Views and interpretations expressed in this document do not necessarily reflect the views of the Embassy of the Netherlands in Kosovo.

This Research Report was prepared under the “Combat Hate in Kosovo” Project, implemented by ATRC in partnership with RADC and supported by the Embassy of the Netherlands in Kosovo

Author: Valon Junuzi

Researchers: Mimozë Veliu and Tringë Sokoli

Pristina, 2021 2 Introduction

This Research Report on the use of hate the research data, main trends per category, speech in the Kosovan media is published and the correlation between the variables. by ATRC in partnership with RADC, with the This is the crucial part of the research findings. support of the Embassy of the Netherlands in The fourth section provides an overview of the Kosovo, under the “Combat Hate in Kosovo” main linguistic characteristics used in media Project under the MATRA Program. The pur- discourse in Kosovo, and an interpretation of pose of this research is to take record of the the findings from the previous section. The use of hate speech on ethnic, gender, racial, fifth section provides recommendations with religious, and other grounds in public and to reference to international findings and good combat this phenomenon which incites hate practices. acts or crimes. Through this research activity, ATRC and RADC have created a database This research does not aim to reach final con- and subsequently compiled this report which clusions on the use of hate speech in the Koso- will serve to inform the public, promote a more van media, nor does it aim to make the data effective institutional response to hate related published in this Report universally applicable cases, and develop local resources to address outside the given time frame and context. it effectively. Rather, the purpose of this research is to pro- mote public research and discussion on this This report is based on textual and audiovisual topic, which is generally perfunctory in Koso- media research for the period January-June vo. Through this Report we aim to promote a 2019, and January-June 2020. In the second healthy public debate on ethics and profes- half of 2021, under the same project, ATRC sional principles in the Kosovan media when and RADC aim to conduct further research reporting on discriminated and easily vulner- and publish related results on a second report, able groups and communities. Recognizing covering the period June-December 2019, the irreplaceable role that the media have not and June-December 2020. Data collection only in reporting the events that make up our and classification was carried out from 1 Oc- collective and public life, but most important- tober 2019 through 15 November 2019. The ly, in shaping public attitudes towards social data presented in the present Report is based phenomena, this Report attempts to pave the on the main database, which is the intellectual way for further research, which would put the property of ATRC. spotlight on this role of the media in post-con- flict societies and societies in transition (not The Report follows the structure of similar only economically, but also in terms of values) research reports, where it begins by providing such as Kosovo. the necessary information on the used defi- nitions, sample selection, and data collection and classification. The second section focus- es on the legal framework in Kosovo and the world governing the prohibition and sanction- ing of hate speech in public discourse. This section provides a comparative perspective between the countries under examination, and good practices that can serve as a model for Kosovo. The third section exclusively presents 3

Introduction 2

Executive Summary 4

Basic Definitions, Methodology Used, Overview of Hate Speech Presence in 1. and Selection of Media for Analysis 6 3.1 Selected Media 24

Dominant Trends by Community and Chosen Definition and Key Elements of 3.2 Relationship Between Relevant Vari- 1.1 Hate Speech 6 ables 32

Content Analysis and 1.2 Analysis Units 9 Putting Findings in Context and Giving 4. Them Meaning 39 Media Selection, Data Collection, and 1.3 Classification of Findings 13 Key Characteristics of Problematic Comparative Aspects between Appli- 4.1 Language by Selected Categories 40 2. cable Standards in Kosovo and Globally 17 Analysis of Used Discourse and Identifi- National Legal Framework Against 4.2 cation of Problematic Language 2.1 Incitement to Hatred and Ethical Stan- Structures 44 dards Applicable in Kosovo 18

Key Research Findings and Identifica- RECOMMENDATIONS 48 3. tion of Representative Trends for 5. BIBLIOGRAPHY 49 Target Communities 23 EXECUTIVE SUMMARY

Report on ‘Research on Hate Speech in Koso- cles) and 201 audiovisual materials (TV shows) vo Media 2019/2020 I’ is the first empirical have been analyzed and transferred to the research on the presence of hate speech in main database. Kosovo media. As such, the report represents a significant step in monitoring the media coverage of events related to communities in Kosovo which are most exposed to vulnerabil- At the aggregate level, findings show that in ity and verbal discrimination. The aim of the the period covered, 12.05% of online articles report is to provide scientific based evidence and 9.30% of analyzed TV shows use hate on the use of hate speech in Kosovo media speech against target communities or cite im- and consequently exposes ways through which properly persons or authorities that use hate ethnic communities, LGBTIQ and migrants speech against a particular group. It is worth are ‘othered’ in Kosovo’s media. To this end, noting that in the online articles the dominant the report delves into a critical examination of frame used in the hate speech is that which media coverage of these communities, by de- depicts the ‘other’ as a violent, and a threat- constructing language used to portray these ening element in the society. This is followed by communities or actions inferred in the use of a offensive reporting towards the traditions and particular discriminatory language. lifestyle of the communities, and to lesser ex- tent by frames that explicitly or implicitly incite violence or call for murder of target communi- ties. On the other hand, the dominant frames On the methodological level, the report relied in the TV shows with hate speech content use on the qualitative method of content analysis offensive and negative representation, fol- to enable the researcher identify dominant lowed by prejudicial and essentialist language, frames used to portray target communities, and violent representation of target communi- and semantic relations which construct the ties. otherness of these groups in relation to the rest of the society. This method is applied in textual analyses of 10 online media articles, and linguistic analysis of 6 TV channels. The At the group level, results show that in both report has used objective criteria to determine online articles and TV shows, com-munities which media is included in the research, and to (namely Roma, Ashkali and Egyp-tians) are increase as much as possible the generaliza- mostly targeted with hate speech, followed tion of the findings. In order to engage in com- closely by Serbs, and to a lesser ex-tent by parative analysis of trends, the examination of migrants and LGBTIQ community. More articles and shows in these media outlets was specifically, media tends to portray com- conducted in two six-month periods, 1 January munities Roma, Ashkali and Egyp-tians as 2019 - 30 June 2019, and 1 January 2020 aggressors (against their own members), - 30 June 2020. The research applied total sexual predators, and their culture is population sample, and simple random sam- associated with lack of hygiene and re- ple to choose articles for textual analysis and sponsibility. Regarding LGBTIQ, the dominant criterion sample for linguistic analysis of TV reporting frames are those targeting mem- shows. In total, 2,512 textual materials (arti- bers of this community as deviant, nationally incompatible, and dangerous. The reporting in this category portrays LGBTIQ persons as focuses on codes of ethics adopted in dem- offenders of the ‘laws of nature/God’ with ocratic countries, and on successful media emphasis on their ‘abnormal’ inclinations. For self-control mechanisms which guarantee fair immigrants/refugees, reports tend to present reporting on social phenomena, without un- them as aggressors, dangerous, and manip- dermining the editorial independence of the ulators. The framing approach that portrays media. The report also gives a list of recom- immigrants/refugees as dangerous in articles mendations for all relevant actors involved in attempts to depict them as factors of insta- public discourse which if implemented could bility in Kosovo and as elements of extremist sensibilize media and other actors about the movements in the Middle East. For Serbs, effects of hate speech in shaping people’s reports associate all Serbs with war crimes opinions and attitudes on particular communi- and atrocities, and highlight the incompatibility ties. between Albanians and Serbs, and portrayals as aggressors are predominant.

The report also includes a section on success- ful international practices about mechanisms put in place for preventing the use and spread of hate speech in public discourse. The section 6 I. Basic Definitions, Methodology Used, and Selection of Media for Analysis 1. Basic Definitions, Methodology Used, and Selection of Media for Analysis

