Hate speech in the EU and the C.O.N.T.A.C.T project

• You may download this work for personal use only. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying this open access version If you believe that this document breaches copyright please contact us providing details and we will investigate your claim. Please direct all enquiries to puresupport@bib.sdu.dk

Migration phenomena characterised by a large influx of populations can question our conception of territories and social relations. Since this conception is part and parcel of our identity, migration has the power to trigger political discourses on identity issues. One such occasion has indeed been unravelling lately, especially since the summer of 2015, with the arrival in the European Union (henceforth EU) of migrants from a variety of places, and in particular from regions in conflict, such as Syria, Libya or Iraq, countries under totalitarian regimes, such as Erythrea, as well as countries with high levels of poverty, such as Pakistan and Bangladesh. As a result, Europe has been politically and socially shaken: photos of thousands of migrants roaming across Europe have made the news, and such media images have been instrumentalised to serve different, often far-right, political agendas.
The question of refugees-and more broadly migrants-and their integration in Europe has been in the spotlight, with media discourse being on the whole alarmist, with an iteration of expressions like a 'huge migration crisis', 'waves of migrants flooding the EU' and a focus on violence and threat as the main outcome of such arrivals (cf. UNHCR 2016). In turn, Europe is witnessing the growth of nationalism, with violent reactions being related to the feelings of insecurity, fear or anger, and several xenophobic political parties, such as Golden Dawn in Greece or AfD (Alternative for Germany) in Germany feeding these feelings of anxiety and resentment to attract voters. Finally, recent reports still indicate that the migration issue continues to be one of the major preoccupations of European citizens (cf. European Commission 2016a).
Indeed, the 2016 report of the European Commission against Racism and Intolerance observed a sharp increase in hate crime while also noting that "racist insults have become increasingly common and xenophobic hate speech has reached unprecedented levels" (ECRI 2017: 9, italics our own). At the same time, both researchers and NGOs have repeatedly noted how Web 2.0 has facilitated the global spread of hate. For example, the latest Shadow Report by the European Network Against Racism (ENAR 2016) has pointed out a rise in racist discourse both on social media and the internet. In response to the situation, the EU has encouraged several initiatives with a view to containing both hate speech and hate crime within its remit. Legal provisions (cf. Sect. 1.1) foresee penalties for those publicly inciting to racial hatred, while the European Agency of Fundamental Rights has defined within the Framework Decision on Racism and Xenophobia the following priorities: • the identification of hate crime, • the increasing use of the internet as a tool of hate and propaganda, • the under-reporting of hate crime, • the rise of extremist groups and political parties in the EU.
The C.O.N.T.A.C.T. 1 project (2015)(2016)(2017), which was co-funded by the Rights, Equality & Citizenship Programme of the European Commission Directorate-General for Justice and Consumers (JUST/2014/RRAC/AG), sought to address the above priorities by combining complementary expertise from academics and experienced NGOs working in the area across a number of EU member states, namely Cyprus, Denmark, Greece, Italy, Lithuania, Malta, Poland, Romania, Spain and the United Kingdom. To this end, under the central coordination of the University of Cyprus and, more specifically, Professor Fabienne H. Baider, C.O.N. T.A.C.T. partners have engaged in a number of activities, which to a great extent follow Ramalingam's (2012: 11-13) categorisation of measures that would effectively target far-right extremism. These include: up-stream preventative measures, such as the collection and scientific analysis of data that will help better understand the context of hate speech online, as well as the development of training sessions targeted at relevant stakeholders (police, youth and media) with a view to building a stronger civil society. reactive measures and response mechanisms, such as the establishment of a dedicated web platform and phone app for reporting hate incidents. intervention through the training of the relevant stakeholders and the organisation of awareness-raising events. 2 Against this background, the present volume is an attempt to collectively report on some research that several C.O.N.T.A.C.T. partners undertook as part of their involvement with the project. Even though hate speech is a hotly debated topic in 1 C.O.N.T.A.C.T. stands for 'Creating an On-line Network, monitoring Team and phone App to Counter hate crime Tactics'. 2 For more information about the C.O.N.T.A.C.T. project, visit our website at: http://www. reportinghate.eu. 1 Introduction and Background legal and policy-making circles, the relatively little attention it has received by researchers of linguistic pragmatics and discourse analysis is arguably disproportionate to its social relevance and importance. In this respect, the main aim of this volume is to showcase that an implementation of certain research methodologies that linguists, and more specifically discourse analysts, have at their disposal can fruitfully contribute to the better understanding of a phenomenon that, as we saw, is becoming increasingly widespread these days. In light of this, the contents of the present volume should be approached as more of a 'proof of concept' demonstration, rather than an exhaustive analysis of hate speech in the EU. The reason for this is simple: as McGonagle (2013: 3) points out even though the term 'hate speech' is often incorporated, at least as a notion, into legal and policy documents, there is still no universally accepted definition for it, which on its own warrants further investigation into the ways in which hate, in the relevant sense, is both expressed and perceived.
