boundaries are drawn – explicitly and implicitly – in relation to freedom of expression, the study finds boundary-drawing principally along the dimensions right/obligation to publish vs. right/obligation to refrain from publishing.

Two main opposites emerge: On the one hand are those who advocate what we may term an *absolute interpretation of freedom of expression.*<sup>9</sup> On this side, there are no legitimate reasons to limit freedom of expression or even to explain the motivations behind a statement. As one individual stated in a letter to the editor of the daily tabloid VG: 'They [Dagbladet] use their freedom of expression, and that does not require any other motive!' (Steingrim Wolland, VG 20.02.2010).

Following the demonstrations and violent protests against the cartoons in Jyllands-Posten in 2006, Norway's political parties debated how to approach the issue. Øyvind Vaksdal, a member of parliament for the Progress Party (Fremskrittspartiet) wrote an editorial in the regional daily Stavanger Aftenblad, arguing for an absolute interpretation of freedom of expression:

For the Progress Party freedom of expression and freedom of the press are absolute. We are happy to discuss politics with other groups, but freedom of speech and press freedom, we will not touch (Stavanger Aftenblad. 24.02.2006). 10

<sup>9</sup> Advocates of this line of argument are occasionally referred to as *free speech fundamentalists*.

<sup>10</sup> This and all subsequent quotes were translated from Norwegian by the author.

In this argument we find the kind of reasoning that tries to put freedom of expression to rest: *We will not touch this!* For Vaksdal, for the Progress Party, and for several other participants in the debate, the Labour Party (Arbeiderpartiet) and its leadership of Prime Minister Jens Stoltenberg and Foreign Minister Jonas Gahr Støre represented a more cautious, dialogue-centric line, for which they were severely criticized.

Gahr Støre in particular has often been placed at the other extreme from the *absolutists*. In this bloc we find those who do not see any particular reason to treat freedom of expression as a special case, as a 'holy cow'. Statements and utterances must be judged by what they contribute to the common good. This position is in line with *consequential* arguments that argue in favor of freedom of expression for its ability to promote truth and democracy (cf. Alexander, 2005 p. 127). In practice, the *consequentialist* arguments most often take the form of a defense of dialogue with protesters in the cartoons controversy or the young Islamists in the Mohyeldeen Mohammad case. The leader of the Socialist Left Party (Sosialistisk Venstreparti), Audun Lysbakken, promoted dialogue in 2010:

As I see it, the line of conflict today is not between Muslims and non-Muslims, but between moderates who wish to build a community and those who mock and ironize dialogue (Dagbladet. 17.02.2010).

The argument that opinions need to be out in the open for us to counter them is expressed frequently in the debates. This is what we may term the marketplace of ideas argument (often attributed to John Stuart Mill) for freedom of expression: That the open exchange of ideas and opinions will lead us to select the most beneficial ones (Gordon, 1997). Journalist Sofie Mathiassen of the business daily Dagens Næringsliv presented an example of the marketplace argument in relation to the 2010 controversy involving young Islamist Mohammad:

After Mohammad came out with his extreme statements, Norwegian Muslims have lined up to distance themselves from them. They have shown the plurality that exists within the Muslim community. Extreme Islamic beliefs are being challenged by Muslims themselves. That is a healthy debate, and at the very core of freedom of expression. We cannot argue against statements which are not heard. It is only when people have the freedom to voice extreme opinions that we can use our liberal democratic freedoms to fight them (Dagens Næringsliv. 20.02.2010).

The polarization between the absolutists and the consequentialists should not lead us to disregard other positions in the debates. There are also many examples of 'on the one hand, on the other hand' arguments, criticizing or sympathizing with both sides and weighing arguments carefully. While the polar opposites are somewhat locked in their positions, not seldom ridiculing the opposite camp, the larger sphere of debate is dynamic, open to interpretations, even at times coming close to a Habermasian idealist notion of persons engaging in a respectful, disinterested and equivalent trial of arguments.

As the editorial in the newspaper Aftenposten expressed it at the height of the Mohammad cartoons controversy in 2006:

More importantly, both governments [Norway and Denmark] need to show the necessary combination of a firm attitude against violence and in favour of freedom of expression on the one hand, and respect for the religious feelings of Muslims on the other (Aftenposten 21.02.2006).

In general, what we have in these debates is not a legalist argument over how to apply the legal principles of freedom of expression, but a quite dynamic field of negotiations – both principled and issue-specific – over what limits may be imposed, who has the power to draw boundaries, and how we can recognize the boundaries.

## Muddling freedom of expression and freedom of the press

The cartoon controversies in 2006, 2010 and 2015 present cases of how distinctions are made between freedom of expression as a human right and freedom of the press as an institutionalized practice. These two notions – related, but different – are often muddled in the debates. The liberal notion of freedom of expression at the individual level grants people the right to say what they want as long as they do not in effect curtail other people's free speech. They do not, however, have a fundamental right to have what they say published in a newspaper. Free speech does not equal free publication, although the Internet and social media have made certain forms of publication widely accessible. The most basic articulation of this muddling of freedom of expression and freedom of the press can be found in the form of a letter to the editor complaining that a previous letter was not published, and thus that freedom of expression had been curtailed. In other instances, the distinctions can be harder to make, and news outlets are easily accused of stifling critical voices when acting as gatekeepers of opinions.

Nonetheless, within a liberal tradition of freedom of the press, it is a newspaper's right to decide whether to publish or not. In the debates analyzed here, the autonomy of the press is recognized in diverging ways by the two main opposites: The absolutist actors in the debate typically highlight a newspaper's *right to publish (offensive materials).* On the other hand, the consequentialist actors tend to emphasize the *right to refrain from publishing (offensive materials)*. It is difficult to see how a free press should not be left to decide for itself in each case. This is pointed out by the Labour Party's then parliamentary leader, Helga Pedersen, in a letter to the newspaper Dagbladet in 2010, following the controversy of the Mohammad-as-a-pig cartoon in the paper:

Some confusion has arisen over recent criticism of newspapers for using their press freedom. This confusion is unnecessary. It is Dagbladet's sovereign right to publish controversial texts, pictures and drawings – even if they represent a poor judgement of taste – a right the paper has exercised on a number of controversial occasions in recent times. At the same time, it is every citizen's sovereign right to criticize Dagbladet's editorial decisions. (February 20, 2010).

While Pedersen's statement can be seen as an example of the capacity of social actors to reflect clearly on the situations they find themselves in, the cartoon controversies include multiple arguments where claims of supporting press freedom seem to coincide with a narrowing down of editorial autonomy.

One aspect of the famous 'Je suis Charlie'-campaign can be said to collapse the distinction between freedom of expression and freedom of the press: In this campaign of solidarity with the Charlie Hebdo cartoonists, ordinary people and news outlets were urged to show compassion with Charlie *by becoming Charlie*. That is to say, people posted 'Je suis Charlie' slogans on social media, while news outlets republished the controversial Charlie Hebdo cartoons not merely to document the case, but in an act of solidarity. As Wessler et al. (2016) have argued, the Charlie Hebdo aftermath presented a case of right to offend vs. deliberative self-restraint. From a deliberative point of view, the close identification with Charlie Hebdo expressed in 'Je suis Charlie' was problematic:

[R]ituals of public solidarity are important but insufficient and the general public should support but not wholly identify with *Charlie Hebdo*. We should thus be with *Charlie*, but not unconditionally be *Charlie* (Wessler et al., 2016 p. 323)*.*

The consequence of being Charlie was for newspapers to publish the Charlie Hebdo cartoons, seemingly setting aside their editorial autonomy. *Right to publish* thus became a *duty to publish*. This act of solidarity can be seen as a way for the journalistic profession under siege to close ranks. The balanced weighing of pros and cons becomes secondary in a 'critical moment', which is seen as a threat to the mission of journalism.

### An age of 'offendedness'?

While the sections above identified and highlighted certain positions in the debates, the following part of the analysis elaborates on the level of the argument and how, precisely, freedom of expression is made relevant to the case in question.

In the coding process, I identified six categories of overall arguments. The first category reasons along historical, philosophical, and religious lines (what I termed *FoE as a historical-philosophical idea)*. In contrast, the second category relates the argument only to a single case and tends to discuss it in isolation (*FoE in relation to a delimited case)*. Another set of items are characterized by legal principles being the central argument (*FoE in relation to law and legal principles)*. In some few instances the frame is activated by means of research, facts or polls (*FoE in relation to empirical facts or research)* or made in the form of satire or humor (*FoE as enabled through a satirical or*  *humoristic response)*. Lastly, I coded items that use emotions as a standard of justification (*FoE in relation to an emotional-affective response).* In the following I shall look more closely at this latter type of argument, addressing claims of an age of 'offendedness'.11

There have been claims and reports that the current debate culture in many societies is centered on emotional responses – e.g. expressed sentiments that emotional boundaries are violated and an inclination to see oneself as a victim. In a Norwegian context, commentators have spoken of an 'age of emotions' (Hobbelstad, 2015) and of a 'tyranny of offendedness' (Toje, 2011).12 In a US context, emerging concepts such as 'trigger warnings', 'microaggressions' and 'safe spaces' have caused concern that public debate is stifled, particularly on college campuses (Haidt & Lukianoff, 2015). In light of this, Campbell & Manning (2014, 2016) see the emergence of a culture of victimhood that is distinct from previous forms of cultures of honor and cultures of dignity.

Clearly, what I term here 'the cartoon controversies' were more than merely heated debates; they were violent conflicts with tragic outcomes, involving deep-seated and very real tensions. The premise of the analysis was to examine what part emotional reactions, quite justified under the circumstances, played in the public debate. Were emotional-affective responses present at all? Were they recognized by other debaters or countered in some way?

The category *FoE in relation to an emotional-affective response* contains articles where I found the main argument to be based on an emotional response of some sort, such as expressions of

<sup>11</sup> The notion of 'offendedness' is an approximate translation of the Norwegian 'krenkethet'.

<sup>12</sup> For a broad Norwegian-context discussion, see also: https://morgenbladet. no/2015/07/foleri-folera

rage, frustration, disappointment, shame or 'offendedness'. While there is reason to point out that the distinctions between rational and emotional styles of reasoning are fuzzy, it was possible to single out news items in which emotions and affective responses were made explicit and were the foremost standard of justification (rather than the other five categories). Nonetheless, the analysis finds little support for the view that emotions and affective responses are dominant in the public sphere. Rather, the mediated contestations analyzed here are based on historical, philosophical and religious lines of reasoning or explicitly connected to a single case only.

In general, 'offendedness' comes across as an inefficient justification standard in the mediated public sphere. However, some examples could be found, particularly from Muslims identifying as the offended party in relation to the Mohammad cartoons in 2005/06 and the protests in Oslo five years later. Well-known radical Islamist Arfan Bhatti penned an op-ed to Dagbladet in 2010:

It is only natural for us Muslims to react against what we perceive as offensive. Reactions to and frustration over caricatures, mockery of Islam and bullying of Muslims have built up for years (VG 16.02.2010).

A young student expressed a similar notion, perceiving a bias in the media:

If you are offended, then you just have to live with it. This is how I feel it has become. Muslims have long been unfairly exposed in the name of freedom of expression. If they criticize the publication [of cartoons], it is regarded as an attack on free speech. The journalists are embraced and made to look like victims, but are never held accountable for their work. Is it carte blanche for them to write and print exactly what they want? What about ethics and social responsibility? (Asjad Mahmood, VG 19.02.2010).

Despite these instances, the comparative dearth and ineffectiveness of emotional-affective arguments emerge as a clear finding in the study presented here. However, other that does not mean that emotional arguments are not valid in 'orders of worth', in Boltanski's sense. The public demonstrations against the Mohammad cartoons do seem to display a kind of emotional outrage, as do many social media comments on the subject. This may indicate the newspaper debates being out of sync with arguments expressed in demonstrations and in social media, and is certainly worthy of more research.

Rather than finding a dominance of emotional arguments, this study finds an abundance of arguments that question the validity of the emotional-affective response. Somewhat paradoxically, this shows the centrality of the notion of 'offendedness' in the mediated public sphere, but mostly as a counterpoint to argue against. The perception of a widespread over-sensitivity in the public sphere arises, as expressed in the words of Einar Gelius, pastor and author, writing in 2015:

One weighs pros and cons, so that nobody should feel either violated or ostracized. The messages and opinions conveyed are eventually nothing but political nonsense, without substance and content (Dagbladet 26.02.2015).

The non-validity or non-applicability of the emotionally grounded argument is also expressed by politician Michael Tetzschner in 2010:

Regarding the delimitation of freedom of speech, we all as individuals would like for the debate to be characterized by broadmindedness, respect, knowledge and respect for others. The best debates are just so. And those who want to win others over will find it easier to gain support if one is factual and not inflammatory. But as a legal boundary for utterances, the requirement that no one should feel "hurt" is completely unsuitable. Yes, worse than that: it presupposes censorship either by the state or the individual, which undermines the foundations of freedom of expression (VG 17.02.2010).

Tetzschner's argument quite explicitly makes the case that different standards of justification are valid in different contexts, arguing that public debate is best when characterized by broadmindedness and respect, but that the legal sphere needs to operate with different standards, where feelings of being hurt or offended have little or no place. While this view is quite prevalent in the newspaper debates, it is reasonable to suggest that other actors than those represented here (many of them journalists or politicians) would argue otherwise, in particular under different *régimes d'action* (cf. also Moe, Thorbjørnsrud and Fladmoe, Ch. 4).

## **Concluding remarks**

In this chapter I have presented and discussed findings from a study of freedom of expression debates in Norway, drawing on both quantitative and qualitative data. Analysis of media coverage in the years 1993-2015 indicates that debates over freedom of expression take the form of short intense bursts of attention followed by a return to relative normalcy. These clearly defined spikes in attention, as seen in the graphs reproduced here, are what we may refer to as 'critical moments'. The themes and issues that were most prevalent in the freedom of expression media coverage were the renditions of cartoon controversies, from the coverage of reactions to the Mohammad cartoons in 2006 to the debate following the attacks on Charlie Hebdo and Krudttønden in 2015. More than merely constituting news events, these 'critical moments' were triggers for wider public discussions and protests. They were calls for reflection, discussion and negotiation of the boundaries of free speech.

While some see the boundaries of freedom of expression as clear-cut, others argue over fuzzy and negotiable borders. In terms of overall positions that can be outlined in the debates, the analysis finds a wide-ranging public debate in Norway, with moderate voices flanked by opposites that either see no cause for restricting freedom of expression (the *absolutists*) or rather find cause to carefully weigh options and motivations for publishing something potentially hurtful (the *consequentialists*).

Moreover, the analysis presented here carries a message concerning the state of mediated debates. Contrary to declarations of an age of 'offendedness', where emotions take center stage in the public sphere, the freedom of expression debates analyzed here are not dominated by emotions and claims of victimhood. Rather, they are characterized by historical, philosophical and legal lines of argument. Nonetheless, the notion of 'offendedness' plays a part in constituting a position (real or perceived) from which to argue. As a standard of justification, 'offendedness' and claims of victimhood or of being hurt fall short in these debates. However, legal principles, striking historical parallels or negations of 'offendedness' cannot put the debates to rest. All in all, the freedom of expression debates of 2006, 2010 and 2015 are unsettled, dynamic and lively in ways that make them similar and recursive. As such, it seems that we are always reinventing the freedom of expression debate.

Finally, we may consider briefly the role of the cartoons that have taken center stage in the freedom of expression debates here. Why have political cartoons become so enmeshed in the debates over freedom of expression during the past decade and stood out so clearly in terms of the size of media coverage? Previous research has shown that the cartoons spurred debate because they lent themselves readily to the aims of strategic interest groups and could fit in with existing and emergent media logic (Eide, Kunelius, & Phillips, 2008; Hussain, 2007; Klausen, 2009; Sniderman et al., 2014). In the cartoon controversies, different justifications rub against each other, spurring reactions, both violent and non-violent. However, I would suggest – perhaps as a starting point for further research into the role of cartoons as symbols of free speech – that there are other dimensions particular to the drawings that make for heated debates. The cartoons are typically ripped from their original contexts (language, culture, original publication etc.), and this loss of context makes the situation ripe for misunderstandings and interpretations (cf. Hussain, 2007). Decontextualization can also be seen as part of the logic noted by Durham Peters (2008) in which particular subject matter tends to be subsumed under the freedom of expression heading. As images, the cartoons are deceptively simple to read (cf. Müller, Özcan, & Seizov, 2009). *If what I see is the face of the holy prophet Mohammad with a bomb in his turban, then what else is there to know?* The loss of context is thus accompanied by a (false) sense of always having the adequate amount of information.

### **References**

Abbott, A. (1995). Things of Boundaries. *Social Research*, *62*(4), 857–882.


Bleich, E. (2011). *The Freedom to Be Racist?* Oxford University Press.

Boland, T. (2007). Critical Subjectivity: Towards a Gnomonic Model of Subject Constitution. *Culture, Theory and Critique*, *48*(2), 123–138. https://doi.org/10.1080/14735780701723058


Hervik, P., Eide, E., & Kunelius, R. (2008). A long and messy event (kap. 2). In *Transnational Media Events. The Mohammed Cartoons and the Imagined Clash of Civilizations* (pp. 29–38). Göteborg: Nordicom.

Hobbelstad, I. M. (2015). Følsomhetens tid [The Age of Emotions]. *Samtiden*, 2.


Molotch, H., & Lester, M. (1974). News as Purposive Behavior: On the Strategic Use of Routine Events, Accidents, and Scandals. *American Sociological Review*, 39(1), 101–112. https://doi.org/10.2307/2094279


Sniderman, P. M., Petersen, M. B., Slothuus, R., & Stubager, R. (2014). *Paradoxes of Liberal Democracy: Islam, Western Europe, and the Danish Cartoon Crisis*. Princeton University Press.

Thornberg, R., & Charmaz, K. (2014). Grounded Theory and Theoretical Coding. In *The SAGE Handbook of Qualitative Data Analysis* (pp. 153–170). London: SAGE Publications Ltd.


### Chapter 7

# Boundaries of free speech in the political field

*Arnfinn H. Midtbøen, Ph.D., Senior Research Fellow, Institute for Social Research*

Building on in-depth interviews with the leaders of Norway's political youth organizations, this chapter focuses on two types of barriers to free speech that are at work in the political field: First, external barriers resulting from harassment and threats related to identity markers like gender, sexuality, disability and ethnic background. Second, internal barriers stemming from informal party cultures characterized by conformity pressure and silencing mechanisms. These barriers constitute boundaries of free speech which influence some politicians more than others. On the one hand, individuals who bear 'marks of difference' seem to be the major recipients of external harassment and threats, raising the cost of engaging publicly in controversial issues. On the other hand, politicians embedded in informal party cultures characterized by 'cultures of expression' which discourage political dissent, seem to face social sanctions potentially leading them to silence their voices. Implications for free speech legislation and the future recruitment to politics are discussed.

chapter 7

### **Introduction**

Being a public figure is risky business. Any public engagement involves the risk of receiving unpleasant comments, harassment or even verbal and physical threats (Meloy et al., 2008). Politicians, in particular, are sometimes subject to extreme exposure in the media (Thompson, 2000). In part, this is a consequence of their deliberate choice of political commitment: In democratic societies, people in positions of power should be exposed to criticism, and they must consequently be expected to handle unpleasant comments and satire. However, the exposure may sometimes be extreme, and it may have unforeseen negative consequences on the personal level (Thorbjørnsrud, 2003). In turn, this can lead to a democratic problem if individuals choose to withdraw from politics or if negative experiences make them less willing to take a stance in controversial issues due to considerations of their own safety or fear of isolation in the political community they belong to.

This chapter centers on the role of young politicians in Norway and their experiences while acting in public. Being the leader of a political youth party entails, of course, a range of positive aspects: It is a testament to broad political involvement which is crucial for democracy; it is a position of power and influence; and it may – as is the case for many previous leaders of political youth organizations in Norway – result in a life-long political career. However, the visibility of political leadership also involves the danger of negative experiences. Indeed, the comprehensive survey in the first round of the *Status of freedom of speech in Norway* project demonstrated that a large share of the Norwegian population tolerate politicians being subjected to negative characterizations, and that they are much more tolerant of negative characterizations of politicians than of minority groups such as Muslims or LGBT people (Enjolras and Steen-Johnsen, 2014). Previous research has also showed that Norwegian politicians are sometimes subject to extreme exposure in the media (Allern, 2001; Thorbjørnsrud, 2009), and that they are significantly more exposed to stalking than ordinary people (Narud and Dahl, 2015). Negative feedback from the outside is, however, not the only factor that can affect politicians' willingness to speak their mind on topics of importance to them. Politicians are embedded in specific party organizations characterized by formal structures and informal cultures which may influence what they say and how they act (Barrling, 2013; Barrling Hermanson, 2004). While formal structures are of obvious importance for understanding the functioning of party organizations, informal party cultures play a crucial role in defining the room for dissent and open conflict – and the social sanctions involved when crossing the line. Both of these factors – one external, related to the outside world, and one internal, linked to the inner life of the party – constitute boundaries of free speech in the political field. To the extent that these barriers negatively affect politicians' willingness to engage in controversial issues or even in politics altogether, they represent challenges to the processes of deliberation and – in the long run – to democracy itself.

This chapter builds on in-depth interviews with the current leaders of the political youth organizations in Norway, as well as with a selection of their predecessors. It considers their subjective experiences acting in the public sphere, distinguishing between, first, their experiences with harassment, libel and threats after speaking in public; and second, whether they find the informal culture in their respective party organizations to be open or closed to points of view which deviate from mainstream opinions in the organization. In the final part of the chapter, the implications of barriers to free speech in the political field are chapter 7

discussed by highlighting how negative experiences may affect the politicians' general views on free speech legislation and their wish to pursue a political career.