1.1 Chosen Definition and Key Elements of Hate Speech

The term ‘hate speech’, together with other historical determination. For this reason, hate terms that can be used interchangeably with speech scholars – unlike concept scholars and it, is one of the most problematic terms for jurists who treat hate speech as a concept, definition and delineation, as well as for estab- rather than a phenomenon – in analyzing the lishing its applicability in concrete cases. This presence of hate speech in public discourse, challenge is not merely technical by nature – always rely on the definitions of International finding a definition that would be comprehen- Organizations, as the only option to escape sive yet not infringe on freedom of expression the arbitrariness of a phenomenon lacking – but also due to the evolutionary character of universally accepted definition. Internation- the term itself over the last 70 years. Regard- al Conventions comprising the International ing the former, it should be noted at the outset Human Rights Law have consistently included that today’s debates on the prohibition of hate the prohibition of hate speech in their enacting speech are constantly taking place in rela- clauses, thereby providing us with a credible tion to freedom of expression, and that these resource to underpin research into the occur- discussions always complicate the problem rence and implications of hate speech in tar- of striking a compromise between the funda- get societies. mental freedom of expression and the right to equality and dignity. As for the latter, ‘hate Among the first international documents ad- speech” is a term that is inherently changeable dressing inter alia hate speech is the Universal and deeply embedded in the context where it Declaration of Human Rights (UDHR, 1948), occurs, thus making the need to define it not while although it does not explicitly establish an effort for universal and trans-historical ‘hate speech’ as a standalone category, stip- definitions, but rather, for fluid, localized, and ulates ‘protection against any discrimination’ I. Basic Definitions, Methodology Used, 7 and Selection of Media for Analysis and ‘any incitement to such discrimination’ as This definition is considered the most compre- provisions which limit the right to freedom of hensive definition4 of ‘hate speech’ because expression. In the same vein, Article 29 of the it is not limited to prohibiting hate speech as UDHR refers to duties to the community and an act (as the ultimate manifestation of hate recognizes the possibility of limitations in ex- speech), but also includes intermediate acts ercising certain rights to enable ‘recognition such as disseminating, inciting, promoting, and and respect for the rights and freedoms’ of justifying hate speech. In this sense, the impli- others. Since then, several international in- cations of this definition are that the responsi- struments have been adopted progressively bility for hate speech falls not only on the per- recognizing the need to eradicate hate speech son who expresses it, but also on the parties and elaborating circumstances that justify that serve as platforms for the dissemination certain limitations on the right to freedom of and promotion of such language. Additionally, expression.1In this spirit, regional human rights the definition provides a broad base of pro- instruments were developed, more specific in tected characteristics, which include a wide addressing hate speech because they respond range of grounds for intolerance. This defini- to the specific and contextual needs of differ- tion was the most appropriate for the purpos- ent regions2. es of this research because the object of study was the language used in the media against In the absence of a standardized definition communities that are considered marginalized used by all international instruments (in fact, in Kosovo – which can be not only the prima- the definition of ‘hate speech’ by various acts ry agents of the expression of hate speech, reflects the debates of the time on the com- but also conveyors of such language by other promise between ‘freedom of expression’ and actors. Accordingly, this definition enables the ‘hate speech’), this research will rely on the analysis of media attitudes towards margin- definition issued by the Committee of Ministers alized communities where these attitudes of the Council of Europe (1997), which though express the editorial policies of the media, but not an official definition, represents one of also where the media serve as platforms for the most comprehensive definitions of hate inciting hatred by third parties. In other words, speech. This CoE definition provides that hate the object of study in this research is textual speech shall be understood as and audiovisual material in relation to select- ed communities, where the classification of “covering all forms of expression which spread, articles and TV shows as problematic content incite, promote or justify racial hatred, xenopho- material is done considering the attitudes of bia, anti-Semitism or other forms of hatred based the media towards the phenomenon subject on intolerance, including intolerance expressed by to reporting (article writer’s attitude), and how aggressive nationalism and ethnocentrism, dis- the media feature other parties’ attitudes to- crimination and hostility against minorities, mi- wards the phenomenon in question (quoted or grants and people of immigrant origin”3 paraphrased statements of third parties).

1 The International Covenant on Civil and Political Rights (1978) for the first time explicitly requires Signatory States to prohibit by law ‘any advocacy of national, racial or religious hatred that constitutes incitement to discrimination, hostility or violence’ (Article 20, para. 2). However, the Convention is exceptionally prudent in authorizing limitations on freedom of expression, presenting specific circumstances which must be met before acting against allegedly problematic language. Likewise, the International Convention on the Elimination of All Forms of Racial Discrimination (1969) requires State Parties to prohibit by law ‘all propaganda and all organizations which are based on ideas or theories of superiority of one race or group of persons of one color or ethnic origin’ (Article 4 (a)). 2 Some of the most renown regional instruments are European Convention on Human Rights (1950), American Convention on Human Rights (1969), African Charter on Human and Peoples’ Rights (1981) 3 Recommendation No. R (97) 20 of the Council of Europe, 1997 For more information, see: https://rm.coe.int/1680505d5b 4 This definition was later expanded to include sexual orientation as a protected characteristic. For more information, see: https://search.coe.int/cm/Pages/ result_details.aspx?ObjectID=09000016805cf40a 8 I. Basic Definitions, Methodology Used, and Selection of Media for Analysis

At the level of analysis, the research has clas- 1. Context, which refers to the social and sified as hate speech materials those articles political circumstances where hate speech or TV shows which express or incite hatred occurs. In practice, this means recognizing the towards persons/communities based on an prevalent conflicts/tensions in the society sub- immutable, fixed characteristic that consti- ject to research, at the time when hate speech tutes a fundamental identity element of the occurred. Based on this test, the research target person/community. On this basis, arti- scrutinized the materials considering the social cles and shows classified as hate speech con- and ethnic tensions in Kosovo, with special fo- tent materials meet this important criterion, cus on recognizing the dominant structures of which enables us to identify the occurrence of social, political, and cultural power that mar- hate speech more easily, but also serves as a ginalize certain communities, and institutional- mechanism to prevent the misclassification of ize and justify discrimination. hate speech content, which can easily fall into the domain of free expression of thought. As 2. Identification of Speaker, or Recognition an illustration, an article that talks about the of the Authority of the Person who utters hate “threat that Muslim immigrants pose to Koso- speech is important to understand the effect vo” can be classified as hate speech against that a certain expression may have on the immigrants, because being a Muslim (a funda- group who perceives the Speaker as an Au- mental part of a person’s/immigrant’s identity) thority that can influence its actions. As will be is represented as problematic and associated seen in the results section, there is a consid- with potential acts of violence. Whereas an erable number of cases where persons in the article that talks about the ‘criminal acts by capacity of religious/political or professional the Syrian President’ does not constitute hate authority incite and promote hate speech speech, because the (criminal) act is individ- against certain communities, influencing how ualized, and does not relate to an immutable followers shape their attitudes. characteristic of the subject (Syrian President), but to his position as a politician, party leader, 3. Intent is another important test for sect leader, etc. (voluntary, mutable affilia- identifying whether incitement to hatred was tions). planned or an unintentional omission by the Speaker. Although it is difficult to find out with In addition to this basic criterion, this research certainty what was the real intent behind an draws on the practices of National Courts expression, judges usually check how often when dealing with cases involving allegations the person in question used the specific term, of hate speech.5 In dealing with these cases, the track record of the Speaker’s opinions, and the courts use clear tests that help determine how explicit was the hatred in the used lan- whether a certain expression qualifies as hate guage. Likewise, this research has classified speech. In researching and classifying data, as problematic materials those articles/shows we have used such tests to determine whether with a recorded trend of the language or ex- materials reach the ‘threshold’ to be classified pression used and the semantic aspects of the as materials that incite hate speech against a article/show in question. certain community. These tests are6:

5 For more information on EU policies against hate speech and the Case Law of the European Court of Human Rights, see: https://www.europarl.europa.eu/ RegData/etudes/STUD/2020/655135/IPOL_STU(2020)655135_EN.pdf 6 These tests are elaborated in more detail in: https://www.refworld.org/pdfid/50bf56ee2.pdf I. Basic Definitions, Methodology Used, 9 and Selection of Media for Analysis

4. Intensity and Likelihood of Harm Oc- expressions, and what is the implicit appeal curring, is another complementary test which that the expression entails against the target analyzes the frequency of occurrence of hate community. speech and the real likelihood of the expres- sion inciting violence against the targeted per- The above tests are at the same time the main son/community. In the case of this research, elements of hate speech, and this research this test has been applied to analyze how the relies on these to analyze and classify the arti- target community is framed by the Speaker’s cles and shows subject to examination.

1.2 Content Analysis and Analysis Units

In data collection and interpretation, this choice of this method consists in recognizing research used the qualitative method of con- the fact that the media represent not only a tent analysis, which enables the identifica- reflection of social reality, but also at the same tion of predominant terms/symbols within time influence the shaping (construction) of any text/show and the analysis of explicit or this reality. And, since ‘language’ used in me- implicit messages in the material subject to dia communications is not neutral in terms of examination.7. The researchers of this Report being detached from the predominant values used content analysis as a tool to determine in the relevant society, content analysis allows the presence of hate speech in the examined the deconstruction of everyday media vo- materials, and the relationships between the cabulary and of the ways how media lexicon main subjects displayed in the material. More reflects and actively contributes to the stabili- specifically, this research is based on rela- zation of the dominant social narratives. 8 For tional content analysis which focuses not only this reason, the content analysis method fits on establishing the presence of problematic best with the purpose of this research, which terms, but also examining the semantic re- is directed not only towards finding explicit lationships built around these terms. In other expressions that incite hatred, but also repre- words, research is not reduced to finding rac- sentative trends on marginalized communities ist/homophobic/xenophobic terms, but also that incite hatred and are embedded in daily attempts to find implicit messages embedded media reporting. in the material, which are particularly import- ant to understand how the ‘other’ is construct- In this sense, this research goes beyond iden- ed and what are ‘his’ relationships with the tifying the disparaging terms/jargon used by ‘other’ part of society. The rationale behind the the media for the targeted communities, but

