- Email: [email protected]
The sociocognitive dimension of hate speech in readers’ comments on Serbian news websites
Discourse, Context & Media 33 (2020) 100366
Contents lists available at ScienceDirect
Discourse, Context & Media journal homepage: www.elsevier.com/locate/dcm
The sociocognitive dimension of hate speech in readers’ comments on Serbian news websites Jasmina P. Ðordevic´ ´ irila i Metodija 2, 18100 Niš, Serbia Faculty of Philosophy, University of Niš, C
a r t i c l e
i n f o
Article history: Received 24 April 2019 Received in revised form 29 November 2019 Accepted 29 November 2019
Keywords: Readers’ comments Serbian news websites Hate speech Discourse analysis Sociocognitive dimension
a b s t r a c t Comment sections on news websites bear no immediate threat of sanctions and users feel free to express open criticism or insult. Other than causing offense and undermining the dignity of readers with a different opinion, a rather dangerous outcome of hate speech is that it affects the sociocognitive interface between discourse and society in the sense that it has a negative impact on the readers’ cognitive notions of information, beliefs and knowledge. Based on the analytical approach proposed in Sociocognitive Discourse Studies, the study examines 939 comments posted in the comment section of the Serbian news website Politika. Besides confirming that hate speech in the readers’ comments is a rather frequent means of establishing a communicative Common Ground among news readers, the analysis shows that hate speech is a reflection of the negative impact of news on the sociocognitive interface between discourse and society. The research also demonstrates that Sociocognitive Discourse Studies may be applied to the analysis of readers’ comments as a specific type of written discourse that reflects the mental models news readers create when exposed to media discourse. Ó 2019 Elsevier Ltd. All rights reserved.
1. Introduction The role of comment sections on news websites has recently become a recurring focus of scholarly research in the field of media discourse analysis (Barnes, 2018; Bruce, 2018; KhosraviNik & Zia, 2014; Tenenboim & Cohen, 2015; Toepfl & Piwoni, 2015; Weizman & Dori-Hacohen, 2017; Yeo et al., 2019). Initially introduced to encourage public discussion and monitor reader engagement, comment sections often become a platform of derogatory and pejorative language rather than serving the purpose of civil debate. Such discourse necessitates closer research. Readers’ comments reflect a reaction to news content, thus creating a specific interface between a news article and its readership (Weizman & Dori-Hacohen, 2017). This interface is dominated by a dimension that relates discourse structures to social structures and shows the immediate effect of this relationship (Bruce, 2018; Toepfl & Piwoni, 2015). Thus readers’ comments demonstrate how the information presented to them in the news affects their attitudes, ideologies and language use. Critical Discourse Studies (CDS) relates the structural properties of text or talk to social structures, enables both the analysis of the discourse structure of a comment and its context and provides the grounds for an explicit viewpoint (van Dijk, 2018). However, it lacks the sociocognitive
E-mail address: [email protected] https://doi.org/10.1016/j.dcm.2019.100366 2211-6958/Ó 2019 Elsevier Ltd. All rights reserved.
dimension to relate the discourse structure to social structures via a sociocognitive interface which is needed to critically describe the cognitive aspects of concepts used by readers (2018). Assuming that particular news provoke a specific response, an important aspect to consider is that of newsworthiness, i.e. news values which are criteria applied by news workers when creating news (Bednarek & Caple, 2014). Any set of news values constructed by journalists is expected to attract readers’ attention, instigate meaning and motivate action, such as posting a comment referring to a specific news item. Given that readers’ comments are a reflection of the interaction between discourse structures and social structures via a sociocognitive dimension, this study refers to the theory of Sociocognitive Discourse Studies (SCDS) proposed by van Dijk (2016, 2018). The sociocognitive approach establishes the discourse-cognition-society triangle and it provides the relevant tools to explore the cognitively mediated relations between discourse and society (van Dijk, 2016) that comments reflect. Since discourse structures and social structures are different, a relationship between them can be established only through the mental representations of language users who are both individuals and social members. Their ideas, knowledge and opinions are an interpretation of the mutual relationship between social structures and discourse structures. Social structures influence discourse structures through people’s interpretations of the social environment they are part of and vice versa, discourse structures can only
2
J.P. Ðordevic´ / Discourse, Context & Media 33 (2020) 100366
influence social structures through the same cognitive interface of mental models, knowledge attitudes and ideologies (2016). The present research is based on a corpus of comments posted by readers on the website of the Serbian news outlet Politika. The aim is to analyse the comments from the point of view of SCDS so as to examine how hate speech (HS) in them is affecting the sociocognitive interface between discourse and society. An additional analysis of the linguistic devices and news values (Bednarek & Caple, 2014) used in the articles that generated the comments will show what news values provoked the readers to post comments. The hypothesis is that HS in readers’ comments on news websites indicates a negative impact of news on the readers’ cognitive notions of information, beliefs and knowledge. Based on a quantitative and a qualitative analysis of the compiled corpus, two objectives will be pursued: (1) examine whether HS in the readers’ comments is a means of establishing a communicative Common Ground among news readers and (2) find evidence that HS in readers’ comments is a reflection of the negative impact of news on the sociocognitive interface between discourse and society.
2. Hate speech as an online discursive practice Despite the lack of a universal definition of HS, a significant number of authors put HS in the centre of their attention (Brown, 2018; Cammaerts, 2009; Lillian, 2007; Vollhardt et al., 2006). HS is assumed to reflect abusive and harassing expressions of violence or discrimination, directed against people on the basis of their race, ethnic origin, religion, gender, disability, sexual orientation, etc. and it is mostly meant to silence opponents, or at least prove their opinions invalid. Critical Discourse Studies (CDS) (Wodak & Meyer, 2016; Fairclough, 2003) has relied on the idea that mass media are ‘‘an obvious powerful site where discursive power is exploited to (re)construct and (re)define social realities” (KhosraviNik & Esposito, 2018, p. 54) and as such it has been preoccupied with the notion of the power of discourse in affecting and shaping society and socio-political orientations (Wodak & Meyer, 2009). However, the development of participatory media and social networks as a new communication environment has resulted in a substantial increase of online discursive practices, HS being a prominent one. This new development has instigated the necessity to (re)evaluate CDS and its applicability to the concept of HS as an online discursive practice which is why CDS should pursue new interdisciplinary synergies with other disciplines (KhosraviNik & Esposito, 2018). As stated by KhosraviNik and Zia (2014), participatory media provide a space for ‘‘non-elite text producers resisting the discourses of established institutions” (p. 757). Thus the participants in the digital space create a valuable data repository which can be compiled from discourse found in the comment sections of shared news and posts on social networks. The quite excessive amount of data compiled from online communication is difficult to manage which is why KhosraviNik and Zia (2014) suggest ‘‘down-sampling” it in order to perform a ‘‘qualitative detailed textual analysis” and to keep it ‘‘under control for CDA contextualisation purposes” (p. 764). Being a discursive practice occurring quite frequently in participatory media, HS may be analysed based on the reactions and debates among readers and users in the interactive digital media initiated by a piece of news, social network post or similar content published online. Given the necessity to explore how the media may shape and influence society, KhosraviNik (2017) introduces Social Media Critical Discourse Studies (SM-CDS) which is ‘‘not only interested in what happens in media per se but in how it may shape and influ-
ence the social and political sphere of our life worlds” (p. 586). SMCDS may be used to investigate ‘‘what goes on online in such a way that the availability of technological affordance per se is assumed to be the ultimate sufficient context and force for actual social, political, and cultural change in society” (p.586). KhosraviNik further suggests that all media practices and the content presented in the media ‘‘should be interpreted within a wider sociopolitical context of a given society” which is where ‘‘a social, historical, cultural, psychological, or political account is provided for explication of the discourse under investigation” (p. 586). In the context of the present research, digital performances, online hostility being a prominent one, may be interpreted by including the sociocognitive dimension proposed by van Dijk (2016, 2018). 2.1. Hate speech in the context of newsworthiness The theory of newsworthiness (Harcup & O’Neill, 2017) may be applied to a corpus of news to confirm what factors influence news selection in a certain audience. An important criterion determining the choice of a particular news article are the news values which are discursively constructed and are defined as the ‘‘newsworthy aspects of actors, happenings and issues as existing in and constructed through discourse” (Bednarek & Caple, 2014, p. 138). Since the HS in readers’ comments is a reaction to a particular news items, certain news values are more provocative than others. Among the various devices based on which news values may be constructed, linguistic devices are most easily to apply within a specific discursive approach. Since the proposed approach in this article is SCDS (van Dijk, 2016, 2018), news values are the socially-shared cognitive representation while linguistic devices are the tools for construing news values that provoke HS in readers’ comments. Bednarek and Caple (2014) suggest a list of key linguistic devices (e.g. negative lexis, quantifiers, intensifiers, evaluative language, etc.) which may be used to analyse news values, such as negativity, proximity, superlativeness, prominence, impact, etc. The list is not universal and it may be further adapted. 2.2. Hate speech in the context of the European Union The EU General Policy Recommendation No. 15: Combating Hate speech from March 2016 states that HS includes ‘‘denigration, hatred, vilification, harassment, insult, negative stereotyping, stigmatization, threat, etc. which are based on a non-exhaustive list of personal characteristics or status that includes ‘race’, colour, language, religion or belief, nationality or national or ethnic origin, as well as descent, age, disability, sex, gender, gender identity and sexual orientation” (Zubcˇevic´, Bender, & Vojvodic´, 2017. p.11). However, regulations regarding HS also implicate that laws on HS in national contexts are subject to objections (Langton, 2016). Restrictive laws on HS undermine the concept of freedom of expression, a right guaranteed in all democratic countries which is why European Union legislation has been established to clearly define HS as opposed to the violation of freedom of speech. The main criteria that the European Court of Human Rights relies on when deciding whether a certain form of expression shows either one or the other are ‘‘the purpose pursued by the applicant as well as the content of the expression and the context in which [hate speech] was disseminated” (Weber, 2009, p. 33). Additional conventions and declarations have been stipulated to support lawful and just handling of HS without the risk of violating freedom of speech (Zubcˇevic´ et al., 2017). The implications of regulations regarding HS need more research because it depends on far too many different factors that
J.P. Ðordevic´ / Discourse, Context & Media 33 (2020) 100366