Generally speaking, hate speech could be described as the expression of hatred towards an individual or group of individuals on the basis of protected characteristics, where the term 'protected characteristics' denotes membership to some specific social group that could, on its own, trigger discrimination (cf. OSCE/ ODIHR 3 2009: 37-46). What these protected characteristics are, however, remains open to interpretation, with different states including different categories under this rubric, as will be discussed in more detail in the following section of this introductory chapter. Just to give an example, the EU definition of hate speech that is put forth in the Council Framework Decision 2008/913/JHA of 2008 confines hate speech to "all conduct publicly inciting to violence or hatred directed against a group of persons or a member of such a group defined by reference to race, colour, religion, descent or national or ethnic origin" (Council of the European Union 2008), essentially leaving out of the equation such characteristics as sex, gender identity and sexual orientation.
As Baider (2017) notes, however, in an attempt to define 'hate speech' more broadly, one could follow the International Covenant on Civil and Political Rights which does not single out any particular protected characteristics and instead proposes that hate speech essentially amounts to an "advocacy of discriminatory hatred which constitutes incitement to hostility, discrimination or violence" (UN General Assembly 1966, our italics; see also OHCHR 2013). While the question of how to exactly interpret the words 'hatred', 'discrimination', 'violence' and 'hostility' in this definition still remains open, it manages to express more concretely the forms that the expression of hatred, in the relevant sense, may take. What is more important here, however, is the word 'incitement', which takes centre stage and renders the intention to trigger potential actions against members of protected groups a precondition for considering a speech act hate speech, assuming, thus, a link between hate speech and hate crime, with the former presumably leading to the latter.
This significance of intention in identifying hate speech should be enough to justify the potentially critical contribution that research in linguistic pragmatics, and more specifically discourse analysis, could make towards delineating the term at hand, since pragmatic inquiry by definition revolves around the specification of speaker-intended meaning. After all, research in the field has shown that implicitly communicated meaning can lead to action as much as-and maybe even more than -overtly expressed meaning. This is precisely why any legal deliberation, both within the remit of hate speech/crime laws and beyond, squarely depends on the way in which a judicial body interprets both law and evidence. 4 This brings us to what is probably the thorniest issue in approaching hate speech from a discourse analytic perspective. This would be the discrepancy between the legal understanding of the term and the multiple-and concealed-forms that the expression of hate can take. Taking, for example, the aforementioned Council Framework Decision, one could isolate the criteria qualifying speech as hate speech in the EU as follows: 1. A call motivated by racial/ethnic/national bias; 2. A call for violence; 3. A call punishable by the criminal law of the country where it occurs.
Legally speaking, it is only speech that lies at the intersection of these three criteria that would qualify as illegal, and thus prosecutable hate speech in this context. Still, there could still be cases of inflammatory, offensive comments or comments characterised by prejudice and intolerance that would not meet the threshold provided in the description above. And even though such cases of general disparagement, vilification and abusive language may not be considered hate speech in the legal sense, they arguably still constitute hate speech in that they may have a devastating effect on their recipients on the grounds of moral harassment-which has, for instance, been conducive to suicide on several occasions. 5 In this regard, there seem to be two different categories of hate speech. On the one hand, there is what could be called hard hate speech, which comprises prosecutable forms that are prohibited by law, and on the other, there is soft hate speech, which is lawful but raises serious concerns in terms of intolerance and discrimination. As we will see in the section that follows, the threshold for distinguishing between hard and soft hate speech (especially in relation to protected characteristics) varies from country to country. On top of this, different democracies have altogether different approaches towards regulating and combating hate speech. So, while the USA, at governmental level, gives priority to the protection of the freedom of expression and opinion, many EU member states do invoke measures to 4 Even though their potential role in Social Justice DG programs has not yet been yet acknowledged, forensic linguistics techniques have repeatedly been used in/applied to court cases related to hate speech and sexist, racist discourse (cf. Carney 2014; Olsson and Luchjenbroers 2013;Coulthard and Johnson 2017). 5 For an in-depth overview of the effects of cyberbullying for LGBTQ youth, see Abreu and Kenny (2017). regulate and combat hate speech. Given this volume's motivation and methodological angle then, we will not be addressing the distinction between legal and illegal hate speech here. Rather, we will be focusing on the features of discourse that encompasses a discriminatory attitude as a means of identifying different ways in which hate, broadly construed, is expressed in spontaneous online comments.