7 For more information on content analysis as a method, see: http://www.umsl.edu/~wilmarthp/mrpc-web-resources/content-analysis.pdf 8 For more on the role of media in creating and constructing social realities, see Fowler, R (1991), Discourse and Ideology in the Press, Routlege, New York 10 I. Basic Definitions, Methodology Used, and Selection of Media for Analysis also puts the spotlight on how media report This method was applied to the collection and on these communities even when there are no interpretation of data on 4 social groups which problematic terms in the material subject to as minorities, are considered the most vul- examination. As shown below, most articles/ nerable in public discourse. These 4 groups, shows classified as hate speech do not use ex- serving as the key units of analysis, are: Serb, plicit terms against the target community, but LGBTIQ, Immigrant/Refugee, and Roma, present (construct/shape) the target commu- Ashkali and Egyptian communities. The main nity in a discriminatory/inferior/violent man- reasons for including these categories in the ner in relation to the majority community. For research are related to the tensions/conflicts this reason, the research used a scale-based in relation with the majority community (the data classification scheme, where articles are case of the Serb community), social taboos on categorized according to how the target com- sexuality munity is constructed, and the implicit appeal (LGBTIQ community), cultural prejudice (Immi- implied by the content. Accordingly, through grants/Refugees), and ethnic prejudice (Roma, content analysis we have classified as incite- Ashkali and Egyptian minorities).9 The results ment to hatred all those articles/shows which section will show how these groups are present the action of the subject because of covered in the media using binary terms such its immutable characteristics. By way of illus- as ‘us’ and ‘them’, and that the actions of tration, media reports that ‘a Roma killed his members of these groups are constantly brother’ are racist and incite hatred against explained as a natu-ral consequence of their members of the Roma minority, because of affiliation (being an immigrant/refugee is all the traits of the perpetrator, the author of associated with violence and looting, and the article chooses to zero in on the perpetra- being a Serb with recent war crimes in tor’s ethnic background while reporting on his Kosovo). Therefore, it is expected that such actions. In this regard, content analysis has prejudice and this deep ignorance of the enabled the understanding of these semiot- ‘other group’ will be manifested in the pub-lic ic structures which create perceptions in the discourse, namely by the media as crucial audience about certain communities, conse- articulators in this discourse. These key ana- quently contributing to the preservation and lytical units have been examined in relation to replication of bias against these communities. the ‘grounds of hate speech’ (bias categories), which serve to identify the underlying causes of hate speech and target a particular social group.

9 Among Roma, Ashkali and Egyptian minorities, it should be noted that articles and shows concerning the Roma, Ashkali, and Egyptian minorities have been examined because of systematic discrimination against these minorities and past tensions with the majority community. The data show that hate speech is exclusively directed at members of the Roma, Ashkali, and Egyptian ethnic groups. The content of such language and the way the ‘other’ is shaped in relation to these three ethnicities is the same, with differences that are negligible for analytical purposes. I. Basic Definitions, Methodology Used, 13 and Selection of Media for Analysis

The main findings show that hate speech against selected groups is based on ethnicity/ race and sexual orientation. There are cases where the underlying causes of bias are two- fold, and the target group (unit of analysis) is subject to hate speech based on both eth- nicity and sexual orientation (grounds of hate speech).

1.3 Media Selection, Data Collection, and Classification of Findings

This research focused on examining the con- These media outlets are: koha.net, gazeta- tent of articles and shows broadcast by the express.com, indeksonline.net, klankosova.tv, Kosovan national and local media. The sample lajmi.net, telegrafi.com, and kallxo.com.12 While selection includes a total of 10 electronic me- the 3 local broadcast media were selected dia for the textual analysis of media articles, according to the following requirements: 1) and 6 TV channels for the linguistic analysis of To have been on air as of 1 January 2019, at TV shows (audiovisual materials). For purpos- minimum, 2) to be as original as possible in es of establishing a comprehensive overview of media coverage, and to copy mainstream me- the potential differences in the Kosovo media dia news as little as possible, and 3) to have a regarding hate speech, this research examined functional digital archive, and 4) to be media electronic articles published in 7 electronic outlets based in the largest centers of Kosovo, media outlets that have an audience at the if possible13. The selected sample included: ve- national level, and 3 local broadcast media. All riu.info, tvprizreni.info, and drenicapress.com. 7 electronic media outlets with national audi- Regarding the analysis of audiovisual materi- ence were selected referring to their ranking als, the analysis included shows broadcast on on the www.alexa.com platform, which pro- mainstream national media: RTK, KTV, RTV 21, vides global rankings and web traffic analysis Klan Kosova, TV Dukagjini, and T7. by country.10 Referring to the list provided by this platform, the sample selection includes The examination of articles and shows in these 7 media outlets top ranked by the audience.11 media outlets was conducted in two six-month

10 As Kosovo is not listed on the platform, Kosovan media appear in the list of media from Albania. 11 Since data collection started in October 2019, the 7 top ranked media outlets at the beginning of October 2019 were selected. The listing was filtered to remove Albanian media outlets from the ranking, thereby leaving only Kosovan media outlets in the running for selection. 12 In the results section, these media outlets are not identified by name for the sake of maintaining their reputation. The media outlets are referred as Media 1 through 10. 13 As will be shown in the data section, the selected local media outlets are hardly able to meet all these requirements. However, compared to other local media outlets, they rated as the most suitable for purposes of this research. 14 I. Basic Definitions, Methodology Used, and Selection of Media for Analysis

periods, 1 January 2019 - 30 June 2019, and As for audiovisual materials (shows), the re- 1 January 2020 - 30 June 2020. The reason search approach used criterion sampling for choosing this monitoring period involves which involves selecting units (shows) that finding and comparing reporting trends be- meet some criterion of importance prede- tween the two semesters and analyzing even- termined by the researcher.15 In this regard, tual changes in media discourse from one research focused on finding shows airing on period to another. The sample used for ana- dates where the likelihood of mentioning a lyzing textual materials was a total population particular community is higher compared to sample, and a simple random sample. The other dates. In other words, shows are exam- total population sampling approach was used ined based on predetermined broadcast dates to examine articles in relation to the LGBTIQ because the researcher has guesstimated that community, Immigrants/Refugees, and Roma, on those dates the 4 selected groups will be Ashkali and Egyptian minorities. This means that subject to discussion.There were two shows all articles selected for each media outlet: 1) prime time (across feature sections) written in relation to news, and 2) one (1) show where political and these categories that are accessible have cultural events are discussed. The table below been examined, thus increasing the credibility presents the sample used for both types of and representativeness of the findings. The analysis, textual and audiovisual. simple random sampling was used for the Serb com-munity, where each article is equally likely to be selected by the researcher.14

14 The main reason for this sampling approach involves the enormous number of articles generated by the application of keyword ‘Serb’, such that the examination of all articles was rendered impracticable. 15 For more information on using this sample in qualitative research, see: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4012002/ 15

Figure 1: Sample Applied in Research Related to Textual and Audiovisual Materials

TARGET GROUPS

Non Serbian Serbian LGBTIQ Immigrants/ minorities Refugees

Textual Materials

GazetaEx- GazetaEx- GazetaEx- GazetaEx- Electronic press.com, press.com, press.com, press.com, medias koha.net, koha.net, klankoso- koha.net, koha.net, klankoso- klankosova.tv, va.tv, telegrafi.com, klankosova.tv, va.tv, telegrafi.com, telegrafi.com, indeksonline.com, telegrafi.com, indeksonline.com, indeksonline.com, lajmi.net, indeksonline.com, lajmi.net, lajmi.net, kallxo.com, lajmi.net, kallxo.com, kallxo.com, tvprizreni.info, kallxo.com, tvprizreni.info, tvprizreni.info, drenicapress.com tvprizreni.info, drenicapress.com drenicapress.com drenicapress.com

All items from the Articles selected based All items from the All items from the Selected period of 1 on random sample period of 1 January - period of 1 January - January - 30 June selection 30 June 2019 and 1 30 June 2019 and 1 Articles 2019 and 1 January - 30 June January - 30 June January - 30 June 2020. 2020. 2020.

Audiovisual Materials

Televisions RTK, Klan RTK, Klan RTK, Klan RTK, Klan Kosova, RTV Kosova, RTV Kosova, RTV Kosova, RTV 21, KTV, T7, 21, KTV, T7, 21, KTV, T7, 21, KTV, T7, TV Dukagjini TV Dukagjini TV Dukagjini TV Dukagjini

TV Shows Main news (x6) Main news (x6) Main news (x6) Main news (x6) IMAZH, Bon IMAZH, Bon IMAZH, Bon IMAZH, Bon Bon, Frontal, Bon, Frontal, Bon, Frontal, Bon, Frontal, Debat Plus Debat Plus Debat Plus Debat Plus

Dates 15 February 6 January 2019/ 17 May 20 june 2019/2021 2020 2019/2020 2019/2020 and 8 April 2019/2020 28 June other dates from 24 June 2019/2020 the chosen 2019/2020 25-25 April materials for 280-31 May 2019 2019/2020 textual analysis 16 I. Basic Definitions, Methodology Used, and Selection of Media for Analysis

The data was collected through electronic After collecting the data, the researcher clas- media archives, or other alternative channels sified the hate speech articles/shows accord- (official channels on YouTube, and Facebook). ing to a 5-color code, where each color had an The generated articles and shows were ana- explanation attached as applicable to articles/ lyzed and the researcher, depending on con- shows subject to classification. This color-cod- tent, decided whether it qualified as an article/ ed scheme included a color for classifying show with hate speech. Each item rated as an articles that do not contain hate speech yet article with problematic content was subject use prejudicial and essentialist language. The to content analysis and was entered into the figure below shows the colors used for article main database, including inter alia the prob- classification. lematic sentence, grounds of bias, the entity suffering from the alleged target community action, and the target community framing ap- proaches by media outlets. Similar filters have been used to collect data from audiovisual materials, with some additional boxes (time/ minutes of the show where the speaker utters hate speech, and the capacity of the speaker).