# **Why youth politicians?**

Politically active youth are the key to the development of democracy. The history and influential position of the Norwegian youth party movement suggest that political youth organizations are of particular interest in this context. In Norway, every political party has its own youth organization and the power and position of these organizations are larger and more central than in most other countries (Halvorsen, 2003). Most leaders of the youth organizations are visible actors on the public scene. All youth organizations are represented – with voting rights – in the parent party's executive committee, and they are active participants in the national congress meetings (Heidar and Saglie, 2002 p. 223). Even though conflict occasionally occurs – the youth organizations are often more ideologically anchored, and have historically tended to propose more radical political solutions to social problems than do their parent parties (Svåsand et al., 1997 p. 111) – there is no real debate about the close relationship between the main parties and their youth fractions in Norway (Heidar and Saglie, 2002 p. 58).

One reason for the preservation of these strong ties is probably that the youth organizations have always been important channels for recruitment to the political elite. The former leaders of the youth organizations tend to gain influential positions in the parent parties at a later time. In the case of the Labour Party (Arbeiderpartiet), for example, all elected leaders of the party until 2014 have had a position of trust in the Labour Party Youth (the current leader, Jonas Gahr Støre, is the exception), and the same goes for all of the Labour Party's prime ministers since World War II (Halvorsen, 2003). Similar ties between the youth organizations and their parent parties are found across the political spectrum. In fact, as Svåsand and colleagues point out (1997 p. 111), the top leadership in Norway's political parties is mostly recruited from the youth organizations. Hence, the general statement that 'youth can be expected to (co-)determine the further evolution of democracy and its institutions' (Forbrig, 2005 p. 13) holds particularly true in the Norwegian context. The leaders of the political youth organizations have the formal power to influence the parent parties while holding their leadership positions, as well as being likely to become important figures – if not leaders – of the main parties in the future. How the young politicians both describe their own public experiences and reflect on how the public 'rules of the game' should be defined in the years to come is consequently of great interest.

### **Boundaries at work in the political field**

Theoretically, this chapter takes as its point of departure that the public sphere is a locus of 'boundary struggles', concerning which groups and what opinions are offered a legitimate space in the public sphere (see Ch. 1). Some individuals are more vulnerable to negative comments and harassment than others, depending on their actual or alleged group membership, as well as on the topics they choose to engage in (Midtbøen and Steen-Johnsen 2016; Midtbøen 2016; see also Nadim, Ch. 8). Additionally, certain points of view are more contested than others. Individuals who choose to take on deviant positions in the public sphere, by challenging mainstream opinions on controversial issues, may experience social sanctions resulting in their voices being silenced, withdrawal from the public sphere or the development of echo chambers (Sunstein 2003; see also Thorbjørnsrud, Ch. 9). Brought together, these barriers to public participation constitute social boundaries of free speech.

How social boundaries of free speech play out in the political field is of great interest. In a democratic perspective, political decision-making should be based on the dissemination of competing perspectives on a given issue, which subsequently should be followed by processes of deliberation. Of course, decisionmaking processes in politics are always characterized by conflict and contestation. As Bourdieu (1991) schematically pointed out, the political field may be defined as a semi-autonomous social field organized around a binary logic in which the heterodox and the orthodox, the transformists and the conservatives, represent the main opposing poles. These poles exist both between parties and within each party organization, and the dynamics between them is crucial for political debates and decision-making. Under ideal conditions, the dissemination of standpoints is made without fear of other consequences than receiving rational counter-arguments and losing a vote. However, politicians may sometimes face barriers to free speech; both externally – by harsh responses received e.g. in social media, and internally – by silencing mechanisms operating within the party organization.

The external barriers to politicians' free speech are linked to the dynamics of the public sphere. Although politicians act and argue in their role as politicians, they are also individuals carrying markers with significance in the wider social context in which they operate. Markers of difference can be, for example, skin colour, ethnic or religious background, gender, disability or sexual orientation. These markers are not objective facts with a given 'effect' on individual identity or life chances, but become meaningful through individual self-identification to group categories, and by the categorization made by others through symbolic and social boundary-work (Lamont and Molnár, 2002; Barth, 1969; Jenkins, 1997; see also Nadim, Ch. 8). Regardless of the role individuals play, markers of identity may affect the experiences of single individuals, making some more vulnerable than others. Analyzing how markers of difference may result in the extreme exposure of some politicians while others can concentrate exclusively on their role as politicians, and how these barriers affect politicians' willingness to disseminate their perspectives or keep engaging in politics, is therefore crucial.

The internal barriers to free speech are located within the party organizations. While political parties are organizations operating in accordance with formal structures (not to be considered in this chapter), they are also characterized by different party cultures and traditions of individuality and conflict, historically developed and intrinsically linked to political ideology (Barrling Hermanson, 2004). Party culture 'determines the actual freedom of group members, the expectations about how they are to act and the means by which they are able to obtain social status' (Barrling, 2013 pp. 178- 179). Whether a party culture is open or closed to dissent and dissemination of deviating points of view may be decisive to politicians' ability or willingness to speak their mind. Of course, challenging the ideological or topical foundation of a political party may be considered taboo in most parties. However, what these taboo areas consist of differs across the political spectrum, and so also might the real or perceived sanctions experienced by those crossing the line. Challenging the core foundation of a given party – what 'we' agree upon – can in some organizational contexts lead to social isolation and deprive individuals of future opportunities. In turn, this chapter 7

can lead to 'spirals of silence' (Noelle-Neumann 1974) keeping others from raising critical questions or presenting deviating perspectives. Such silencing mechanisms may result in the absence of open debates, which would benefit the public and democracy itself (Sunstein, 2003), because it can lead to a cementation of party organizations and programs that are not adapted to rapidly changing societies.

In sum, politicians may experience both external and internal barriers to free speech, potentially affecting whether they engage in controversial issues in the public sphere and how they adjust to or challenge the real or perceived party culture. This chapter explores both of these dimensions by analyzing interviews with the leaders of Norway's political youth organizations.

### **Data, method and ethics**

The data underlying the analysis in this chapter consists of indepth interviews with the current leaders of the eight political youth organizations whose parent parties are currently represented in the Norwegian Parliament (in the period 2013-2017): Labour Party Youth (AUF), Norwegian Young Conservatives (Unge Høyre), Progress Party's Youth (FpU), Socialist Youth League of Norway (SU), Young Christian Democrats (KrFU), Young Liberals of Norway (Unge Venstre), Center Youth (Senterungdommen) and Young Greens of Norway (Grønn Ungdom). To include the entire spectrum from left to right, the current leader of the left-wing socialist youth party Red Youth (Rød Ungdom) was also included. Additionally, two former leaders – of the Socialist Youth League of Norway and the Center Youth – were interviewed. The latter two were included because they were unusually visible and controversial even within their own party community throughout their periods as leaders, indicating that their experiences are of particular relevance to the questions asked in this chapter.1

The interviews lasted between one and one and a half hours, were tape recorded and transcribed in full. The conversations were semi-structured using an interview guide. All interviews started by asking the informants to give an account of their political involvement and what topics they are particularly interested in, as well as the pathways to their leading positions. Further, they were asked to elaborate on the youth organization's relationship to the parent party; their own experiences participating in public debates; what topics they eventually avoid or are reluctant to engage in; and how they deal with questions that are considered controversial, 'difficult' or even taboo for the mainstream public and within their own organization.

The informants were between 20 and 29 years old when interviewed. All of them are well-known actors in the Norwegian public sphere. Some regularly participate in national and local media while others are somewhat less exposed. The extent of their visibility partly reflects the size and influence of the youth organization and how the parent party is positioned in current politics. However, visibility is also related to a distinct personal dimension in which some leaders tend to receive more attention, either because they take strong stances in controversial issues or because they have a background that fuels reactions from the media, mainstream society or particular social groups. A striking variety of

<sup>1</sup> Unfortunately, two other former leaders who would have been relevant to include in the analysis – Eskil Pedersen, leader of the Labour Party Youth during the terrorist attack on July 22, 2011, and Himanshu Gulati, the first leader of the Progress Party's Youth who has immigrant parents – did not wish to participate.

### chapter 7

backgrounds characterizes the current and former leaders of the political youth organizations in Norway. Among the eleven politicians interviewed, there are four women and seven men and they represent diversity in terms of ethnicity, sexuality and physical disability. Although the sample is small, the interview material presented in this chapter is well suited to explore whether public visibility in itself is a sufficient basis for negative experiences, or rather if public exposure needs to be combined with other markers of difference to create adverse conditions for public participation.

The interviews are to be regarded as elite interviews, as all informants have influential positions and are chosen because of who they are and the position they occupy rather than randomly or anonymously (Hochschild, 2009). As such, the standard anonymization offered to informants in qualitative studies was not possible to achieve. For this reason, I have chosen to use their full names when presenting direct quotes, while otherwise referring to them by the name of the political organization. This is acknowledged by all informants and my use of the interview material is made in agreement with all of them.

In the following sections, I analyze the interviews with the leaders of Norway's political youth organizations. The first section explores how markers of difference influence the experiences of the informants when they act as public figures, and how they handle and reflect on these experiences. The second section focuses on the significance of party cultures and how 'cultures of expression' affect the perceived freedom of speech offered to politicians in different parties. In the third and final section, I discuss the implications of the findings for free speech legislation and the future of democracy.

## **Markers of difference**

Previous survey research from Norway has demonstrated that while individuals with an ethnic majority background usually experience negative comments related to the content of what they write and their political standpoint, individuals with an immigrant background report that negative comments are often related to religion, ethnicity, national origin or skin colour, and, having such experiences, they are far more likely to be hesitant towards public participation in the future (Midtbøen, 2016; Midtbøen and Steen-Johnsen, 2016; Staksrud et al., 2014, Ch. 5). These findings suggest that being ascribed membership in ethnic or religious minority groups leads to less favourable conditions for participation in the public sphere. However, experiences of this kind are not limited to ethnic or religious minorities. Negative comments, harassment and threats may be directed at other typical target groups or at particularly visible public figures, like journalists (Hagen, 2015) and politicians (Narud and Dahl, 2015).

Based on the interviews with young political leaders in Norway, this study suggests that there is great variation in the experiences individual politicians have when acting in the public sphere in the sense that some politicians seem to be more exposed to harassment and threats than others. Quite strikingly, none of the white, straight, male interviewees report experiences that suggest that their public participation led to serious forms of harassment or threats. To the extent that these politicians have negative experiences at all they are not linked to core identity features, but rather to their political points of view which – according to their own assertions – is 'part of the game' of political involvement. As Atle Simonsen, the leader of the Progress Party's Youth, said when asked about his own

### chapter 7

experiences: 'When you are in the media, you get a lot of shit, a lot of negative stuff. But that's part of the game and I think you should handle it.'

The leaders that are either female or have a minority background, report receiving external responses of a different kind than their majority male counterparts. Consider, for example, the current leader of the Labour Party Youth, Mani Hussaini. Hussaini was born in Syria and arrived in Norway in 1999 when his parents, political dissidents, had to flee the country for safety reasons. Although having a Christian background, Hussaini is continuously believed to be a Muslim. According to his own statement, he normally receives loads of negative comments and threats especially when debating immigration and asylum policies – not least in the context of the 'refugee crisis' which dominated Norwegian public debates in the fall of 2015. These experiences have led to a feeling of discomfort when discussing these issues:

I'm not comfortable with debating asylum policies, I must be careful about what I say. The reason is that people immediately make connections; Mani Hussaini, an asylum seeker, a Muslim. I am not a Muslim, but this is the assumption, that I'm a Muslim. That I'm going to introduce Sharia in our country, and that that's why I'm the leader of the Labour Party Youth. It hasn't prevented me from engaging in asylum issues. But I must be prepared for the shit I'm going to get, because of my name, my skin color, my alleged motivations for talking about immigration at all. And, also, because I'm from Syria. I thought it would be a good thing in the current situation [the refugee crisis], that I have a background from Syria. But it is perceived as if I'm helping my own people. For example, at the National Congress in April last year, we [the Labour Party Youth] proposed bringing ten thousand refugees from Syria, or from neighboring countries, to Norway. It was like... I've never gotten so much shit. […] People wrote the most extreme things. There were death threats and... […] In 2014, when I was nominated as leader of the Labour Party Youth, a man wrote that the entire election committee should be executed, because they had nominated a Muslim.

The case of Mani Hussaini illustrates some important features of public participation in Norway. First of all, his story suggests that Islam has come to be an important, if not the dominating, demarcation line in the public sphere, defining who are entitled the privileges of unquestionable belonging, and who are not. The role of Islam in Europe has been compared to the role of race in the US context and claimed to be *the* bright boundary that hinders individuals from achieving parity with majority peers (Alba, 2005; Alba and Foner, 2015). Empirical research from Norway has pointed in the same direction, suggesting that individuals of Muslim background – and particularly those who are religiously conservative or who have chosen to use their voice to criticize racism and discriminatory practices in Norwegian society – are more exposed to harassment and critique than others (Bangstad, 2015).

Second, Hussaini's experiences show that even individuals who have an *alleged* Muslim background may face the same barriers as those who in fact are Muslims. On the one hand, this underscores the role of Islam as an important lens through which ethnic minorities in the public sphere are observed and evaluated. On the other hand, as Hussaini points out in the interview excerpt, different features of identity – name, skin colour, alleged religious background – may in fact overlap and in combination shape individual experiences (Jenkins, 1997). Previous research suggests that individuals of various ethnic and religious minority backgrounds – including those who are not assumed to be Muslims – may have severe negative experiences participating in the public sphere, for example if they take chapter 7

a strong stance in controversial political issues, but also that individuals of Muslim background do not necessarily have negative experiences (Midtbøen, 2016). While not downplaying the role of Islam as an important demarcation line in the public sphere, research on the conditions of public participation must be aware of the danger of 'methodological Islamism' (Brubaker, 2013 p. 13), implicitly assuming that the barriers facing individuals of Muslim background by default are greater than for other ethnic and religious groups.

Related to this latter point is the fact that other markers of difference, besides ethnicity and religion, also may be important determinants for the individual experience of public participation. Being gay, Nicholas Wilkinson, leader of the Socialist Youth League of Norway, for example, has not received direct harassment linked to his sexual orientation, but says that he avoids talking too much about LGBT issues in public. Wilkinson explicitly states that this strategy is a way of avoiding being transformed into '*the* gay politician'. This rationale is common among politicians with minority backgrounds and has previously been called 'the curse of representation' (Midtbøen and Steen-Johnsen, 2016 p. 25): In order to be able to express themselves as individuals and not be 'locked' in a minority category, these politicians tend to avoid commenting on minority-related policy issues (see also Nadim, Ch. 8).

The female political leaders in this study also report receiving negative comments related to their gender, personality and intelligence to an extent that their male counterparts, according to their own statements, do not. Especially when engaging in debates about gender equality or feminism, the female leaders report having negative experiences. Anna Serafima Svendsen Kvam, spokesperson of the Young Greens of Norway, for example, uses her experience of writing an article in a national daily newspaper about the use of gender-neutral pronouns to illustrate this problem:

That was when I've gotten the most shit, both on Facebook and in comment fields. A fairly large amount of comments were only harassment, not to the point at all. Very much like, "you're stupid", you know, that I'm unintelligent and stupid. And then there was some characterizing of my looks or my personality or ... yes, of me being a woman.

Similarly, Linn-Elise Øhn Mehlen, leader of the Red Youth, reports getting comments like 'she has her period', 'she hates men', 'she is jealous', when talking about feminism in public debates. On a direct question of whether she has received concrete threats she confirms:

Yes. I got a text message when I was at summer camp in 2014, from a man who had read an article that I'd published in the newspaper. And he was like, "I know you're at summer camp" ... He threatened me and said that we were red bastards, and, like, 'enjoy yourselves at summer camp". After Utøya, stuff like that is a bit uncomfortable.

Threats targeted at youth politicians at summer camp in Norway cannot be separated from the horrors of July 22, 2011, when 69 politically active youth at the Labour Party Youth meeting on Utøya, as well as eight individuals in the government offices in the Oslo city center, were brutally assassinated. In fact, the terror attack is mentioned by every informant in this study, although no direct questions about the terror attack were asked in the interviews. Often the reference to July 22nd was made when the interviewees were asked if they had observed other politicians being targets of hate speech, harassment or threats. Eskil Pedersen, the leader of the Labour Party Youth when the terrorist attack occurred, was then regularly chapter 7

mentioned as an example due to his experiences in the days and weeks following the terror attacks.2

Although July 22nd remains as a key reference point for the Norwegian public and among the interviewees in this study, both as an extremely violent attack on political youth in Norway and as an example of the vulnerability attached to holding a political leadership position, the role as leader of a youth organization does not in itself determine how single politicians experience the public sphere. Rather, core identity features like the leader's gender, ethnicity or sexuality seem to be the target of much negative response, particularly when combined with a controversial style or when specific topics – like immigration and gender equality – are under discussion.

An example where identity features and a controversial style are combined is Sandra Borch, the former leader of the Center Youth. Being a woman and the first person of short stature to lead a political organization in Norway, while simultaneously being a strikingly outspoken politician at a point in time when the parent party, the Center Party (Senterpartiet), was undergoing massive debates about its leader and the future direction of the party, Borch reports several incidents of serious threats during her period as a leader:

I've never cared much about what others say and do. And at first it was okay, I did not care so much about what was stated in the comment fields online. But there comes a point when... The condition

<sup>2</sup> Pedersen managed to escape Utøya immediately after the terrorist started shooting, while most of the other camp participants were left behind. These circumstances were never discussed critically in the professional media, but the incident received much attention on social media, including a range of severe attacks on Pedersen's person, his role as a leader and – being gay – his sexuality (Thorbjørnsrud & Figenschou, 2016).

I have made the comments include many more than just myself, like my parents and grandparents. My mom got e-mails and comments on Facebook stating that I never should have been born. [...] In the election campaign in 2013, things were pretty rough. There were death threats, people called from hidden phone numbers and submitted threats. PST [The Norwegian Police Security Service] chose to surveil where I was at all times. And that was kind of unpleasant, because I had never thought of it that way. I had to go different routes home at night by order of the police. It was really a lot.

Borch's experiences are extreme, and demonstrate the potential costs of public involvement. Indeed, she was a controversial politician who enjoyed provoking party colleagues as well as political opponents, and taking on such a role does entail a certain amount of resistance. However, in her case, the line was crossed: 'It went too far. When my family at home cried because they received so many messages, it was just not worth it'. Striking, too, is her experience of receiving little support from the party organization, particularly not from the parent party. This stands in stark contrast to the experiences of the other youth leaders who describe the party apparatuses as crucial in providing protective shelters when the external pressure gets too high. For Borch, the consequence of her experiences was a temporary withdrawal from politics on the national level. Although she continued to be active in local politics in Northern Norway, she chose not to run for a new period as leader of the Center Youth in the 2013 election, and she has kept a low profile in national media ever since. However, she has recently stated in public that she wants a comeback in national politics and in the Parliamentary election in 2017 she is the top candidate for the Center Party in Tromsø County.

Summing up, there can be few doubts that political leadership may result in extreme exposure in the media. chapter 7

However, the price paid for public engagement does not seem evenly distributed among youth politicians. Some are more exposed than others due to the size, influence or level of conflict in the parent party, but their individual characteristics seem to be decisive. The female politicians and politicians with a minority background interviewed in this study report having received harassment and threats targeted at their core identities, or that they avoid discussing in public topics that can be linked to their minority background in an effort to avoid being locked in a minority category. By contrast, the male politicians with majority backgrounds report no similar experiences.

Of course, being female or having a minority background could result in a higher awareness of the potential risks of negative experiences, indicating that male politicians with majority backgrounds may interpret otherwise similar situations not as incidents of harassment, but of criticism that follows naturally from public exposure (see Hagen and Drange 2016 for an interesting discussion of male journalists who experience sexual harassment). That being said, most of the male political leaders pointed out in the interviews that they are probably faced with different and less severe barriers to participation in the public sphere compared to their female and minority colleagues. The differences in experiences reported by the political youth leaders in this study are also reflected in recent survey research, demonstrating striking differences in the types of comments received by men and women, and by ethnic minorities compared to individuals of ethnic majority background (Midtbøen & Steen-Johnsen, 2016; Staksrud et al., Ch. 5). Although political leadership involves running the risk of extreme exposure, the stakes seem higher for some than others.

# **'Cultures of expression': The significance of party cultures**

While experiences of receiving harassment or threats represent the most clear-cut examples of how single politicians face barriers when engaging in the public sphere, boundaries of free speech are also set by the organizational cultures in which politicians operate. Political parties are established to promote particular group interests or policy issues. Although many topics are open for negotiation and internal struggles, some core ideas about society represent a party's backbone. Whether or not parties formally or informally allow for public debates about these core ideas will vary, and so will the personal consequences of challenging the party line.

Political youth organizations represent an interesting case when assessing how party cultures define the boundaries of free speech for political leaders. On the one hand, the role of youth organizations is to serve as a radical or ideologically 'pure' opposition to the parent parties and their leaders are elected to translate this opposition into political practice. On the other hand, the leaders of political youth organizations will often have political ambitions of their own, which may make them cautious in challenging mainstream opinions in the parent party. Indeed, the informants in this study point to the existence of such considerations. However, the various parties seem to be characterized by distinct 'cultures of expression', defined by the ways in which deviating points of view are sanctioned.

All informants confirm that the parties they represent have some core ideas which constitute their identity and ideological basis. The current leader of the Center Youth, for example, claims that no person in the organization would argue that Norway should apply for membership in the European Union. In this organizational context, EU resistance is part of the party chapter 7

identity which is seldom challenged. However, not all issues have this 'sacred' character. Sometimes political leaders may be uncertain of where the party line goes – particularly in the case of new proposals or issues in which the party does not have any formulated policy. Tord Hustveit, leader of the Young Liberals of Norway, describes how he deals with such situations: 'To me this is about what the organization thinks. I visualize four to five heads or faces in the organization, and then I think, like, how will they react to this?' According to Hustveit, these 'faces in the organization' could be county leaders, particular party members or members of the party's executive committee, whose reactions to a given proposition he tries to imagine. This way, he will cover the different viewpoints in the party organization and feel secure before making a public statement.

Kristian Tonning Riise, leader of the Norwegian Young Conservatives, states that when he says something in public about controversial issues, like immigration, he considers what words he uses to avoid being misunderstood. Similar to the leader of the Young Liberals of Norway, Riise thinks first and foremost about his own organization before making a public statement:

My role is to be a spokesperson for my members and it is important that my message appears in line with their opinions and values. And I am extra cautious in the immigration debate because you're very easily misunderstood. If things come out in a different way than I meant it, my members will react to it and think what on earth is our leader up to now? And then I'll think of how Høyre [the parent party, the Conservative party] will react. That's the next thing I think about.