Discrimination has been a widely studied topic in discourse-analytic theorising, which investigates the significance of language in the production, maintenance, resistance and change of social relations of power, through mainly the ideological workings of political and media discourse (Fairclough 1989;van Leeuwen and Wodak 1999;Halliday 1989). Through its iteration, discriminatory discourse 'manufactures' assumptions, legitimises dominance and naturalises inequality. Different approaches in discourse analysis such as discursive psychology or critical discourse analysis have developed concepts that can be particularly useful in understanding the relationship between linguistic practices and social structures, and help provide links between language use and processes of social change that take place outside discourse. At the same time, these latter processes have been shown to be substantively shaped by relevant discourses (Chouliaraki and Fairclough 1999: 4). In this respect, discourse analysis is key when it comes to social change, as discourse shapes political decisions and defines what WE are (i.e. who we are and what we can do or not), as well as what is acceptable or not by linguistically attributing characteristics to people, events or practices, and in effect bringing people to accept or at least rationalise the unacceptable (like, for example, the use of metaphors like COCKROACHES or PARASITES when discussing migrants). Fairclough (1989), for example, blends Foucault's (1971Foucault's ( , 1975 formulations of "orders of discourse" and "power-knowledge", Gramsci's notion (1971) of "hegemony" and Althusser's (1971) concept of "ideological state apparatuses" to describe discourse as an accepted flow of common knowledge (discourse) about which we have assumptions (thoughts) and on which we make decisions (actions). In this perspective, a discourse-analytic approach to Othering processes is fundamental for an understanding of the actions taken against minorities, whether these are sexual or social.
At the same time, critical discourse analysis has as its focus the relationship between ideology, inequality, and power through discourse, analysing them on the basis of "opaque as well as transparent structural relationships of dominance, discrimination, power and control as manifested in language" (Wodak 1995: 204). One of its main tenets is that social interaction (partially) takes a linguistic form. This critical approach is distinct from other approaches to discourse analysis in its view of (a) the relationship between language and society and (b) the relationship between analysis and the practices analysed (Wodak 1997: 173). It places the focus on the linguistic features and organisation of concrete instances of discourse, such as the choices and patterns in vocabulary or rhetorical figures (e.g. metaphors, wording), grammar (e.g. transitivity, modality), cohesion (e.g. conjunctions, anaphors, etc.). For example, the use of passive voice in news reporting the deportation of migrants or an assault to a transgender person can have the effect of obscuring the agent(s) of the relevant processes and therefore minimise accountability. Some critical discourse analysts combine (quantitative) corpus linguistics and (qualitative) textual analysis techniques. Their addition of quantitative measures is motivated by the belief that a focus on the distribution of linguistic forms is an empirically reliable means for uncovering the linguistic processes through which Othering is socially materialised, as such quantitative data can help understand the relationship between "social structure and individual subjectivity and the ways in which language mediates between the two" (Levon and Mendes 2017: 15).
Wodak and her associates have also developed the critical and historical discourse analysis strand with the intention of tracing the (intertextual) history of phrases and arguments on a given topic (Wodak 1995;van Leeuwen and Wodak 1999). The method consists in triangulating sources, i.e. in using different documents to analyse the same phenomenon, ethnographic research and analysing news reporting. This triangulation aims to understand a particular phenomenon from different standpoints.
The analyses which follow in the following chapters are mostly based on such discourse analytic approaches. For example, as will become evident in the remainder of this volume, the triangulation methodology has been used as a basis for the research carried out within the CONTACT project. More specifically, taking into account the relevant EU laws on discriminatory discourse and hate speech, we analysed comments posted on main news portals, and carried out interviews and administered questionnaires so as to understand the public perception of discriminatory statements with a view to reaching a broader understanding of the kinds of Othering discourses that are circulated in the European space.