Figure 2: Color Coded Classification of Hate Speech and Prejudicial Language

There is hate There is hate There is hate There is There is no There is no The group There is no There is hate There is hate speech and speech and speech hate speech hate speech, mention of is hate speech, speech speech killing is violence is through through however there the group mentioned, however there through through violent directly or implied directly violent negative but are subject to but there is are negative but representation metaphorically or metaphori- representa- not violent prejudicial/es- examina- no hate prejudicial/es- not violent of the implied in cally in relation tion of the representa- sentialist tion speech/- sentialist representa- community relation to the to community community tion of the notions in neutral notions in tion of the subject to community subject to subject to community relation to the language relation to the community examination subject to examination examination subject to community community subject to examination examina- subject to subject to examination tionm examination examination

*Colors for textual data collection *Colors for audiovisual data collection 17

2. Comparative Aspects between Applicable Standards in Kosovo and Globally 18 II. Comparative Aspects between Applicable Standards in Kosovo and Globally 2. Comparative Aspects between Applicable Standards in Kosovo and Globally

2.1 National Legal Framework Against Incitement to Hatred and Ethical Standards Applicable in Kosovo

The national framework against incitement dom of expression reflects international hate to hatred in Kosovo consists of several acts, speech prohibition practices, which attempt which separately address different aspects of to avoid treating hate speech as an ‘a priori the prohibition and sanctioning of incitement regulation’ of freedom of expression. For this to hatred. The Constitution of the Republic of reason, the deliberate prohibition of incitement Kosovo addresses hate speech in the article to hatred is formulated as a limitation that that provides on the right “to express oneself, responds to necessary circumstances (in cases to disseminate and receive information, opin- where national security or social cohesion is ions and other messages without impediment”. threatened). This constitutional provision is di- 16 In the subsequent paragraph, the Constitu- rectly reflected in the Criminal Code of Kosovo tion allows for the limitation of the freedom of providing that incitement to hatred is subject expression “by law in cases when it is neces- to criminal sanctions. Moreover, the Criminal sary to prevent encouragement or provocation Code of Kosovo provides a broader list of pro- of violence and hostility on grounds of race, tected characteristics (Article 70, paragraph nationality, ethnicity or religion.”17 This inclusion 2, subparagraph 2.12), including sexual orien- of hate speech under limitations of the free- tation, language, disability, economic status,

16 Article 40, paragraph 1 of the Constitution of the Republic of Kosovo. 17 Article 40, paragraph 2 of the Constitution of the Republic of Kosovo II. Comparative Aspects between Applicable 19 Standards in Kosovo and Globally etc. This Code is progressive in the sense that to review appeals about breaches of license it establishes as an aggravating circumstance conditions, code of conduct or other legal acts if the criminal offense is a hate act, motivated by broadcasters (Article 35, paragraph 1, and upon the grounds provided therein. Regarding paragraph 3). In case the IMC finds that the the sanctioning of hate speech, this Code in broadcaster has acted in violation of the pro- Article 141 (Inciting Discord and Intolerance), visions provided by law, then it has the right to paragraph 1, stipulates: impose fines or even terminate the broadcast- ing license (Article 30, paragraph 1). 21 “Whoever publicly incites or publicly spreads hatred, discord and intolerance between Regarding the specific rules of media broad- national, racial, religious, ethnic, and other casting, the IMC has issued a Code of Ethics groups or based on sexual orientation, gender for Media Service Providers (MSPs) in the identity and other personal characteristics, in Republic of Kosovo, which aims to determine a manner which is likely to disturb the public the rules of ethics for broadcasters to ensure order shall be punished by a fine or impris- that media reporting shall not endanger the onment of up to five (5) years” 18 [emphasis national security and territorial integrity, and added] ensure due respect for the “diversity of ethnic, cultural and religious heritage in the Republic Other paragraphs in this article criminalize of Kosovo” (Article 1).22 It is important to note incitement and discord leading to acts of vio- that the Code also refers to human rights and lence and social unrest (paragraphs 2, 3 and dignity (according to the Conventions and the 4).19 In addition to the Criminal Code, provid- Universal Declaration of Human Rights), as a ing comprehensive sanctions for inciting and basis for setting ethical rules for media report- promoting hate speech, Kosovo has adopt- ing. In addition to prohibiting Vulgar and Of- ed Law No. 05/L-021 on Protection Against fensive Language, the Code of Ethics specifi- Discrimination, ensuring that persons with cally addresses Incitement to Hatred, in Article characteristics protected by law shall not be 5: discriminated against on such grounds. This law is not applicable to the prohibition of dis- “MSPs should not encourage or promote, criminatory language as rhetoric, but discrim- intentionally or indirectly, any form of discrim- ination as a practice in employment, access ination and intolerance and must not broad- to public services, education, social services, cast any material that denigrates an ethnic etc.20 While on the monitoring of hate speech or religious group or implies that an ethnic or in media discourse (which coincides with the religious group is responsible for criminal ac- object of study in this research), Law No. 04/L- tivity” 23 [emphasis added] 44 on the Independent Media Commission (IMC) provides, inter alia, for the oversight that Although in this paragraph the Code is limited the IMC should exercise over Media Service to citing ethnicity and religion as prohibited Providers in relation to broadcast content. grounds for discrimination and intolerance, the In addition to licensing, the IMC is authorized following paragraph lists a wide range of char-

18 Article 141 of the Criminal Code of the Republic of Kosovo, see: https://gzk.rks-gov.net/ActDetail.aspx?ActID=2834 19 Ibid. 20 Like the Criminal Code, this Law provides an extended list of protected characteristics. The explanation for this is 21 For more information, see: https://gzk.rks-gov.net/ActDetail.aspx?ActID=2809 22 Code of Ethics for Media Service Providers in the Republic of Kosovo, see: http://www.kpm-ks.org/lajmi_i_plot/2297 23 Ibid. 20 II. Comparative Aspects between Applicable Standards in Kosovo and Globally acteristics which should not serve as grounds the Press Code recognizes the crucial role that for hate speech (race, color, gender, religion, the media have not merely in informing the political opinion, sexual orientation, disability, public, but also in shaping public opinion on etc.). The key elements from this paragraph, social phenomena. For this reason, this type of and the following paragraphs, which are rele- organization aims to set minimum standards vant to the findings of this research are: 1) hate for media reporting without compromising the speech is prohibited even when it is implicit, 2) editorial independence of member media. Key generalization and collective blame for crim- points from the Press Code that are relevant inal activities is qualified as hate speech, 3) to the findings of this research are: 1) Journal- language that denigrates, puts in a disadvan- ists and editors must make a distinction be- tageous position, or mocks a certain group,24 tween news and opinion. News means “infor- 4) references to racial, ethnic, religious back- mation about facts and data, while opinions ground, sexual orientation, etc. while report- convey thoughts, ideas, beliefs or value judg- ing on an event must be made only when ments on the part of publishers, journalists such information is of direct relevance to the and editors.”26 2) Ethnicity, sexual orientation, event being reported. As will be shown in the race, religion, and other characteristics shall research results section, most hate speech be reported only when fundamentally relevant materials involve such elements. to the news, 3) While reporting on crime, the media shall not treat any individual as guilty of In addition to this Code which has been issued a crime prior to a court decision to that effect. and is administered by the IMC, the print me- These standards, same as those set out in the dia through the Press Council of Kosovo25 have IMC Code have been applied to data classi- issued the Press Code for Kosovo (2019), as fication, and the results show that non-com- a self-regulatory instrument for the editorial pliance with these rules is quite prevalent in policies of active media outlets in Kosovo. As media reporting. a self-regulatory body, this Council through

24 Article 4, paragraph 2 of the Code for MSPs. 25 For more, see: https://presscouncil-ks.org/ 26 Article 2, paragraph 4 of the Code for Print Media. II. Comparative Aspects between Applicable 21 Standards in Kosovo and Globally

2.2 Overview of Successful Global Practices and Self-Control Mechanisms Against Hate Speech in Print and Audiovisual Media

Democratic countries with longstanding ex- of Ethics in Britain, Ireland, and Australia have perience of dealing with the phenomenon developed specific rules which consider ex- of hate speech, have developed successful isting narratives in society, vulnerable groups mechanisms for preventing the occurrence in society, and general social dynamics. Ac- and spread of hate speech in public discourse. cordingly, these three states in their Codes of Special attention is paid to creating an envi- Ethics in Media are constantly updating media ronment where the public sphere is not only reporting practices to respond to hate speech kept safe from extremist elements in society, against migrants/refugees at a time when but also resources are mobilized in creating extremist groups are spreading propaganda counter-narratives to deconstruct the myths against them, or when certain ethnic groups that provide grounds for hate speech.27 Since are the target of verbal attacks because of a the mechanisms created for hate speech are certain event.28 These practices instruct jour- exhaustive, this section focuses on addressing nalists to consider the prevailing tendencies some codes of ethics regulating media report- in the representation of groups that are vul- ing in Europe and Australia and on elaborat- nerable in relation to extremist groups before ing successful media self-control mechanisms reporting and to neutralize such tendencies which guarantee fair reporting on social phe- by providing facts that disrupt the extremist nomena, without compromising the editorial discourse. In Australia, for example, following independence of the media. the terrorist attacks on some members of the Muslim community in New Zealand, the media The main features that distinguish Codes of were instructed to exercise caution when re- Ethics in Media in democratic countries are: 1) porting on the assassin’s Manifesto29 and not Development of rules that correspond to the to report on the views of groups or individuals specific context, 2) Recognition of the spe- who support such actions. Likewise, this Code cifics of the hate speech target communities, was updated during the so-called ‘Refugee and 3) Extra due care in the terms used and Crisis’, where the media were cautious in using in affirming controversial opinions of charac- terms such as ‘illegal immigrants’ and ‘fake ters recognized as hate speech promoters. asylum seekers’, as these may cause bias in Regarding the former, as an illustration, Codes the audience. Concerning recognizing the spe-