Considerations such as these are prevalent across the political spectrum. This should come as no surprise: The leaders of political organizations are elected to represent the will of the organization and the extent to which they succeed in this representation defines their legitimacy as leaders. However, digging deeper into how the politicians in this study describe the organizational tolerance of open debates about controversial issues reveals some interesting differences.

In terms of organizational tolerance within parties, the dividing line in these interviews goes between the left and the right of the political center. While the politicians representing the youth fractions of the right-wing or center-to-right political parties describe their parent parties as relatively tolerant concerning what they as youth leaders can say in public, the leftwing parties describe a culture of expression characterized by a lack of room for internal critics. Particularly, this is the case when immigration is on the table. Nicholas Wilkinson, leader of the Socialist Youth League of Norway, for example, claims that challenging mainstream opinions on the political left, in an effort to create new policies through open discussions in the public, is very difficult:

There's a very strong internal justice on the left, where the right way to think is in accordance with what we've believed before. Take immigration, for example. I love immigration. I'm so glad that I can eat kebab when I go home from town, and not stock fish. But that's also why we can raise important issues without being accused of racism. But we keep our mouths shut. And that makes me angry. I think it's absolutely terrible. There's much disagreement on the left, about what is right and wrong, and there are a lot of emotions. The idea is that, 'This is like the rhetoric of FrP" [the Progress Party]. And then it's vicious and dangerous by default. I think that's a pretty weak logic, to put it nicely. I'm really against the policies of FrP, but that doesn't mean that anything that sounds like something they could have said is automatically wrong and evil.

chapter 7

Another example of this same phenomenon stems from Andreas Halse, the former leader of the Socialist Youth League of Norway. Being an outspoken internal critic throughout his period as leader of the youth organization, Halse claims that the political left in Norway is characterized by a striking conformity:

We [the political left] have for a long time steered clear of topics that have been difficult or divisive. I think there are two things here. One is a fear of conflict. This is something we share with the rest of Norway. Norwegians don't like conflict; we don't like too much disagreement and we're always looking for compromises. And if someone thinks something's unpleasant, it's better not to say it out loud. The second is the legacy of an ideology where there is one right answer. The further you go to the left, the more pronounced is the idea that there is an answer that is right and that opinions deviating from the correct answer should be rectified. I often meet party members who see it as their role to correct other members' opinions.

Like his party colleague, Nicholas Wilkinson, Halse finds that immigration is the most difficult topic to discuss openly. Rather than debating existing challenges to immigrant integration, for example, he claims that members of the Socialist Left Party, as well the political left in general, refrain from speaking their minds in fear of being sanctioned by the use of labels:

I believe that there are many people in the Socialist Party who keep opinions that they think are unpopular to themselves, that they are either reluctant to express themselves, or that they simply do not front these opinions in the open. And that's because you'll encounter some resistance that is not always based on facts. The left is very good at labelling. If discussing topics like racism and Islamophobia, everything is right-wing and reactionary, not solidaristic. In a number of issues, we simply hand out labels instead of discussing political solutions and what's actually on the table. I've always felt that large fractions of the left are controlled by emotions. And if something *feels* wrong, it's very difficult to discuss rationally.

To be sure, this alleged intolerance of deviating opinions on the political left did not keep Halse from acting like an internal critic when he served as leader of the Socialist Youth League of Norway. It is also important to point out that leaders of other, non-Socialist youth organizations describe taboo issues which are difficult to debate openly. The current leader of the Young Christian Democrats, Ida Lindtveit, for example, mentions same-sex marriages as such a topic, in which liberal views within the party are 'dangerous' to voice in public. Still, only the politicians on the political left describe entire party *cultures* as conformist with little room for deviating opinions. Like Wilkinson and Halse, Linn-Elise Øhn Mehlen, leader of the Red Youth, finds that the political left is characterized by a conformist culture of expression:

You tend to get labelled as a right-wing deviationist [høyreavviker] if you are critical of your own people or have new ideas. It's probably a form of conformist thinking. I'm not sure if it extends as far as the Labour Party, I think maybe not, but at least in SV [the Socialist Left party] and Rødt [the Red Party], I think it is very like, 'This is how you should think". And if you don't, then it's like, 'You can't sit with us".

There are some paradoxes related to the left-wing politicians' statements about conformist cultures of expression in their respective parties. First of all, they are strikingly open about the issue in the interviews, which suggests that there actually is some room for voicing internal criticism. Indeed, both the current and former leader of the Socialist Youth League of Norway have made controversial statements in national media while still

### chapter 7

being part of the party, pointing in the same direction. However, that internal critique is possible does not dismiss the possibility that conformist cultures of expression in fact may be present in these parties. Especially when the topics at hand are immigration and minority rights, there are even plausible reasons why it may be difficult to express critical perspectives as left-wing politicians. The parties on the left in Norwegian politics have been important in recruiting minorities to politics. They are generally concerned with discrimination, racism and hate speech, and they have traditionally argued in favour of liberal, inclusive immigration and integration policies, which is also probably a main reason why immigrants in Norway have tended to vote for the parties left of center (Bjørklund & Bergh 2013). The flip-side of this inclusive approach may be a fear that open discussions about the challenges of immigration to Norwegian society would feed into the rationale of immigrant-hostile social forces, resulting in a conformist culture of expression, at least on this specific question.

Of course, the limited set of informants in this study suggests that one should be cautious in making firm conclusions about the significance of party cultures in defining boundaries of free speech. Due to its salience in public debates, the immigration issue – which was used as an example of a controversial topic in the interviews – may also represent an extreme case which makes the differences between parties look more striking than would be the case if other topics had been in focus. One could even imagine that other topics would turn the findings up-sidedown, demonstrating similar conformity pressure on the political right as was reported on the political left in this study. Finally, as only leaders of youth organizations were interviewed, the relevance of these findings for the cultures in parent parties is unknown. All of these precautions warrant future studies. Still, the findings in these interviews suggest that party cultures may influence politicians' ability to speak their mind and this should be taken seriously. Fear of isolation and for being punished for failing to toe the party line may lead politicians to silence their voices. As such, conformist cultures of expression represent a type of boundary to free speech with important implications for political decision-making.

# **Implications for free speech legislation and democracy**

The former two sections have suggested that individual markers of difference and informal party cultures represent two distinct factors which influence politicians' willingness to express their opinions openly. What might be the implications of these findings – for legal requirements and recruitment to politics? Do politicians who have severe negative experiences tend to engage in a stronger regulation of free speech? And may the negative experiences of some political leaders, from external threats or internal opposition, lead others to silence their voices or refrain from engaging in politics because they cannot bear the potential costs?

In terms of the regulation of free speech, one could expect that the leaders of Norway's political youth organizations, who represent the entire political spectrum from right-wing to leftwing, would display a variety of opinions on where the legal boundaries of free speech should be drawn. However, they are generally consistent in their approach to free speech: From right to left, all informants are in line with a liberal approach, arguing that they as politicians and potential future legislators should not restrict public utterances unless they explicitly encourage the use of physical violence. This even goes for leaders who have severe personal experiences. Mani Hussaini, for example, uses his own family history to argue why freedom of speech is important: 'My family had to flee from Syria because we could not express what we wanted. It was forbidden to do so. So you can say that I have inherited the belief that freedom of speech is inviolable. You don't mess with freedom of speech'.

The only informant who argues that more legal protection against hate speech is necessary is the former leader of the Center Youth, Sandra Borch. Although educated as a legal scholar and principally in favour of a liberal approach to free speech, Borch believes that the current situation creates barriers to public involvement. As the only informant in this study who has decided to withdraw from national politics because of harassment and threats, Borch uses her personal experiences when arguing for a stronger legal protection against hate speech in Norway. Besides Borch, however, there is a strong consensus concerning the current legal boundaries of free speech. Although the politicians on the political left seem to have more difficulty in providing an ideological answer to why state regulation is not the solution to the dilemmas occurring when the execution of free speech by some may create barriers to participation by others, there is striking support for a liberal approach to free speech across the political spectrum.

Is this consensus surprising? On the one hand, free speech has been on the agenda throughout the political socialization of these young political leaders, and at times the debate has been intense (see Colbjørnsen, Ch. 6). In that sense one could expect that different points of view would be represented among the young politicians in this study. The fact that a wider range of positions does not seem to be reflected in the opinions of young political leaders in Norway is thus worth noting. On the other hand, consensus on this key principle of liberal democracy may also indicate that debates over free speech usually are not about legal provisions, but about tone and conduct in the public sphere and the harsh climate that sometimes characterizes public debates in Norway, not least in social media and in the comment fields.

While the personal costs of public participation may not lead to a restrictive view on free speech legislation, another important question relates to the consequences of negative experiences for the willingness to engage in controversial issues – in public and in internal party processes of policy development – and for future recruitment to politics. Indeed, several of the political leaders interviewed in this study worry that young people, especially women and individuals with minority backgrounds, may be discouraged from engaging in politics because they observe the personal costs of engagement in politics. For Sandra Borch, who chose to withdraw temporarily from the public spotlight due to the extreme pressure she had experienced, the implication of the harsh debate climate for future recruitment to politics is an important reason why she argues that a stronger regulation of free speech is necessary. 'I have received messages from people who do not want to get involved in politics because they cannot bear the consequences.' Several of the other interviewees have had similar experiences, claiming that they know of young people, not least young women, who have chosen not to pursue their political commitment, at least not in party politics, either because of their own negative experiences or because they have witnessed what others go through.

Most of the interviewees in this study are aware of the potential risks of participating in public debates, and several of them have implemented structures within their own organization to protect and support party colleagues. Some leaders systematically send a supporting email or personal message on Facebook to young colleagues who have participated in a public debate or published a feature article. Others have themselves received strong support from the parent party when they have had unpleasant experiences, referring to this support when explaining why they have implemented similar structures in their own organization. Such structures represent important bulwarks against the personal costs of public participation.

There seems to be far less attention to the silencing mechanisms operating internally in party organizations, however. This is probably a reflection of the fact that the external and internal barriers to free speech differ in at least one fundamental respect: External barriers – that is, experiences of harassment and threats coming from forces outside of the party organization – may create a feeling of internal solidarity and cohesion, strengthened by support provided by the political leadership or party colleagues. Internal barriers, stemming from conformity pressure within the party organization itself, on the other hand, represent a much more subtle type of barrier to free speech. If individuals fear the social sanctions involved in challenging mainstream opinions within the party, the consequence is more likely that they will act in accordance with the mainstream view, avoid controversial topics or simply opt out of politics, rather than addressing the sources of such sanctions. Challenging the internal barriers to free speech involves running the risk of creating conflict with friends and colleagues and potentially experiencing social isolation – that is, the opposite of internal solidarity and cohesion.

## **Conclusion**

Building on in-depth interviews with the leaders of Norway's political youth organizations, this chapter has shown that politicians may experience both external and internal barriers to free speech. On the one hand, political leaders are required to take an active role in the public sphere, by representing their members and fighting for their points of view. As such, political leadership is synonymous with power and influence, but public visibility also exposes politicians to harassment and threats. However, this chapter has shown that not all politicians are equally exposed to harassment and threats when acting in public. Despite their professional role as politicians, political leaders are individual bearers of identity markers which to a large extent seem to determine whether their path towards political influence will be easy or hard. Although most of the political leaders interviewed in this study have learned to live with being an exposed public figure – and many of them probably enjoy it – the chapter has shown that leaders who are either female or have a minority background of some kind do report having experiences with harassment and threats to a far greater extent than their majority male counterparts. This might suggest that women and minorities are more aware of the potential risks of public participation, but it probably also suggests that these politicians are in fact more exposed to harassment and threats than others – and quite surely that harassing comments directed at core identity features like gender, ethnicity or physical disability, have more severe consequences than comments related to political points of view.

On the other hand, boundaries of free speech also stem from the internal life of party organizations. Political parties are characterized by informal cultures with varying traditions for dissent and open conflict. In this study, an interesting distinction between the left-wing parties on the one hand and the center and right-wing parties on the other hand has come to the fore. The former and current political leaders of the left-wing political

### chapter 7

youth organizations describe the party culture as conformist and labelling, downplaying internal differences and sanctioning deviant opinions on topics perceived as taboo. By contrast, the leaders of the center and right-wing political youth organizations describe their party cultures as more open and that internal conflicts are allowed to be played out in public. Of course, the limited number of interviewees in this study warrants future studies of the significance of party cultures in creating barriers to free speech in the political field. Not least, the focus on immigration in these interviews should make one cautious in assuming that the conformism holds for all taboo areas, or that focusing on other issues would not reveal a similar conformism on the political right. The key finding here is not necessarily that it is difficult to discuss openly the challenges of immigration in Norway's left-wing political youth parties, but that different cultures of expression in fact are present in the political field. Such cultures may prevent open debate and silence perspectives or arguments that deviate from mainstream opinions and as such function as *de facto* barriers to free speech.

What are the implications of barriers to free speech in the political field? Liberal democracies are preconditioned to the individual right and ability to participate in the public sphere unhindered by social markers, and political decision-making should be based on viewpoints disclosed without fear of social isolation. Barriers to free speech, either through harassment or threats in the public sphere or by silencing mechanisms within the party organization, may result in the withdrawal of certain groups or opinions from politics. Harassment or threats from the outside world may keep individuals from addressing topics of importance to them or from pursuing a political career. Social isolation within the party organization, or the fear of this dynamic, may silence critical voices. The consequence of both barriers may be that groups and perspectives are excluded from political decision-making and as such serve as barriers to deliberation. Studying how boundaries of free speech play out in the political field is consequently of interest for the functioning of liberal democracy.

# **References**


Allern, S. (2001) *Flokkdyr på Løvebakken.* Oslo: Pax Forlag.

Bangstad, S. (2015). *The Politics of Mediated Presence: Exploring the Voices of Muslims in Norway's Mediated Public Spheres.* Oslo: Scandinavian Academic Press.

Bangstad, S. & Vetlesen, A. J. (2011). Ytringsfrihet og ytringsansvar. *Nytt Norsk Tidsskrift, 4*, 334–346.

Barrling Hermanson, K. (2004). *Partikulturer: Kollektiva självbilder och normer i Sveriges riksdag.* PhD Thesis. Uppsala University.

Barrling, K. (2013). Exploring the Inner Life of the Party: A Framework for Analysing Elite Party Culture. *Scandinavian Political Studies 36*, 177–199.


Chapter 8

# Ascribed representation: Ethnic and religious minorities in the mediated public sphere

*Marjan Nadim, PhD, Senior Research Fellow, Institute for Social Research*

This chapter explores how people of ethnic or religious minority backgrounds perceive and experience the conditions for accessing public debate, and more specifically, unpacks the role of ascribed identities and ascribed representation. The analysis is based on indepth interviews with individuals with a minority background, in addition to previous empirical studies, and shows that accessing public debate is not perceived as a challenge in itself. However, the conditions for access, most notably, who they are allowed to be in public debate can be a barrier for participation. Public participation is associated with a risk of becoming reduced and fixed to their minority status, and becoming a representative of their (assumed) minority group. In order to ensure a diversity of experiences and perspectives in public debate, it appears necessary to create a space for individual rather than group representation for minorities.

### **Introduction**

Not all people have or wish access to public discourse. But if specific voices, perspectives and experiences are systematically absent from public debate, it can become a democratic problem. To gain a better understanding of the boundaries of free speech, it is therefore relevant to examine the possibility of accessing public debate for individuals, in particular for minorities, whose voices are often not equitably represented (cf. Retriever, 2015). This chapter explores the conditions for participating in public debate in Norway for people with an ethnic or religious minority background, and examines one aspect of these conditions in particular: being ascribed the role of representing the group they (seemingly) belong to.

Many of the traditional media outlets have become concerned with presenting a diversity of voices including minorities (Bangstad, 2013). However, minorities who are granted access to the media often experience being *ethnicified*. They are ascribed and fixed to an identity as an ethnic other, and this one aspect of their identity comes to determine how they are portrayed and what topics they can address (see Bangstad, 2013, 2014; Eide, 2010a, 2010b; Midtbøen, 2016). At the same time, there appears to be space – at least for some individuals – to transcend ethnic and religious categorization and develop individual identities in public debate (Midtbøen, 2016).

The questions asked in this chapter are: What is the role of ascribed identities, and *ascribed representation*, in relation to participation in the mediated public sphere for people of ethnic or religious minority backgrounds? And how are these issues dealt with by (potential) participants? The analysis is based on an empirical investigation of ten potential, and two experienced, participants in public debate, all with diverse ethnic and religious minority backgrounds. In addition, the analysis draws on previous empirical studies of active media participants with a minority background (Bangstad, 2013, 2014; Eide, 2010a, 2010b; Midtbøen, 2016; Midtbøen & Steen-Johnsen, 2016). The chapter concentrates on participation in the mediated public sphere, i.e. traditional media outlets where editors and reporters act as gatekeepers (e.g. Shoemaker & Vos, 2009, see also Ihlebæk & Thorsen, Ch. 5 this book). This is still considered a forum for public debate which has more impact and higher legitimacy than others (e.g. Carpentier, 2011), and issues of ascribed identities and ascribed representation become particularly relevant here, as the individual is at the mercy of editors and reporters when it comes to access and how they are portrayed.

The study adds to the literature on minorities in the public sphere in two ways. First, it contributes to broadening the understanding of the barriers for participation by also examining the experiences and perspectives of people who are not (yet) an established part of public debate. Research on minorities' perspectives on media participation in Norway has so far concentrated on experienced media actors (Bangstad, 2013; Midtbøen, 2016). Although these actors describe challenges and barriers in accessing the mediated public sphere, they have in fact already overcome the main of barrier of access.

Second, the chapter unpacks the role of ascribed identities and ascribed representation for minorities' participation in public debate. Inspired by Phillips (2009) the chapter introduces an analytical distinction between group and individual representation, and highlights the fact that while ascribed representation poses an obligation to represent a certain group in public, it also creates an obligation towards the group, to represent them in a manner they recognize and accept. By analyzing the experiences and perspectives of people who are not a part of public debate, while also drawing on empirical studies of experienced media actors, the chapter contributes to knowledge about the conditions securing minorities (un)equal access to the mediated public sphere.

### **Minorities in the mediated public sphere**

The presence of minorities in the mediated public sphere has been studied in several ways. In media studies there is a long tradition of examining how minorities are represented and portrayed in the media. These studies show that immigrants and other minority groups are often ascribed stereotypical attributes based on their group membership (see Cottle, 2000b; Gullestad, 2006; Hall, 1997; Retriever, 2015). Recent studies, however, indicate that there are tendencies towards more complex portrayals of minorities in the news (Cottle, 2007, 2000b; Retriever, 2015). Others have studied the presence of minorities in the media. Studies of Norwegian media find that people with immigrant backgrounds are underrepresented in news media (Retriever, 2015), also in news about immigrants (Figenschou & Beyer, 2014, see also Cottle 2000b).

Despite underrepresentation, there appears to be an increasing presence of minority voices in Norwegian media, and some of the main media institutions have publically declared that they aim to increase the share of immigrants in their outlets (Bangstad, 2013; Midtbøen, 2016). A few studies have examined the experiences of these 'minority voices' in Norway (Bangstad, 2013, 2014; Eide, 2010a, 2010b; Midtbøen, 2016; Midtbøen & Steen-Johnsen, 2016). The studies show that the active media participants experience being *ethnicified* or subject to culturalization by reporters (Eide, 2010b). This media ethnification takes the form of a 'one-sided, dominant media focus on a person or group as an *ethnic other*, an emphasis on her difference (from a presumed 'us'), based on her being (more or less) visibly different or on a tacitly presumed *background* that differs from the mainstream' (Eide, 2010b: 66).

The profiled media actors experience Norwegian media as being mostly interested in them in their capacity as minorities, and in relation to minority-related topics, like religion, immigration and integration (Eide, 2010b; Midtbøen, 2016). Thus, certain topics and ways of presenting oneself seem to grant more ready access to the mediated public sphere.

The empirical contributions differ in how they interpret the possibility for minority actors to create a space for themselves as individuals rather than as representatives of a group. While Bangstad (2014) implies that race, ethnicity and religion have wide-ranging significance, Midtbøen (2016) argues that there is a space for individuals with a minority background to transcend ethnic boundaries to the extent that they can participate in public debate as individuals, regardless of their minority background.

## **Ascribed identities and ascribed representation**

The empirical studies of minorities in the Norwegian mediated public sphere illustrate the importance of ethnic boundaries in shaping the experiences of minority participants in public debate. Fredrik Barth's (1969) influential work emphasizes the relational nature of ethnic boundaries. Ethnicity, but also identities more generally, concerns both *self-identification* – one may see oneself as a member of a particular ethnic group – and *ascription* – others may identify a person or a group of persons as a member of a particular ethnic group (Barth, 1969; Jenkins, 1997 p. 53). The boundaries between ethnic groups are defined, not by objective or 'real' differences between groups, but by what are considered to be *socially relevant* differences. However, not all people are able to choose how they are categorized (Jenkins, 1997). Visible markers of difference, like skin colour, can lead to immediate categorization and the ascription of an ethnic identity, regardless of how one understands oneself. It is not possible to opt out of this categorization because the identity is simply ascribed, the visual marker of category membership poses an external obligation to adopt the identity in question (Duveen & Lloyd, 1990; Jenkins, 2014).

Alba (2005) emphasizes that all boundaries do not operate in the same manner, and makes a distinction between bright and blurred ethnic boundaries. When ethnic boundaries are bright there is no ambiguity in who belongs to the ethnic group. But when boundaries become blurred, the location of the boundaries is indeterminate or ambiguous, at least for some sets of individuals (Alba, 2005). Ethnic boundaries are thus not given once and for all, but are rather the result of everyday boundarymaking processes – both internal and external to the group – that are dynamic and can change over time (Barth, 1969; Jenkins, 1997).