Since this volume focuses on the EU, however, it seems necessary to first briefly outline some of the differences that countries that are represented in the C.O.N.T.A. C.T. project exhibit in their understanding and regulation of hate speech issues, before moving on to the particularities of the online setting as a locus for the expression of hate.

Natalie Alkiviadou
Notwithstanding the perplexities associated with defining hate speech as a result of the free speech debate, the EU managed, after seven long years of negotiations (European Commission 2014: 1), to take a major leap forward in 2008 with its Framework Decision on Combatting Racism and Xenophobia through Criminal Law (Council of the European Union 2008). As is reflected in its title, this is not a document dealing with hate speech per se but, instead, with some of the phenomena underlying such speech. However, it was hate speech that kept the negotiations going for so many years and, particularly, the significant divergences in the legal traditions of EU member states vis-à-vis free speech (European Commission 2014: 1). These varying understandings of hate speech also mean that, regardless of the Framework Decision at the EU level, there is little coherence amongst EU member states on the definition of hate speech. To this end, in February 2017, the European Parliament put forth a motion for a resolution on establishing a common legal definition of hate speech in the EU (European Parliament 2017).
In light of this, this section will consider the main characteristics of the legal frameworks of the ten countries participating in the C.O.N.T.A.C.T. project. 6 This will allow us to see how hate speech is approached on a decentralised (member-state) level and determine possible convergences and divergences amongst the member states themselves. Before moving on, however, it is worth noting that the term 'hate speech' is not found in any of the legislations of the C.O.N.T.A.C.T. project partner countries; rather, all these countries transposed or acceded to the United Nation's ICCPR (UN General Assembly 1966) and ICERD (UN General Assembly 1965), with the UK making a reservation to the relevant articles on the grounds of free speech. As will be demonstrated below, regardless of the ratification or accession to the aforementioned UN documents, the transposing laws are not the ones habitually relied upon to tackle hate speech. A relevant example is Denmark, where a court was faced with the statement 'negroes are less intelligent than Europeans', which falls within the framework of statements pertaining to racial superiority, prohibited by the ICERD; yet, this was deemed to be permissible speech, as it was made as part of a political debate. 7 With this in mind, we can now turn to the legal provisions of each C.O.N.T.A.C.T. partner country in alphabetical order below.
The main anti-hate speech legislation in Cyprus is The Combatting Certain Forms and Expressions of Racism and Xenophobia by means of Criminal Law 134 (I) of 2011, which transposed the Framework Decision into national law. Cyprus chose to incorporate the provision of punishing only conduct which is either carried out in a manner likely to disturb public order or which is threatening, abusive or insulting. Cyprus went a step further from the protected characteristics of the supra-national level and also passed Law 87 (I)/2015 amending the Criminal Code. This amendment incorporates Article 99A into the Criminal Code, which punishes hate speech targeted at a person or person's sexual orientation or gender identity. In sum, there is no explicit definition of hate speech in Cyprus but, instead, a transposition of supra-national documents which offer their own appraisals of hate speech and which set out varying thresholds. This results in a discordant legal setting which, nevertheless, has the positive feature of going beyond the hierarchy of hate embraced by the supra-national framework by incorporating the grounds of sexual orientation and gender identity as protected characteristics in the sphere of hate speech. Still, the above legislation has not yet been used in Court and there is no national case-law relevant to the issue of hate speech. 6 It should be noted that the information provided in this section in relation to each member state's national context has been synthesised from the desktop research conducted by C.O.N.T.A.C.T. partners in each member state during the first stages of the project, rather than this section's author.