27 For more information on EU policies on dealing with hate speech, see: https://www.europarl.europa.eu/thinktank/en/document.html?reference=IPOL_ STU(2020)655135 28 This approach has been practiced especially after alleged sexual assaults by refugees in Western European countries. Far-right groups have attempted to exploit these events to organize propaganda campaigns over the alleged ‘threat’ posed by refugees from the Middle East. 29 Immediately after the attack in Christchurch, New Zealand, the assassin’s Manifesto, explaining the reasons and motives for the attack, took the social media by storm. 22 II. Comparative Aspects between Applicable Standards in Kosovo and Globally cifics of marginalized communities, the media lic in relation to the materials they broadcast in Britain and Ireland have established rules and the values they promote. In the countries for reporting news which may be sensitive to of the European Union this practice is quite certain communities. In practice, this has been widespread and has proven effective in en- translated into specific rules on how to report suring the implementation of Codes of Ethics. on ethnic minorities, religious minorities, and Arranging this mechanism requires the orga- the LGBTIQ community. In cases of reporting nization of the media through the format of on migrant-related events, the media were the Council of Journalists, which is responsible also advised to provide the perspectives of the for issuing, updating, and monitoring the Code immigrants/refugees themselves and to re- of Ethics.32 At a time when the use of social port on the stories of suffering and challenges media is incredibly widespread, this type of they faced while traveling to their destination self-regulation enables the media to control countries.30 In cases where a group or char- the comments of other parties in the materials acter known for extremist rhetoric expresses they distribute and to react more quickly to problematic attitudes, the media have the remove problematic comments. In this regard, right not to become a platform for dissemi- the 2016 initiative of the European Commis- nating such attitudes to the public, and must sion to issue a Code of Conduct, which in co- take the side of the vulnerable group by pub- operation with various media representative licly condemning the advocacy of violence and organizations requires major platforms, such hatred. 31 as Facebook, Twitter, YouTube, and Microsoft, to enable the reporting of hate speech and es- The second important mechanism developed tablish mechanisms for evaluating such com- in democratic countries to control hate speech ments, and if a particular comment is deemed is the creation of self-regulatory capacities to contain hate speech, it must be removed of the media, which avoid the interference of from the platform. The implementation of this state authorities in editorial policies, enabling Code is monitored by a network of civil society the media to establish their own reporting organizations from various EU countries, which rules. The idea behind such mechanisms is for test the response efficiency of these compa- the media to self-organize voluntary struc- nies and recommend changes if necessary.33 tures that hold them accountable to the pub-

30 For more on rules of media reporting on race, see: Updated NUJ race reporting guidelines - National Union of Journalists 31 Ibid. 32 For more, see: http://europeanjournalists.org/mediaagainsthate/wp-content/uploads/2018/02/Self-regulation-and-%E2%80%98hate-speech%E2%80%99- on-social-media-platforms_final_digital.pdf 33 For more information, see: https://ec.europa.eu/info/policies/justice-and-fundamental-rights/combatting-discrimination/racism-and-xenophobia/eu-code- conduct-countering-illegal-hate-speech-online_en II. Aspekte krahasuese mes standardeve 23 të aplikueshme në Kosovë dhe në botë

3. Key Research Findings and Identification of Representative Trends for Target Communities 3. Key Research Findings and Identification of Representative Trends for Target Communities

3.1 Overview of Hate Speech Presence in Selected Media

As elaborated in the data collection section, main unaddressed. E.g., the search for articles electronic archives of official media channels about the Roma community was conducted by on their online platforms were used to find the using the word ‘rom’ (Roma) and not the words articles and TV show broadcasts subject to ‘rome’ (Roma woman), ‘romë’ (Roma plural) examination. To find articles addressing topics which would exclude from the research articles on selected communities, researchers have that only contain the word ‘rom’. Below are the used some predetermined keywords for each keywords used to find articles for each com- community. Word search through the search munity. option was performed by applying the rule of the shortest common denominator per social Figure 3: Keywords Used to Find Articles group, ensuring that no article referring to the social group subject to examination shall re-

Communities Serbian LGBTIQ Non-Serb Commu- Immigrant/Refu- nities gee

Keywords: ‘Serb’, ‘shka’, ’peder’ [homosex- rom, ashkali, [Immigrant, ’shkije’, ‘shkin(ë)’ ual, bisexual, egjiptian, egjiptas, Refugee, [derogatory terms transexual, lesbi- ‘magjup’, ‘gabel’, asylum(seeker), for Serb, Serbs, an, gay, ‘faggot’] ‘jevg’, RAE [Roma, Syrian, Afghan, Serb woman] Ashkali, Egyptian, Iraqi, Moroccan, ‘gypsy’, ‘gyppo’, Arab, Vranidoll, ‘vagrant’, RAE] Magure] III. Key Research Findings and Identification of 25 Representative Trends for Target Communities

As can be seen in the table, the used key- problematic language (hate speech + prejudi- words were selected to return articles where cial language), of which 314 (12.05%) contain hate speech is explicit (containing direct hate hate speech only. While out of 201 shows ex- speech such as ‘shka’, ‘peder’, ‘magjup’, ‘gabel’, amined, in 86 of them involve topics on the se- etc.) and implicit (that can be expressed indi- lected communities, where 12 (13.95%) contain rectly when mentioning terms such as ‘serb’, prejudicial language (hate speech + prejudicial ‘homoseksual’, ‘rom’, ‘refugjat’, etc.). language), and 8 (9.30%) contain hate speech.

According to the sampling approach explained The figures below show the statistics on the earlier and the keywords listed here, the total distribution of these materials according to the number of textual materials (articles) subject color code assigned for problematic language to examination is 2,512, while the number of classification (elaborated in section 1.3 of this audiovisual materials (shows) is 201. Of the Report). 2,512 articles handled, 453 (18.03%) contain

2,512 Textual Materi- als (Articles) Examined

453 (18.03%) Problematic Lan- 314 (12.05%) guage Materials Of which, with Hate Speech

8 (9.30%) 12 (13.95%) Of which, with Hate Speech Problematic Language 201 Materials Audiovisual Materials (Shows) Communities Serbian LGBTIQ Non-Serb Commu- Immigrant/Refu- Examined nities gee

Keywords: ‘Serb’, ‘shka’, ’peder’ [homosex- rom, ashkali, [Immigrant, ’shkije’, ‘shkin(ë)’ ual, bisexual, egjiptian, egjiptas, Refugee, [derogatory terms transexual, lesbi- ‘magjup’, ‘gabel’, asylum(seeker), for Serb, Serbs, an, gay, ‘faggot’] ‘jevg’, RAE [Roma, Syrian, Afghan, Serb woman] Ashkali, Egyptian, Iraqi, Moroccan, ‘gypsy’, ‘gyppo’, Figure 4: Distribution of Problematic Language by Color Arab, Vranidoll, ‘vagrant’, RAE] Magure] Code for Textual Materials (Total) 26 III. Key Research Findings and Identification of Representative Trends for Target Communities

Total (including Prejudicial Language) For clarifying the color spectrum refer to page 16. 224

139

77

7 6

blue green yellow red black

The results presented above show that the prevalent in these articles replicates existing largest number of articles classified to have social stereotypes towards communities and problematic content are in relation to lan- portrays essentialist views in relation to eth- guage providing a violent representation of nicity or sexual orientation. This type of lan- the target community (yellow). This means guage notes the ethnic or sexual orientation of that the prevailing discourse portrays others, the target person without any relevance to the whose actions, presence, or beliefs are re- news, with essentialist attributions to tradition jected, vilified, and discriminated against, as as an innate part of the group or by publishing a threat to the society. As will be shown in the photos of victims in violation of their privacy following sections, the main context of report- and endangering their well-being. Negative ing in these articles is crime, felonious conduct, representation continues with offensive dis- war wrongdoings, offenses involving moral course (green), appeals for violence (red), and turpitude, and the like. Prejudicial language fewer calls for murder (black) for selected follows next in prevailing representations of communities. target communities. The language that is III. Key Research Findings and Identification of 27 Representative Trends for Target Communities

Figure 5: Distribution of Problematic Language by Color Code for Audiovisual Materials (Total)

Total (by Color Code) For clarifying the color spectrum refer to page 16.

73

6 4 2

violet blue green yellow

In the audiovisual materials (shows), it can be noted that the number of shows involving the se- lected communities on relevant dates is relatively low (42.78% or 86 of the 201 shows examined). There is a prevalence of offensive and negative representation of target communities in the classified shows, where the conduct of members of a particular community is subject to offense and prejudice, and disparaging terms are used in referring to them. Next comes prejudicial and essentialist language (blue), and violent representation of communities subject to examination (yellow).