When boundaries are bright, ascribed identities can become essentializing, in that certain traits are seen as fixed and shared for all members of a certain group, whereas individual variation and change are disregarded. Essentialism means that one trait, for instance your skin colour, or ethnic or religious background, comes to determine the whole definition of your identity, and you become reduced to that one trait, regardless of how you identify yourself (cf. Hall, 1997; Mansbridge, 1999). When individuals are seen solely as members or representatives of a specific group, ascribed identities can entail imposed or *ascribed representation*.

In the context of political representation, Anne Philips makes a distinction between a corporatist representation on the one hand, where 'individuals serve as the authorized representatives of their group and are regarded as its authentic voice', and on the other hand, 'looser measures that seek to increase representation of people sharing the markers and experiences of these groups' (Phillips, 2009 p. 168). Philips warns that when individuals are seen as representatives of a group, it invokes a reified understanding of *the* group, *the* culture, or *the* community that is being represented. This type of representation – which I will refer to as *group representation* – can reinforce essentializing tendencies, especially if it is imposed. Insisting that ethnic or religious minorities represent 'their group', regardless of whether they take on this role, implies that there is an essential quality to being, for instance, Jewish that all Jews share, and that gives them common interests despite what might divide them (cf. Mansbridge, 1999 p. 637). In the extreme, this implies that any person of Jewish background represents all Jewish people, regardless of their political beliefs, gender or other differences (cf. Mansbridge, 1999 p. 638).

The 'looser' type of representation that Phillips addresses can be termed *individual representation*. This type of representation allows for the multiplicity of identities and the unique interests and experiences of the individual to be recognized. Although categories such as ethnicity and religion do not determine individuals, these categories have substantial material and discursive significance (cf. McCall, 2005). An individual's minority status will thus – to some degree at least – shape their interests and experiences, making it a democratic necessity to have individuals with different group attributes represented in public debate and in the decision-making process (Phillips, 2009 p. 168). Individual representation means that individuals are allowed to participate with the full spectrum of their experiences and traits (for instance as an economist that happens to be Muslim), without being reduced to one aspect of their identity (i.e. being seen only as a Muslim).

This chapter explores the perspectives of individuals with an ethnic or religious minority background. Ethnic and religious identities are not always easy to disentangle. They often overlap, and are to varying degrees conflated in public debate. Alba (2005) argues that while race is a bright and salient boundary in the US, religion, specifically Islam, plays a parallel role in the European context. The category 'Muslim' has become racialized, so that religion, ethnic origin and skin colour largely become conflated, to the extent that dark-skinned individuals with immigrant backgrounds are immediately categorised as Muslim (see Midtbøen, ch. 7 for a striking example). In contrast, the category Jew is much less salient, and often less visible in the Norwegian context. In this chapter, I do not unpack the distinctions between ethnic and religious categories, but I treat them as minority statuses that can become salient in different ways.

# **About the study**

The chapter is based on qualitative interviews of people with an ethnic or religious minority background, who are *potential* – but not established – participants in public debate. Because the aim of the study is to capture potential barriers to participation in public debate, it was important to identify individuals for whom participation is somehow experienced as a relevant option. It is not obvious how to define this criterion, but I operationalized it as individuals who are relevant for public debate either through their (formal or informal) position in an organization, the ethnic community or professionally, or more generally through their social commitments (e.g. being highly engaged in issues such as gender equality, immigrant integration or religious rights).

The study includes twelve interviews: ten in-depth interviews with people of ethnic or religious minority backgrounds in Norway and two more informal interviews with experienced media actors with minority backgrounds. The main sample consists of five women and five men, within a wide age range (from 17 to 55 years). The participants have different kinds of minority backgrounds. Two have a Norwegian-Jewish background. Six identify as Muslim, with immigrant backgrounds from Pakistan, Sri Lanka and Somalia. The remaining two have an immigrant background (Tamil and Turkish), but did not declare a religious affiliation. All but two of the interviewees have grown up in Norway.

The participants were recruited through the two experienced media actors, and through already established networks in different minority communities with subsequent snowball sampling. I made an active effort to get participants from different sources. Seven of the interviewees have (or have previous experience from) formal positions in organizations built around an ethnic/religious community or dealing with minority issues. The rest have an informal position in the community or a strong social commitment that makes media participation a distinct option. The *potential* participants in public debate recruited for this study all turned out to have some sort of experience with the media, either as sources or through participating in the opinion pages of a newspaper. Still, none of them should be considered as experienced media actors, neither did they understand themselves as such.

The interviews were semi-structured, and the topics of the interviews were to some degree adjusted to the situation of each individual. Generally, the interviews covered the following issues: Experiences with actual participation in public debate (both in traditional and social media), experiences and perceptions of the conditions for accessing public debate, and perception of the possibility for them to express their opinions.

### **Ascribed identities as part of the game**

How do ethnic and religious minorities perceive the possibility to access the mediated public sphere? Perhaps unsurprisingly, the studies of experienced media actors find that they do not consider *access* to the media a challenge in itself (Bangstad, 2013; Eide, 2010a; Midtbøen, 2016; Midtbøen & Steen-Johnsen, 2016). However, the interviewees in this study, who do not have much experience with participating in public debate, perceive that they too have access to the media and public debate, in case they should wish and try to participate. The interviewees consider the Norwegian media to be interested in people of minority background, describing them as having 'quotas for people with a minority background'. They regularly witness others with a minority background in the media, and the presence of wellestablished and visible media actors with a minority background reinforces the perception of Norwegian media as accessible for minorities.

Although access is not seen as a great challenge, '[...] the way you reach out, that can be a small problem', as one of the interviewees puts it. In other words, the *conditions* for access can be problematic (cf. Bangstad, 2013; Midtbøen & Steen-Johnsen, 2016). And a central 'problem' of access is that the gatekeepers of the media – editors and reporters – are often interested in them only as minorities, and not in their other capacities. One of the interviewees describes an illustrative situation:

[A major online newspaper] calls me and says: "We need some people who are not visible, but who are competent and want to be visible". And I say that I'm an economist. I can talk about my topic or I can talk about career and education, if you're interested in that. And he says: "Yes, [this person] said that all Muslims are terrorists. What's your comment on that?" I say that I can't comment on this in the media. It has no relevance. Okay, I have a Muslim and Pakistani background, but I don't feel that I'm a spokesperson for that topic, and I'm not going to comment on it.

Although the reporter extended a seemingly open invitation to contribute to the public debate, it soon became clear that he was not interested in the interviewee in his professional capacity, but as a representative of Muslims. Another interviewee explains that gaining access to the media is unproblematic as long as you stick to minority-related issues:

As soon as I have something to say about Islam, Muslims, something like that, ISIS, it's so easy to get access. Journalists love to write about it. […] As soon as I play my minority card or religion card, there's no problem getting an issue in the media.

The interviewees have a clear perception that their minority background is what interests the media, and shapes what topics they are most readily granted access to address. This conditional access to public debate is also described by more experienced media actors, who point out that Norwegian media are mostly interested in minorities engaging in specific – and minorityrelated – topics, typically questions related to religion, integration and immigration (Eide, 2010b; Midtbøen, 2016; Midtbøen &

#### chapter 8

Steen-Johnsen, 2016). Thus, they experience being interesting primarily as representatives of their minority group.

Furthermore, the interviewees state that not all positions are equally attractive to the media. Bangstad (2013) argues that the Norwegian media include and privilege the voices of individuals of Muslim background who engage in criticizing Islam, while they often exclude Muslims who are not prepared to engage in such critique (see also Gullestad, 2002). The interviewees in this study perceive that it is polarized views – including critique of one's own group, but also for instance extreme religious beliefs – that easily gain access, whereas positive stories and nuances are more often excluded.

The interviewees see the focus on their minority background and the privileging of polarized positions as part of 'the game' of media participation. These experienced conditions for access partly reflect common journalistic conventions. The use of 'cases' is a common feature of present-day journalism. It means that individuals are used as an illustrative example of the issue at hand, portraying them as representatives of a group to show that their story has relevance beyond themselves as individuals (Hågvar, 2016 p. 292). Using polarization, at the expense of nuances, to create debates is also not specific to minorities in the media (e.g. Ihlebæk & Thorseth, ch. 5). Still, such common features of journalism can contribute to reinforcing the tendency towards reducing media participants to the minority aspect of their identity.

It is not only the gatekeepers of the mediated public sphere who ascribe identities, and have a one-sided focus on their minority status. The interviewees also believe that, within the general public, visible minorities inevitably become reduced to their minority status, regardless of whether it is relevant to the issue at hand, in particular in the discussions in the comment sections. One interviewee describes the response to a news article about his outstanding educational achievements: 'There was a lot of focus on religion, which wasn't even mentioned [in the article]. And there was a lot of focus on ethnicity, which didn't have anything to do with it, right?' Although he found his ethnic and religious background to be of little relevance to the news story – and in fact these characteristics were not highlighted in the article – his Pakistani and (assumed) Muslim background were made 'bright' in the discussions that followed.

Although accessing public debate in itself is not seen as a great challenge, the conditions for access frame *who* you are allowed to be in the public sphere. The interviewees have a clear notion that participating in public debate entails having a minority identity ascribed, regardless of how they wish to identify themselves. Put crudely, this happens – indirectly and directly – through which topics they are seen as eligible to discuss, through what positions are available to them, and through how they are received by the general public.

## **Ascribed representation as a barrier for participation**

Some of the interviewees find the conditions for access detailed above – and the ascription of identity that comes with it – highly problematic. The challenge, as the interviewees describe it, is not so much that they are ascribed a minority identity regardless of how they present themselves. It is that they are *reduced* to that identity only. They articulate a concern that by participating in public debate, they risk becoming '*the* Minority' for the Norwegian public. For instance, a participant with a Jewish background describes her main reason for not wanting to participate in public debate: 'Although you might want to participate a tiny bit, you don't want to become, well, the Norwegian Jew'.

Another interviewee with a Jewish background has a similar reflection. Although she feels that she has something to contribute to the public debate, she does not participate because she is reluctant to become a 'public Jew':

I agree that the voice [that I represent] is perhaps needed in some way or another, but I don't want to assume that role. And that's mostly about… What is it mostly about? I think it's mostly about somehow taking on the role of a public Jewish person.

The interviewees imply that participating in public debate means, not only being ascribed a Jewish identity, but being positioned as representing Norwegian Jews. Both interviewees with a Jewish background are white and come from families who have lived in Norway for generations. Thus, they do not stand out as visible minorities. The question of whether or not to participate as a *Jew* in public debates is for them, therefore, also a question of whether they wish to draw attention to their minority status.

However, the fear of becoming *the* representative of 'their group' is not limited to the non-visible minority interviewees. Also those with an immigrant background, who are dark-skinned, and therefore more visible minorities, articulate a concern for becoming *the* Muslim in Norwegian public debate. As one of the experienced media actors in the study explains: 'If you go out [in public] you become a representative of Muslims'. While several of the interviewees explicitly refer to the fear of becoming '*the* Muslim', no one mentions the fear of being reduced to a representative of their ethnic group. I will return to the specific role religion seems to play later in the chapter.

When Midtbøen and Steen-Johnsen (2016 p. 25) describe the 'curse of representation' for their media-active participants, they refer to how the participants experience being limited with regard to what *topics* they are invited or even permitted to discuss. However, for the interviewees in this study, who are in the periphery of public debate at best, there seems to be more at stake. They problematize being ascribed a public minority status in itself. When they talk about the fear of becoming *the* Minority in public, it implies that their religious or ethnic identity is not only made visible, but that it comes to define them entirely. This means being ascribed a role as what I, following Phillips (2009), have termed group representatives. They are seen solely as members of a specific group, speaking on behalf of that group. The contrasting form of representation would be to be able to participate as individuals, with distinct experiences and interests that might be shaped by their minority status, but without being reduced to being only a minority.

Although the participants understand the risk of representation in broader terms than what topics they gain access to talk about, topics also matter. Those who wish to participate in public debate in order to communicate their experiences and perspectives as a minority can get caught in the tension between representing an individual with a minority perspective and becoming a representative of the group. One of the Jewish participants articulates this tension when she explains why she does not want to participate in debates about anti-Semitism or the conflict in the Middle East:

[…] the debates where I have something I wish to say, are often debates in which I don't want to… At the same time as I want to get to say what I want to say, I don't really want to get the stamp that you often get if you participate in those kinds of debates. [...] It's those debates where I often sit and bite my tongue about things I would like to say, but don't dare to say. Or choose not to say. It's not really that I don't dare, but that I don't want the kind of attention you get. Or the attention I think you maybe get.

The attention she is afraid of getting, is partly related to her notion that her opinions go against the established positions in Norwegian society and that they will therefore be challenging to voice. But it is also about the type of representation she thinks comes with participating in these debates, namely being cast as 'The Young Norwegian Jew', as she articulates it. Thus, she believes that it is not possible merely to be a voice that represents minority perspectives and experiences in the public debate, without being cast as *the* Minority. As Eide (2010b: 73) finds in her study of experienced media actors: 'If you (sometimes) speak as or on behalf of a [minority] group, you are deemed to *be* that group' [emphasis in original]. For some interviewees the concern about being ascribed a position as *the* Minority is a substantial barrier to participation in the media and public debate. But why are they so reluctant to risk ascribed representation?

## Feared consequences of becoming The Minority

What is perceived to be at stake in ascribed representation centres around a concern for three types of consequences: 1) hateful reactions, 2) ascribed opinions and beliefs, and 3) professional consequences.

The interviewees worried about negative reactions and hate speech as a consequence of public participation, a concern that is also present among the experienced media actors (see Bangstad, 2013; Eide, 2010b; Midtbøen & Steen-Johnsen, 2016). Some have experienced negative and frightening reactions after participating in public debate, while some have witnessed others receiving negative comments, hate speech, and even serious threats. There is a shared perception that people with a Muslim background are especially at risk of getting such reactions (see Fladmoe & Nadim ch. 2 for analyses of experiences of receiving hate speech among people of immigrant background). One of the interviewees explains:

I know several people who found this to be a great strain. Everybody who writes and has a Muslim background, no matter how well they write or how badly they write, they experience the same thing. These trolls. […] I've warned the whole youth group: "Just don't spend time reading that, it'll only make you want to go and kill yourself ".

His dramatic wording probably reflects how degrading he finds the comments you can find in online debates and comment sections. For some, the fear of these kinds of negative reactions is their main reason for staying away from public debate, as one interviewee describes: 'I'm not one of the strong ones who dares to be in the media and receive online hate and the many strange comments you see in the comment sections, regardless of what article you're reading.' Others, however, describe degrading comments as an unpleasant aspect of public participation, but not as something that would hinder them in participating altogether.

A different type of consequence is more directly related to ascribed identities, namely ascribed opinions and beliefs. Categories, such as 'Muslim', 'Jew', 'immigrant', are imbued with meaning. In being ascribed an identity you are also being ascribed certain attributes, opinions and beliefs (cf. Duveen & Lloyd, 1986). One of the experienced media actors explains that Muslims are automatically seen as orthodox, and are expected to comment on, or denounce, any negative act that is done in the name of Islam (see also Bangstad, 2013; Eide, 2010b). Another interviewee explains that you 'have to use a lot of energy on positioning yourself differently than people perhaps assume'.

However, ascribed identities and ascribed opinions are significant mostly because the interviewees see them as a clear threat to their professional lives. A recurring concern is that being cast as *the* Minority will overshadow their professional competence, to the extent that it will have consequences for their career. For instance, the two participants with a Jewish background work in different fields, but both assert that becoming 'a public Jew', as they phrase it, will impair their professional credibility. One of them says: 'I also feel I choose not to take that position in public debate to not close any doors – in my professional life'.

Another participant, who has a Muslim background, elaborates on this concern:

Once you say something about religion in the media… I don't feel I have anything whatsoever to gain from it, no matter how positive my statements are. Because of the extreme secularism [in Norwegian society], it will backfire on me. And it will backfire against what is typically my professional competence. And that makes me very conscious of saying anything about religion in the media, for example. Because I don't want it to overrun my professional competence. That you become defined as the Muslim in [the company] instead of the economist in [the company].

Interestingly, the concerns about the consequences of becoming a hyper-visible minority are expressed in relation to religious minority status, but not in relation to an *ethnic* minority status. This might reflect the prominent position and 'brightness' of religion as a boundary in current public debate (cf. Alba, 2005; Alba & Foner, 2015). As the interviewees point out, in a highly secular context like Norway, the role of religion in the public sphere is contested. There is an implicit hierarchy of worth, and perhaps a hierarchy of rationality, of religions, where Islam seems to be at the bottom (cf. Bleich, Stonebraker, Nisar, & Abdelhamid, 2015), being portrayed as a threat to liberal societies and rational thinking (cf. Huntington, 1996). Although one could assume that the position of the Muslim is more stigmatized than that of other minorities, it is noteworthy that the interviewees with a Jewish background appear equally reluctant to take on the role of a public representative.

## **Strategies to deal with ascribed representation**

There are different strategies to deal with the risk of ascribed representation and essentialization. As I have shown above, for some of the actors on the periphery of public debate, the answer is to avoid participation in the mediated public sphere. Another way this is handled is by strategically playing by the rules of game, playing the 'minority card or religion card' as one of the interviewees phrased it. A young interviewee, who is engaged in politics, describes using this strategy:

I think there are two ways of seeing it [the media's focus on minority background], and I only see it as an advantage. Yes, yes, okay, so maybe I'm there because of my skin colour, that's why I'm on that news story on NRK [The Norwegian Broadcasting Corporation]. But I'm in. So I don't see it as something negative. If that's what gives me the opportunity, then of course I'm going to use it, right? I mean, we all have our qualities and attributes so we can use them.

She sees how other young people struggle to be heard, and although she recognizes that it can be problematic that the basis for her participation is her skin colour, she argues that her minority status gives her an opportunity that she should embrace. Similarly, Eide (2010b) finds that some of the experienced media actors with minority backgrounds engage in *strategic* 

### chapter 8

*essentialism* to obtain media attention and recognition. This implies complying with the conventions of the journalists and temporarily downplaying internal differences, and accepting a simplified and essentializing image of the group (Eide, 2010b; Spivak, 1996). This can be understood as a form of *strategic*  group representation.

While strategic representation, or 'playing the minority card' means accepting ascribed representation, most of the interviewees in this study, and the experienced media actors studied by others (Bangstad, 2013, 2014; Eide, 2010a, 2010b; Midtbøen, 2016), problematize and contest this condition for participation. Some of the interviewees portray ascribed representation as more or less inevitable – either through an emphasis on how the readership reduces everything to religion and ethnicity, or through their perception that the mediated public sphere is only interested in them as minorities. While others hold that there are ways to circumvent this, and actively try to negotiate and challenge this condition for media access.

The main strategy of challenge is avoiding minority-related topics (see also Midtbøen, 2016). As one of the interviewees describes: 'So I've been conscious about this, right, everything I'm going to comment on or things like that, should be directed at my professional competence, and not background.' He deliberately does not talk about religion, integration or related topics. By strictly participating as an expert in his discipline, he argues that he can avoid the minority label. This strategy is even more pronounced for the interviewees with a non-visible minority background.

One of the interviewees with a Jewish background explains that she occasionally participates in debates or interviews on television and in the printed press through her work. However, she is very reluctant to participate 'as a Jew':

I've sometimes been asked by editors or others to say something about how it is to be a Jew in Norway, how one, as a Jew, sees the conflict in the Middle East and things like that, where I have said no. On the one hand, I feel that if I were to comment on that, I would have wanted a much more professional perspective on it […] But when it becomes a personal matter, I don't feel that… Or I don't want to speak about my personal experiences in the media.

Her standpoint is very similar to that of the interviewee described above. Both emphasize that it is fine to speak about religion or the experiences of the minority group, if this is your area of expertise. But it is not relevant for them to participate as a member of a minority who speaks only in the capacity of being a minority. The difference between the two interviewees is, however, that one is not a visible minority and is thus freer than the one with a more visible minority background to choose and control the extent to which her minority background is emphasized.

Nevertheless, avoiding minority-related questions can come at a cost, because they might not be granted access on the terms they insist on, in the form of individual representation. Several of the interviewees say they have tried to gain attention on non-minority related issues, without success, while others do not consider themselves competent enough in their professional field to wish to take a public role in that capacity. Midtbøen's (2016) study of experienced media actors, however, finds that some individuals are in fact able to transcend ethnic and religious boundaries, and participate in public debate based on individual merits and preferences, on their own terms. He finds that second generation immigrants particularly, born and raised in Norway, seem to be able to challenge and overcome ethnic boundaries.

To sum up, this study reveals different strategies employed by ethnic or religious minorities in the face of the risk of ascribed representation: *avoidance* – sidestepping the risk by avoiding media participation altogether; *acceptance* and strategic accommodation to the conventional conditions for media access by emphasizing their minority status, thus accepting group representation; and *challenge* by insisting on individual representation.

### **The legitimacy to represent**

So far I have examined the question of representation in terms of who participants in the mediated public sphere are seen to be by those outside the minority group (e.g. the media and the general public). However, representation does not only concern your role externally, it also concerns how the group you are supposedly representing understand and identify you (cf. Barth, 1969; Jenkins, 1997). People have different positions within a given group, and a challenge with (being ascribed) group representation is the question of with what legitimacy one can take (or be given) such a role.

One of the interviewees articulates how speaking as a minority can also pose challenges within the community:

We are so few who say anything [in the media] […] that you can be a bit frightened of the reactions from the community. Or you can be a bit sensitive to them. It increases because you're the only one perhaps who says something. Because suddenly you're supposed to look after everybody's standpoint. And that's not possible, right? Here you have people from the far right to the far left, and everything in between. You can't make a statement that everybody can support. And then you might get an uneasy feeling that people in the community think you've done a bad job with your statements. And you definitely cannot be bothered with that. The people you're supposedly in the same box with don't agree with you or think you've done a bad job. Then you perhaps might as well not do the job.

The interviewee implies that it is difficult to insist that you participate in the public debate as an individual, also for the community itself, when there are few others with the same background visible in the media. Also within the group, you are seen to become a group representative – or at least as trying to take on this role. Furthermore, she points to a central difficulty in taking on such a role: There is no unified standpoint to communicate, and you only risk disappointing people.