In Denmark, hate speech is connected to Section 266b of the Danish Penal Code which criminalises expressions that "publicly or with intent to disseminate to a wider circle, threaten, insult or degrade a group of persons on the basis of race, skin colour, nationality, ethnicity, faith or sexual orientation". Evidently, this definition is more extensive than its supra-national counterparts, as it includes grounds such as sexual orientation. Important to this understanding of hate speech is that expressions must be made publicly or with an intention to disseminate to a wider circle, and, therefore, private conversations do not fall within the prohibited sphere. Unlike Cyprus, Denmark has relevant case-law which, inter alia, sheds light on the meaning of terms used in Section 266b. For example, the statement 'coloured people like you are not allowed in my parents' apartment' which was uttered in a nursing home, was not considered by a District Court to be punishable, as the nursing home was deemed as not constituting a public place. 8 In Greece, the main national legislation is Law No 927/1797 on punishing acts or activities aimed at racial discrimination, as amended by Law 4285/2014 that implements the Framework Decision. Article 1 deals with public incitement to violence, hatred or discrimination against a person or group of persons due to their race, colour, religion, status, ethnic origin, sexual orientation, gender identity or disability if this poses a danger to public order or constitutes a threat to the life, liberty or physical integrity of the person or persons involved and is punished with a prison sentence ranging from three months to three years and with a monetary fine of five thousand to twenty thousand euros. The scope of protected characteristics of this law is, together with Lithuania and Spain, discussed below, one of the most extensive in the C.O.N.T.A.C.T. partner countries, incorporating grounds such as disability, which is not found elsewhere. While there have been several relevant cases before Greek courts, one characteristic example which demonstrates a threshold that needs to be met, in terms of the impact of the speech and its publicity, involved a Golden Dawn member. In this case, the defendant stated on camera that 'we are ready to open the kilns. To make soaps. Not for the people, since … we may fall sick …' These were some of the phrases he used to refer to migrants. The court decided that, even if these phrases were exaggerations, they demonstrated the accused's intention publicly to provoke people to cause harm to migrants, so that the rest of them would be convinced to abandon Greece. 9 The main relevant Italian Law is Law 205/1993 which makes it a crime to "propagate ideas based on racial superiority or racial or ethnic hatred, or to instigate to commit or commit acts of discrimination for racial, ethnic, national or religious motives." The law also punishes those who "instigate in any way or commit violence or acts of provocation to violence for racist, ethnic, national or religious motives." Although there are no strict thresholds to meet, such as public order, as is the case of Cyprus for example, Italy limits itself to the protected characteristics of ethnicity and religion, as provided for by the supra-national level.
In Lithuania, the central provision dealing with this issue is Article 170 of the Criminal Code entitled 'Incitement against Any National, Racial, Ethnic, Religious or Other Group of Persons.' This article punishes the handling or distribution of impugned material and expression, which incites hatred, violence, discrimination or contempt for a person or persons belonging to a group defined by sex, sexual orientation, race, nationality, language, descent, social status, religion, convictions or views. This definition is particularly broad including grounds such as sex but also convictions, which are not necessarily affiliated with religion. Its threshold is also low, with discriminatory expression also falling in the net of prohibited expression. Interestingly, in relation to the punishment of expression (rather than material), the article also renders ridiculing expression a punishable offence. It also punishes a person who publicly incites violence against a person or persons of a particular group. To give an example from case law, a defendant was found guilty for publicly mocking a person of Asian origin in front of others with obscene epithets saying that 'foreigners are not welcome here.' 10 This demonstrates the low threshold necessary in Lithuania for finding speech hateful.
The central provision in Malta is Article 82 of the Maltese Criminal Code, which punishes any person who uses any threatening, abusive or insulting words or behaviour, or displays any written or printed material which is threatening, abusive or insulting or otherwise conducts himself in such a manner, with intent to stir up violence or racial hatred against another person or group on the grounds of gender, gender identity, sexual orientation, race, colour, language, ethnic origin, religion or belief or political or other opinion.
The protected characteristics are also broad in Malta, although not as broad as, for example, Greece, which also incorporates the grounds of disability, Lithuania, which also includes sex or as Romania and Spain discussed below.
In Romania, Article 369 of the Criminal Code prohibits "public incitement by any means, hatred or discrimination against a class of persons." Order 137 of 2000 sets outs the protected characteristics which are race, nationality, ethnicity, language, religion, social, belief, sex, sexual orientation, age, disability, non-contagious chronic disease, HIV infection and membership of a disadvantaged group. This is the only country to incorporate HIV positive persons as protected by hate speech legislation and which incorporates a broad ground of disadvantaged groups. Moreover, by incorporating discrimination, the threshold of prohibition remains low.
As for Spain, although, like for other countries, there is no legislative definition of hate speech, the Constitutional Court held that hate speech is a "heavy burden of hostility that incites, directly or indirectly, violence by way of humiliation." 11 The main piece of legislation is Article 510 of the Criminal Code on the incitement to hate crime, violence and discrimination. This punishes those who provoke discrimination, hate or violence against groups or associations due to racist, 10 Criminal case No. 1A-407-337/2009, Panevėžys district court. 11 The Constitutional Court in its STC 176/1995 (Case Makoki).