need to be explored and explained. HS is a product of a specific historical and cultural context and as such the content and meanings are related to that context. What is HS in one context may not be considered HS in another one (Parekh, 2012) which is why legislators, linguists, social actors, political actors and other participants in society, politics and culture need to explore the concept of HS from various points of views to determine specific circumstances, occurrences, forms and expressions with respect to a specific context and time. 3. Comment sections on Serbian websites The media landscape in Serbia is specific because of Serbia’s application to join the European Union. About 2000 media outlets are currently registered in the Serbian Business Registry Agency and they often depend on financial support from the government so that they are exposed to political influence (Ðordevic´, 2019; Matic´, 2014; RWB, 2018). The European Commission performs an annual in-depth scrutiny regarding EU values, freedom of speech being crucial in this respect. The reports confirm that there is no progress regarding the freedom of expression and that this is a matter of increasing concern (Annual Progress Report of the European Commission for Serbia, 2019). The same concern was expressed by Reporters Without Borders (RWB) who stated that Serbia ‘‘utterly fails to meet EU press freedom standards” (RWB, 2018). Online news outlets with daily news and with a major influence in Serbia include blic.rs, telegraf.rs, b92.net, novosti.rs, n1info.com, danas.rs and politika.rs whereby only n1info.com and danas.rs are perceived as neutral (the number of average daily real users is available on Gemius Audience, 2018). All the listed outlets have comment sections on their websites and readers use the possibility to post their comments in large numbers. Comment sections may not always reflect opinions of actual readers as a significant number of bots are employed by the government to comment news on a daily basis. According to unofficial resources (Kulacˇanin, 2018) the Serbian Progressive Party employs more than 3000 people who are part of the so-called Internet team. Their job description includes commenting news on various websites, posting opinions on social networks, sharing news and posts, tweeting and retweeting tweets, reacting to comments and posts by clicking the relevant option provided by a news website comment section or by clicking the appropriate emoji below a post on social networks. Topics that are most often commented on refer to government officials’ activities. The Serbian President’s activities dominate both the news and the comment sections. The language used in comments ranges from polite and civilized to spiteful and offensive so that HS is a common occurrence on all Serbian news websites. Readers openly show disrespect of other readers’ opinions, argue about facts, accuse each other of lying and misrepresenting ideas, pass insults and use sarcasm with the aim to diminish opinions. Therefore, comment sections reflect personal frustration and aggression rather than public opinion and civil discussion. 4. The sociocognitive dimension of discourse analysis As stated in the introduction, Sociocognitive Discourse Studies (SCDS) ‘‘relates discourse structures to social structures via a complex sociocognitive interface” (van Dijk, 2016, 2018). The main subjects of SCDS are the ongoing communicative Common Ground, the shared knowledge and the attitudes and ideologies of language users who are participants of a particular communicative situation and members of social groups and communities (van Dijk, 2018). SCDS makes ‘‘explicit the fundamental role of mental representations” and it ‘‘shows that many structures of discourse itself can
3
only (completely) be described in terms of various cognitive notions” (p. 28). Therefore, SCDS relies on ‘‘explicit psychological theories of mental representations of journalists or other language users” as well as on ‘‘the ways these models mediate between shared social cognition (knowledge, attitudes, ideologies), social structures and actual text and talk” (p. 28). In addition, SCDS includes the cognitive interface between discourse and society. The sociocognitive approach does not accept that there is a direct link between the structures of discourse and society, but supports the notion that there is a cognitive interface where the two meet and interact. The result of the interaction between discourse and society via the cognitive interface constitutes the sociocognitive dimension of the sociocognitive approach to discourse analysis. This means that the sociocognitive approach assumes that structures of society can only ‘‘affect text and talk through the minds of language users” because ‘‘social members represent both social structures as well as discourse structures in their minds, and thus are able to relate these mentally before expressing them in actual text and talk” (van Dijk, 2018, p. 28). Given the limited scope of this research, only a brief review of the most important notions about SCDS will be provided here. First, the notions of mind, memory and discourse processing are crucial since ‘‘cognitive processes take place in the mind or memory of individual social actors as members of social groups and communities” (van Dijk, 2018, p. 29). The way discourse is produced and comprehended is related to the way linguistic and discursive knowledge systems are applied by language users so that the processing of lexical items, syntactic structures and various meanings of discourse is closely related to grammar, the lexicon, semantics and different forms of interaction. Since local and global meanings attributed to discourse are based on underlying mental models as subjective representations of events or situations, a crucial aspect of the sociocognitive dimension is how the production and comprehension of discourse will involve the particular expression of a mental model regarding an event, person, opinion, belief, etc. The direct communicative intention of discourse is mainly to transmit the mental model of a speaker or writer. However, the way the hearer or reader will understand the transmitted discourse depends on how they will interpret it based on their own mental models which can be ‘‘individual, personal, subjective and multimodal” (van Dijk, 2018, p. 30). These models go beyond the subjective representation of a situation or event and they include opinions and emotions of both the ones who create discourse and those who read or hear it. The models reflect the social cognition of the members of the linguistic and discursive communities who share their knowledge of language and discourse, their sociocultural knowledge of the world as well as attitudes, ideologies, norms and values. To conclude, the sociocognitive dimension that SCDS relies on constitutes a complex interface that establishes a common space between discourse and society where different elements interact. These elements combine various forms of social cognition (attitudes, ideologies, norms and values based on sociocultural knowledge of the world and knowledge of language and discourse) that is being created in that commons space. At the same time the social cognition of the members of linguistic and discursive communities (social actors, news readers, participants in events, etc.) influences and shapes that very same space. 4.1. Discourse structures indicating hate speech in readers’ comments Being a product of discourse that reflects the knowledge, opinions and beliefs of readers who directly react to the discourse in the news, readers’ comments are in fact a result of the interaction between discourse and society via the sociocognitive interface (van
4
J.P. Ðordevic´ / Discourse, Context & Media 33 (2020) 100366
Dijk, 2016, 2018). Readers’ comments on news websites reflect the mental models that readers as members of society and discursive communities create as a reaction to the discourse they have been exposed to. Therefore, readers’ comments are discourse that comprises discourse structures based on which the sociocognitive dimension can be identified, described and analysed. The identification, description and analysis of discourse structures proposed by van Dijk (2018) that may indicate HS include: stress, intonation, word order; meanings of words, sentences or sequences of sentences; coherence; opinion and emotion words; global topics or themes; deictic or indexical expressions; speech acts; evidentials; conventional, schematic, canonical structures; metaphors and ideological polarizations. The identification of discourse structures is based on lexical items that are identified as individual instances referring to a certain discourse structure within a broader context. For instance, lexical items may refer to social actors (e.g. politicians, victims, attackers, police, government officials, etc.) or their representation in a context (e.g. individual, collective or impersonal). The mentioning or description of a social event may refer to a global topic or event (e.g. demonstrations, riots, elections, etc.), while lexical items indicating evaluative and emotional representations may refer to ideological polarizations (e.g. verbs such as urge, warn, state, etc., or adjectives such as important, imperative, dangerous, etc.). The identified discourse structures serve the purpose of both a quantitative and a qualitative analysis enabling the determination of the extent to which HS occurs in a particular discourse and the broader analysis of the mental representations and the socioculturally shared knowledge. The outcome is that the relationship between discourse and society is both quantified and qualified through the sociocognitive interface between them. 5. The research 5.1. Methodology and corpus The present research is based on the hypothesis that HS in readers’ comments on news websites reflects the way news affects the readers’ cognitive notions of information, beliefs and knowledge. It is assumed that users feel free to express open criticism or insult so that derogatory and pejorative language in comments may be expected to indicate that news affects the sociocognitive interface between discourse and society in a negative way. Since it is obvious that certain news are more provocative than others, the linguistic devices that occur in the article may indicate what type of news values provoke the readers to comment the news item. Therefore, the first objective of this research was to confirm that HS in the readers’ comments is a rather frequent means of establishing a communicative Common Ground among news readers. In addition, a list of key linguistic devices used in the article to construe news values has been identified so as to determine the types of news values that provoked the readers’ comments. The second objective was to find evidence that HS is a reflection of the negative impact of news on the sociocognitive interface between discourse and society. Starting with the assumption that readers freely express criticism or insult in comment sections when reacting to news articles (Tenenboim & Cohen, 2015), the textual analysis of the compiled corpus of comments was expected to indicate the role of comments as a reflection of the sociocognitive dimension (KhosraviNik & Zia, 2014; van Dijk, 2018) existing between the news values in the discourse presented in the news (Bednarek & Caple, 2014) and the readers as members of society. In order to identify the discourse structures that reflect the sociocognitive dimension in the selected comments, the analysis relied on the identification of lexical items explicitly reflecting instances of HS.
The corpus compiled for this research comprises 939 comments written in response to 77 news articles published on politika.rs (the website of the news outlet Politika) on 14 December 2018, the day after Priština announced the transformation of the Kosovo Security Forces (KSF) into a regular army. On average, the comments consist of 40 words, the shortest one consisting of one word only and the longest of 200 words. The website’s comment section poses a limit of 1000 characters per comment which is why the maximum length of 200 words is considered imposed as some readers might have wanted to write longer comments. The website politika.rs was used as a source for the corpus because Politika is the oldest newspaper in Serbia with the longest tradition of news publication and a circulation of the print edition estimated at around 100,000 copies a day. The website’s readership (Gemius Audience, 2018) counts almost 50,000 average daily real users and the profile of the readership includes mainly highly educated middle-class members, most of whom have been loyal to Politika for decades. In comparison to politika.rs, other influential news websites (e.g. blic.rs, telegraf.rs, b92.net, novosti.rs), despite a rather high number of readers, have not been chosen for this research because the administrators of their comment sections quite frequently tolerate profanity and vulgar language. Since the aim of this research was to explore how news affects the readers’ cognitive notions of information, beliefs and knowledge, examples of profanity and vulgar language would not have provided a relevant corpus. Two other highly influential news outlets in Serbia (n1info.com and danas.rs) were not chosen for the research because they are considered anti-government and a large number of bots post comments in their comment sections. This fact makes it difficult to distinguish real comments from the comments posted by bots. The comments were first read, coded, analysed and categorized (KhosraviNik & Zia, 2014). Then an analysis of the linguistic devices indicating news values in the articles was performed (Bednarek & Caple, 2014). A detailed presentation of all the news values and linguistic devices identified in the 77 articles would exceed the scope of this article; therefore, only the most frequent values and devices discovered in the articles will be presented here in English (Table 1). As can be seen, the most frequent news values and linguistic devices could be identified mainly in the articles from the sections Politics, World, Society and Region. Since this section is meant to introduce the corpus analysed in this research, the implications of this analysis will be presented within the qualitative analysis in Section 5.2.2. The analysis of explicit HS in the comments revealed a total of 182 instances in the entire corpus (Table 2). Lexical items containing sarcastic remarks or ironic statements were not marked as instances of HS because the sarcasm and irony identified in the comments did not explicitly reflect a clear intention of hatred, offense or abuse. Similar to comment sections on other websites, politika.rs also includes a mechanism to measure reader engagement. The readers can click one of two options below each comment: a) ‘reply’ encouraging the readers of comments to engage in discussion with the authors of the comments and b) ‘recommend’ based on which readers can show agreement with the opinion expressed in a particular comment. In the corpus analysed here (Table 2), a total of 205 direct replies to separate comments and a total of 2255 recommendations could be identified whereby the total number of minimum recommendations of each comment (0 or 1) was 131 and the total number of maximum recommendations of each comment (ranging from 3 to 231) was 2124. 14 comments had no recommendations at all while 21 comments had more than 30 recommendations each, out of which 14 comments had more than 50 recommendations each.