An important factor in how the interviewees relate to the question of representation is their understanding of what position they speak from. Some have formal positions that allow them to easily take on the role of a group representative that they think is seen as legitimate in their own community. Others find the position of a group representative problematic precisely because they feel they lack such legitimacy.

For instance, one of the interviewees used to be the spokesperson of a mosque. Now that he no longer has this role, he is reluctant to participate in public debate. He explains this in terms of how time-consuming and straining it is to get the media to portray 'their cause' in a proper manner. He does not, however, seem to question his qualification to speak on behalf of his mosque or on behalf of Muslims more generally.

In contrast, another interviewee feels she has no legitimacy to represent or speak on behalf of the others in her community. She used to have a central role in an ethnic organization, but she explains that she was essentially excluded after asking critical questions. She believes she has a lot to contribute to public debate, but it is not an option for her to speak without a position to speak from:

You have to get in a position where I am the spokesperson for [the ethnic group], speaking for their issues. That's the only option I can see. And how I'm going to get to that position, it's actually not easy.

It is not an option for her to participate in public debate unless she somehow has a mandate to represent her community, even though she thinks many in the community agree with her standpoints. Because she has felt the consequences of speaking against people she considered to be on her side, she is very cautious about what she says in public. Her story illustrates the fact that limitations on minorities' access to freedom of speech not only come from the majority society or the conditions for accessing the mediated public sphere. Internal social control mechanisms are also a factor, as people strive as much for acceptance in their own community as in public. Ascribed representation poses an obligation to represent a certain group in public, but it also creates an obligation towards the group, to represent them in a manner that they recognize and accept.

## **Discussion and conclusion**

As immigration and diversity have become an integral part of Norwegian society, and are no longer new phenomena, do ethnic and religious minorities still face specific barriers in accessing public debate? This chapter has explored the experiences and perspectives of individuals with an ethnic or religious minority background who are (potential) participants in public debate. While access to the media is not seen as a challenge in itself, the conditions for access, that frame who they are allowed to be in the public sphere, are considered to be more problematic. The interviewees experience public participation as coming with a risk of not only being ascribed a minority identity, but being fixed and reduced to that attribute only. They articulate a fear of becoming *the* Minority in Norwegian public debate, ascribed a position as a representative of 'their group'. Being ascribed group representation is seen as problematic both in what it communicates to the outside world (e.g. undermining their professional competence) and in what it communicates to the community they are supposedly representing (e.g. the legitimacy they have to do so). The analysis suggests several strategies to deal with ascribed representation: avoiding participation in the mediated public sphere, accepting the rules of the game and using one's minority status strategically to gain access, or challenging the conditions for media access by strictly avoiding minority-related topics.

When discussing ethnic and religious minorities' access to the mediated public sphere, it is important to keep in mind that only a small fraction of the general population ever participate in public debate. And those who do are subject to a news media logic that does not necessarily leave much room for nuance and complexity. Nevertheless, as Cottle (2000a p. 21) argues, the question is not whether the news media values are exclusive to reporting on minorities, because they clearly inform other news stories as well, but to what extent they figure disproportionately when minorities are involved. The experience that the media ascribes an identity and presents you as belonging to a group is not specific to ethnic or religious minorities (e.g. Eide, 2010a p. 75). However, not all identities become equally fixed on all individuals (cf. Skeggs, 2004). Ethnic and religious identities appear as identities that – perhaps to a greater extent than other identities – can reduce the individual to only that attribute, reflecting the brightness of ethnicity and religion as symbolic boundaries in contemporary Norway and Europe (e.g. Alba, 2005). Thus, ascribed group representation appears as a challenge particularly for minorities' participation in public debate. This suggests that symbolic boundaries, in terms of ethnic and religious distinctions, translate into social differences (e.g. Lamont & Molnár, 2002), in terms of differential access to public debate and to the exercise of freedom of speech.

While the media perform an important role in the public representation of symbolic group boundaries, and thus contribute to reinforcing such boundaries, it can also affirm diversity and provide important spaces in and through which imposed identities can be resisted, challenged and changed (Cottle, 2000a p. 2). This study, together with the literature review, suggests that the 'old story' of ascribed identities and minority-specific barriers for participation has not lost its relevance. At the same time, Norwegian media seem to become increasingly conscious of how to handle diversity, and the empirical investigation provides examples of how individuals challenge the conditions for access, and both expect and demand to participate in public debate on their own terms, in the form of individual representation (see also Midtbøen, 2016). However, in a democratic perspective, it is not only important that individuals with a minority background can participate in public debate, it is also vital that the perspectives they have as minorities are voiced. It is not only a matter of which individuals participate, but to what extent a diversity of experiences, perspectives and interests is represented in the debate. The challenge is therefore to create conditions that allow individuals with minority backgrounds to participate in public debate, not as representatives of a minority per se, but as individuals who have unique experiences and perspectives, shaped by their minority status.

### **References**


Barth, F. (1969). *Ethnic groups and boundaries: the social organization of culture difference*. Bergen: Universitetsforlaget.

Bleich, E., Stonebraker, H., Nisar, H., & Abdelhamid, R. (2015). Media Portrayals of Minorities: Muslims in British Newspaper Headlines, 2001–2012. *Journal of Ethnic and Migration Studies, 41*(6), 942–962.

Carpentier, N. (2011). *Media and participation: A site of ideologicaldemocratic struggle*. Bristol: Intellect.


Huntington, S. P. (1996). *The clash of civilizations and the remaking of world order*. New York: Simon & Schuster.


### Chapter 9

# Immigration critique: Moral boundaries, silence and polarization

*Kjersti Thorbjørnsrud, PhD, Senior Research Fellow, Institute for Social Research*

Building on theories of symbolic boundaries and the civil sphere, this chapter explores the limits of the Norwegian immigration debate seen from the perspectives of immigration critics. It asks if and why people subdue their views on immigration and immigration policies, and how opinions on immigration relate to moral stigma. The study is based on qualitative interviews with both informants who refrain from uttering their opinions in public and individuals who take an active part in the immigration debate. They all share stories of stigma and social exclusion, expressing the power of moral judgments on their willingness or refusal to express their opinions. Peer effects stand out as vital, and the closer one associates with or has relations with milieus associated with the liberal left, the more painful are accusations of immorality. The chapter finally relates these findings to processes of polarization and echo chambers.

## **Introduction**

The debate about immigration and integration looms large in contemporary society. It reflects conflicts over values, resources and an increasing cleavage between elites and the general populace (Freeman, Hansen, & Leal, 2013). In the wake of globalization, economic crises and recent unanticipated immigration flows (PEW 2016), the US and Europe have seen the rise of antiimmigration and anti-Muslim political movements, and negative attitudes to immigration are on the rise (IOM, 2015). Concomitantly, the worry of established political parties and liberal elites is growing, their mobilization against illiberal attitudes is intensified, followed by warnings that extreme views will gradually infest main stream debate in a manner that resembles a dark European past (Mudde, 2016). Others maintain that an inclusive debate with room for controversial statements is the best way to avoid that people, for fear of stigma, leave the public sphere and turn to closed groups of likeminded discussants. Such echo chambers could boost extremism in the absence of counter voices, the argument goes (Sunstein, 2003).

This chapter explores the effect of a polarized debate climate on people's willingness to express their views on immigration and integration. It studies the boundaries of immigration debates from the perspective of immigration critics; that is, people who are concerned over the perceived negative impact of immigration on society and oppose current immigration levels. The study departs from central findings in the comprehensive surveys of the *Status of freedom of speech in Norway* project from 2013 and 2016 (see introduction, this book). These surveys show that the issue of immigration touches a sensitive nerve in public debate. In general, many people hold back their opinions for fear of offending or hurting others, or to avoid ridicule and social isolation. The fear of being perceived as racist is one central factor that makes people stay silent about their views, and people who are negative to immigration are more prone to self-sensor to avoid social stigma than others (Steen-Johnsen & Enjolras, 2016; Steen-Johnsen, Fladmoe, & Midtbøen, 2016). Why is this, when negative attitudes to immigration are so common (IOM, 2015)? A tentative answer is that negative attitudes are widespread, but so is the disapproval of these attitudes in public debate. Opposition to immigration is associated with dubious and illegitimate positions. People with a liberal view on immigration and higher levels of education, among them journalists and those who associate with the liberal left, are more inclined to support restrictions on immigration critique in public debate than others (Steen-Johnsen & Enjolras, 2016; Steen-Johnsen et al., 2016). To get a deeper understanding of what types of moral stigma and self-censoring are related to negative views on immigration, the present study relies on qualitative interviews with informants who have experienced barriers and costs related to expressing such views.

With the growth of populist right wing movements that propagate anti-immigration ideologies at odds with the ground rules of constitutional liberal democracies, there has been a growing interest in studies of right-wing extremism (e.g. Hainsworth, 2016; Horgan, Altier, Shortland, & Taylor, 2016; Ivarsflaten & Stubager, 2012; Mudde, 2016). The present study follows a different path. Rather than researching the arguments at the extreme ends of the debate, the focus is rather on how a polarized debate climate affects more moderate immigration critics. In this context moderate is defined as respecting and identifying with the ground rules of democratic processes and debates. The study does not include individuals who operate outside or in a grey zone in relation to the law: Extremists with a racist ideology are not included, neither are people who defend

### chapter 9

undemocratic means, violent actions or hate speech. As such they represent viewpoints that, in a formal sense, are legitimate parts of democratic debate. Nevertheless, their perspectives are in the crossfire of struggles over which perspectives are morally acceptable and which arguments pose a danger to civil society, warranting collective condemnation.

The 14 in-depth interviews in this study represent a tiny first step towards an understanding of how opinions on immigration relate to deeper moral virtues and vices in the public sphere, and how these moral boundaries affect people's motivation to speak up. To get a glimpse of different types of barriers in the immigration debate, the interviews rely on two types of informants: people who in general refrain from, or to a limited degree, utter their opinions on immigration in public; and individuals who take an active part in the public immigration debate. Theoretically, the analysis builds on Jeffrey Alexander's theory of boundary formations in the civil sphere (2006) paired with theories of silencing and peer effects (Noelle-Neumann, 1974), outlined in the next section.

## **The moral boundaries of the immigration debate**

### Boundary formation in the civil sphere

The basic theoretical premise of this analysis is that debates over immigration take place within a normative framework of moral values. These are principles that take the form of *symbolic boundaries* that categorize people and practices; they separate people into groups and generate feelings of similarity and group membership (Lamont & Molnar, 2002). The theory of Jeffrey Alexander (2006) on the civil sphere, provides a scheme that captures how such boundary formations are tied to binary

immigration critique

values of right and wrong, good and bad, separating legitimate actors, relations and institutions from their uncivil counterparts in liberal societies.

Alexander defines the civil sphere as a moral community based on a shared set of universal values and institutions (the legal system, mass media, civil associations) (Alexander, 2006, p. 31). These are values inherited from a long history of Western philosophy, religious thought and political struggle, expressed in the founding documents of democratic societies, like laws, constitutions and bills of rights (Alexander, 2006, p. 60). According to Alexander, these constitutive values have complementary positive and negative values. On the positive side are the values of autonomy, reason and sanity, built on relations that are open, trusting, critical and truthful. Their complementary uncivil side subsumes dependence, irrationality and madness, based on secretive, suspicious, self-interested and deceitful relations. Civil institutions are defined by rule of law, equality and justice; their uncivil antidotes are hierarchic, arbitrary and based on personal power (Alexander, 2006, p. 57-59).

According to Alexander, these binary codes provide the structure for the everyday stories that guide taken-for-granted political life. Those who are considered worthy members of a civic community are defined in terms of the positive side of this symbolic set; those who are termed unworthy are defined in terms of the negative side. The positive side forms a *discourse of liberty*, the negative *a discourse of repression*.

The constructions of public virtue and public vice tend to be widely accepted even in societies characterized by high levels of conflict. What is contested is how the antithetical sides of this discourse will be applied to particular actors and groups. When defined in terms of the negative codes of the civil sphere, the deepest moral integrity and rationality of an actor or a movement are put into question. People judged to represent these public vices are regarded as profoundly threatening to the civil community, hence their activities, practices and opinions need – like a contagious disease – to be isolated, silenced, repressed or displaced.

It is vital to note that the theory of Alexander implies that the values of the civil sphere are never actually fulfilled in reality. They represent higher values, a *secular faith*. Real civil societies are contradictory and fragmented, created by social actors at a particular time in a particular place. Arbitrary qualities (e.g. gender, race, nationality) are transformed into necessary qualifications for inclusion in the civil sphere. It is a premise of the theory that the discourse of repression is extended to groups and persons whether they actually are 'really' evil or not. A central argument is, however, that insofar as the founding values of democratic societies are universalistic, they are open to inclusions of new groups and actors who can argue their way in as new members of the civil sphere based on a reference to the universal.

As hypothesized by Alexander, symbolic boundaries gain power when they are defined and maintained by elites, e.g. decision makers, intellectuals, media professionals and leaders of civil associations. Other theories of opinion formation add, that people are most receptive to the values and perspectives of peers, i.e. groups and persons an individual identifies and associates with. Individuals need the fellowship of others, and to be socially isolated because of deviant opinions is frightening for most people (Glynn, Hayes, & Shanahan, 1997). Based on this premise, the theory of Noelle-Neumann (1974) argues that people with minority views tend to hold back their opinions and adapt to a dominant climate of opinion. The propensity to defend unpopular standpoints is actually atypical, it is reserved for an 'Avant Garde': those few who create change by opposing consensus and tradition.

### The dual nature of immigration debates

The framework of Alexander captures the existential and dual nature of public debates concerned with the ground rules of inclusion and exclusion - like current immigration debates. These debates are based on a system of classification with a double face, expressing both the power and limits of the universal values of the civil sphere. The regulation of immigration is marked by a contradictory and ambiguous co-existence of idealism and realpolitik (Brochmann & Kjeldstadli, 2008, p. 16). Nation states monopolize, organize and distribute rights and duties, entitlement and responsibilities based on national membership (Tilly, 1998). At the same time, nation states based on constitutional democracy, adhere to and are limited by universal values of human rights, individual freedom and equality before the law, expressed in international conventions and national constitutions. Their universal rhetoric might conflate the divide between in-groups and out-groups based on nationality in the current world order. They do nevertheless make a difference. The asylum principle, immigrants' claims and minority rights are defended with reference to these higher moral principles (Borchgrevink, 1999; Brochmann, 2002; Vertovec, 2011). They regulate public debate in the sense that debaters who wish to take part in the mainstream, democratic public sphere cannot ignore them.

Debates about the scale of immigration, the closure of borders, the limits of tolerance for differences, and inequality take place in a climate where contenders on both sides depict each other as threats to the very existence of civil society as they know it. This debate then, while discussing the principles for inclusion and exclusion of groups with geographically and culturally foreign origins, at the same time defines who are moral insiders and outsiders *within* the national community. The subsequent analysis of the experiences of immigration critics, explores how these informants react to an ascribed status as uncivil outsiders with illegitimate and potentially dangerous and contaminating views.

## **Design: Informants and interviews**

The informants in this study shared a deep worry over the consequences of immigration to Norway and Europe. They were critical towards current immigration policies and concerned over the perceived lack of successful integration of immigrants. They defended more restrictive policies, ranging from total border closure to a reform of the asylum system and more active integration policies. They were, in particular, critical to the scale and consequences of non-Western immigration. Most informants referred to the negative influence of conservative religious practices connected to Islam, particularly related to gender. Others focused on the assumed negative consequences for the welfare state, pointing to the challenges following low-skilled immigrants from clan based societies. The prospect of increased social instability, insecurity, violence and crime were often mentioned. All informants considered the media coverage of issues related to immigration and integration as severely deficient, and pointed to a muting of vital information and voices in the public.

The informants were recruited through different approaches during the year 2016. A few were contacted through personal networks. They were asked if they knew potential informants who had experienced some type of barrier or cost related to the expression of their views. Through them, new informants were recruited. Additionally, I monitored Facebook discussions on immigration, and got a good overview of different levels of engagement. New informants were contacted based on their activities there, often in the form of a personal message. A last couple of informants were recruited to represent the front players in this debate; they were contacted directly with reference to their public role.

The informants were selected to represent gender (5 women, 9 men) and age variation and different types of occupation in the public and private sectors. Their education varied from low to high. Most of the informants voted for the party with the most restrictive immigration policy, the Progress Party; others did not have a clear party affiliation or voted for parties on the moderate left or center of the political spectrum. Most importantly, the informants were selected to represent both people who were reluctant to share their views on immigration in public, and main actors in the public debate.

The informants can be divided into different groups based on their participation or lack thereof in open debates. These groups range from those who refrained from uttering their opinions in most forums (4), to those who uttered their opinions in social media only (2), individuals who occasionally entered public debate (6), and finally, full time debaters in public forums (2). The most active debaters tended to have expert skills relating to the media and professional communication. The interviews were based on a semi-structured interview guide, conducted face-to-face, and lasted from 60 to 90 minutes. All interviews were transcribed and analyzed through qualitative analysis software. For many informants, full anonymity was a premise for their participation, and was fully secured. For those who are front players in public debate, full anonymity was not a prerequisite. Their stories might make them recognizable to some readers, even if personal details are omitted.

The aim of this study is to understand the lifeworld of the participants and to give them a voice in a non-judgmental way. This approach does not imply the absence of critical questions. Taking people seriously, involves challenging their views by chapter 9

probing the implications of their opinions. Their stories are told through extensive quotes that allow their own expressions to come forward. This approach, even if common in qualitative studies, is not often used vis-à-vis the group of people in focus here (but see the recent and extensive fieldwork of Arlie Hochschild on the Tea Party Movement (2016)).

# **Self-censoring and stigma in the immigration debate**

The informants in this study had in various ways experienced the social costs of uttering criticism related to immigration and integration. However, they chose different strategies to tackle them, ranging from choosing not to discuss the issue to taking part as full time information providers and opinion leaders. In the subsequent analysis they are grouped along this passiveactive dimension, involving the *silent*, those going *semi-public* in social media, and finally the actors that engage part time or full time in the *public* media debate.

### The silent

The group of informants presented in the following, largely remain silent about their views in public. Two of them are women living in Oslo in areas with many immigrants, with children in schools having a large proportion of pupils with Muslim backgrounds. They describe themselves as initially positive to their diverse neighborhoods. But, gradually, experiences of what they regard as repressive Muslim gender practices and religiously based in-group loyalty changed their minds. One, a librarian, describes herself as a 'dedicated atheist', against any type of increased religious influence in society. In recent elections she voted for the Progress Party, not because she agrees with their ideology in general, but because of their position on the immigration issue. She gives this account of why she believes current immigration patterns are problematic:

Basically, I know about it through my kids: they go to a really multicultural school. It's things like kids not coming to birthday parties, single-sex swimming lessons, and girls not being allowed to go on school trips and take part in the social life outside school. There are immigrants from all over the world, and it's fine, mostly. But those with Muslim backgrounds have problems. They're the ones who aren't allowed to take part.

This informant explains that she does not have anything against particular individuals, it is the overall influence of Muslim norms on society and what she sees as a changed social environment for girls that concerns her. She believes that gender equality should be a core focus in school, rather than what, in her opinion, is an exaggerated focus on religious feelings. When asked what development she fears most she answers:

I fear a development where the control of girls increases. I see it in our neighborhood, you do not show your belly or wear short skirts on the street, and if you do, you can blame yourself for any unwanted attention.

The other informant works in public administration and has been active in local politics in the social democratic party. For her, it does not really feel as if immigrant groups represent a vulnerable minority and that she herself belongs to the dominant majority. Rather, she sees the rise of a Muslim identity as a reversed form of othering at the expense of those who are not part of the Muslim community. Aware that she will easily be chapter 9

judged as too overly generalizing, she hesitates often, stops herself and starts again:

In the last few years this religion thing has gotten much stronger. Before, it was like Norwegian-Pakistanis or Norwegian-Iranians, or... now, it's like, "We're Muslims". And I see it in our school as well, from an early age. They're brothers and sisters, you know? And those who aren't Muslims, well they're not brothers and sisters. There is something a bit unsettling about it. It becomes a way of excluding. And especially this thing with girls, the views on women. That's what you notice the most. Covering up girls, with scarves terribly early. It becomes a marker that, yeah, we're different. Equality and gender, those are really important values, but it's not like they are carved in stone.

Neither of the two women find that they can discuss their concerns about a changed local community freely, whether it be in their neighborhood, in meetings at their children's schools or in their workplaces. They feel that to be considered legitimate, criticism must be directed at the conventional majority. 'It is always the same bias. It is all about hate speech from Norwegians. It worries me when legitimate criticism is defined as hate speech,' explains one. She points to how many families in the neighborhood avoid the nearest school because of a high share of immigrant boys. This is not discussed in the open: Families make their decisions in private, but in public 'everything is fine' she explains. Stories of kids being bullied because they are white non-Muslims are kept secret, or only mentioned in private. She feels alone with these experiences in many settings, like her workplace, she explains:

At work no one has kids in a school with such a high proportion of minorities as I do. Nonetheless, they have no interest whatsoever in hearing about it. They are very politically correct, some of them. I know about all these stories that were really bad. But no one cares, because it's bad on the wrong side. And I said that to my boss once. And he says, well, after all, we are the majority. So I said, but when you go to our school, you don't think of yourself as a majority.

The other informant explains that if she talks openly about her views, she is met with 'embarrassed silence' even if some 'whisper to me that, actually they agree quite a bit'. In school, her disapproval of gender segregated activities like swimming has been met with ridicule and lifted eyebrows from the principal. In general she feels that people avoid issues related to religious suppression and Islam. She gives this example:

This weekend, I shared a status on Facebook about enjoying a glass of wine and eating cashew nuts. It got lots of likes. And then I shared a link about the fight against circumcision of girls. Then it is all silent. Nothing. It is telling I think. I mean, no one is really *for* circumcision. But it is like they don't want to touch it.

It is the fear of being perceived to overly generalize, to seem prejudiced, or to be looked at as an outright racist that keeps these informants from uttering their thoughts in different forums. This is a type of subtle stigma that works through silence more than outspoken counter arguments. It is communicated through evasive body language, downcast eyes, uneasy laughter or simply silence.