1.2 Regulating Hate Speech in the EU anti-Semitic reasons or any other reasons related to ideology, religion or belief, family situation, belonging to an ethnic group or race, national origin, gender, sexual preference, illness or handicap. The grounds for protected characteristics in Spain are extensive and the thresholds low, incorporating, for example, discrimination and not requiring, for example, the disturbance of public order.
Turning to the UK, the Public Order Act 1986 provides that acts intended or likely to stir up racial hatred include the use of words or behaviour or display of written material, the publishing or distribution of written material, the public performance of plays, the distribution, showing or playing of a recording and/or the broadcasting of a programme in a cable programme service. The offence of stirring up religious hatred has been defined and incorporated into the 1986 Public Order Act by the Racial and Religious Hatred Act 2006, with Sections 29B-F of the latter addressing the issue of stirring up religious hatred in the same way as it does its racial hatred counterpart. However, in relation to religious hatred, Section 29J of the Racial and Religious Hatred Act stipulates that nothing in this Part shall be read or given effect in a way which prohibits or restricts discussion criticism or expressions of antipathy, dislike, ridicule, insult or abuse of particular religions or the beliefs or practices of their adherents, or of any other belief system or the beliefs or practices of its adherents, or proselytising or urging adherents of a different religion or belief system to cease practising their religion or belief system. Therefore, in relation to religious hatred, the threshold is higher, since expression such as insulting a particular religion is deemed permissible.
From the above approaches to hate speech and the variations therein, it could be argued that, although some common elements can be discerned, "hate speech seems to be whatever people choose it to mean" (Kiska 2012: 110) As we have seen in the previous section, at the supra-national EU level, protected groups are limited to ethnic and religious groups, demonstrating an adoption of a hierarchy of hate in such arenas, with some characteristics perceived as simply being more important than others. At the national level, countries such as Lithuania, Romania, Spain and Malta have an extensive conceptualisation of protected groups whilst others such as Italy limit themselves to those set out by the UN and the EU. The thresholds of what is considered prohibited speech also varies amongst countries, with Italy having a lower threshold, prohibiting, for example, ideas of racial superiority, and Cyprus incorporating safety nets such as the impact of public disorder. On a last but important note, these conceptual variations of definitions render effective challenging of online hate on the borderless medium known as the internet particularly complex.

Hate Speech in the Online Setting César Arroyo López and Roberto Moreno López
Following the technological revolution that began in the 1960s, the ever-growing expansion of the internet since the 1990s has had considerable impact across the globe. Ultimately, we have gone from a system of information transmission dominated by the mass media, state and lobbies, to a knowledge society where citizens are not just information transmitters themselves but can also assume a more active role, as creators and co-creators of new content. In the online world, a place of global relations characterised by a dilution of space-time limitations, anyone with online access can offer their opinion, contribute to dialogue and put forth their knowledge and perceptions for the gestation of modern culture or "cyberculture" (Sacristán 2013: 126). It is thus hard to dispute that the rapid expansion of the internet has impacted and continues to impact societies at a micro-, meso-and macro-scale.
Communication, including the production and sharing of information content, is one of the core features of the internet. Yet, this type of digital communication is marked by a number of particularities: the internet is a space that provides users with the capacity for expressing their views and communicating without limits, and typically (though not always) without control; the online setting makes it easy for users to hide their identity (in whole or in part) and, in some cases, even to hide their location and activity. As de Salvador Carrasco discusses, this anonymity is "the ability to perform any access, communication or publication in the network without third parties having the possibility to identify or locate the author of said action," although it is also true that such anonymity can only become a possibility through the implementation of specific strategies and tools usually not known to most educated laymen who use the internet (de Salvador Carrasco 2012: 2). Still, even though most of the public communication that is produced online is essentially traceable in origin, most users perceive the internet as a platform where they can express themselves freely and anonymously. Interestingly, research conducted by Childnet International in over 68 countries revealed that the experience of anonymous communication is one of the elements most sought after by young people, to such an extent that they feel that the anonymous use of the internet should be safeguarded, despite its potential dangers (Childnet 2013).
These characteristics of the worldwide web have encouraged a breeding ground for the phenomenon of cyberhate, understood (in a non-restrictive way) as any use of electronic communications technology to spread anti-Semitic, racist, bigoted, extremist or terrorist messages or information. These electronic communications technologies include the internet (i.e., web-sites, social networking sites, 'Web 2.0' user-generated content, dating sites, blogs, online games, instant messages, and e-mail) as well as other computer -and cell phone-based information technologies (Anti-Defamation League 2010: 4).