J.P. Ðordevic´ / Discourse, Context & Media 33 (2020) 100366
5
Table 1 The most frequent news values and linguistic devices identified in the articles. News value
Linguistic device
Number of occurrences of linguistic devices
Section
Example
Negativity Proximity
26 38
Superlativeness
Negative evaluative language References to place Nation Inclusive 1st pers. pl. Intensifiers Metaphors
Prominence
High status role labels
27
torture, ruin, crazy, Nazi terror, risk, illegal Kosovo, Serbia, EU, USA, UK, Germany, France, Brussels Albanians, Serbs, Americans We (Serbian government) extremely, heated, urgent, meek Kosovo sponsors, little Serbia, backbencher, ‘‘titans” Church, Patriarch, President, Ambassador
Impact
Descriptions of significant/ relevant consequences References to emotion
29
Politics, World Politics, World, Region, Society Politics, World, Region, Society Politics, World, Society Region, Politics, World Politics
Personalization
22
13
gradual increase of minimum wage, sort out situation within the church, will defend their own people frighten, calm, tranquil, dignified, proud
Table 2 The corpus used in the research. Sections
News articles
Comments
21
77
939
Recommendations Min.
Max.
Total
131
2124
2255
5.2. Results and discussion This section presents the results gathered from the quantitative and the qualitative analyses of the analysed corpus. The quantitative analysis shows that HS in readers’ comments is a rather frequent means of establishing a communicative Common Ground among news readers in Serbia which was the first objective of the research. The qualitative analysis provides evidence that derogatory and pejorative language are a reflection of the negative impact of news on the cognitive interface between discourse and the Serbian society which was the second objective of this research. Examples from the corpus will be provided to illustrate the kind of expressions of HS that were identified. All examples are in Serbian with a gloss translation in English. 5.2.1. Quantitative analysis Given that the corpus consisted of comments (C) posted on a single news website on a single day, 182 instances of HS expressions may be considered excessive. The identified instances occurred as reactions to 41 articles published in 11 sections (Table 3). Five articles on that day generated the highest number of C and the highest number of hate speech expressions (HSE). Out of these, three articles were published in the section Politics, one article in the section My life abroad and one in the section World. The article with the highest number of C (122) and the highest number of HSE (23) was an account of President Vucˇic´’s view of the latest developments regarding Priština’s decision to transform KSF into an army. The second highest number of C and HSE was related to the article about France accepting the transformation of KSF (76 C and 11 HSE). The article that generated 33 C and 18 instances of HSE was related to reactions coming from the European Union. One article from the section My life abroad about teachers punishing students generated 24 C and nine instances of HSE. Finally, the article that generated 21 C and eight instances of HSE was about Bishop Filaret asking the Serbian Orthodox Church to rehabilitate him. Three articles generated seven instances HSE each. The first was from the section Politics with 47 C about Chepurin’s visit to Vucˇic´, the second again from the section Politics with 25 C about Germany asking Serbia and Kosovo to calm down the tensions after Priština’s decision to create an army and the third was from the section World with 25 C about the German Bundestag approving the registration of a third gender.
Direct replies to comments
Instances of explicit HS
205
182
To sum up, articles referring to Priština’s decision to transform KSF into a regular army dominated the news on the selected day and they generated the highest number of C and instances of HSE. The Common Ground created among the readers reflects a high level of concern that a regular Kosovo army is a threat to Serbia. However, the individual instances of HSE, as will be presented in the next section, refer mainly to the inability of the Serbian government to convince the European Union of this threat. In addition, the HSE identified in the comments also indicate a strong dissatisfaction with the reactions coming from the European Union as well as the belief that EU authorities are tolerating Kosovo authorities while forcing Serbia into accepting whatever Kosovo decides. 5.2.2. Qualitative analysis Since the starting point of the analysis of HS in this research were van Dijk’s (2018) discourse structures identified in the readers’ comments, the main objective of the quantitative analysis was to provide evidence that derogatory and pejorative language are a reflection of the negative impact of news on the cognitive interface between the online discourse presented in the news and the Serbian society. Thus, van Dijk’s discourse structures were further analysed in reference to individual actors, collective actors, political relationships, social/cultural/economic issues and country (Table 4). Two reasons imposed this decision. The first is that the relationship between social actors in Serbia and the specific context that the relationship is occurring within is ideologically based (van Dijk, 2016). A common way to present this relationship is through the media whereby the audience, as a recipient, has the right to react to the news or not. One way of reacting to it is by posting their opinion about this relationship in the comment section of the website that the news was presented on. In the corpus analysed here, the readers frequently opted for HS in their comments. This choice indicates that the relationship between the social actors and the context that it is occurring in was not only provocative enough to motivate the readers to communicate their attitudes in their comments but it also implicates that the said relationship impacted them in a negative way. The second reason for the broader analysis of van Dijk’s discourse structures was that the attitudes of the readers who post comments occur in an online space where there are neither constraints nor repercussions (Erjavec & Kovacˇicˇ, 2012; Tenenboim & Cohen, 2015). The online environment practically serves as a
J.P. Ðordevic´ / Discourse, Context & Media 33 (2020) 100366
6 Table 3 Total number of instances of HS. Section
Total no. of news articles in the sections
Total no. of news articles with instances of HS
Total no. of C
Instances of HS
Among us
4
3
Commentaries Latest news My life abroad Region Serbia Economy Chronicle Society
2 1 1 1 3 3 1 4
1 1 1 1 1 1 1 2
Politics
30
24
World
10
5
7 3 4 4 7 24 26 3 6 5 21 7 4 47 122 9 3 20 18 13 18 25 14 33 15 5 20 23 21 6 10 20 10 83 54 16 76 15 6 7 25
3 1 3 2 3 9 4 1 1 3 8 1 2 7 23 4 1 3 2 3 2 7 6 18 3 1 5 2 4 3 1 4 1 6 4 3 11 5 4 1 7
11
60
41
855
182
Table 4 The most frequent discourse structures identified in the C reflecting HS. Discourse structure
Reference Actors (individual)
Actors (collective)
Political relationships
Social/cultural/economic issue
Country
Opinion and Emotion Words
Vucˇic´, Tacˇi, Macron, Chepurin, Kocijancˇic´, Sela, Zaev
EU, USA, NATO, Russia
Serbia-Kosovo, Vucˇic´-EU, Vucˇic´-Merkel, EU-Serbia
TV subscription, university teachers, education system, salaries, religion, gender
Kosovo, Serbia
Metaphors
Vucˇic´ Merkel
EU, USA, NATO, Kosovo government, Serbian government
Kosovo-Serbia
education system, gender
Kosovo, Serbia
platform for the instigation of hatred that the readers communicate in their comments (Cammaerts, 2009). Therefore, the discourse structures identified in this research, along with the cognitive basis expressed within the anonymity of online comment sections, provide evidence of the sociocognitive dimension via which the discourse of the presented news is connected to the readers as representatives of the social group that the news are being written for and directed at. The most frequent discourse structures identified in the corpus analysed here were Metaphors and Opinions and Emotion Words (Table 4).
Since metaphors are ‘‘based on multimodal structures of mental models of experience” (van Dijk, 2018, p. 11), they are indicative not only of the readers’ creative use of language but also of the general sociocultural knowledge the readers share. The identified instances of HSE in the corpus analysed here mainly refer to Kosovo, Serbia, President Vucˇic´ and his relationship to the EU as may be seen in the following examples: Kosovo Dzˇamahirija [Jamahiriya], šiptarski brlog [shiptar lair], Tacˇijeva prokleta avlija [Thaçi’s damned yard], Tacˇijev vilajet [Thaçi’s
J.P. Ðordevic´ / Discourse, Context & Media 33 (2020) 100366
ˇ K hirurgija [UÇK (Kosovo Liberation Army) surgery vilayet], UC ward], britanska trojanska koza [British Trojan horse], Frankenštajn [Frankenstein]
Table 5 Articles in the section Politics with the highest number of C and HSE. Article
Topic
Total number of C
Instances of HSE
1
Serbs in Mitrovica listened together to Vucˇic´’s speech Chepurin visiting Vucˇic´, carrying a message from Moscow They had one intention – to trample down Serbia (Vucˇic´) Sela: Zaev’s statement directed against Kosovo Albanians and USA Kneissl: Peace cannot be obtained by forming an army Vucˇic´ visited the Serbian Army at an unknown location New York Times: Priština defying Serbia and NATO Kosovo is not entitled to an army Cyprus worried about Priština’s moves Serbian government condemned the decision of the Priština government Selakovic´: International law has died Germany seeks that tensions after Kosovo decision calms down Opposition signed joint conditions for free and fair elections EU like a broken record: Change of mandate KSF only in compliance with the constitution Ðuric´: Kosovo army is an illegal occupation formation Opposition asking for Assembly meeting regarding Kosovo Tacˇi: Kosovo army – a holiday gift Gojkovic´: Priština decision against the Serbian people Haradinaj again provocative: The army will be stationed all over Kosovo The transformation of Kosovo Security Force may make the integration of Kosovo into NATO more difficult Rakic´: By forming the army, Priština has attacked peace and security Stoltenberg: I am sorry about the decision made by Kosovo, NATO supports the dialogue US embassy in Priština repeated they supported the transformation of the KSF into armed forces Vulin: Serbs in Kosovo have the President’s support and the support of the Serbian Army Kosovo Army is the greatest single threat to peace in the Balkans Albanians celebrating with fireworks and music BBC: Kosovo has deviated significantly form the Euro Atlantic course Kosovo parliament without MPs from the Serbian list adopted the law on the army Bloomberg: Next conflict between NATO and Russia will be in the Balkans D-day for Kosovo army 30
4
2
47
7
122
23
9
4
0
0
3
1
20
3
18 9 13
2 0 3
18 25
2 7
14
6
33
18
15
3
5
1
20 6
5 0
23
2
2
0
5
0
21
4
6
3
10
1
20
4
10
1
8
0
83
6
54
4
16 639
3 115
Serbia tragikomedija [tragicomedy], cirkus [circus], banana drzˇava [banana country], brlog [lair] EU politicˇki patuljak [political dwarf], briselske(NATO)birokrate [Brussel(NATO)bureaucrats], NATO kolonizator [NATO colonizer], sponzori srpske vlade [Serbian government sponsors] Vucˇic´ veliki voda [great leader], prestolonaslednik [crown prince], vrhovni komandant [supreme commander-in-chief], gazda [boss], cˇovek na mestu predsednika [the man in the position of the president], veleizdajnik [worst traitor]
2 3 4 5 6 7 8 9 10