A male informant shares the concern over an increasingly segregated society with the female informants presented above. But in contrast to them, he lives in a white middle class environment in the western part of Oslo, and is part of a milieu of media professionals and creative professions. He sees himself as open minded and individualistic, without strong ties to any political side. He expresses his point of view in the following:

Norway has been so uniform - culturally speaking. It felt very safe and then there were loads of reasons to make fun of it as well. A bit stuffy... boring. But anyway, I think that a successful society is a society where the vast majority accepts a few basic principles and where there is a good mix across ethnicities, where within a generation immigrants have Norwegian boyfriends and girlfriends. But now it is pretty watertight. I fear segregation and a class based society.

Working as a freelancer in the media, he does not share these views in public. He can discuss with friends, but finds that he is quickly placed to the far right politically. To speak openly about how he looks at the influence of Muslim immigration, involves being associated with attitudes he feels no familiarity with. Politically it means to be grouped with the extreme right; personally it means that you have a callous racist personality. He explains:

You can't say anything without being branded. You have to, like, make 500 qualifying statements if you just want to say how things are. I find it so much easier to be grouped with left wing people, then you can be ridiculed as nice and naive, that's the worst that can happen to you. It's far worse to be stigmatized as racist and evil.

He follows discussions on immigration in social media, but never gives his own opinion, even if he gladly discusses other political questions. He is also very careful not to 'like' anything from profiled immigration critics on Facebook, even if he agrees with them. He says it is a question of social stigma, but also about fear of losing his job:

I don't have a permanent job. Workwise, it can be risky, someone could report me, you never know what might happen. "Do you know what he believes?" That sort of thing. Media companies want to protect their reputations: you have to be very careful in that branch.

Like other informants in this study he argues that the issue of immigration policy is so delicate because it is intrinsically linked to morality and humanity, a complex policy field is reduced to a good or bad side, he asserts.

The young feminist activist presented in the following shares the frustration over a perceived dominance of feelings rather than rationality in the debate. She has experienced what those who hold back their opinions fear: In a post on Facebook she referred to an incident involving the harassment of women by male asylum seekers. As a feminist she has been fighting many types of oppression of women; she has been offended by adversaries, but has always been supported by her own peers. This time, it was all different she tells:

And so I write that it's enough now. Women are unsafe enough, with Norwegian men. We should not import even more abusers who can treat them like dirt. And people are just... Oh, my God. Everyone at my university course was just, ah "racist"! In and of itself, being in favor of a restrictive immigration policy, which I haven't been until now, is not the same as being a racist. And it just amazes me how everyone is willing to sacrifice the struggle for women's rights in the fight against racism. I mean, people in the feminist movement come up and say, "You shouldn't say that because you are paving the way for fascists".

To her mind, religion is intimately related to discrimination of women; Islam, like Christianity and Hinduism, is 'hatred of women embodied' she declares. She refers to herself as someone who in general speaks out about any type of subject, but not on this issue. In social media there are posts she would like to share, but she stops herself. Disappointed by the Women's Movement she has been a part of, she feels abandoned and has resigned from organizational duties. She has experienced assaults, and even threats from Norwegian men after demonstrations against repression of women. In a certain way, that actually gave her some credit she explains, it was a sign that she had done something right. But the lack of support from her own group hurts much more than harassment from angry men.

### In the semipublic space of social media

As opposed to the young feminist who experienced a 'one time moment of public shame' and went silent afterwards, the group of informants presented in the following regularly use social media to discuss immigration. But they keep a low profile with family and friends, and are not active debaters in mainstream media. One of them was very much in doubt as to whether it was a good idea to participate in this study, his wife told him not to. The reason is that he runs a small firm, and feels vulnerable: He cannot risk losing customers because of his opinions. He is an active debater on Facebook, but does not in general share his views with his old friends in person. Apart from the concrete economic risk associated with going public, he believes that what he calls the risk of 'intellectual murder' stops people from speaking their mind. In his opinion there are people who act as consensus guardians, who attack the few who dare to speak up:

They signal to all of society that it if you say this, then it will cost you a bloody lot. Like taking quotes out of context and sending them to your aunts and grandmother, and anything that will do maximum damage.

He, like the other informants, feels that there is no 'ceiling of blunders', if you have uttered something that can be used against you, it will always stick to you he claims:

You must be allowed to say something stupid. You must be allowed to share something, whatever, conspiratorial, and then say, "Yeah, those are good counter arguments, I was convinced, but I don't believe in what I said anymore'. But that's not the way it is. There's no undo button.

He is familiar with arguments implying that his views are connected to Nazism and fascism, but says he chooses to confront these types of allegations up front. He refers to debates during the refugee crisis in the autumn of 2015 in the following:

In that period Nazi references were rife from those who thought that with every tightening of asylum policy we were well on our way to a holocaust. But there are very few who say it right out. They say something like, "Yeah, we haven't seen that policy since the 1930s". "Yeah, if you mean Nazi then say Nazi," I write.

This informant, with a background in technology, has had many positions of trust through his work but has never been active in politics. He has voted for the Norwegian Communist Party 'to get some critical voices into parliament,' but now votes for the Progress Party. In his view, to be an immigrant is a much tougher destiny than people are aware of, and he fears what he calls a 'client state' where many asylum seekers never succeed in taking an active part in society through work:

I don't think that people, as individuals or a group or a race or nationality, are lazy or useless or sly or anything. I just think it is really tiresome and difficult. We know that half of those who come here will never get a real job. And I think any realism about the whole thing is just totally lacking.

This informant expresses a feeling of deep unrest. He was brought up with a belief in the United Nations and the fraternity of people, but has lost faith in international organizations. He fears the breakdown of a generous welfare state, and a society characterized by ethnic and religious conflict. His conclusion is that immigration must stop altogether. He always expresses his views politely he says, but is scared by this development; in his view, those with power do not seem to listen or understand.

Another informant engaged in the semipublic sphere of social media. He has lived for years in the eastern part of Oslo, with chapter 9

kids in a school dominated by minorities. He does not fear Islam so much; rather he is worried about immigration from what he calls dysfunctional states where kinship is central to social security. He fears that Norwegian society will not live up to the challenges, and does not take the welfare state for granted. Having seen how it is to live without one, he explains:

I really do believe that immigration in many ways has been a good thing. But at the same time I see many challenges in the wake of the arrival of people from very different cultures. Integration is not straightforward. That is why I support very restrictive immigration policies. I have travelled a lot, met many great people. But meetings between cultures are not always easy. "Norwegian" has for instance become an insult among groups of immigrants in Norway. Like, "You have become too Norwegian." We didn't think it would be like that when the first Pakistanis came to Norway in the 70s.

He underscores that he has not in any sense been threatened due to his opinions, and will not appear as a victim, but nevertheless expresses sadness over the social exclusion that follows from perspectives like his. He has mainly used social media as the arena where he discusses issues related to immigration policies. It has been, and still is, a disagreeable experience, he conveys:

It's not okay at all. A colleague said he couldn't work at the same school as me, he considered resigning. It happened in a discussion on Facebook.

The more typical reaction however is no response at all. He meets silence from former friends and fellow discussants more than counter arguments, he tells:

I have lots of friends on Facebook - very few have unfriended me, but they never like anything I post and they never comment on my posts anymore. It's dead, it's sad. And I should add that I have never said anything racist, nothing disparaging about other cultures, never said a word about Islam. But, it's just that if you are for a restrictive immigration policy, that's enough in itself. Some people who I know pretty well, I never see anymore. I never hear from them, there is quite a lot of that.

He has many contacts among journalists and in academia. He describes this milieu as avoidant, immigration is a 'none issue' all together. He adds, 'And really – where they live they don't experience the consequences of immigration – or if they do only the positive effects – exotic restaurants, cheap labor'.

## Part time on the public stage of the immigration debate

The next group of informants are engaged in the public debate about immigration through professional or semiprofessional writing, in the mainstream media and the editing or writing of books. For them, the perceived lack of transparence in the debate is a main motivator. These informants have close connections to a milieu dominated by media professionals and academics, associated with broadly leftwing liberal attitudes to immigration. On the one hand this makes them vulnerable to condemnations from this group. On the other hand they are angered by the alleged hesitancy of the political left to criticize illiberal movements and practices when ethnic minorities and immigrants are involved. Their professional position to some extent makes them protected from economic repercussions; they have multiple professional identities and competencies. However, in spite of their public profile as critics of immigration policies they, like the informants outside the public spotlight, find the issue of immigration too troubling and delicate to be suitable as a subject over lunch or in conversations with neighbors. A prerequisite for their engagement is the support, or at least the absence of disapproval, from their own family.

One of these informants has a background in business and academia, and writes regularly for selected newspapers. She is often involved in debates on controversial issues, and often takes what she calls 'super contrary' positions. She is used to public scolding but has also been given credit for her alternative perspectives. When she conveys her analysis on immigration and integration however, it feels very different, she recounts:

If I say something about immigration, it's like I've got some disease or something that makes people avoid me. Someone wrote on Twitter that he had long had a suspicion about what I was like as a person now he had it confirmed. There hasn't been much of that - but what there is hurts terribly. Some crazy person on Facebook said I should think about my responsibility when kids in asylum centers burn to death. I saw who had liked it, and there were many well-known people within culture and the media. I take that sort of thing very hard.

Like the young feminist, it is not the moral condemnation so much in itself, but whom it comes from that is painful. She explains that these experiences make her strategic and careful, without silencing her completely – she considers the issue far too important for that. But she intentionally avoids writing about immigration too often – a couple of times each year is the maximum. And she is careful to make her arguments as acceptable as possible. She criticizes, for instance, the Progress Party when she finds that their arguments lack statistical underpinnings, and she always refers to minority voices to provide examples of successful integration: 'It's important to me to lift the good voices that exist, strategically, so I'm not called a racist, but also because integration is vital. I want things to go well for those who are here,' she explains. But sometimes she feels dishonest. As an example she mentions that she was a source in a news story where she agreed with a claim that the tables are turned now, it is no longer a problem to support opponents of Muslim practices and vigilant critics of immigration policies in public. She elaborates:

But it was a lie! I said it for two reasons: Firstly, to be optimistic. But mostly to be accepted by the left. And I wanted to puke when I said it. For the thing is, to support these actors in public actually gave me lots of problems. Like, my boss, who otherwise is a wonderful person, told me she thought it must be difficult for Muslim students to have a lecturer like me. And what she is really saying by that is: my job is at risk. But I have *never* talked about these things with my students, I am professional. If we discuss discrimination, hijabs, it is strictly pros and cons. It is not about my personal views at all.

Another informant works as a journalist in a niche newspaper. He points to his multi-professional background as an important premise for his critical pieces on the economic consequences of immigration: His journalist identity is not all that important to him he claims. He refers to statistical analyses of population growth and migration trends as decisive for his position. In his view, disinformation and intended lack of openness about statistical facts, be it from top politicians, researchers or the Norwegian Census Bureau make it all the more worthwhile to do the necessary research and math himself:

If you think that the best thing for, for example, Somalis is to bring them to Norway, then do it. But don't come to me and say that it is so terribly profitable. There are some very worrying trends when you look at employment statistics, which are very low for non-Westerners and even lower if you group them by Islamic countries, and that is what we have done in my newspaper. That gets people really worked up, right? But as we say, it would be unfair to Tamils not to group them. Because, after all, they work an incredible amount. And we did it a bit just for the pure hell of it. When people tell me that I can't write something, it makes me sort of angry and "yeah, we'll see about that" sort of.

He has received many reactions to his reports, from the top levels in the Norwegian government to critical colleagues and economists. Accusations of fascism and Nazism are familiar to him, even threats to his personal safety have occurred. In the beginning, the reactions from his own colleagues were strong too, they found his focus inappropriate and indecent. He was also criticized for the absence of *cases* – of stories of individual immigrants in his reports. He elaborates:

Journalism must have a case, right? But it is obvious that you just find the positive cases. And so I've always said that, if these are the statistics, then I'm not going to use a positive case. And it would be totally unethical to hang an individual out to dry who represents those who have failed.

He does not agree with the notion that discussions of the negative effects of immigration might lead to prejudice and a more polarized society. Rather he refers to free debate as the founding principal of open societies.

It is a kind of banal post-modern theory that I despise, the idea that words are actions and that as long as we don't talk about things then everything will be fine, right? The entire West is based on the idea that we talk about things. That is what an open society means.

Like other part timers in the immigration debate, he has some strategies when he writes about immigration. He is careful not to do it too often, and he takes care not to be obsessed with the topic, a type of monomania he thinks characterizes some who engage in immigration critique.

The informants in this group of public debaters, all point to a fascination with the unsayable. They understand themselves as the one who sits in the back of the classroom, raises their hand and says what others might be thinking but do not dare to express. An example is the informant presented in the following, who is relatively new on the stage of the immigration debate, but has a status as an enfant terrible in general public debate. Like the other part time debaters, he has an independent position, with several sources of income. He has an academic degree, 'but is not a face in the corridors of the university'. He contends that until recently, the economic cost of immigration has been under communicated, the threat from Islam as an ideology has been underestimated, and the breaks with basic rights to freedom within minority groups neglected. His public engagement in these issues started when he defended a controversial Norwegian Islam critic and activist in public:

My point was that the criticism against her was completely exaggerated and unfair. And symptomatic for a perspective that perceives immigration critics or Islam critics to be a bigger problem than the Islamists themselves.

The response was massive. There was a pile of emails and personal messages from people who thanked him for saying aloud what they were thinking, some of them in academia, some related to the political left. The *public* response from these milieus however, was shocked disapproval. He became the new representative of the 'dark side' as he calls it, condemned for instigating prejudice, exclusion, hatred and even violence against Muslims. He asserts that he regularly meets invalid arguments, of the type 'ad hominem', guilt by association and straw man:

You always have to explain that no, I never said that. No, this is not correct. I have never criticized 1.5 billion Muslims and so on. You chapter 9

have to repeat it again and again. And sometimes you just don't have the energy.

He refers to the stigma of being associated with the wrong sources and actors, independent of what these sources are actually arguing for in a particular text. They have a status of being contaminated – and thus contaminate those who refer to them, he explains. He has himself become a person whose postings others hesitate before they like or share in social media, and even he hesitates before sharing articles from actors defined as illegitimate. Like other informants, he points to the role of emotions and morals as important in explaining why the debate is experienced as being so sensitive. He believes attitudes to immigration define not so much who people are, but who they are *not*:

If you have higher education, like from the social sciences or the humanities, and identify with the broader left, then it goes without saying that you are *not* against immigration. We might not be able to define who we are, but at least we know who we are not. Even if you have this awareness that, hell, things are not as simple as I thought, things are going in the wrong direction…Even then, it takes a lot to make concessions to the dark side.

### Full time in the immigration debate

The last two informants presented in this study, have very different backgrounds, but are both more or less occupied full time with issues related to immigration and Islam critique. Due to their public roles in the immigration debate, they find that their career choices and professional opportunities have become limited. One is an intellectual writer and former editor. His story starts with his personal confrontation with the orthodox Christianity of his childhood. Gradually he became aware of the spread and power of an orthodox form of Islam, fueled by key events like the fatwa against Salman Rushdie and Islamist terror attacks. 'From criticizing the oppression of Christianity, I simply moved on to criticizing the oppression of Islam, just more strongly,' he states. When he was a professional editor he published a book on the issue. It was the start of what he calls a big 'social fall':

I didn't know it was that bad. I could hold a contrarian stance and so on before, but people didn't link it to my moral character. There was no stigma linked to my viewpoints. I enjoyed great respect and recognition. And I lost that in extensive parts of the milieu I was a part of.

Coming from a well-connected position in a network of authors and editors, he now has no formal professional position. Gradually more isolated from his old network, he has intensified his own writing on the negative impact of Muslim orthodoxy and cultural segregation. He writes on these issues daily in social media, with a large group of followers, and has written a book on the issue. It frustrates him deeply when he is associated with right wing extremism and totalitarian ideologies. He makes many references to a Western philosophical canon, describing his engagement as part of a long critical discourse. Like other informants he finds what he calls the sentimentalization of public debate as a vital barrier to rational deliberation. The importance paid to protecting feelings stops viable arguments he contends:

There has arisen a sentimentalization of new groups in society to whom you are not allowed to apply normal critical sense. I did it anyway, and I crossed some sort of decency line. What was or had been normal debate was suddenly subject to loads of sort of moral responsibility norms that are pretty alien to a Western culture.

Immigration policies are related to morality in a very different manner than other policy areas in his view:

And if you don't manage to deliver the phrases that save your skin, you can very quickly end up in total darkness. I don't want to overdo it, but I think it has cost quite a lot, including friendships and not least acquaintances.

Debaters with a liberal approach to immigration, some with influential positions in the Norwegian or Swedish debates, have characterized him as 'brown', connoting Nazism. They might not be that many, but he feels it deeply when no one comes to his support:

These are people who express themselves from a very superior position, and even though I believe that many think "those were curiously harsh words", there are in fact very few who go in and show solidarity with those who are picked out as right-wing extremists or nationalistic.

Turning from academia, the informant presented in the following is not connected to Norwegian cultural elites. When entering the immigration debate as a writer and editor of an internet site dedicated in full to immigration and Islam critique 10 years ago, she 'came from nowhere, with nothing to lose,' she claims. That said, in a passing remark, she mentions that when she recently changed her occupation, she did not really have the opportunity to opt for an ordinary job anymore. She has a diverse professional background and was part of the anti-racist movement in her youth. She votes for the social democratic party, but insists that the established parties have lost contact with ordinary people and lack the ability to take their opinions seriously, in particular with regard to immigration policies. After several years working full time with what other interviewees describe as a 'contaminated' source of information, she is used to harassment and assaulting emails late at night. She elaborates:

Much of what I receive would probably have frightened me ten years ago. I often get it in the neck because I'm not educated, there is a lot of that. Like, "Why in the world should we listen to this type of idiot, right? You don't even have an education". And there are a few who say something like, "You better watch out", but I have a big dog and my husband's from the country, so we'll be fine.

She tells however about a traumatic time in the aftermath of the Oslo terror attacks July 22, 2011. The terrorist was a male ethnic Norwegian with an extreme anti Islam right wing ideology. Her milieu soon came into the spotlight, partly as a possible network for the perpetrator, partly by being blamed for his extreme ideas. Stressing that there are others who are the real victims here, she recounts:

It was July 22nd and when you are sitting there and are a totally unknown writer, and then the day after you are suddenly having problems getting the BBC not to come and film your house, right? They called my kids' mobile phones, and it was... I think I got 54 calls before noon... It was like the world had just torn down a wall and came crashing in. I still feel a bit unwell when I think about it. Obviously, when you are defined as sort of "insulation" for this terrorist, it's catastrophic really. And being hung out to dry as a racist and Nazi and God knows what. I received lots of threats and had to contact the police. I had kids who were suddenly not welcome in their classmates' homes. But after that, after we got through that somehow or other, then it would take a lot.

People who have worked close to her have suffered much more in the wake of Nazist, racist and fascist characteristics, she explains. Some have lost their job, friends and public reputation. Some live with protection from the police. She herself has been advised to no longer have a public phone number. But as opposed to friends and colleagues, she started out with no connections to Norwegian academia or the journalistic milieu.

If you are in this journalist environment and that's where your friends are, it's difficult. I have seen how much it pains them when they walk into a room where there are people who they have perhaps known all their lives and worked with, and people turn their backs. If someone turns their back on me, I just think, that's because they don't know me, right?

She adds that even if she works full time with these issues, she is not engaged 24/7. Home, family and friends are another place, she does not bring her job with her.

### Polarization, isolation and echo chambers

The participants in the Norwegian immigration debate are used to characterizations that position them on the wrong side of the border between the civil and the uncivil, the good and the bad, in the public sphere. At the same time, they are involved in boundary work to separate their own position from perspectives they themselves find illegitimate, facing movements and arguments that scare them. The editor presented above is well aware of the presence of extreme attitudes directed towards individual immigrants – and Muslims as persons. She distinguishes her own position by pointing to levels of generalization she finds unacceptable, and denounces any type of conspiracy theory. She works together with people of Muslim background, and it is important to her to communicate that they are as varied as anyone else. It worries her when people do not see the individual, but only threats and enemies. She elaborates:

It's fine to say what the heck you want about Islam, because it's a religion. But you can't call Muslims a cancer. It's this dehumanization that lies beneath what they are doing. And that's not okay. We have premoderated comments, but especially when there has been a terrorist attack, we have to delete seventy percent, they just can't go out.

I: What do you delete then?

P: Well, it's that sort of, right "Chase them out and set fire to them," people go completely crazy. I have noticed that before it mostly came from anonymous people. Now, you often see people writing stuff using their real name, and that is a pretty new phenomenon.

We believe that those who want to contribute to a better immigration debate cannot just go ahead and shriek like that. But then they think that it's censorship. And I answer, yeah, because clearly I want Fredrik1234 to write really racist things that I, under my full name, have to take responsibility for, yeah?

Informants like her fear a development in which right-wing extremism, violence and racism grow. But they wholeheartedly believe in an inclusive debate. They do not accept the argument that criticism of immigration policies, religious practices or cultural norms leads to racism, violence and extremism. Rather, they believe that silence and silencing are the main conduits to right wing populism and extremism, expressed by this informant:

I believe that it is those who paint a rosy picture, who deny reality who spur right-wing populism, by branding those who are skeptical, putting very ugly labels on them. That, with good reason, makes good people frustrated, sad and angry. Not least angry. I understand that anger, I have felt it myself to some degree. In my opinion it's quite the contrary. It's fact-oriented criticism built on liberal values that restrains right-wing populism.

Other interviewees describe how people who engage in the immigration debate experience heavy stigma and high costs, and cannot avoid being affected by it. Processes of exclusion and the formation of likeminded groups might push people towards less flexible and pragmatic views. They see a tendency where people in the absence of inclusive debate forums, gather together with likeminded people. On the one hand they refer to it as a boost to finally be able to discuss the topic. On the other hand, processes of reinforcement might lead to a type of monomaniac absorption in the issue.