Figure 6: Distribution of Problematic Language Derived from Textual Materials, by Community

Total (Hate Speech only)

133 125

45

11

LGBTIQ immigrant Roma, Ashkali serbian and Egyptian Com-munities 28 III. Key Research Findings and Identification of Representative Trends for Target Communities

Distribution of Prejudicial Language and Hate Speech by Rhetoric and Type of Appeal

0 5 serb 99 21 42

0 1 80 Roma, 52 Ashkali and Egyptian 66 Communities

0 0 immigrant 44 1 21

3 1 1 LGBTIQ 6 10

blue green yellow red black

For clarifying the color spectrum refer to page 16. III. Key Research Findings and Identification of 29 Representative Trends for Target Communities

The data presented above shows the distribu- periods when their visibility is higher (e.g., May tion of problematic language in textual mate- 17th when the media reports on the LGBTIQ rials, and the type of language used for each community). Regarding the distribution of hate community. As indicated, taken as a whole, speech and prejudicial language by communi- minorities Roma, Ashkali, and Egyp-tian are ty, the data shows that the Serb minority the most exposed to hate speech leads comfortably in violent representation, (133 articles) in media articles, followed close- while Roma, Ashkali and Egyptian minorities lead ly by the Serb minority (125 articles). Of the in terms of prejudi-cial language used against examined articles on immigrants/refugees and them. Language directed at immigrants/ the LGTBIQ community, 45 articles used hate refugees prevails in their violent speech in relation to immigrants/refugees, representation. The most interest-ing finding and 11 articles in relation to members of the may be that the LGBTIQ commu-nity is the LGBTIQ community. It should be noted that only community experiencing calls to murder the number of articles generated for these two language. Of the 6 articles calling for the communities was smaller than for the Serb murder of the ‘other’, all are aimed at and Roma, Ashkali and Egyptian minorities. This members of the LGBTIQ community, which does not nec-essarily mean that LGBTIQ points to the radical and intimidating persons and im-migrants/refugees are less discourse that exists in relation to this vulnerable to hate speech, but rather these community. groups are less fre-quently covered by reporting and the specific

Total (by Hate Speech)

9 8

3

1 1

Roma, Figure 7: Distribution of Problematic LGBTIQ immigrant Ashkali serbian Language Derived from Audiovisual and Materials, by Community Egyptian comunities

Distribution of Prejudicial Language and Hate Speech by Rhetoric and Type of Appeal

40 35 30 28 20 6 10 0 1 0 4 0 0 1 3 04 1 1 1 0 Roma, LGBTIQ immigrant Ashkali and serb Egyptian minorities violet blue green yellow The group is mentioned, There is no hate speech, There is hate speech There is hate speech but there is no hate however there are through negative but not through violent representa- speech/neutral language prejudicial/essentialist violent representation of tion of the community notions in relation to the the community subject to subject to examination community subject to examination examination 30 III. Key Research Findings and Identification of Representative Trends for Target Communities

The data generated by examined shows indi- character, both of racial and sexual orienta- cate that members of the Roma, Ashkali, and tion discrimination. Regarding the distribution Egyptian communities are more vulnerable to by type of hate speech, it is noted that for hate speech in the media broadcasts subject the Roma, Ashkali, and Egyptian and Serbian to research. This number of hate speech communities there is greater coverage on the shows against minority communities focused selected dates. As indicated, offensive and especially on reporting related to the case of a prejudicial language prevails in the coverage Roma woman who was subject to fear on the Roma, Ashkali, and Egyptian commu- mongering and subse-quently harassed and nities, while Serbs and immigrants are repre- brutally beaten in May last year. In this case, sented as violent and framed as dangerous to the media not only re-ported unprofessionally the majority community. The case of the offen- on the events related to the beating of the sive show on the LGBTIQ community involves woman in question, but also incited the mobs the above case of the Roma woman, where against her. It should be noted that hate the sexual orientation of the person is raised speech in this case was not exclusively as the reason that makes the victim suspicious directed at the victim’s ethnicity, but also at and an object of speculation. suspicions that she may be trans-gender or homosexual. In this case, the fear mongering propaganda against her has a dual

Incidence of Prejudicial Language and Hate Speech in both semesters 46 47 41 50 41 37 40 26 27 30 24 22 14 17 22 13 20 12 6 11 6 10 5 0 0 media 1 media 2 media 3 media 4 media 5 media 6 media 7 media 8 media 9 media 10

2019 2020

Figure 8: Incidence of Problematic Language in Textual Materials, Both Semesters III. Key Research Findings and Identification of 31 Representative Trends for Target Communities

Figure 8.1: Incidence of Problematic Language in Relation to Examined Articles (by Medium), Both Semesters

Incidence of Prejudicial Language and Hate Speech in relation to the reviewed articles (for each medium) in both semesters 50 36.9% 41.4% 35.3% 40 32.6% 29.9% 26.1% 27.0% 30 20.2% 23.9% 16.5% 15.5% 16.2% 17.9% 18.8% 17.6% 11.8% 20 10.3% 9.6% 10.1% 10 0.0% 0 media 1 media 2 media 3 media 4 media 5 media 6 media 7 media 8 media 9 media 10

2019 2020

The two figures presented above are gener- ic language. For 2020, the data shows that ated to draw comparisons between the oc- Media 4, 1, 5, and 7 have the most articles with currence of hate speech in the media selected problematic language, a trend which shows for textual material analysis. The first figure that hate speech is more prominent in some shows the number of articles with problematic media outlets than in others (3 out of 4 media language in the first half of 2019 and 2020 outlets are at the top of the list in both semes- respectively, while the second figure shows ters). In relation to this fact, it should be noted the relative incidence which is based on the that it is not important to only identify media number of articles with problematic language outlets that constantly use problematic lan- Incidence of Prejudicial Language and Hate Speech in both semesters in relation to the number of articles examined. guage, but also the content of the language 46 47 41 It is necessary to present relative incidence used. In other words, there are media outlets 50 because for some media (marked with *) it that do not have many articles published on a 41 37 40 was impossible to generate articles for certain particular community, but when they do write 26 27 30 24 keywords (due to problems in their electronic about them, the use harsh language. For these 22 14 17 22 13 20 archives), consequently the smaller number reasons, quantitative research is not enough, 12 6 11 6 of articles can give them an edge over media because despite the accuracy it may provide 10 5 0 without such problems. From what is present- regarding emerging trends, it is insufficient to 0 ed in these figures, we can argue that there is understand the reasons why certain language media 1 media 2 media 3 media 4 media 5 media 6 media 7 media 8 media 9 media 10 no consistent trend which could give us con- is used. 2019 2020 fidence in drawing the conclusion that in one 6-month period there are more occurrences of problematic language than in the other period. While for the media outlets where problem- atic language has appeared with increased frequency, the data suggests that for 2019, Media 1, 10, 7, and 4 are the media outlets that have published the bulk of problemat- 32 III. Key Research Findings and Identification of Representative Trends for Target Communities

3.2 Dominant Trends by Community and Relationship Between Relevant Variables

To better understand how the media report ties that the article alleges suffer from the ‘ac- on selected communities, this research has tions/attitudes’ of the target group. The same paid special attention, during data collection, is applicable for shows, researchers identified to contextual aspects that help identify trends these contextual aspects that are also includ- in the occurrence of hate speech. To under- ed in the graphic representations below. stand how the fear of the ‘other’ is construct- ed, researchers, while entering data into the Figure 9: Distribution of Problematic Lan- database, identified the feature section of the guage in Textual Materials by Feature Section article with problematic content, the framing (LGBTIQ and Immigrant) of the target group by the article, and the enti-

LGBTIQ immigrant

0 1 2 3 4 5 6 7 8 9 10 0 10 20 30 40 50 4

lifestyle region 3

curiosities daily news 47

black show biz 1 chronicles

daily news

sport

world

The data presented in the figure on the In the news, problematic anti-LGBTIQ lan- LGBTIQ and Immigrant/Refugee communities guage takes the form of transmitting ho- show that there is a broader distribution of mophobic statements by third parties (es- reporting events involving the LGBTIQ commu- pecially religious representatives), whereby nity. As shown, in relation to this community, the media outlets that broadcast such atti- problematic language is prevalent in the re- tudes become a platform for disseminating porting of current affairs in Kosovo, followed homophobic attitudes. Most media outlets by world news, and human-interest features. broadcast extremist opinions without any III. Key Research Findings and Identification of 33 Representative Trends for Target Communities warning that these statements are homopho- group is spoken of with problematic language bic and discriminatory. The same trends are about their (il)legality, actions against the ma- observed in world news features, where the jority community, and their movements inside media share the homophobic opinions of world and outside the borders. It should be noted leaders, without applying any editorial filters that although the figure indicates only 1 article or warning the audience about the upcom- appearing in the crime chronicles feature, the ing opinion content. It is interesting to note, ‘criminality’ of immigrants/refugees is featured as will be shown in section 3, the way homo- in the daily news section. Regional Reporting is sexuality is reported in the human-interest mostly alarming and laden with sensational- features, where the traits of homosexuality ism, alerting the public of the impending arriv- are analyzed, interrogated, and objectifying al of refugees in Kosovo. conclusions are given on being homosexual. Concerning immigrants/refugees, articles that Figure 10: Distribution of Problematic Lan- follow daily events related to this group domi- guage in Textual Materials by Feature Section nate the field, reporting on their location, their (LGBTIQ and Immigrant) actions, and their plans to stay in Kosovo. This

Serb non-serbian 120 minorities

114

99 98 95

2 1 1

dark main main news sport dark show biz chronicles news chronicles 34 III. Key Research Findings and Identification of Representative Trends for Target Communities