Actors (collective) šiptari [shiptars], zˇuti [yellow (referring to the Serbian oposition)], NATOvci [NATOplayers], štetocˇinski kadrovi [pestilential staff (Serbian government)], Vucˇic´evi poslušnici [Vucˇic´’s servants (Serbian government)], isterivacˇ pravde [exorcist of justice (USA)], svetske grabljivice [world predators (USA)] Actors (individual) šefica [head, f (Merkel)], ucˇiteljica [teacher, f (Merkel)], rusofili [Russophiles], sendvicˇari [sandwich eaters (bots)], dosmanlije [dosmanlias (reference to the Othoman empire – Serb. Osmanlijsko carstvo)]. Examples of Opinion and Emotion Words could be identified mostly in reference to the way Vucˇic´ speaks when addressing the press, as may be seen in the following examples: jadikovanje [lamentation], kukumavkanje [wailing], kukanje [sobbing], patetika [pathetic], zˇalopojka ucveljenog deteta [dirge of a bereaved child], prenemaganje [affectation], zamajavanje [time-wasting], Kalimero‘‘ politika [‘‘Calimero” politics]. The presented examples show a generally negative impact of news on the cognitive interface between discourse and society which is reflected in the fact that the readers share a negative opinion about almost all actors and relationships involved in the decision about the transformation of KSF. Even more prominent is that the Kosovo government as well as those who support them are seen as mutual enemies. It may be concluded that the sociocognitive dimension in the analysed corpus establishes a negative common space between the analysed online discourse and the Serbian audience as the representatives of society. The format of online comments may be seen as a means of communication used by readers to freely express their negative opinion. The various elements that interact in the identified sociocognitive dimension combine explicit forms of a negative social cognition, including attitudes about political issues, norms and values referring to historical, social and political problems and ideologies beyond individual problems. A more detailed analysis of the content of the readers’ comments showed that most comments that reflected some form of HSE occurred in the section Politics (30 articles with a total of 639 C and 115 instances of HSE) (Table 5). As presented in Table 5 (the headlines are displayed only in English to avoid an excessive word count), all articles except one (Article 16) in this section were about the transformation of KSF. Six articles (5, 9, 12, 20, 22 and 27) presented the views of foreign authorities regarding the transformation and they all expressed their disagreement with this decision. These six articles generated
7
11 12 13 14
15 16 17 18 19 20
21 22
23
24
25 26 27 28 29 30
a total of 65 C and 11 instances of HSE most of which occurred as a reaction to Article 12 (a review of Germany’s opinion regarding the transformation). Article 3, the longest of all articles in the entire corpus, generated most readers’ C. The article is a review of all media statements that President Vucˇic´ made after Priština announced their decision regarding the transformation of KSF. The significant number of comments in this section indicates that reader engagement was most prominent in relation to the news about the transformation of KSF so that the step that Kosovo
J.P. Ðordevic´ / Discourse, Context & Media 33 (2020) 100366
8
had taken was a matter of concern almost all readers’ comments shared. The analysis of the comments showed that readers mostly discussed three issues: (1) the fact that the forming of a regular Kosovo army is against all international agreements related to Kosovo (254 C contained historical, political and social facts about Kosovo reflecting thus a high level of knowledge about the topic), (2) the obvious inability or lack of willingness among foreign authorities to prevent an army in Kosovo (538 C reflected the readers’ dissatisfaction that foreign authorities would not prevent the creation of a Kosovo army) and (3) the lack of authority of the Serbian government to do anything about it (597 comments reflected the readers’ dissatisfaction with the Serbian government’s incapacity to do anything about the situation). The role of comments among online news readers in Serbia seems to be twofold. On the one hand, the mediated communication enabled via the comment sections on the websites of Serbian news outlets transforms comments into an expression of the freedom of speech guaranteed to democratic societies and a means of promoting a democratic public sphere, thus facilitating unconstrained deliberations of opinions, beliefs, knowledge and ideologies (Cammaerts, 2009). Obviously, people in Serbia want to preserve this right and online comment sections provide it. On the other hand, readers’ comments reflect the interaction between discourse structures and social structures via a cognitive dimension (van Dijk, 2018), thus exhibiting the mental models, knowledge or ideologies, as part of the actual production and comprehension by Serbian language users. In other words, readers of news in Serbia deliberately choose the online space of mediated interaction provided by the comment sections on websites of news outlets. One reason is that this type of communication is anonymous. However, a more important reason established earlier (Section 3) is that Serbia ‘‘utterly fails to meet EU press freedom standards” (RWB, 2018) and almost all media are exposed to political influence primarily exerted by the Serbian government (Ðordevic´, 2019; Matic´, 2014; RWB, 2018). This leads to the conclusion that communication via comment sections seems to be the only way for readers of news in Serbia to actively participate in the political reality of Serbia. The final analysis focusing on the correlation between the linguistic devices identified in the articles and the instances of HS identified in the readers’ comments indicates a significant relationship between these two variables (r = 0.567, p = .000) (Table 6). The conclusion is that a certain set of news values (Table 1) motivated the readers not only to post comments but also to use offensive and abusive language. The present research shows that communication via online comment sections reflects a negative impact of particular news and news values (Bednarek & Caple, 2014) on the sociocognitive interface between the online discourse on Serbian websites and the readers of the news as representatives of the Serbian society (van Dijk, 2018). Such finding highlights the need for further research with three primary foci: (1) the influence of the media in the context of freedom of speech in societies where this democratic right is not guaranteed, (2) the importance of comment sections and the necessity to preserve them as a forum of expressing Table 6 Correlation between Linguistic devices identified in articles vs. instances of HS.
Linguistic devices discovered in articles Instances of HS **
Linguistic devices discovered in articles
Instances of HS
Pearson Correlation
1
0.567(**)
Sig. (2-tailed) Pearson Correlation Sig. (2-tailed)
0.567(**) 0.000
Correlation is significant at the level 0.01 (2-tailed).
0.000 1
public opinion and (3) the relationship between online discourse and society in countries with a political context similar to the one in Serbia.
6. Conclusion Based on the analysis of discourse structures in readers’ comments, the present research has confirmed that HS in the readers’ comments on the news website Politika is a rather frequent means of establishing a communicative Common Ground among news readers. The analysis has also shown that HS is a reflection of the negative impact of news and news values on the sociocognitive interface between discourse and society. Various negative forms of social cognition are created in the common space between discourse and society while at the same time the social cognition of the members of linguistic and discursive communities influences and shapes that very same space. From a methodological point of view, the present research has shown that SCDS (van Dijk, 2016, 2018) may be applied to the analysis of hate speech in readers’ comments, thus enabling not only the analysis of hate speech as a discursive practice but also the identification of the social cognition created through hate speech. CDS has undoubtedly established itself as a discipline investigating the power of discourse in affecting and shaping society and socio-political orientations (Wodak & Meyer, 2009) as well as a valuable contextualization tool (KhosraviNik and Zia, 2014). Nevertheless, the development of participatory media and social networks as a new communication environment has instigated the necessity to (re)evaluate the applicability of CDS to specific online discursive practices, hate speech in the comment sections of news websites being one of them. Therefore, this research has shown that SCDS facilitates a new interdisciplinary synergy with CDS (KhosraviNik & Esposito, 2018) to analyse hate speech as a specific type of discourse in mediated forms of communication. That is why it may also be concluded that SCDS is not only a new approach within CDS but also an analytical tool which will facilitate CDS to evolve as a discipline since it may be used for the critical description of the cognitive aspects of concepts used by readers while relating the structural properties of text or talk to social structures. Van Dijk’s discourse structures (2018) may easily be applied within critical discourse analyses related to other types of digital media discourse, such as discourse occurring in social media where users tend to express their thoughts and attitudes quite freely. Obviously, this is an ideal setting for the investigation of the role of such discourse as a reflection of the sociocognitive dimension (KhosraviNik & Zia, 2014; van Dijk, 2018) existing between a certain type of discourse and the readers as members of society. The newly-established synergy between SCDS and CDS, along with more specific approaches to digital media discourse, such as SM-CDS (Khosravinik, 2017), provides elements for the discourse-cognition-society triangle, thus supplying the relevant tools to explore the cognitively mediated relations between discourse and society (van Dijk, 2018) in mediated forms of communication (KhosraviNik, 2017). Although the negative impact on mental models that readers construe when exposed to the discourse presented to them in the news is limited to specific news content, the fact that readers opt for HS when expressing their opinions clearly indicates that more research is needed to explore the immediate effect of news on the sociocognitive interface between discourse and society.
Declaration of Competing Interest The author declare that there is no conflict of interest.
J.P. Ðordevic´ / Discourse, Context & Media 33 (2020) 100366
Acknowledgements This paper benefited from the project The Dynamics of Structures of the Contemporary Serbian Language (178014) supported by the Ministry of Education, Science and Technological Development of the Republic of Serbia and the project Credibility, Honesty, Ethics, and Politeness in Academic and Journalistic Writing (CHEP 2018), supported by DAAD (March 2018 – February 2019).
Funding sources This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Appendix A. Supplementary material Supplementary data to this article can be found online at https://doi.org/10.1016/j.dcm.2019.100366. References Annual Progress Report of the European Commission for Serbia, 2019. Retrieved from (19 June 2019). Ðordevic´, J., 2019. Translation in Serbian media discourse: The discursive strategy of argumentation as an adaptation technique. Perspectives: Studies in Translation Theory and Practice. https://doi.org/10.1080/0907676X.2019. 1595068. Barnes, R., 2018. Uncovering online commenting culture: Trolls, fanboys and lurkers. Cham: Palgrave Macmillan. https://doi.org/10.1007/978-3-319-702353. Bednarek, M., Caple, H., 2014. Why do news values matter? Towards a new methodological framework for analysing news discourse in Critical Discourse Analysis and beyond. Discourse. Brown, A., 2018. What is so special about online (as compared to offline) hate speech?. Ethnicities 18 (3), 297–326. https://doi.org/10.1177/ 1468796817709846. Bruce, T., 2018. New technologies, continuing ideologies: Online reader comments as a support for media perspectives of minority religions. Discourse Context Media 24, 53–75. https://doi.org/10.1016/j.dcm.2017.10.001. Cammaerts, B., 2009. Radical pluralism and free speech in online public spaces. Int. J. Cult. Stud. 12 (6), 555–575. https://doi.org/10.1177/1367877909342479. Erjavec, K., Kovacˇicˇ, M.P., 2012. ‘‘You don’t understand, this is a new war!” Analysis of hate speech in news websites’ comments. Mass Commun. Soc. 15 (6), 899– 920. https://doi.org/10.1080/15205436.2011.619679. Fairclough, N., 2003. Analysing Discourse: Textual Analysis for Social Research. Routledge, London. Gemius Audience (2018). Retrieved from (10 April 2019). Harcup, T., O’Neill, D., 2017. What is news? Journalism Studies 18 (12), 1470–1488.