Some see an opening up of the debate in the wake of the migration crisis in 2015; more people are engaged and it is easier than before to present critical analysis. Others describe the debates in social media as increasingly similar to echo chambers where opinion leaders are surrounded by fan groups. A paradox arises: debaters might find themselves in a situation where people they would like to discuss with abandon them, while they themselves back away from the embrace of people with a racist or conspiratorial worldview.

# **Concluding discussion**

This chapter has explored the moral boundaries (Lamont et al., 2002) of the immigration debate seen from the perspectives of immigration critics. The analysis reveals that informants relate themselves, their values and arguments to key values of the civil sphere to explain who they are and why they think and feel as they do. In line with the theory of the binary principles of the civil sphere by Jeffrey Alexander (2006), they express how their positions are deemed uncivil, evil, immoral and potentially dangerous by counterparts in public debate. Informants refer to the elites of the civil sphere, like intellectuals, journalists, and leaders of civil organizations, as key players in this boundary work, creating a climate of opinion where the fear of social exclusion and moral condemnation stops people from conveying their opinions openly. Echoing the central concepts of Alexander's theory, they describe how some actors and sources are depicted as contagious and untouchable, in which a mere association with them is enough to be compromised in the public sphere.

Immigration critics are often related to a type of populism where emotion rather than reason, dramaturgy and rhetoric rather than facts and arguments, authoritarian rather than liberal values are defining characteristics (Muller, 2016). Indeed Alexander's theory stipulates that affect and frenzy are the uncivil antidote to the civil values of rationality and calmness. Interestingly, the informants in this study do not concur with this type of psychological diagnosis. Instead they reverse the binary classifications by referring to their own values and virtues as those based on openness, reason, individual freedom and truth. Further, they point to the liberal 'good side' of immigration debates as the uncivil, describing *their* arguments as irrational, emotional, secretive, and dishonest. Those informants who actively engage in the public debate on immigration share some traits worth noting. They have in common that they regard themselves as opponents and dissenters with an inclination to go against the common crowd.

While grounding their views in the core of classic values of Western liberal societies, informants in this study *do* criticize a cosmopolitan ideal of diversity and tolerance related to relativism (Calhoun, 2008). Instead they defend the primacy of a type of secular individualism and lifestyle they feel the need to protect against norms based on religion and traditional culture. These perspectives position them in the heart of current negotiations over where to strike the right balance between the assimilative forces of universal values versus the value of diversity and acceptance of difference (Haidt, 2012).

This study has explored the experiences of people with very different connections to public debate, from those who refrain from conveying their views openly to front players in the current Norwegian immigration debate. It is worth noting that they all describe the threat of social exclusion from people one identify with as the main factor that leads to withdrawal from a forum of discussion, be it public or private. This is a type of peer effect (Glynn et al., 1997) that warrants more research: The stigma associated with immigration critique seems to be stronger the closer people identify with a broad liberal and leftist mindset, a phenomenon also discussed in Midtbøen, Ch. 7. This finding points to an important implication. Processes of silencing and repression might prevent some actors from entering public debates, while representatives of movements further away from liberal values and the established public sphere, might grow and dominate the debate untouched by the constraints of a liberal discourse turned repressive.

One could call it an irony that those who identify the most with the normative canon, the positive side of the codes of Alexander, might be the ones who find it most troubling to convey their opinions on immigration policies and integration regimes. To be defined as, or even associated with, the 'bad' side of the civil sphere involves social sanctions that cannot but influence – or change those who experience it. Bitterness, feelings of alienation and a search for alternative support that again strengthen a feeling of alienation from their usual crowd might ensue. The consequence could be a debate climate that silences the more nuanced, principled and reflective critical voices, resulting in a polarized and one-sided debate.

### **References**


### Chapter 10

# Boundary work in the public sphere

*Bernard Enjolras, PhD, Research Professor, Institute for social research*

Contentions about freedom of speech aim at the boundaries of this freedom, not its core. The objective of this chapter is to recast and interpret the findings of the preceding chapters within a theoretical framework, combining the insights of two separate fields of scholarship: the sociology of the public sphere and the sociology of social boundaries. This chapter develops an understanding of the public sphere as a social sphere, being both a sphere of cultural and symbolic integration, as well as of conflict and power struggle. It emphasizes the need to extend our understanding of the public sphere beyond its role as a space for rational discussion and deliberative politics. It continues by spelling out the criteria that an extended concept of the public sphere should meet. The chapter gives an interpretation, in terms of symbolic boundary-making processes, of the public debates related to immigration and freedom of speech in Norway. Public debates about freedom of speech are concerned not only with the limits of freedom of speech, but also with the symbolic recognition and integration of identity groups. Both types of boundaries (of freedom of speech and identity groups) can be understood as a power struggle for the position of these identity groups in the political community. What is at stake in these debates is the inclusion or exclusion of different identities in a multicultural society. The social definition of these symbolic boundaries impacts society's 'moral order' and society's social stratification.

## **Introduction**

In democratic polities the public and scholarly discourse on freedom of speech does not tend to focus on the core aspects of this freedom, but on its boundaries, on the borderline cases where disagreements come to the fore about where the limits of free speech must be drawn. Most of the literature on freedom of speech has traditionally been of a legal and philosophical nature (being mainly the work of philosophers of law, political theorists, and constitutional lawyers), and has emanated from the need to justify the principle of freedom of speech and to delineate its content and limits. In contrast to this tradition, a sociological perspective on freedom of speech is not concerned with elaborating normative principles for justifying the right of free speech or for limiting this right. It is concerned with the social practices of expression in public, i.e. the ways social, cultural, and institutional processes and structures *de facto*  enable and limit the exercise of free speech. It also focuses on the social, cultural and institutional stakes, and the roles played in debates and discourses on freedom of speech.

The focus of this book has been the making of symbolic boundaries in public debates, with an emphasis on the public debates related to freedom of speech and immigration in Norway. Insofar as the right to freedom of speech would not have any concrete existence without the presence of a social space in which free speech is made public, making sense of these debates entails situating them within a broader understanding of the public sphere as a space where symbolic struggles and processes of symbolic integration and exclusion take place. From this viewpoint, the public sphere is considered to be a space for the affirmation and contestation of society's moral order, and not only as a space of rational discourse and deliberative politics.

As stated above, the objective of this concluding chapter is to recast and interpret the findings of the preceding chapters within a theoretical framework combining the insights of these two separate fields of scholarship: the sociology of the public sphere and the sociology of social boundaries. Hence, the first task that the chapter seeks to achieve is to develop a sociological perspective on the public sphere, understood as a social space of struggle and integration. Starting with Habermas' understanding of the public sphere, and its criticisms by political theorists from different philosophical positions, it emphasizes the need to extend our understanding of the public sphere beyond its role as a space for rational discussion and deliberative politics. The chapter then sketches out the elements of such an extended conception of the public sphere, emphasizing the material and conflicting dimensions, as well as the cultural-symbolic and integrative dimensions of the social world. Such a conception will enable us to understand how the public sphere contributes to the creation and maintenance of society's moral order and social stratification.

Equipped with this understanding of the public sphere, the chapter turns to an analysis of the public debates about freedom of speech in Norway, conceived as a process of symbolic boundary-making in public. The empirical findings presented in the previous chapters of this book shed light on how the symbolic boundaries of freedom of speech are made and contested. They also illustrate how the public sphere functions as a locus where universal claims about symbolic boundaries are made, contested, legitimized or marginalized. Finally, they illuminate some of the cultural processes linking symbolic boundaries to social stratification.

## **The public sphere as a space for struggle and integration**

Modern democracy is usually thought of as a product of the Enlightenment, which raised the idea of publicity to a fundamental principle. In his treatise 'What Is Enlightenment?' Immanuel Kant ([1784], 1991) puts the freedom to make public use of one's reason at the core of the process of Enlightenment, i.e. man's emergence from his self-incurred immaturity. The principle of publicity also constitutes the foundation of public opinion and people's sovereignty, a major legitimizing basis of modern democracy. While pre-modern systems of government legitimized themselves by referring to divine will, modern democracies - where power is based on the consent of the governed - refer to public opinion. In contemporary democracies, the idea of publicity indicates the public sphere primarily, in which the public use of reason or public discussion of free and equal citizens can take place, and public opinion is formed and expressed. In modern democracies, political consent is generated through continuous discursive activity in the public sphere.

Jürgen Habermas has formulated the idea of the public sphere as a site where public opinion is formed through rational discourse in which private individuals forge a common understanding of public goals and exercise scrutiny over the state. Interest in the public sphere, at least in the English-speaking scholarly community, was renewed with the translation of Habermas' *Strukturwandel der Öffentlichkeit* (first published in 1962) into English as *Structural Transformation of the Public Sphere* (1989). Habermas' concept is central to any discussion of the public sphere, but the fact that Habermas revised his own ideas in his *Theory of Communicative Action* ([1981], 1984) and subsequent works makes it important to distinguish between the different versions.

In *The Structural Transformation of the Public Sphere*, Habermas defines the public sphere as the realm of social life in which public opinion is formed. Public spheres are created when private citizens come together and form a public body through dialogue. Habermas differentiates between the political public sphere, which, in contrast to the literate public sphere, is oriented towards the state's activities. The media are necessary in order to disseminate information to a large public body. However, for Habermas the mass media put the public sphere at risk of manipulation and propaganda. Habermas follows Adorno and Horkheimer's assessment of mass media as authoritarian media, broadcasting massive and identical messages, and having the power to reverse the project of Enlightenment. As emphasized by Craig Calhoun (2011), Habermas in *Structural Transformation of the Public Sphere* tends to idealize 18th century English parliamentarianism, newspapers and coffee house conversations. Such an idealization is often at risk of ushering in golden age concepts and narratives of decline, as public discourse mediated by mass media is thought of by Habermas in terms of the loss of the rational-critical capacity.

In the *Theory of Communicative Action* (1984) and *Between Facts and Norms* (1996), Habermas developed his concept of the public sphere on the basis of his theory of symbolic interaction with the *lifeworld* – the background environment of competences, practices, and attitudes where communication and understanding take place according to the rules of practical rationality, in contrast to technical or instrumental rationality that characterizes the *system* – as a critical point. *Reciprocal understanding* demands *universal validity claims,* which are inherent to all speech situations. These speech situations constitute the foundations of a democratic public sphere. The media contribute to the enlargement of the potential for communicative action: 'Writing, the printing press and electronic media mark the significant innovations… by these means speech acts are freed from spatiotemporal contextual limitations and made available for multiple and future contexts' (Habermas, 1984 vol. 2, p. 184). Communicative action 'is raised to a higher power by the electronic media of mass communication'. Despite the fact that the media are now given a role as distributors of communication, Habermas condemns the media for not permitting validity claims to emerge. The media are not an *ideal speech situation* or a democratic public sphere. Mass media are also part of the *system* and threaten to invade the *lifeworld* of intersubjective and communicative interaction. Habermas (2006) has moderated this conclusion in a more recent appreciation of the role of mediated political communication in the public sphere. In this later contribution, which is influenced by the works of Bernhard Peters ([1993], 2008), mediated communication is seen as a means of facilitating 'deliberative legitimation processes in complex societies' (Habermas 2006).

Habermas' approach to the public sphere merits consideration because he accurately conceptualizes the nature of the public sphere, the shift from opinion to public opinion by the development of the public sphere's preeminent institution, the mass-media. However, Habermas' conception of the public sphere has been criticized from different standpoints. Two main traditions – radical democracy and political realism –challenge the normative foundations of public reason, communicative rationality and deliberative democracy on which Habermas' understanding of the public sphere builds.

The radical democratic tradition (Tønder & Thomassen, 2006) emphasizes the principle of participatory parity as a fundamental democratic principle in a culturally plural modern society and criticizes the Habermasian ideal of communicative rationality. Indeed, this ideal demands that deliberations in the public sphere take the form of fully rational and impartial reasoning. This entails, for the participants in public deliberation, the imperative of agreeing with the best argument independently of their particular interests or identities. However, discourse in the public sphere may be characterized by a purposive and instrumental orientation, as well as by other expressive and emotional communicative strategies involving irony, personal narrative, aesthetic interventions, and theatricality. Additionally, a strict focus on rational deliberation disqualifies everyday talk and its relevance for democracy. Young (1990), for example, criticizes the ideal of impartiality in public deliberations for reducing difference to unity, and consequently impeding a genuine participatory parity. The stances of detachment and dispassion attached to the ideas of public reason and communicative rationality have the consequence, for Young (1990), of forcing individuals acting in the public sphere to abstract their feelings, affiliations, and points of view, generating 'a dichotomy between universal and particular, between public and private, reason and passion' (Young 1990 p. 97). In the same vein, Behabib (1996) insists on the need for normative theories of public deliberation to recognize that different visions of the good life and different collective identities play a central role for individuals acting in the public sphere.

The realist tradition underscores the public sphere as a site of public contestation and the enduring and creative nature of conflict. Political realism (Galston, 2010) designates a heterogeneous set of approaches that have in common the development of a critique of Rawls' and Habermas' 'ideal theory'. Following Bernard Williams (2005), distinguishing between political moralism and political realism, the characteristic of 'ideal theories' such as those advanced by Rawls and Habermas, is to make the moral prior to the political. Political realism affirms the autonomy of politics in relation to morality (the right), denies the possibility of achieving coordination through consent, and considers that coordination will require coercion or the threat of coercion. Political realists 'see political conflict as ubiquitous, perennial, ineradicable, and they regard political moralists as being far too sanguine about the possibility of achieving either normative or practical consensus' (Galston, 2010 p. 393). They insist that political disagreements are of a different nature than intellectual disagreements, since in political disagreements our interests and cultural identities are at stake. In her critique of the Habermasian tradition, Mouffe (2000) mobilizes Wittgenstein's notion of the 'language game' pointing to the fact that agreement on language necessitates agreement on 'forms of life', a fact that entails, in pluralist societies, the prevalence of antagonist conceptions that develop into power struggles. Mouffe (2000) develops a perspective of 'agonistic pluralism' entailing a concept of the public sphere in which conflicts and power struggles are compatible with democratic values. Hence, political realism leads to a conception of the public sphere as a space, among other institutional spaces, where political struggles take place, since political struggles are also struggles for being heard (Rasmussen, 2016).

Taken together, these criticisms of Habermas' theories of the public sphere point towards three points of contention: the *kind of talk* that ideally should characterize public deliberation in the public sphere, the *role of rationality and emotions* in public deliberations, and the *power mechanisms* that are at play in public deliberations.

Concerning the first dimension of contention— the *kind of talk* acceptable in public discourse *–* the Habermasian ideal of communicative rationality conceives public discourse as fully rational and impartial reasoning, entailing the imperative that participants agree with the best argument independent of their particular interests or identities. However, political utterances are aimed at finding solutions for conflicts and have a purposive and instrumental orientation. In addition, a strict focus on rational deliberation disqualifies everyday talk and its relevance for democracy.

The second issue of contention relates to the role of rationality in public discourse. The normative demands of public reason and communicative rationality are seen as excluding from public debates other expressive and emotional communicative strategies such as irony, personal narrative, aesthetic interventions, and theatricality, which are necessary for motivating and maintaining engagement in the public sphere.

The third issue has to do with discursive and social power. The ideals of communicative rationality in the public sphere suppose that all participants are equal. However, participation in public deliberation often correlates with power and cultural capital. As pointed out by Young (1990), public settings that require universal, neutral and egalitarian discursive modalities may reflect the *habitus* of the privileged class and constitute a form of symbolic power. Further, there exists a contradiction between the ideals of communicative action and the nature of politics involving power and conflict relations.

These issues of contention reflect higher-level disagreement about the telos (consensus vs. conflict), the mechanisms of power, and the nature of agency in the public sphere. Concerning these three issues of disagreement, positions are formed around some basic differences. A first differentiation can be drawn between perspectives emphasizing consensus and social integration through common values as the result of the deliberation process taking place in the public sphere (Habermas 1984, 1996; Rawls 1993), and those considering that conflict and radical disagreement (agonistic pluralism) are fundamental characteristics of the public sphere (Mouffe, 2000; Williams, 2005; Luhmann, 2000). A second differentiation operates between realist theories (Williams 2005, Luhmann 2000), considering debates in the public sphere as expressing power and interest struggles, and idealist theories, for which public debates are concerned with values and cultural representations (Habermas, 1996, Mouffe, 2000). A final differentiation is related to different conceptions of agency in the public sphere. Whereas Habermas and Rawls conceive agency in the public sphere in the Kantian tradition, based on rationality and public reason, others (Mouffe, 2000; Young, 1990; Benhabib, 1996) emphasize that public debates involve individuals who are 'situated selves', including their identities and emotions, and not just rational agents detached from their concrete situation. Additionally, common to these approaches is the fact that, in addition to being philosophical elaborations – not analyses of how the 'real' public spheres work - they tend to underscore either the integrative and legitimating capacity of consensus reached through public communication, or the fundamental nature of disagreement in politics and the historical nature of discourse (Rasmussen, 2016). Finally, these normative theories tend to limit the role of the public sphere as contributing to opinion-formation and the legitimacy of political decision-making in a democratic polity, obscuring the role played by the public sphere in making and maintaining society's social integration, moral order and social stratification.

In spite of the limitations of these different normative perspectives, the concept of the public sphere remains a central analytical tool to help us make sense of the relationship between the media and democracy (civic engagement). In contrast to these normative theories, a sociology of the public sphere will offer us a more empirically grounded understanding of public communication, including a wide range of social behaviors. A sociology of the public sphere, will additionally be multidimensional, reflecting the fact that social reality consists of both material and cultural elements, and is characterized by conflicts as well as by social integration and solidarity.

## **A sociology of the public sphere**

A sociological understanding of the functioning of the public sphere has to be empirically grounded in the analysis of historical societies, and must develop a conceptual apparatus adapted to the task of analyzing the public sphere as a social space. Furthermore, it needs to consider a wide range of social behaviors and motivations for social action, beyond rational agency and moral principles. As Gueuss (2008) reminds us, normative theories in politics tend to be conceived in terms of applied ethics, the best-known instance of this approach being Kantianism, focusing on a few general and abstract principles to be applied universally and independent of historical and social contexts. In contrast, a sociological approach to the public sphere has to be primarily concerned with how people and institutions actually operate in society and not how they ought 'ideally' to operate.

A sociology of the public sphere also needs to be multidimensional, considering the material and conflict dimensions as well

### CHAPTER 10

as the cultural-symbolic and integrative dimensions of the social world. Following Bourdieu (2000 p. 187), the social world can be considered as

Both the product and the stake in inseparably cognitive and political symbolic struggles over knowledge and recognition, in which each individual pursues not only the imposition of an advantageous representation of himself or herself […] but also the power to impose as legitimate the principles for the construction of a social reality most favorable to his or her being – individual and collective, with […] struggles over the boundaries of groups.

Hence, from this viewpoint, the public sphere may be seen as a privileged social space where cognitive and symbolic struggles over recognition – entailing struggles over the symbolic boundaries delimiting group belonging and exclusion, as well as social worth –take place.

However, at the same time, we need not lose sight of the insights of Durkheimian sociology because, as put by Alexander (2006 p. 3):

Societies are not governed by power alone and are not fueled only by the pursuit of self-interest. Feelings for others matter, and they are structured by the boundaries of solidarity. How solidarity is structured, how far it extends, what it is composed of – these are critical issues for every social order, and especially for orders that aim for the good life. Solidarity is possible because people are oriented not only to the here and now but to the ideal, to the transcendent, to what they hope will be everlasting.

Consequently, the public sphere is not only and uniquely a space of symbolic and political struggle, but is, by the same token, a space of solidarity and social integration. In public discourse and debate, the boundaries of solidarity are actualized (and sometimes contested and redefined) and common ideals and values are mobilized and enacted.

Finally, we need to understand the social role of the public sphere beyond being a deliberative space contributing to opinion formation and producing 'inputs' to the political system, also thus contributing to the political system's efficacy and legitimacy. The public sphere, in modern differentiated and mediatized societies, plays a crucial role not only in producing and reproducing society's 'moral order' (Wuthnow, 1987) by universalizing cultural boundaries that sustain people's commitment to morally valued activities, but also by producing and reproducing social inequalities (Lamont, Beljean, & Clair, 2014).

The public sphere is the sphere where different conceptions of justice, the common good and solidarity, i.e. the boundaries and finalities of the community, are confronted and are objects of struggle for *universal recognition*. This entails both struggles among competing values and value-orientations (i.e. struggles over the symbolic categories or boundaries defining and delimitating the good and the evil, the worthy and the worthless), and struggles for the recognition of individuals' and groups' identities (including the symbolic boundaries between these groups, the assertion of their social worth or status, and the perimeters of solidarity). The public sphere is also the social space where demands for justice and regulation emanate, are negotiated and pushed through the state and the political system, insofar as the state has the legitimate capacity and power to universalize and to coerce.

From such a perspective, freedom of speech may be conceptualized as a feature or dimension of civil society in democratic liberal societies, i.e. as an institution of justice, an institutionalized condition for the functioning of democracy, but also as an object of political contestation and political decisions (Williams, 2005 p. 26). Indeed, as an institution of justice, Human Rights, of which the right to freedom of speech is one, stand against 'people using power to coerce other people against their will to secure what the first people want simply because they want it' (Williams, 2005 p. 23). However, what, in a given historical context, counts as injustice is not invariant, and how this right has to be limited is a matter of political contestation and social struggle. Because the right to freedom of speech is both a means by which social and political struggles are fought and a locus for these struggles, this right is threatened by non-civil practices (violence, secrecy, hate-speech, threats, libel, bullying, censorship and self-censorship).

Conceiving the public sphere as a social space where cultural struggles are fought, where the moral order is shaped, maintained, and contested, where symbolic boundaries are publicly enacted and struggled with, entails shifting the focus of the analysis of the public sphere from its role as the site of public-opinion formation to its role as a privileged locus of the social and cultural fabric of society. From this viewpoint, much of the public debate about freedom of speech, during the last decade, can be analyzed as the result of a process of boundary-making where individuals and groups struggle over the legitimate 'principles of vision and division of the world', their recognition and universalization, especially about which social divisions in terms of identity groups are to be recognized as legitimate, and about whether the right to freedom of speech has to be limited in order to protect these identity groups.