Regarding the Serb community and the Roma, that reporting in the crime chronicles features Ashkali, and Egyptian minorities, it is interest- is the most prominent for these communities. ing to examine the concentrated distribution The content of such reporting is quite prob- of problematic language related to these two lematic, where members of these communities groups. In relation to Serbs, reporting is mostly are constantly presented as violators, rob- provided in daily news, where the nature of re- bers, and abusers. The same is not observed porting is versatile. Most prevalent is reporting while reporting on violations and offenses on war crimes, where crimes related to the last committed by members of the majority com- war are generally attributed to all Serbs, the munity. Additionally, the daily news features ethnicity of the key person is noted without any predominantly cover the ‘immoral’ behavior by relevance to the news, and the ‘differences’ members of these communities and domestic between Serbs and Albanians are underlined. violence. Sexual abuse and sexually deviant Concerning reporting on the Roma, Ashka- behavior occupy a special place in the daily li, and Egyptian communities, it is incredibly reporting of selected media. telling that reporting on these communities is almost equally distributed between daily Figure 11: Target Group Framing Approaches news and crime chronicles. The data shows in Textual Materials

non-serbian LGBTIQ minorities

100

87

10 20 8 6 4 3

Aggressor Deceitful Deviant Deviant No compatibility Dangerous between sexual orientation and nationality III. Key Research Findings and Identification of 35 Representative Trends for Target Communities

immigrant Serb 80

71 30 28

23

28 2 16 15

4

Aggressor Deceitful Dangerous War Unwanted Cruel Aggressor Deceitful criminal

The figure above shows the principal framing Albanian is portrayed as controversial, mean- approach used by media while reporting in ing a priori incompatibility between the two problematic language on the selected groups. identities that appear to be mutually exclusive. Of the four (4) groups, reporting on the Serb Whereas LGBTIQ persons are portrayed as community employs more versatile framing dangerous when their role in society is under- approaches, while reporting on the other three lined, with prevailing voices that attempt to (3) groups is more consistent. Disaggregated paint the members of this community as ‘fami- data by group shows that when reporting on ly destroyers’ and ‘moral turpitude offenders’. the Roma, Ashkali, and Egyptian communities, the predominant framing approaches por- For immigrants/refugees, reports tend to tray characters from these communities as present them as aggressors, dangerous, and aggressors, deviants, and con men. Articles con men. They are portrayed as aggressors portraying members of these communities as through reports underlining their ethnicity 1) aggressors, are those reporting on alleged when it comes to any criminal offense com- crimes committed by these members, 2) de- mitted in Kosovo. The framing approach that viants, those reporting on ‘abnormal’ acts/ portrays immigrants/refugees as dangerous in behaviors such as early pregnancy or sexual articles that attempt to depict them as factors abuse, 3) con men, in cases covering the forg- of instability in Kosovo and as elements of ex- ery of documents or actions that cause harm tremist movements in the Middle East. to the victim. For Serbs, reports that frame being Serb with Concerning the LGBTIQ community, the dom- war crimes and atrocities, incompatibility be- inant reporting frames are those targeting tween Albanians and (unwanted) Serbs, and members of this community as deviant, na- portrayals as aggressors are predominant. tionally incompatible, and dangerous. Re- They are featured as war criminals through garding the former, the reporting in this cat- generalizations in relation to war crimes and egory portrays LGBTIQ persons as offenders ethnic cleansing of Albanians in the ‘90s and against ‘laws of nature/God’ with emphasis in terms of the legacy of past hostilities. The on their ‘abnormal’ inclination. Concerning the framing approach that portrays them as cruel incompatibility between sexual orientation is evoked through disparaging designations for and nationality, sexual orientation and being Serb(s) as ‘shka’ or ‘shkije’ respectively, a term 36 III. Key Research Findings and Identification of Representative Trends for Target Communities always used in referring to situations where This figure presents the data from the analysis Serbs are shown as ruthless and murderous of audiovisual materials for the 4 groups sub- to Albanians. Regarding the incompatibility ject to examination. Unlike textual materials, between Serbs and Albanians, media reports in TV shows it was observed that the Roma, tend to highlight the cultural and historical Ashkali, and Egyptian communities are framed divisions between the two ethnicities. This in more than one reporting frame. The domi- is most evident when reporting on potential nant framing approach in TV shows involving cooperation, cultural exchanges, and sports these communities portrays them as unclean, activities between the two parties. with emphasis on waste and the dirty envi- ronment. Whereas they were portrayed as Figure 12: Target Group Framing Approaches deviant and dangerous especially when the in Audiovisual Materials object of reporting was a Roma woman, where certain shows in a glaring lack of profession-

Non-serbianNon-serbian LGBTIQLGBTIQ minoritiesminorities

2 2 2 2

1 1 1 1

DeviantDeviant DangerousDangerous DeceitfulDeceitful DangerousDangerous /deviant/deviant

immigrantimmigrant serbserb

2 2

1 1

TerroristTerrorist AgressorAgressor III. Key Research Findings and Identification of 37 Representative Trends for Target Communities alism fueled panic among the population on a security issue, because these people come the possible ‘threat’ from this woman with a from conflict-ridden countries and terror- ‘strange’ appearance. The same framing ap- ist attacks can also target Kosovo. As for the proach is applied to the LGBTIQ community, Serb community, shows that use problematic where the physical appearance of the Roma language in relation to this community frame woman was covered controversially, and her Serbs as aggressors against Albanians with- gender and sexual identity was discussed in out distinguishing between crimes commit- relation to being her ‘dangerous’. Anytime the ted by the Serbian state apparatus and Serb population’s heinous reaction to her was dis- citizens. In both cases, disparaging terms such cussed, the hosts or guest commentators on as ‘shka’ and ‘shkije’ are used by third persons, the show mention the woman’s appearance and without any warning that such attribu- and her ‘suspicious’ orientation. Concerning tions are racist and inappropriate. immigrants/refugees, the show dedicated to their arrival in Kosovo presents the opinion of Figure 12: Entities Suffering from Alleged Tar- a security expert, who says that the arrival of get Group Actions (Textual Materials) these people in Kosovo must be perceived as

Subjects suffering from the alleged actions of target groups

Outside the group

Inside the group 120

100

67

51

42

11 4 5

0

LGBTIQ immigrant non-serbian serb minorities

The figure above presents data on how me- the article implies that the entity suffering dia articles have constructed the entities that from the target group is someone outside suffer from the actions/omissions of the target the group, and ‘in-group’ – where the article group member(s). In other words, the research presents the suffering entity’s group to be the has focused not only on how certain groups same as that of the perpetrator. Accordingly, are represented, but also on the entities that for example, problematic articles on LGBTIQ these articles tend to portray as ‘victims’ of persons present the heterosexual community these groups. The classification of data com- in Kosovo as a suffering entity. The ‘offenses prises of two categories: ‘out-group’ – where involving moral turpitude’ allegedly caused by 38 III. Key Research Findings and Identification of Representative Trends for Target Communities the promotion of LGBTIQ rights are present- interesting sets of data were generated for ed as a problem for the heterosexual family, the Roma, Ashkali and Egyptian communi- which will find it difficult to raise children ac- ties, where the suffering entity in most cases cording to ‘sound moral principles’. Also, in turns out to be ‘in-group’, and less ‘out-group’. many of these articles is mentioned the ‘exis- Accordingly, most of the problematic articles tential threat’ posed by people with different present the members of these communities sexual orientation for the future of the Alba- themselves as victims of their cohorts. The nian nation. Coverage of immigrants/refugees main topics covered in these articles relate is dominated by reporting that portrays them to sexual abuse of minors, early pregnancies, as a threat to the local population (out-group), domestic crimes, and parental negligence and in a few cases as a threat to themselves (e.g., non-vaccination of children). Coverage of (in-group). The local population is presented the Serb community is dominated by articles as the suffering entity from their ‘illegal’ ac- presenting others as victims of Serbs as an tions, the religious and ethnic background of ethnic group. As noted earlier, such reporting immigrants/refugees, and the potential they is largely dominated by war stories where the possess to cause instability in Kosovo. Cases blame for the committed crimes tends to be where they pose a threat to themselves are attributed to all Serbs. reported in articles covering brawls at the Asylum Center in Magurë or Vranidoll. In these Figure 13: Entities Suffering from Alleged Tar- cases, the victim of the actions of an asylum get Group Actions (Audiovisual Materials) seeker is another asylum seeker who shares the same space with the aggressor. The most

Subjects suffering from the alleged actions of target groups

Outside the group

Inside the group

3 2 2 2

1 1

0 0 0

LGBTIQ immigrant non-serbian serb minorities

In audiovisual materials, although the number number of shows is small enough to produce of hate speech shows turns out to be lower, consistent trends, it turns out that the same the trends are almost replicated. LGBTIQ per- as with the written articles, these communities sons are portrayed as a threat to the hetero- are portrayed more as a threat to themselves sexual community, while immigrants/refugees than to others. The topics of hygiene and lack and Serbs are framed in relation to the ma- of awareness are most prominent when these jority community. Concerning the Roma, Ash- communities are portrayed as self-victimizing. kali and Egyptian communities, although the 39

4. Putting Findings in Context and Giving Them Meaning 40

4. Putting Findings in Context and Giving Them Meaning

4.1 Key Characteristics of Problematic Language by Selected Categories

The table below summarizes the print media reporting for the selected groups.