9
KhosraviNik, M., 2017. Social media critical discourse studies (SM-CDS). In: Flowerdew, J., Richardson, J. (Eds.), Handbook of Critical Discourse Analysis (pp. 582–596). Routledge. KhosraviNik, M., Zia, M., 2014. Persian nationalism, identity and anti-Arab sentiments in Iranian Facebook Discourses. J. Lang. Polit. 13 (4), 754–780. KhosraviNik, M., Esposito, E., 2018. Online hate, digital discourse and critique: Exploring digitally-mediated discursive practices of gender-based hostility. Lodz Papers in Pragmatics 14 (1), 45–68. Kulacˇanin, N., 2018. 3.456 botova [3,456 bots]. Danas. Retrieved from (10 April 2019). Tenenboim, O., Cohen, A.A., 2015. What prompts users to click and comment: a longitudinal study of online news. Journalism 16 (2), 198–217. https://doi.org/ 10.1177/1464884913513996. Toepfl, F., Piwoni, E., 2015. Public spheres in interaction: comment sections of news websites as counterpublic spaces. J. Commun. 65 (3), 465–488. https://doi.org/ 10.1111/jcom.12156. van Dijk, T.A., 2016. Critical discourse studies: A sociocognitive approach. In: Wodak, R., Meyer, M. (Eds.), Methods of critical discourse studies. SAGE Publishing, Amsterdam, pp. 63–85. van Dijk, T.A., 2018. Socio-cognitive discourse studies. In: Flowerdew, J., Richardson, J.E. (Eds.), The Routledge handbook of critical discourse studies. Routledge, pp. 26–44. Vollhardt, J.R., Coutin, M., Staub, E., 2006. Deconstructing hate speech in the DRC: A psychological media sensitization campaign. J. Hate Stud. 5 (1), 15–35. Weber, A., 2009. Manual on Hate Speech. Council of Europe Publishing, Strasbourg. Weizman, E., Dori-Hacohen, G., 2017. On-line commenting on opinion editorials: a cross-cultural examination of face work in the Washington Post (USA) and NRG (Israel). Discourse Context Media 19, 39–48. https://doi.org/10.1016/j. dcm.2017.02.001. Wodak, R., Meyer, M., 2009. Methods for Critical Discourse Analysis. SAGE Publications. Wodak, R., Meyer, M., 2016. Methods of Critical Discourse Analysis. Sage, London. Yeo, S.K., Su, L.Y.F., Scheufele, D.A., Brossard, D., Xenos, M.A., Corley, E.A., 2019. The effect of comment moderation on perceived bias in science news. Inform. Commun. Soc. 22 (1), 129–146. https://doi.org/10.1080/ 1369118X.2017.1356861. Zubcˇevic´, A.R., Bender, S., Vojvodic´, J., 2017. Media regulatory authorities and hate speech. Council of Europe. Jasmina P. Ðordevic´ (Assistant Professor at the Foreign Language Centre, Faculty of Philosophy, University of Niš, Serbia) teaches English for Specific Purposes. Her academic and scientific interests include Discourse Analysis, the Study of Translation and Computer-Assisted Language Learning. She has published and presented extensively, she is active as a mentor and she is a member of several editorial boards and associations.
Contents lists available at ScienceDirect
Discourse, Context & Media journal homepage: www.elsevier.com/locate/dcm
The sociocognitive dimension of hate speech in readers’ comments on Serbian news websites Jasmina P. Ðordevic´ ´ irila i Metodija 2, 18100 Niš, Serbia Faculty of Philosophy, University of Niš, C
a r t i c l e
i n f o
Article history: Received 24 April 2019 Received in revised form 29 November 2019 Accepted 29 November 2019
Keywords: Readers’ comments Serbian news websites Hate speech Discourse analysis Sociocognitive dimension
a b s t r a c t Comment sections on news websites bear no immediate threat of sanctions and users feel free to express open criticism or insult. Other than causing offense and undermining the dignity of readers with a different opinion, a rather dangerous outcome of hate speech is that it affects the sociocognitive interface between discourse and society in the sense that it has a negative impact on the readers’ cognitive notions of information, beliefs and knowledge. Based on the analytical approach proposed in Sociocognitive Discourse Studies, the study examines 939 comments posted in the comment section of the Serbian news website Politika. Besides confirming that hate speech in the readers’ comments is a rather frequent means of establishing a communicative Common Ground among news readers, the analysis shows that hate speech is a reflection of the negative impact of news on the sociocognitive interface between discourse and society. The research also demonstrates that Sociocognitive Discourse Studies may be applied to the analysis of readers’ comments as a specific type of written discourse that reflects the mental models news readers create when exposed to media discourse. Ó 2019 Elsevier Ltd. All rights reserved.
1. Introduction The role of comment sections on news websites has recently become a recurring focus of scholarly research in the field of media discourse analysis (Barnes, 2018; Bruce, 2018; KhosraviNik & Zia, 2014; Tenenboim & Cohen, 2015; Toepfl & Piwoni, 2015; Weizman & Dori-Hacohen, 2017; Yeo et al., 2019). Initially introduced to encourage public discussion and monitor reader engagement, comment sections often become a platform of derogatory and pejorative language rather than serving the purpose of civil debate. Such discourse necessitates closer research. Readers’ comments reflect a reaction to news content, thus creating a specific interface between a news article and its readership (Weizman & Dori-Hacohen, 2017). This interface is dominated by a dimension that relates discourse structures to social structures and shows the immediate effect of this relationship (Bruce, 2018; Toepfl & Piwoni, 2015). Thus readers’ comments demonstrate how the information presented to them in the news affects their attitudes, ideologies and language use. Critical Discourse Studies (CDS) relates the structural properties of text or talk to social structures, enables both the analysis of the discourse structure of a comment and its context and provides the grounds for an explicit viewpoint (van Dijk, 2018). However, it lacks the sociocognitive
E-mail address: [email protected] https://doi.org/10.1016/j.dcm.2019.100366 2211-6958/Ó 2019 Elsevier Ltd. All rights reserved.
dimension to relate the discourse structure to social structures via a sociocognitive interface which is needed to critically describe the cognitive aspects of concepts used by readers (2018). Assuming that particular news provoke a specific response, an important aspect to consider is that of newsworthiness, i.e. news values which are criteria applied by news workers when creating news (Bednarek & Caple, 2014). Any set of news values constructed by journalists is expected to attract readers’ attention, instigate meaning and motivate action, such as posting a comment referring to a specific news item. Given that readers’ comments are a reflection of the interaction between discourse structures and social structures via a sociocognitive dimension, this study refers to the theory of Sociocognitive Discourse Studies (SCDS) proposed by van Dijk (2016, 2018). The sociocognitive approach establishes the discourse-cognition-society triangle and it provides the relevant tools to explore the cognitively mediated relations between discourse and society (van Dijk, 2016) that comments reflect. Since discourse structures and social structures are different, a relationship between them can be established only through the mental representations of language users who are both individuals and social members. Their ideas, knowledge and opinions are an interpretation of the mutual relationship between social structures and discourse structures. Social structures influence discourse structures through people’s interpretations of the social environment they are part of and vice versa, discourse structures can only
2
J.P. Ðordevic´ / Discourse, Context & Media 33 (2020) 100366
influence social structures through the same cognitive interface of mental models, knowledge attitudes and ideologies (2016). The present research is based on a corpus of comments posted by readers on the website of the Serbian news outlet Politika. The aim is to analyse the comments from the point of view of SCDS so as to examine how hate speech (HS) in them is affecting the sociocognitive interface between discourse and society. An additional analysis of the linguistic devices and news values (Bednarek & Caple, 2014) used in the articles that generated the comments will show what news values provoked the readers to post comments. The hypothesis is that HS in readers’ comments on news websites indicates a negative impact of news on the readers’ cognitive notions of information, beliefs and knowledge. Based on a quantitative and a qualitative analysis of the compiled corpus, two objectives will be pursued: (1) examine whether HS in the readers’ comments is a means of establishing a communicative Common Ground among news readers and (2) find evidence that HS in readers’ comments is a reflection of the negative impact of news on the sociocognitive interface between discourse and society.
2. Hate speech as an online discursive practice Despite the lack of a universal definition of HS, a significant number of authors put HS in the centre of their attention (Brown, 2018; Cammaerts, 2009; Lillian, 2007; Vollhardt et al., 2006). HS is assumed to reflect abusive and harassing expressions of violence or discrimination, directed against people on the basis of their race, ethnic origin, religion, gender, disability, sexual orientation, etc. and it is mostly meant to silence opponents, or at least prove their opinions invalid. Critical Discourse Studies (CDS) (Wodak & Meyer, 2016; Fairclough, 2003) has relied on the idea that mass media are ‘‘an obvious powerful site where discursive power is exploited to (re)construct and (re)define social realities” (KhosraviNik & Esposito, 2018, p. 54) and as such it has been preoccupied with the notion of the power of discourse in affecting and shaping society and socio-political orientations (Wodak & Meyer, 2009). However, the development of participatory media and social networks as a new communication environment has resulted in a substantial increase of online discursive practices, HS being a prominent one. This new development has instigated the necessity to (re)evaluate CDS and its applicability to the concept of HS as an online discursive practice which is why CDS should pursue new interdisciplinary synergies with other disciplines (KhosraviNik & Esposito, 2018). As stated by KhosraviNik and Zia (2014), participatory media provide a space for ‘‘non-elite text producers resisting the discourses of established institutions” (p. 757). Thus the participants in the digital space create a valuable data repository which can be compiled from discourse found in the comment sections of shared news and posts on social networks. The quite excessive amount of data compiled from online communication is difficult to manage which is why KhosraviNik and Zia (2014) suggest ‘‘down-sampling” it in order to perform a ‘‘qualitative detailed textual analysis” and to keep it ‘‘under control for CDA contextualisation purposes” (p. 764). Being a discursive practice occurring quite frequently in participatory media, HS may be analysed based on the reactions and debates among readers and users in the interactive digital media initiated by a piece of news, social network post or similar content published online. Given the necessity to explore how the media may shape and influence society, KhosraviNik (2017) introduces Social Media Critical Discourse Studies (SM-CDS) which is ‘‘not only interested in what happens in media per se but in how it may shape and influ-
ence the social and political sphere of our life worlds” (p. 586). SMCDS may be used to investigate ‘‘what goes on online in such a way that the availability of technological affordance per se is assumed to be the ultimate sufficient context and force for actual social, political, and cultural change in society” (p.586). KhosraviNik further suggests that all media practices and the content presented in the media ‘‘should be interpreted within a wider sociopolitical context of a given society” which is where ‘‘a social, historical, cultural, psychological, or political account is provided for explication of the discourse under investigation” (p. 586). In the context of the present research, digital performances, online hostility being a prominent one, may be interpreted by including the sociocognitive dimension proposed by van Dijk (2016, 2018). 2.1. Hate speech in the context of newsworthiness The theory of newsworthiness (Harcup & O’Neill, 2017) may be applied to a corpus of news to confirm what factors influence news selection in a certain audience. An important criterion determining the choice of a particular news article are the news values which are discursively constructed and are defined as the ‘‘newsworthy aspects of actors, happenings and issues as existing in and constructed through discourse” (Bednarek & Caple, 2014, p. 138). Since the HS in readers’ comments is a reaction to a particular news items, certain news values are more provocative than others. Among the various devices based on which news values may be constructed, linguistic devices are most easily to apply within a specific discursive approach. Since the proposed approach in this article is SCDS (van Dijk, 2016, 2018), news values are the socially-shared cognitive representation while linguistic devices are the tools for construing news values that provoke HS in readers’ comments. Bednarek and Caple (2014) suggest a list of key linguistic devices (e.g. negative lexis, quantifiers, intensifiers, evaluative language, etc.) which may be used to analyse news values, such as negativity, proximity, superlativeness, prominence, impact, etc. The list is not universal and it may be further adapted. 2.2. Hate speech in the context of the European Union The EU General Policy Recommendation No. 15: Combating Hate speech from March 2016 states that HS includes ‘‘denigration, hatred, vilification, harassment, insult, negative stereotyping, stigmatization, threat, etc. which are based on a non-exhaustive list of personal characteristics or status that includes ‘race’, colour, language, religion or belief, nationality or national or ethnic origin, as well as descent, age, disability, sex, gender, gender identity and sexual orientation” (Zubcˇevic´, Bender, & Vojvodic´, 2017. p.11). However, regulations regarding HS also implicate that laws on HS in national contexts are subject to objections (Langton, 2016). Restrictive laws on HS undermine the concept of freedom of expression, a right guaranteed in all democratic countries which is why European Union legislation has been established to clearly define HS as opposed to the violation of freedom of speech. The main criteria that the European Court of Human Rights relies on when deciding whether a certain form of expression shows either one or the other are ‘‘the purpose pursued by the applicant as well as the content of the expression and the context in which [hate speech] was disseminated” (Weber, 2009, p. 33). Additional conventions and declarations have been stipulated to support lawful and just handling of HS without the risk of violating freedom of speech (Zubcˇevic´ et al., 2017). The implications of regulations regarding HS need more research because it depends on far too many different factors that