Insofar as Norwegian public debate about freedom of speech is closely intertwined with issues of immigration and integration of minorities, it can be interpreted in terms of symbolic processes taking place in the public sphere – both collective rituals and symbolic boundary struggles where universal cultural claims are made, that contribute to the constitution, enactment and transformation of the moral order. The rest of this chapter is consequently devoted to the tasks of elaborating a framework to help us understand the public debates about freedom of speech in terms of collective rituals, moral order, boundary struggles, and universal cultural claims, and of assessing the social and structural consequences of these struggles.

# **Two dimensions of the public debates about freedom of speech**

During the last decade (2005-2015), freedom of speech has been problematized, on the one hand, within an increasingly globalized and transnational context, in Norwegian public debates and in social media, mostly in relation to the 'Muhammad Cartoon Crisis' (2005/2006), and in 2015 to the attacks on *Charlie Hebdo* in Paris and Krudttønden in Copenhagen (see Colbjørnsen, chapter 6). On the other hand, during the same period, freedom of speech has been problematized, in relation to the content and tone of public debates about immigration and the integration of migrants, in a period marked by increasing immigration (see Ihlebæk & Thorseth, chapter 5). When it comes to the issue of freedom of speech, a particular point of contention concerned the need to protect and respect ethnic and religious minorities, as opposed to subjecting these minorities to criticism when their cultural practices are in contradiction to democratic values and rights. Additionally, the issues of the representation and participation of minorities in the public sphere have been a constant concern.

Public debate about the exercise of freedom of speech in Norway relates mainly to issues involving, directly or indirectly, individuals' and groups' identities, as well as the regulation of free speech on the basis of these identities. The public debate consists, firstly, of a meta-debate on freedom of speech, whereby the right and duty to publish materials (cartoons, religious criticism, critical discourse about immigration) is discussed in a context where identity-based or religious groups (mainly Muslims) consider these materials to be offensive, and in relation to which fundamentalists and extremist groups (abroad) have used violence in order to silence and sensor the publishing of such materials. It has been argued that the right and duty to publish these materials is not absolute and has to be balanced with the need to recognize cultural and religious identities, and to show consideration for the sensibility of minorities and the violation of their rights and identities (Fladmoe & Nadim, chapter 2; Bangstad & Vetlesen, 2011; Midtbøen & Steen-Johnsen, 2015; Midtbøen 2016).

Secondly, the debate has been extended to include the issue of hate speech and the limits and sanctions of hate speech, insofar as hate speech is mainly directed towards individuals by reason of their belonging to identity and religious groups. What are the stakes in these public debates about the exercise of free speech? Why do the issues of identity and identity groups become framed as debates about freedom of speech in the Norwegian public sphere? These two sets of publicly debated issues (metadebate about freedom of speech and self-censorship) may be understood as reflecting two types of simultaneous and intertwined symbolic processes taking place in the public sphere. On the one hand, these debates can be seen as a *collective ritual* in response to boundary crises in the moral order. On the other hand, these debates are constitutive *symbolic power struggles* in the public sphere. What is at stake in these struggles is the *universalization* of the symbolic boundaries – the categories of cognition or discourse about differences— defining both the limits of inclusion for identity groups and of acceptable expressions about these groups. Indeed, it is not sufficient for identity groups to define symbolic boundaries that differentiate group members from outsiders. These boundaries have to be universally acknowledged and recognized by society (i.e. valid globally for all members of society, not only locally for a subgroup of society, even if the boundary is contested) as being efficacious identity markers. Similarly, the symbolic boundaries delimiting which discourses about these identity groups are not acceptable have to be universally valid (even if contested) in order to restrict public expression. Members of ascribed as well as freely chosen identity groups struggle with mainstream citizens and institutions for the recognition and universalization of the symbolic boundaries they want to prevail as part of society's moral order.

There is a ritualistic element characterizing the recurrent debates about freedom of speech in Norway, especially when they are related to the 'Muhammad Cartoons'. Collective rituals can be thought of as cultural practices that emerge when society's moral order is challenged, and when symbolic boundaries are blurred and in crisis (Wuthnow, 1987 p. 115). Following Durkheim, it can be argued that societies develop collective identities that define boundaries of membership, and distinguish the collectivity from outsiders. When these boundaries change and become blurred – as a result of internal disagreements, ambiguities about values, the need to include newcomers, or external threats— uncertainties about membership, authority and shared values occur.

The violence against cartoonists, the violent manifestations against the Muhammad cartoons, and the terrorist attacks against *Charlie Hebdo* were transgressions of shared values of freedom and freedom of speech in the democratic world. The public debates following these events in 2005/2006 and 2015

were about the press's right and obligation to publish the cartoons in contrast to the press's need to be sensitive to the dignity of minorities. These debates can be seen as the expression of a collective ritual, whose function was to clarify the collective boundaries to free expression and to reaffirm the moral order. The ritualized public debate, mainly driven by journalists and media professionals, served to mobilize collective sentiments and solidarity around the fundamental, but sometimes conflicting, democratic values of freedom and tolerance. The public debate following the attacks against *Charlie Hebdo* has been characterized by a confrontation between those who stressed the value of free speech and the obligation to publish controversial materials, and those who argued for being tolerant and sensitive to cultural and religious differences, and for refraining from publishing these materials. Through this debate, the values of freedom and tolerance have been re-established and consolidated and the symbolic boundaries clarified.

Indeed, collective rituals are essentially social and dramaturgic. They dramatize the moral order and communicate, through symbolic expression, fundamental features of social relations. In the case of the Muhammad cartoons and the attacks against *Charlie Hebdo*, external threats were combined with the difficult process of the inclusion of new members (migrants, especially Muslims) into the community, and provoked a blurring of the boundaries of both the community and free-speech, endangering the established moral order. The public debate following this boundary crisis, by dramatizing and provoking discussions about central values in Norwegian society, helped restore and reaffirm the endangered moral order, and reinforced shared values in the face of moral uncertainties. The democratic values of freedom and equality have been publicly and symbolically reaffirmed. This does not mean that the underlying tensions and conflicts characterizing the integration of newcomers with different cultural and religious backgrounds have been resolved. On the contrary, the ritualistic symbolic reenactment of these values seems to have contributed to making these tensions even more visible, leading to an exacerbation of symbolic boundary struggles in the public sphere.

The two waves of freedom of speech debates in the Norwegian public sphere, in 2005/6 and in 2015, were, however, not only a form of collective ritual contributing to the reaffirmation of the endangered symbolic order. Because they referred to and concerned immigrant minorities, and especially Muslim immigrants, at stake within these debates were also the symbolic definitions of the identity groups that make up Norwegian society and the conditions of inclusion for these groups. These debates were also symbolic boundary struggles, where the issue of drawing the symbolic limits between acceptable and unacceptable speech, as well as the nature of the recognition of minority identity groups were simultaneously contested.

Symbolic boundary struggles taking place in the public sphere are power struggles about the universalization and institutionalization of social boundaries, in order to obtain universal currency in society. When it comes to ethnicity, the emergence of social and symbolic boundaries are the result of the co-occurrence of distinctions made by actors (symbolic distinctions, categories), and differentiated treatment of the members of such categories (social differentiation). As underscored by Wimmer (2013 pp. 4-5), social and symbolic boundaries involve the struggle over power and prestige – group honor, moral dignity, and personal identity, combined with material preoccupations such as access to material benefits or political power. Social and symbolic boundary struggles are not exclusively about 'interests' or 'identity', about 'material' benefits or 'ideals', but mix these various resources into an intertwined struggle over who legitimately should occupy a given position in the hierarchical structure of society.

Drawing identity group boundaries and the boundaries of freedom of speech may be understood as two instances of a power struggle concerning the position and the social conditions of identity groups in the polity, where different normative and cultural conceptions of society and public expression and the public sphere are confronted. What is at stake in the debates about freedom of speech and hate speech is the legitimacy of the political organization (how society should be structured and ruled and how people under government should live; i.e. conceptions of justice, right and good) of an increasingly pluralistic society, including the criteria of inclusion and recognition of identity-based groups.

Identity groups–whether categorized as 'immigration critics' (Thorbjørnsrud, chapter 9) or being representative of ethnic and religious minorities (Nadim, chapter 8) –struggle for both access to the public sphere (experiencing ascription of their identities and devaluation of their social worth, as well as a lack of recognition of their positions as legitimate), and for universal recognition of their identities and political positions. Both groups struggle to transform the symbolic identity boundaries that are imposed on them by others. Existing identity boundaries limit their freedom of speech, and sometimes lead them to selfcensoring insofar as these identity boundaries assign to them a given worth and position in public debates.

The stigma of racism, the public shame, and the moral condemnation that are attached to public expressions critical to immigration imply that public expression of those positions are psychologically and socially costly and lead to social and cultural devaluation and exclusion. As pointed out by Lamont et al. (2016 pp. 281-282) 'cultural membership is given to those who meet the standards of shared definitions of who is worthy in a symbolic community'. The lack of recognition experienced by immigration critics is embedded in the existing cultural repertoire or semiotic code of civil society (Alexander, 1992). Indeed, recognition concerns the social acknowledgement of worth across differences, and is relational, inasmuch as it is a status provided by others in a community (Lamont & al., 2016 p. 282). In the public sphere, the social estimation of worth is culturally articulated in terms of the 'binary code of civil society' (Alexander, 1992; 2006), which is mobilized in order to determine the worth of the persons acting in public, and to delimit the perimeter of the symbolic community of legitimate participants (inclusion).

The struggle for recognition and its afferent symbolic boundary struggle assume different forms for ethnic and religious minorities active in the public sphere (Nadim, chapter 8), who seek to escape their ascribed identity as representatives of a minority. The cultural repertoires and institutionalized scripts patterning the ways in which religious and ethnic groups are identified and classified in the Norwegian context – combined with a media logic emphasizing a diverse representation of minorities— contribute to the *ethnicification* and *culturalization*  of individuals with an ethnic or religious minority background, when participating in the public sphere. These individuals struggle for the recognition of their singularity and individuality, as bearers of plural identities, and against the ascription of a particular identity as the representative of a minority.

The issue of hate speech (Fladmoe & Nadim, chapter 2) also brings together these two types of social boundaries – the boundaries of belonging (identities) and the boundaries of unacceptable speech. Making unacceptable certain types of speech,

### CHAPTER 10

whose focal points (gender, ethnicity, sexuality, religion, etc.) are related to identity groups, might be understood as a strategic move within a more general strategic struggle over social boundaries and their benefits (group honor, moral dignity, and personal identity combined with more mundane preoccupations such as access to professions, public goods, or political power).

Debates about the exercise of free speech are not just about the regulation of public speech, but are inscribed in a network of contested issues about the political integration of immigrant minorities (justice, rights, recognition, pluralism and tolerance), in a context characterized by increased cultural and religious pluralism, and the rise of fundamentalist ideologies and terrorism. These debates are also constitutive of collective rituals where central societal values are reestablished when the collectivity faces external threats and changes. As such, they are cultural phenomena operating in the symbolic realm. However, cultural processes are part of the social fabric and produce real effects on the social structure of society and the distribution of social inequalities.

### **Symbolic boundaries' structural effects**

Symbolic boundaries have effects, not only in the cultural realm, the realm of meaning, but also on the social structure of society and its social stratification. Boundary struggles about the nexus intertwining freedom of speech and identity groups in contemporary Norway affect both the 'moral order' of society and its social stratification. In other words, the categorization and labeling of identity groups constitutive of Norwegian society— in terms of ethnic Norwegians, Muslims, minorities, immigration critics, etc.— contribute to the elaboration of a moral hierarchy, where certain groups are symbolically more worthy than others and more or less included within the national community. This moral hierarchy, in turn, has a bearing on the social hierarchy of the national community, its social stratification, by means of the play of cultural processes that affect these groups' access to material resources and power.

Indeed, social inequalities do not result uniquely from the distribution of material resources, but are also the result of the unequal distribution of symbolic resources and recognition, which is mediated by a series of cultural processes, shaped by the use of shared categories, classification systems, cultural scripts and repertoires (Lamont et al., 2014). These cultural processes operate on the individual level through cognitive activities and on the inter-subjective level, as individuals mobilize shared cultural scripts and structures in order to make sense of their social environment. The use of objectified shared categories for defining group boundaries and sorting people entails the relative stabilization of the hierarchy of categories. (Lamont et al., 2014) Lamont thus distinguishes several cultural processes that generate social inequalities.

One of those cultural processes, *identification*— the process through which individuals and groups identify themselves, and are identified by others, as members of a larger collective – reinforces the stabilization of symbolic boundaries and hierarchies based on group identities. Within this process of identification, two processes, *racialization*<sup>1</sup> and *stigmatization*<sup>2</sup> , are more likely to generate social inequalities as they limit access to material, social, and cultural resources for the members of the groups.

A second cultural process, the process of *evaluation*, concerned with the definition and stabilization of value in social life,

<sup>1</sup> The process by which biological and phenotypic differences between human bodies are attributed social significance.

<sup>2</sup> The negative stereotyping and separation of groups who are labeled as different.

is also relevant for our analysis. According to Lamont (2012 p. 206), this process involves several sub-processes, most importantly *categorization* ('determining in which group the entity [. . .] under consideration belongs') and *legitimation* ('recognition by oneself and others of the value of an entity'). Because members of different social groups are constantly subject to evaluation, based on inter-subjective criteria, their social status and worth are stabilized within a hierarchy of recognition, which influences their opportunity to access material, social and cultural resources.

Several cultural processes have been identified in the previous chapters. Midtbøen (chapter 7) and Thorbjørnsrud (chapter 9) have shown how cultural processes of *stigmatization* operate for young political leaders acting in the public sphere as well as immigration critics. Group categorizations and evaluations, that are part of the established moral order, function as markers of difference, based on ethnicity, religion, disability or sexual orientation, and lead young political leaders to avoid discussing given topics, and in certain instances to censor themselves. As shown by Ihlebæk and Thorseth (chapter 5), the moral order may be enhanced or transformed as the result of *evaluations* and editorial decisions made by newspapers' editors in chief — even if their gatekeeping power is eroded by the rise of social media – concerning which types of opinions on immigration may be published. The stickiness and reinforcement of existing symbolic boundaries constitutive of the moral order (about acceptable speech and identity groups) are also enhanced by the phenomenon of spirals of silence (Fladmoe & Steen-Johnsen, chapter 3). When individuals (mis-)perceive their opinion as a minority opinion and fear social exclusion, they censor themselves and abstain from expressing their views, thus contributing to the reinforcement of the dominant opinion.

The phenomenon of hate speech (Fladmoe & Nadim, chapter 2) illustrates how group-boundaries and the creation of hierarchies interact. Without relatively stable and inter-subjectively constituted symbolic boundaries and categories, which sort and ascribe given identities to people on the basis of gender, religion, ethnicity, or sexuality, the degradation of identity and dignity (the symbolic violence which is inherent to hate speech) would not be possible. Similarly, without the social hierarchy of worth that is associated with group identities, hate speech would be meaningless. Indeed, hate speech takes advantage of the existing symbolic boundaries defining identity groups and their social hierarchy in order to degrade these identities, denying them legitimacy, value, worth, dignity and membership in the political community. Hate speech transgresses a major symbolic boundary of democratic societies – that of equality of dignity and worth of human beings – in order to degrade a person taking a position in the public sphere, negating the worth of her identity and the legitimacy of her opinion as an equal member of the political community and as a participant in public debate.

Debates in the public sphere, because they encompass cultural processes of identification and evaluation, and because they contribute to the establishment and stabilizing of social hierarchies, have long-term consequences for the social structure of Norwegian society and the distribution of material, social and cultural resources. The public debate about freedom of speech might be thought of as a collective ritual in which a society confronted with external threats and uncertainties reestablishes its symbolic boundaries and reaffirms its moral order by affirming for itself, in an act of self-reflexivity, the value of free speech and the intolerance of violence aimed at intimidating free speech, thus also clarifying the limits of freedom of speech. But at the same time, because most of the uncertainties and threats that challenge the moral order are directly or indirectly related to minority identity groups and religion, this debate is also part of a more general symbolic boundary struggle in which what is at stake is the inclusion of these minority groups in the national community. As these struggles involve cultural operations of classification, identification and evaluation, such public debates incur the risk of generating processes of racialization, stigmatization and misrecognition, which, in the long-run, might generate and cement social inequalities based on cultural and ethnic markers.

## **Conclusion**

Despite being limited by its empirical focus—contemporary debates about freedom of speech in Norway – the analytical framework developed in this book, analyzing the dynamics of the public sphere in terms of symbolic boundary struggles taking place in public, is of broader relevance. The literature on symbolic boundaries is large and diversified (Lamont et al., forthcoming). This book's original contribution to this literature has been to explore the processes of boundary making as they occur in public, within the mediated public sphere. Debates in the public sphere do not exclusively consist of rational argumentation about public policies but also have a legitimizing function for the political system. Nor are they uniquely a reflection of diverging interests disguised in ideologies and power struggles between these interests and ideologies. Debates in the public sphere are also symbolic struggles over the boundaries constitutive of the moral order of society, mobilizing the resources embedded in societies' cultural repertoires, affecting the identities, positions, worth, and recognition of diverse social groups, and ultimately producing – by the play of an array of cultural processes— social, structural and institutional effects that contribute to the differentiation and stratification of society in terms of access to resources and power, and having, consequently, real effects on people's lives.

Furthermore, a social and symbolic boundary-making perspective on free speech has allowed us to better understand the past decade of debate about freedom of speech in Norway. These debates are not just about the confrontation of ideological preferences relative to an abstract right, but crystallize one of the most important challenges confronting contemporary Norwegian society, namely that of the social and political integration of immigrants and religious minorities.

A sociological perspective on free speech and the public sphere reminds us that the categories (the symbolic boundaries) which are constructed, mobilized and fought over in public debates, are not just intellectual or ideal constructions and concepts, but have social effects. They become *real* to the extent to which the *social* is as real as the material reality, and they produce real effects on individuals and groups. By considering how the actors struggle over which social boundaries should be considered relevant and legitimate, we endeavor to make visible what the consequences of being an X versus being a Y should entail. Perhaps Norwegian public debate would gain relevance if the participants, instead of debating which speech utterances are acceptable or not, would debate the modes of social and political integration in the Norwegian polity, and become more aware of which social and symbolic boundaries they are drawing and which real consequences these boundaries produce.

## **References**


Habermas, J. (2006). Political Communication in Media Society: Does Democracy Still Enjoy an Epistemic Dimension? The Impact of Normative Theory on Empirical Research. *Communication Theory, 16*(2006), 411–426.


Rawls, J. (1993). *Political Liberalism.* New York: Columbia University Press.

Tønder, L., & Thomassen, L. (eds). (2006). *Radical Democracy.* New York: Palgrave.

### CHAPTER 10


# About the Authors

**Arnfinn H. Midtbøen** is a Senior Research Fellow at the Institute for Social Research. He earned his PhD in Sociology from the University of Oslo in 2013. Midtbøen works on a broad set of research issues related to immigration, citizenship, free speech, integration and patterns of exclusion and inclusion in the labour market. He is head of the project The Status of Freedom of Speech in Norway. arnfinn.h.midtboen@samfunnsforskning.no

**Kari Steen-Johnsen** is a Research Professor and Research Director for the group Politics, Democracy, Civil Society at the Institute for Social Research. Her research interests include political mobilization and participation and public debate and the freedom of speech, with an emphasis on the consequences of digitalization and the emergence of social media. kari.steenjohnsen@samfunnsforskning.no

**Kjersti Thorbjørnsrud** is a Senior Research Fellow at the Institute for Social Research. She earned her PhD in Media and Journalism from the University of Oslo in 2008. Her research concerns political communication, media, migration, journalism, public policy and administration. kjersti.thorbjornsrud@samfunnsforskning.no

**Audun Fladmoe** is a Senior Research Fellow at the Institute for Social Research. He has a PhD in Political Science from the Norwegian University of Science and Technology (NTNU, 2012). His main research interests include civic and political engagement, public opinion, freedom of speech, and hate speech. audun.fladmoe@samfunnsforskning.no

#### about the authors

**Marjan Nadim** is a Senior Resarch Fellow at the Institute for Social Research. She has a PhD in Sociology from the University of Oslo (2014), and her research interests include immigrants and children of immigrants in education and the labour market, gender equality, hate speech and free speech. marjan.nadim@ samfunnsforskning.no

**Hallvard Moe** is a Professor of Media Studies, University of Bergen, where he also earned his PhD in 2009. Moe is interested in the role of media in democracy, and is currently researching citizens' public connection. Hallvard.Moe@uib.no

**Karoline Andrea Ihlebæk** holds a PhD from the Department of media and communication at the University of Oslo, where she is currently employed as a Postdoctoral fellow. She is also an Associate professor at the Norwegian University of Science and Technology, and an affiliated researcher at the Center for Research on Extremism: The Extreme Right, Hate Crime and Political Violence (C-REX). Political communication, journalism, public debate, audience participation, migration, free speech, media power, and editorial strategies are among her research interests. k.a.ihlebak@media.uio.no

**Ingrid Endresen Thorseth** has been working as a Research assistant for the Department of media and communication at the University of Oslo. She is currently enrolled in the Master's program Nordic Media, writing her thesis on marketing, job advertising and where people are looking for jobs in Norway. Thorseth has held various positions in fields such as HR, journalism, communication and marketing. i.e.thorseth@ media.uio.no

**Terje Colbjørnsen** has a PhD from the Department of media and communication, University of Oslo. He has worked on research issues related to digitalization and innovation in the book industry and freedom of speech. Colbjørnsen is currently working as an adviser for The Norwegian Association of Researcher (Forskerforbundet). terjeco@gmail.com

**Bernard Enjolras** is a Research Professor at the Institute for social research and Director of the Center for Research on Civil Society and Voluntary Sector. He received his PhD in Economics from Université de Paris I. Panthéon Sorbonne, Paris in 1994 and his PhD in Sociology from Université du Québec à Montréal in 2006*.* His research interests are related to different aspects of civil society and the third sector, public policies and governance issues, civic and political engagement, social media and digitalization, terror and trust, and freedom of speech. bernard.enjolras@samfunnsforskning.no