Hence, due to its global, immediate and participatory nature, the internet has become a space for both the expression and dissemination of intolerant ideas and beliefs (Isasi and Juanatey 2016), offering an additional means of facilitating the advocacy and spread of discrimination that can potentially even lead to hate crime. Such attitudes and their expression reject difference and intend to deprive persons and groups of their dignity by denying and attacking their identity. It is these intolerant attitudes that constitute one of the main manifestations of hate speech as a social phenomenon, at least as far as the research reported in this volume is concerned. Such soft hate speech as spread online can have a devastating effect on the fabric of social order, as it potentially not only negatively affects the groups or individuals that it targets; it also negatively impacts those who speak out for freedom, tolerance and non-discrimination in our open societies and has a chilling effect on the democratic discourse on online platforms (European Commission 2016b: 1).
For example, a recent report in Spain (Ministerio del Interior 2016) pointed out that, in the year 2016 alone, the Criminal Statistical Service identified 123 cases related to hate speech which were passed on to law enforcement bodies, with more than 75% of these cases occurring on the internet or other ICT platforms. In a similar vein, the Proxi Observatory analysed almost 5000 comments in three major digital newspapers in Spain and concluded that more than half of the user comments that appeared in response to news reported therein were intolerant in character (Cabo et al. 2015: 16-23). All this was occurring at the same time when both the internet and social networks were being used in Spain for explicit incitement to violence against people on the basis of both their ethnic group (e.g. El Diario.es 2016) and their sexual orientation (e.g. elPeriodico 2016).
Of course, this situation is not exclusive to Spain. Similar examples that can be found in most countries around the globe suggest that intolerance and hate can flourish on the internet, taking advantage of its very nature (Gagliardone et al. 2015). And even though the 'terms of service' of most relevant platforms, such as Facebook, Yahoo! or Twitter do stipulate that it is prohibited to post content that is "unlawful, harmful, libellous, vulgar, defamatory, obscene, tortuous, invasive of one's privacy, hateful, or racially ethnically or otherwise objectionable" (Cohen-Almagor 2015: 163), the time it usually takes to remove such content has been an issue of growing concern. This has recently led the EU Commission and various social media giants to agree on a Code of conduct specifically targeting illegal hate speech online (European Commission 2016b).

The C.O.N.T.A.C.T. Research Workstream Stavros Assimakopoulos, Fabienne H. Baider and Sharon Millar
Having justified the focus of the present volume on online discourse in the EU, it is now time to turn to the research on which it reports. As we will see in the following chapter, which outlines the C.O.N.T.A.C.T. methodology, the basic source of data for the more substantial part of our research was comments posted online in reaction to news reports related to migrants and members of the LGBTIQ community. While a comparative discussion of the results obtained in the different national contexts is beyond the scope of the present volume, it seems necessary at least provide a quick reference to our collective results so as to see whether discriminatory discourse is an issue to look out for in the countries of the C.O.N.T.A. C.T. consortium. As is evident from Table 1.1 above, which provides an overview of the results obtained through our analysis of the relevant comments, it certainly seems that both homophobia, and to a far greater extent, xenophobia are quite prevalent in the EU. With the sole exception of Malta, where comments that view members of the LGBTIQ community in a positive light outnumber comments that reveal a negative disposition towards this group, all other national corpora show that the commenter's attitude towards both groups that were researched is more negative than positive.
It is against this backdrop that the analytical chapters, which follow the methodological overview in Chap. 2, are to be understood: Chap. 3 deals with the analysis of online comments to news reports across a number of EU countries, while Chap. 4 discusses some of our findings regarding the folk perception of hate speech on the basis of a qualitative analysis of interviews that several C.O.N.T.A.C. T. partners conducted with members of the general population. Since, as we have already noted, the aim of this volume is to offer a panorama of the strategies most commonly used to express what we have termed soft hate speech as well as an overview of topics central to the way in which the general public perceives such speech, the remarks put forth in each section of the analytical chapters are far from conclusive; yet, they should be enough to justify the usefulness of insights from linguistic pragmatics and discourse analysis when it comes to the analysis of hate speech. And while the discussion of each topic therein is based on data collected in a particular country's context, it should easily become clear to the reader that it also applies to the discussion of hate speech, broadly construed, transnationally.