J.P. Ðordevic´ / Discourse, Context & Media 33 (2020) 100366
need to be explored and explained. HS is a product of a specific historical and cultural context and as such the content and meanings are related to that context. What is HS in one context may not be considered HS in another one (Parekh, 2012) which is why legislators, linguists, social actors, political actors and other participants in society, politics and culture need to explore the concept of HS from various points of views to determine specific circumstances, occurrences, forms and expressions with respect to a specific context and time. 3. Comment sections on Serbian websites The media landscape in Serbia is specific because of Serbia’s application to join the European Union. About 2000 media outlets are currently registered in the Serbian Business Registry Agency and they often depend on financial support from the government so that they are exposed to political influence (Ðordevic´, 2019; Matic´, 2014; RWB, 2018). The European Commission performs an annual in-depth scrutiny regarding EU values, freedom of speech being crucial in this respect. The reports confirm that there is no progress regarding the freedom of expression and that this is a matter of increasing concern (Annual Progress Report of the European Commission for Serbia, 2019). The same concern was expressed by Reporters Without Borders (RWB) who stated that Serbia ‘‘utterly fails to meet EU press freedom standards” (RWB, 2018). Online news outlets with daily news and with a major influence in Serbia include blic.rs, telegraf.rs, b92.net, novosti.rs, n1info.com, danas.rs and politika.rs whereby only n1info.com and danas.rs are perceived as neutral (the number of average daily real users is available on Gemius Audience, 2018). All the listed outlets have comment sections on their websites and readers use the possibility to post their comments in large numbers. Comment sections may not always reflect opinions of actual readers as a significant number of bots are employed by the government to comment news on a daily basis. According to unofficial resources (Kulacˇanin, 2018) the Serbian Progressive Party employs more than 3000 people who are part of the so-called Internet team. Their job description includes commenting news on various websites, posting opinions on social networks, sharing news and posts, tweeting and retweeting tweets, reacting to comments and posts by clicking the relevant option provided by a news website comment section or by clicking the appropriate emoji below a post on social networks. Topics that are most often commented on refer to government officials’ activities. The Serbian President’s activities dominate both the news and the comment sections. The language used in comments ranges from polite and civilized to spiteful and offensive so that HS is a common occurrence on all Serbian news websites. Readers openly show disrespect of other readers’ opinions, argue about facts, accuse each other of lying and misrepresenting ideas, pass insults and use sarcasm with the aim to diminish opinions. Therefore, comment sections reflect personal frustration and aggression rather than public opinion and civil discussion. 4. The sociocognitive dimension of discourse analysis As stated in the introduction, Sociocognitive Discourse Studies (SCDS) ‘‘relates discourse structures to social structures via a complex sociocognitive interface” (van Dijk, 2016, 2018). The main subjects of SCDS are the ongoing communicative Common Ground, the shared knowledge and the attitudes and ideologies of language users who are participants of a particular communicative situation and members of social groups and communities (van Dijk, 2018). SCDS makes ‘‘explicit the fundamental role of mental representations” and it ‘‘shows that many structures of discourse itself can
3
only (completely) be described in terms of various cognitive notions” (p. 28). Therefore, SCDS relies on ‘‘explicit psychological theories of mental representations of journalists or other language users” as well as on ‘‘the ways these models mediate between shared social cognition (knowledge, attitudes, ideologies), social structures and actual text and talk” (p. 28). In addition, SCDS includes the cognitive interface between discourse and society. The sociocognitive approach does not accept that there is a direct link between the structures of discourse and society, but supports the notion that there is a cognitive interface where the two meet and interact. The result of the interaction between discourse and society via the cognitive interface constitutes the sociocognitive dimension of the sociocognitive approach to discourse analysis. This means that the sociocognitive approach assumes that structures of society can only ‘‘affect text and talk through the minds of language users” because ‘‘social members represent both social structures as well as discourse structures in their minds, and thus are able to relate these mentally before expressing them in actual text and talk” (van Dijk, 2018, p. 28). Given the limited scope of this research, only a brief review of the most important notions about SCDS will be provided here. First, the notions of mind, memory and discourse processing are crucial since ‘‘cognitive processes take place in the mind or memory of individual social actors as members of social groups and communities” (van Dijk, 2018, p. 29). The way discourse is produced and comprehended is related to the way linguistic and discursive knowledge systems are applied by language users so that the processing of lexical items, syntactic structures and various meanings of discourse is closely related to grammar, the lexicon, semantics and different forms of interaction. Since local and global meanings attributed to discourse are based on underlying mental models as subjective representations of events or situations, a crucial aspect of the sociocognitive dimension is how the production and comprehension of discourse will involve the particular expression of a mental model regarding an event, person, opinion, belief, etc. The direct communicative intention of discourse is mainly to transmit the mental model of a speaker or writer. However, the way the hearer or reader will understand the transmitted discourse depends on how they will interpret it based on their own mental models which can be ‘‘individual, personal, subjective and multimodal” (van Dijk, 2018, p. 30). These models go beyond the subjective representation of a situation or event and they include opinions and emotions of both the ones who create discourse and those who read or hear it. The models reflect the social cognition of the members of the linguistic and discursive communities who share their knowledge of language and discourse, their sociocultural knowledge of the world as well as attitudes, ideologies, norms and values. To conclude, the sociocognitive dimension that SCDS relies on constitutes a complex interface that establishes a common space between discourse and society where different elements interact. These elements combine various forms of social cognition (attitudes, ideologies, norms and values based on sociocultural knowledge of the world and knowledge of language and discourse) that is being created in that commons space. At the same time the social cognition of the members of linguistic and discursive communities (social actors, news readers, participants in events, etc.) influences and shapes that very same space. 4.1. Discourse structures indicating hate speech in readers’ comments Being a product of discourse that reflects the knowledge, opinions and beliefs of readers who directly react to the discourse in the news, readers’ comments are in fact a result of the interaction between discourse and society via the sociocognitive interface (van
4
J.P. Ðordevic´ / Discourse, Context & Media 33 (2020) 100366
Dijk, 2016, 2018). Readers’ comments on news websites reflect the mental models that readers as members of society and discursive communities create as a reaction to the discourse they have been exposed to. Therefore, readers’ comments are discourse that comprises discourse structures based on which the sociocognitive dimension can be identified, described and analysed. The identification, description and analysis of discourse structures proposed by van Dijk (2018) that may indicate HS include: stress, intonation, word order; meanings of words, sentences or sequences of sentences; coherence; opinion and emotion words; global topics or themes; deictic or indexical expressions; speech acts; evidentials; conventional, schematic, canonical structures; metaphors and ideological polarizations. The identification of discourse structures is based on lexical items that are identified as individual instances referring to a certain discourse structure within a broader context. For instance, lexical items may refer to social actors (e.g. politicians, victims, attackers, police, government officials, etc.) or their representation in a context (e.g. individual, collective or impersonal). The mentioning or description of a social event may refer to a global topic or event (e.g. demonstrations, riots, elections, etc.), while lexical items indicating evaluative and emotional representations may refer to ideological polarizations (e.g. verbs such as urge, warn, state, etc., or adjectives such as important, imperative, dangerous, etc.). The identified discourse structures serve the purpose of both a quantitative and a qualitative analysis enabling the determination of the extent to which HS occurs in a particular discourse and the broader analysis of the mental representations and the socioculturally shared knowledge. The outcome is that the relationship between discourse and society is both quantified and qualified through the sociocognitive interface between them. 5. The research 5.1. Methodology and corpus The present research is based on the hypothesis that HS in readers’ comments on news websites reflects the way news affects the readers’ cognitive notions of information, beliefs and knowledge. It is assumed that users feel free to express open criticism or insult so that derogatory and pejorative language in comments may be expected to indicate that news affects the sociocognitive interface between discourse and society in a negative way. Since it is obvious that certain news are more provocative than others, the linguistic devices that occur in the article may indicate what type of news values provoke the readers to comment the news item. Therefore, the first objective of this research was to confirm that HS in the readers’ comments is a rather frequent means of establishing a communicative Common Ground among news readers. In addition, a list of key linguistic devices used in the article to construe news values has been identified so as to determine the types of news values that provoked the readers’ comments. The second objective was to find evidence that HS is a reflection of the negative impact of news on the sociocognitive interface between discourse and society. Starting with the assumption that readers freely express criticism or insult in comment sections when reacting to news articles (Tenenboim & Cohen, 2015), the textual analysis of the compiled corpus of comments was expected to indicate the role of comments as a reflection of the sociocognitive dimension (KhosraviNik & Zia, 2014; van Dijk, 2018) existing between the news values in the discourse presented in the news (Bednarek & Caple, 2014) and the readers as members of society. In order to identify the discourse structures that reflect the sociocognitive dimension in the selected comments, the analysis relied on the identification of lexical items explicitly reflecting instances of HS.
The corpus compiled for this research comprises 939 comments written in response to 77 news articles published on politika.rs (the website of the news outlet Politika) on 14 December 2018, the day after Priština announced the transformation of the Kosovo Security Forces (KSF) into a regular army. On average, the comments consist of 40 words, the shortest one consisting of one word only and the longest of 200 words. The website’s comment section poses a limit of 1000 characters per comment which is why the maximum length of 200 words is considered imposed as some readers might have wanted to write longer comments. The website politika.rs was used as a source for the corpus because Politika is the oldest newspaper in Serbia with the longest tradition of news publication and a circulation of the print edition estimated at around 100,000 copies a day. The website’s readership (Gemius Audience, 2018) counts almost 50,000 average daily real users and the profile of the readership includes mainly highly educated middle-class members, most of whom have been loyal to Politika for decades. In comparison to politika.rs, other influential news websites (e.g. blic.rs, telegraf.rs, b92.net, novosti.rs), despite a rather high number of readers, have not been chosen for this research because the administrators of their comment sections quite frequently tolerate profanity and vulgar language. Since the aim of this research was to explore how news affects the readers’ cognitive notions of information, beliefs and knowledge, examples of profanity and vulgar language would not have provided a relevant corpus. Two other highly influential news outlets in Serbia (n1info.com and danas.rs) were not chosen for the research because they are considered anti-government and a large number of bots post comments in their comment sections. This fact makes it difficult to distinguish real comments from the comments posted by bots. The comments were first read, coded, analysed and categorized (KhosraviNik & Zia, 2014). Then an analysis of the linguistic devices indicating news values in the articles was performed (Bednarek & Caple, 2014). A detailed presentation of all the news values and linguistic devices identified in the 77 articles would exceed the scope of this article; therefore, only the most frequent values and devices discovered in the articles will be presented here in English (Table 1). As can be seen, the most frequent news values and linguistic devices could be identified mainly in the articles from the sections Politics, World, Society and Region. Since this section is meant to introduce the corpus analysed in this research, the implications of this analysis will be presented within the qualitative analysis in Section 5.2.2. The analysis of explicit HS in the comments revealed a total of 182 instances in the entire corpus (Table 2). Lexical items containing sarcastic remarks or ironic statements were not marked as instances of HS because the sarcasm and irony identified in the comments did not explicitly reflect a clear intention of hatred, offense or abuse. Similar to comment sections on other websites, politika.rs also includes a mechanism to measure reader engagement. The readers can click one of two options below each comment: a) ‘reply’ encouraging the readers of comments to engage in discussion with the authors of the comments and b) ‘recommend’ based on which readers can show agreement with the opinion expressed in a particular comment. In the corpus analysed here (Table 2), a total of 205 direct replies to separate comments and a total of 2255 recommendations could be identified whereby the total number of minimum recommendations of each comment (0 or 1) was 131 and the total number of maximum recommendations of each comment (ranging from 3 to 231) was 2124. 14 comments had no recommendations at all while 21 comments had more than 30 recommendations each, out of which 14 comments had more than 50 recommendations each.
J.P. Ðordevic´ / Discourse, Context & Media 33 (2020) 100366
5
Table 1 The most frequent news values and linguistic devices identified in the articles. News value
Linguistic device
Number of occurrences of linguistic devices
Section
Example
Negativity Proximity
26 38
Superlativeness
Negative evaluative language References to place Nation Inclusive 1st pers. pl. Intensifiers Metaphors
Prominence
High status role labels
27
torture, ruin, crazy, Nazi terror, risk, illegal Kosovo, Serbia, EU, USA, UK, Germany, France, Brussels Albanians, Serbs, Americans We (Serbian government) extremely, heated, urgent, meek Kosovo sponsors, little Serbia, backbencher, ‘‘titans” Church, Patriarch, President, Ambassador
Impact
Descriptions of significant/ relevant consequences References to emotion
29
Politics, World Politics, World, Region, Society Politics, World, Region, Society Politics, World, Society Region, Politics, World Politics
Personalization
22
13
gradual increase of minimum wage, sort out situation within the church, will defend their own people frighten, calm, tranquil, dignified, proud
Table 2 The corpus used in the research. Sections
News articles
Comments
21
77
939
Recommendations Min.
Max.
Total
131
2124
2255
5.2. Results and discussion This section presents the results gathered from the quantitative and the qualitative analyses of the analysed corpus. The quantitative analysis shows that HS in readers’ comments is a rather frequent means of establishing a communicative Common Ground among news readers in Serbia which was the first objective of the research. The qualitative analysis provides evidence that derogatory and pejorative language are a reflection of the negative impact of news on the cognitive interface between discourse and the Serbian society which was the second objective of this research. Examples from the corpus will be provided to illustrate the kind of expressions of HS that were identified. All examples are in Serbian with a gloss translation in English. 5.2.1. Quantitative analysis Given that the corpus consisted of comments (C) posted on a single news website on a single day, 182 instances of HS expressions may be considered excessive. The identified instances occurred as reactions to 41 articles published in 11 sections (Table 3). Five articles on that day generated the highest number of C and the highest number of hate speech expressions (HSE). Out of these, three articles were published in the section Politics, one article in the section My life abroad and one in the section World. The article with the highest number of C (122) and the highest number of HSE (23) was an account of President Vucˇic´’s view of the latest developments regarding Priština’s decision to transform KSF into an army. The second highest number of C and HSE was related to the article about France accepting the transformation of KSF (76 C and 11 HSE). The article that generated 33 C and 18 instances of HSE was related to reactions coming from the European Union. One article from the section My life abroad about teachers punishing students generated 24 C and nine instances of HSE. Finally, the article that generated 21 C and eight instances of HSE was about Bishop Filaret asking the Serbian Orthodox Church to rehabilitate him. Three articles generated seven instances HSE each. The first was from the section Politics with 47 C about Chepurin’s visit to Vucˇic´, the second again from the section Politics with 25 C about Germany asking Serbia and Kosovo to calm down the tensions after Priština’s decision to create an army and the third was from the section World with 25 C about the German Bundestag approving the registration of a third gender.
Direct replies to comments
Instances of explicit HS
205
182
To sum up, articles referring to Priština’s decision to transform KSF into a regular army dominated the news on the selected day and they generated the highest number of C and instances of HSE. The Common Ground created among the readers reflects a high level of concern that a regular Kosovo army is a threat to Serbia. However, the individual instances of HSE, as will be presented in the next section, refer mainly to the inability of the Serbian government to convince the European Union of this threat. In addition, the HSE identified in the comments also indicate a strong dissatisfaction with the reactions coming from the European Union as well as the belief that EU authorities are tolerating Kosovo authorities while forcing Serbia into accepting whatever Kosovo decides. 5.2.2. Qualitative analysis Since the starting point of the analysis of HS in this research were van Dijk’s (2018) discourse structures identified in the readers’ comments, the main objective of the quantitative analysis was to provide evidence that derogatory and pejorative language are a reflection of the negative impact of news on the cognitive interface between the online discourse presented in the news and the Serbian society. Thus, van Dijk’s discourse structures were further analysed in reference to individual actors, collective actors, political relationships, social/cultural/economic issues and country (Table 4). Two reasons imposed this decision. The first is that the relationship between social actors in Serbia and the specific context that the relationship is occurring within is ideologically based (van Dijk, 2016). A common way to present this relationship is through the media whereby the audience, as a recipient, has the right to react to the news or not. One way of reacting to it is by posting their opinion about this relationship in the comment section of the website that the news was presented on. In the corpus analysed here, the readers frequently opted for HS in their comments. This choice indicates that the relationship between the social actors and the context that it is occurring in was not only provocative enough to motivate the readers to communicate their attitudes in their comments but it also implicates that the said relationship impacted them in a negative way. The second reason for the broader analysis of van Dijk’s discourse structures was that the attitudes of the readers who post comments occur in an online space where there are neither constraints nor repercussions (Erjavec & Kovacˇicˇ, 2012; Tenenboim & Cohen, 2015). The online environment practically serves as a
J.P. Ðordevic´ / Discourse, Context & Media 33 (2020) 100366
6 Table 3 Total number of instances of HS. Section
Total no. of news articles in the sections
Total no. of news articles with instances of HS
Total no. of C
Instances of HS
Among us
4
3
Commentaries Latest news My life abroad Region Serbia Economy Chronicle Society
2 1 1 1 3 3 1 4
1 1 1 1 1 1 1 2
Politics
30
24
World
10
5
7 3 4 4 7 24 26 3 6 5 21 7 4 47 122 9 3 20 18 13 18 25 14 33 15 5 20 23 21 6 10 20 10 83 54 16 76 15 6 7 25
3 1 3 2 3 9 4 1 1 3 8 1 2 7 23 4 1 3 2 3 2 7 6 18 3 1 5 2 4 3 1 4 1 6 4 3 11 5 4 1 7
11
60
41
855
182
Table 4 The most frequent discourse structures identified in the C reflecting HS. Discourse structure
Reference Actors (individual)
Actors (collective)
Political relationships
Social/cultural/economic issue
Country
Opinion and Emotion Words
Vucˇic´, Tacˇi, Macron, Chepurin, Kocijancˇic´, Sela, Zaev
EU, USA, NATO, Russia
Serbia-Kosovo, Vucˇic´-EU, Vucˇic´-Merkel, EU-Serbia
TV subscription, university teachers, education system, salaries, religion, gender
Kosovo, Serbia
Metaphors
Vucˇic´ Merkel
EU, USA, NATO, Kosovo government, Serbian government
Kosovo-Serbia
education system, gender
Kosovo, Serbia
platform for the instigation of hatred that the readers communicate in their comments (Cammaerts, 2009). Therefore, the discourse structures identified in this research, along with the cognitive basis expressed within the anonymity of online comment sections, provide evidence of the sociocognitive dimension via which the discourse of the presented news is connected to the readers as representatives of the social group that the news are being written for and directed at. The most frequent discourse structures identified in the corpus analysed here were Metaphors and Opinions and Emotion Words (Table 4).
Since metaphors are ‘‘based on multimodal structures of mental models of experience” (van Dijk, 2018, p. 11), they are indicative not only of the readers’ creative use of language but also of the general sociocultural knowledge the readers share. The identified instances of HSE in the corpus analysed here mainly refer to Kosovo, Serbia, President Vucˇic´ and his relationship to the EU as may be seen in the following examples: Kosovo Dzˇamahirija [Jamahiriya], šiptarski brlog [shiptar lair], Tacˇijeva prokleta avlija [Thaçi’s damned yard], Tacˇijev vilajet [Thaçi’s
J.P. Ðordevic´ / Discourse, Context & Media 33 (2020) 100366
ˇ K hirurgija [UÇK (Kosovo Liberation Army) surgery vilayet], UC ward], britanska trojanska koza [British Trojan horse], Frankenštajn [Frankenstein]
Table 5 Articles in the section Politics with the highest number of C and HSE. Article
Topic
Total number of C
Instances of HSE
1
Serbs in Mitrovica listened together to Vucˇic´’s speech Chepurin visiting Vucˇic´, carrying a message from Moscow They had one intention – to trample down Serbia (Vucˇic´) Sela: Zaev’s statement directed against Kosovo Albanians and USA Kneissl: Peace cannot be obtained by forming an army Vucˇic´ visited the Serbian Army at an unknown location New York Times: Priština defying Serbia and NATO Kosovo is not entitled to an army Cyprus worried about Priština’s moves Serbian government condemned the decision of the Priština government Selakovic´: International law has died Germany seeks that tensions after Kosovo decision calms down Opposition signed joint conditions for free and fair elections EU like a broken record: Change of mandate KSF only in compliance with the constitution Ðuric´: Kosovo army is an illegal occupation formation Opposition asking for Assembly meeting regarding Kosovo Tacˇi: Kosovo army – a holiday gift Gojkovic´: Priština decision against the Serbian people Haradinaj again provocative: The army will be stationed all over Kosovo The transformation of Kosovo Security Force may make the integration of Kosovo into NATO more difficult Rakic´: By forming the army, Priština has attacked peace and security Stoltenberg: I am sorry about the decision made by Kosovo, NATO supports the dialogue US embassy in Priština repeated they supported the transformation of the KSF into armed forces Vulin: Serbs in Kosovo have the President’s support and the support of the Serbian Army Kosovo Army is the greatest single threat to peace in the Balkans Albanians celebrating with fireworks and music BBC: Kosovo has deviated significantly form the Euro Atlantic course Kosovo parliament without MPs from the Serbian list adopted the law on the army Bloomberg: Next conflict between NATO and Russia will be in the Balkans D-day for Kosovo army 30
4
2
47
7
122
23
9
4
0
0
3
1
20
3
18 9 13
2 0 3
18 25
2 7
14
6
33
18
15
3
5
1
20 6
5 0
23
2
2
0
5
0
21
4
6
3
10
1
20
4
10
1
8
0
83
6
54
4
16 639
3 115
Serbia tragikomedija [tragicomedy], cirkus [circus], banana drzˇava [banana country], brlog [lair] EU politicˇki patuljak [political dwarf], briselske(NATO)birokrate [Brussel(NATO)bureaucrats], NATO kolonizator [NATO colonizer], sponzori srpske vlade [Serbian government sponsors] Vucˇic´ veliki voda [great leader], prestolonaslednik [crown prince], vrhovni komandant [supreme commander-in-chief], gazda [boss], cˇovek na mestu predsednika [the man in the position of the president], veleizdajnik [worst traitor]
2 3 4 5 6 7 8 9 10
Actors (collective) šiptari [shiptars], zˇuti [yellow (referring to the Serbian oposition)], NATOvci [NATOplayers], štetocˇinski kadrovi [pestilential staff (Serbian government)], Vucˇic´evi poslušnici [Vucˇic´’s servants (Serbian government)], isterivacˇ pravde [exorcist of justice (USA)], svetske grabljivice [world predators (USA)] Actors (individual) šefica [head, f (Merkel)], ucˇiteljica [teacher, f (Merkel)], rusofili [Russophiles], sendvicˇari [sandwich eaters (bots)], dosmanlije [dosmanlias (reference to the Othoman empire – Serb. Osmanlijsko carstvo)]. Examples of Opinion and Emotion Words could be identified mostly in reference to the way Vucˇic´ speaks when addressing the press, as may be seen in the following examples: jadikovanje [lamentation], kukumavkanje [wailing], kukanje [sobbing], patetika [pathetic], zˇalopojka ucveljenog deteta [dirge of a bereaved child], prenemaganje [affectation], zamajavanje [time-wasting], Kalimero‘‘ politika [‘‘Calimero” politics]. The presented examples show a generally negative impact of news on the cognitive interface between discourse and society which is reflected in the fact that the readers share a negative opinion about almost all actors and relationships involved in the decision about the transformation of KSF. Even more prominent is that the Kosovo government as well as those who support them are seen as mutual enemies. It may be concluded that the sociocognitive dimension in the analysed corpus establishes a negative common space between the analysed online discourse and the Serbian audience as the representatives of society. The format of online comments may be seen as a means of communication used by readers to freely express their negative opinion. The various elements that interact in the identified sociocognitive dimension combine explicit forms of a negative social cognition, including attitudes about political issues, norms and values referring to historical, social and political problems and ideologies beyond individual problems. A more detailed analysis of the content of the readers’ comments showed that most comments that reflected some form of HSE occurred in the section Politics (30 articles with a total of 639 C and 115 instances of HSE) (Table 5). As presented in Table 5 (the headlines are displayed only in English to avoid an excessive word count), all articles except one (Article 16) in this section were about the transformation of KSF. Six articles (5, 9, 12, 20, 22 and 27) presented the views of foreign authorities regarding the transformation and they all expressed their disagreement with this decision. These six articles generated
7
11 12 13 14
15 16 17 18 19 20
21 22
23
24
25 26 27 28 29 30
a total of 65 C and 11 instances of HSE most of which occurred as a reaction to Article 12 (a review of Germany’s opinion regarding the transformation). Article 3, the longest of all articles in the entire corpus, generated most readers’ C. The article is a review of all media statements that President Vucˇic´ made after Priština announced their decision regarding the transformation of KSF. The significant number of comments in this section indicates that reader engagement was most prominent in relation to the news about the transformation of KSF so that the step that Kosovo
J.P. Ðordevic´ / Discourse, Context & Media 33 (2020) 100366
8
had taken was a matter of concern almost all readers’ comments shared. The analysis of the comments showed that readers mostly discussed three issues: (1) the fact that the forming of a regular Kosovo army is against all international agreements related to Kosovo (254 C contained historical, political and social facts about Kosovo reflecting thus a high level of knowledge about the topic), (2) the obvious inability or lack of willingness among foreign authorities to prevent an army in Kosovo (538 C reflected the readers’ dissatisfaction that foreign authorities would not prevent the creation of a Kosovo army) and (3) the lack of authority of the Serbian government to do anything about it (597 comments reflected the readers’ dissatisfaction with the Serbian government’s incapacity to do anything about the situation). The role of comments among online news readers in Serbia seems to be twofold. On the one hand, the mediated communication enabled via the comment sections on the websites of Serbian news outlets transforms comments into an expression of the freedom of speech guaranteed to democratic societies and a means of promoting a democratic public sphere, thus facilitating unconstrained deliberations of opinions, beliefs, knowledge and ideologies (Cammaerts, 2009). Obviously, people in Serbia want to preserve this right and online comment sections provide it. On the other hand, readers’ comments reflect the interaction between discourse structures and social structures via a cognitive dimension (van Dijk, 2018), thus exhibiting the mental models, knowledge or ideologies, as part of the actual production and comprehension by Serbian language users. In other words, readers of news in Serbia deliberately choose the online space of mediated interaction provided by the comment sections on websites of news outlets. One reason is that this type of communication is anonymous. However, a more important reason established earlier (Section 3) is that Serbia ‘‘utterly fails to meet EU press freedom standards” (RWB, 2018) and almost all media are exposed to political influence primarily exerted by the Serbian government (Ðordevic´, 2019; Matic´, 2014; RWB, 2018). This leads to the conclusion that communication via comment sections seems to be the only way for readers of news in Serbia to actively participate in the political reality of Serbia. The final analysis focusing on the correlation between the linguistic devices identified in the articles and the instances of HS identified in the readers’ comments indicates a significant relationship between these two variables (r = 0.567, p = .000) (Table 6). The conclusion is that a certain set of news values (Table 1) motivated the readers not only to post comments but also to use offensive and abusive language. The present research shows that communication via online comment sections reflects a negative impact of particular news and news values (Bednarek & Caple, 2014) on the sociocognitive interface between the online discourse on Serbian websites and the readers of the news as representatives of the Serbian society (van Dijk, 2018). Such finding highlights the need for further research with three primary foci: (1) the influence of the media in the context of freedom of speech in societies where this democratic right is not guaranteed, (2) the importance of comment sections and the necessity to preserve them as a forum of expressing Table 6 Correlation between Linguistic devices identified in articles vs. instances of HS.
Linguistic devices discovered in articles Instances of HS **
Linguistic devices discovered in articles
Instances of HS
Pearson Correlation
1
0.567(**)
Sig. (2-tailed) Pearson Correlation Sig. (2-tailed)
0.567(**) 0.000
Correlation is significant at the level 0.01 (2-tailed).
0.000 1
public opinion and (3) the relationship between online discourse and society in countries with a political context similar to the one in Serbia.
6. Conclusion Based on the analysis of discourse structures in readers’ comments, the present research has confirmed that HS in the readers’ comments on the news website Politika is a rather frequent means of establishing a communicative Common Ground among news readers. The analysis has also shown that HS is a reflection of the negative impact of news and news values on the sociocognitive interface between discourse and society. Various negative forms of social cognition are created in the common space between discourse and society while at the same time the social cognition of the members of linguistic and discursive communities influences and shapes that very same space. From a methodological point of view, the present research has shown that SCDS (van Dijk, 2016, 2018) may be applied to the analysis of hate speech in readers’ comments, thus enabling not only the analysis of hate speech as a discursive practice but also the identification of the social cognition created through hate speech. CDS has undoubtedly established itself as a discipline investigating the power of discourse in affecting and shaping society and socio-political orientations (Wodak & Meyer, 2009) as well as a valuable contextualization tool (KhosraviNik and Zia, 2014). Nevertheless, the development of participatory media and social networks as a new communication environment has instigated the necessity to (re)evaluate the applicability of CDS to specific online discursive practices, hate speech in the comment sections of news websites being one of them. Therefore, this research has shown that SCDS facilitates a new interdisciplinary synergy with CDS (KhosraviNik & Esposito, 2018) to analyse hate speech as a specific type of discourse in mediated forms of communication. That is why it may also be concluded that SCDS is not only a new approach within CDS but also an analytical tool which will facilitate CDS to evolve as a discipline since it may be used for the critical description of the cognitive aspects of concepts used by readers while relating the structural properties of text or talk to social structures. Van Dijk’s discourse structures (2018) may easily be applied within critical discourse analyses related to other types of digital media discourse, such as discourse occurring in social media where users tend to express their thoughts and attitudes quite freely. Obviously, this is an ideal setting for the investigation of the role of such discourse as a reflection of the sociocognitive dimension (KhosraviNik & Zia, 2014; van Dijk, 2018) existing between a certain type of discourse and the readers as members of society. The newly-established synergy between SCDS and CDS, along with more specific approaches to digital media discourse, such as SM-CDS (Khosravinik, 2017), provides elements for the discourse-cognition-society triangle, thus supplying the relevant tools to explore the cognitively mediated relations between discourse and society (van Dijk, 2018) in mediated forms of communication (KhosraviNik, 2017). Although the negative impact on mental models that readers construe when exposed to the discourse presented to them in the news is limited to specific news content, the fact that readers opt for HS when expressing their opinions clearly indicates that more research is needed to explore the immediate effect of news on the sociocognitive interface between discourse and society.
Declaration of Competing Interest The author declare that there is no conflict of interest.
J.P. Ðordevic´ / Discourse, Context & Media 33 (2020) 100366
Acknowledgements This paper benefited from the project The Dynamics of Structures of the Contemporary Serbian Language (178014) supported by the Ministry of Education, Science and Technological Development of the Republic of Serbia and the project Credibility, Honesty, Ethics, and Politeness in Academic and Journalistic Writing (CHEP 2018), supported by DAAD (March 2018 – February 2019).
Funding sources This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Appendix A. Supplementary material Supplementary data to this article can be found online at https://doi.org/10.1016/j.dcm.2019.100366. References Annual Progress Report of the European Commission for Serbia, 2019. Retrieved from
9
KhosraviNik, M., 2017. Social media critical discourse studies (SM-CDS). In: Flowerdew, J., Richardson, J. (Eds.), Handbook of Critical Discourse Analysis (pp. 582–596). Routledge. KhosraviNik, M., Zia, M., 2014. Persian nationalism, identity and anti-Arab sentiments in Iranian Facebook Discourses. J. Lang. Polit. 13 (4), 754–780. KhosraviNik, M., Esposito, E., 2018. Online hate, digital discourse and critique: Exploring digitally-mediated discursive practices of gender-based hostility. Lodz Papers in Pragmatics 14 (1), 45–68. Kulacˇanin, N., 2018. 3.456 botova [3,456 bots]. Danas. Retrieved from