Regarding audiovisual materials, it has not been possible to prepare such a table because the number of problematic language shows was not concentrated in one or two media, rather in many of them. This has made it im- possible to derive trends and generalizations on the content of the shows subject to exam- ination. S E L E C T E D G R O U P S

Media Serbian Non-Serb Minorities Immigrant/Refugee LGBTIQ

Media 1 - Objectionable action - The objection- - Refugees from the - Homosexuality is towards the majority able/criminal conduct of Middle East are objectified, mystified community is correlated to persons from the portrayed as dangerous and homosexual the Serbian ethnicity of the Roma, Ashkali, and people, whom the behaviors are put under perpetrator. Egyptian communities citizens should fear. the magnifying glass of - Situations/events where is constantly represent- - Refugees are journalistic ‘research’. Serbs ‘face off’ against ed in association with portrayed as getting Reporting which defines Albanians are constantly their respective benefits from the state homosexuality as a framed as a clash ethnicities. of Kosovo (parasites). choice rather than as a between two enemies and - Negligent acts are - Their ‘illegality’ is biological determination take on the connotation of associated with the highlighted, and their is prevalent. a physical confrontation, Roma, Ashkali, and ethnicity is constantly - Media outlets become where the victory of one Egyptian ethnicities noted when reporting platforms for airing party means loss for the (arson, inebriation, on their actions (brawls, homophobic calls by other ethnicity. In confron- self-harm, etc.). thefts, assaults). local imams inciting tations with Serbs, the - Incest and early hatred and violence individual is always parenthood in minors against the LGTBIQ posited in relation to the are represented as a community. Members of other as an Albanian; - phenomenon associat- the LGBTIQ community Uncensored and unan- ed with being part of the are labeled as terror- nounced use of disparag- Roma, Ashkali, and ists, ‘obscene’, and ing terms for Serbs such Egyptian communities. degenerate, and their as ‘shka, ‘shkina’, ‘shkijet’; speeches are broadcast - Generalization of Serb uncensored. crimes by blaming Serbs for crimes committed by Serbian police, paramili- tary, and military forces. Prevalent use of phrases such as ‘Serb crimes’, instead of language specifying that the crimes were committed by the state authorities of Serbia, and not by the entire Serbian people.

Media 2 - Noting the Serbian Same as Media 1. Reporting on crimes - Positive approach to ethnicity where ethnicity that are not usually the challenges of and has no relevance to the covered when commit- discrimination against news, except for the ted by members of the the LGBTIQ community tendency to associate majority community action with ethnicity. - Uncensored use of disparaging terms for Serbs such as ‘shka’, ‘shkina’, ‘shkijet’ (even when there are calls for violence against Serbs). - Generalization of Serb crimes by blaming Serbs for crimes committed by Serbian police, paramilitary and military forces. Preva- lent use of phrases such as ‘Serb crimes’, instead of language specifying that the crimes were committed by the state authorities of Serbia, and not by the entire Serbian people. Media 3 Same as Media 2 - Waste, pollution, - Use of metaphors such - Positive approach to uncleanliness is as ‘river of Syrians’ and the challenges of and constantly noted in other bombastic discrimination against relation to the neighbor- headlines that are the LGBTIQ community hoods and settlements alarming to the local of members of the population. Roma, Ashkali and - Their ‘illegality’ is Egyptian communities. highlighted, and their - Same as Media 1 ethnicity is constantly noted when reporting on their actions (brawls, thefts, assaults). - Reporting on crimes that are not usually covered when commit- ted by members of the majority community

Media 4 Same as Media 2 Same as Media 1 and 3 - Headlines are alarming - Impropriety of being and warn of threats from gay and Albanian is refugees. implied (State and - Same as Media 1 national symbols and LGBTIQ community symbols are mutually exclusive).

Media 5 - Uncensored use of Same as Media 1 - Refugees from the - Homosexuality is disparaging terms for Middle East are represented as abnor- Serbs such as ‘shka’, portrayed as dangerous mal sexual behavior. ‘shkinë’, ‘shkije’. people, whom the - Same as Media 1 citizens should fear.

Media 6 - Situations/events Same as Media 1 Same as Media 1 There are not many where Serbs ‘face off’ reports about this against Albanians are community constantly framed as a clash between two enemies and take on the connotation of a physical confrontation, where the victory of one party means loss for the other ethnicity. In confronta- tions with Serbs, the individual is always posited in relation to the other as an Albanian. Media 7 Not enough items for - Unclean and negligent; Same as Media 1 There are not many this community were - Same as Media 1 reports about this found in the archive community - Articles refer to situations where third parties exchange offenses by labeling each other as homo- sexual

Media 8 - Noting the Serbian - Uncensored use of - Alarming headlines - This media outlet ethnicity where ethnicity disparaging terms such about the large number published a racist fear has no relevance to the as ‘magjup’, ‘gabel’. of refugees from the mongering article news, except for the - Violation of privacy Middle East who will against a person tendency to associate and objectification of descend into Albanian portrayed as “a man action with ethnicity. the Roma woman who territories. who looks like a 4. ‘Normal’ relations was attacked in multiple woman, and snatches between Serbs and municipalities of children”. Such articles Albanians are portrayed Kosovo from this media outlet, as abnormal, suspi- but also others, caused cious, and weird the perpetration of violence against the woman, who was later established to be

Media 9 Same as Media 2 Same as Media 1 - Use of metaphors Not enough items for such as ‘river of this community were Syrians’ and other found in the archive bombastic headlines that are alarming to the local population

Media 10 Same as Media 1 -There is an article that - Alarming headlines Same as Media 1 incites hatred and about the large number engages in fear monger- of refugees from the ing against a Roma Middle East who will woman, who was descend into Albanian attacked in multiple territories municipalities of Kosovo. This article misinforms that ‘a woman from Llapi possesses a weapon and is part of a criminal network that kidnaps children’ (paraphrasing). Such reporting led to violence and savagery against this woman.. 44 IV. Putting Findings in Context and Giving Them Meaning

4.2 Analysis of Used Discourse and Identification of Problematic Language Structures

In his well-known article ‘Objectivity as Stra- and provide some examples of how the media tegic Ritual: An Examination of Newsmen’s structure reporting on certain events, which Notions of Objectivity’, Tuchman (1972) posits are indicative of the power relations between that the idea of objective reporting by the me- social groups in Kosovo. dia is more of a strategy used to protect itself from criticism over the truthfulness of reported Starting with the LGBTIQ community, media news and the credibility of news sources. Sim- coverage of issues related to this community ilarly, Cohen-Almagon (2008) argues against in Kosovo is centered around important dates objective reporting that would involve mor- for members of this community (on or about al neutrality in covering events, and prefers May 17th). Stories on events or characters re- instead ‘subjective’ journalism, which main- lated to this community outside of Kosovo are tains morally sound attitudes while covering reported throughout the year, without added anti-democratic phenomena or calls against frequency in any period. Regarding the analy- certain communities. The prevailing principles sis of the discourse used by the media for this of journalistic reporting presuppose a ‘detach- community, it should be noted that in most of ment’ of the journalist from the phenomenon the examined media (6 out of 10 print media that is the object of reporting, making it im- outlets) there was one article with content in- possible for the journalist to take a stand on cluding calls for the killing of LGBTIQ members. the event he/she is reporting. Accordingly, ob- Relying on the ‘objectivity’ of reporting, these jectivity as a form of detachment in journalism media outlets have become platforms for becomes an instrument for maintaining the vilifying members of the LGBTIQ community, social status quo and the power relations in where religious authorities promote and incite the given society. This is because access to the violence against this community unchecked. sphere where public opinion is shaped is not The content of these reports speaks of the the same for everyone and powerful groups incompatibility between being Albanian and (harnessing more than material power) dom- Muslim on one side and being LGBTIQ on the inate the public space and use this privileged other side. The media outlets express these access to maintain the asymmetric relations opinions without applying any standard for fil- with marginalized groups. In this sense, media tering information, nor obtaining the opinion of objectivity becomes a ‘mask to disguise or le- the party affected by such problematic calls. gitimize the authority of powerful social groups The rest of the media discourse on this com- over the weakest groups in that society’ (Ibid, munity is objectifying and stigmatizing, where p. 9). The following section will examine the being gay is used as an insult to the other per- discourse used against the selected groups son. Some of these news pieces boast head- IV. Putting Findings in Context and 45 Giving Them Meaning lines such as (headlines are paraphrased)34: lem is the indiscriminate use of terms such as “Fitness Exercising is a Sign of Being Gay”, ‘illegal immigrants’ or ’illegal stay’ to indicate “Scientific Study: “Lefthanded Men Are More the status of immigrants/refugees, although Likely to Prefer the Same Sex”, “Albanian MP no information is provided on whether these Attacks the President of the EC: Alcoholic and persons have applied for asylum and if their Homosexual”, “A Man Dressed as a Woman asylum application has been rejected. In some is Snatching Children in Kosovo”. This type of other cases, water related metaphors are discourse treats homosexual behavior as an used to report on the arrival of immigrants/ object of study and attempts to mystify the refugees. Terms such as ‘river of refugees’, but actions of a member of this community. The also ‘landing of refugees’, or ’refugees head last headline showcases the glaring lack of to borders’ are constantly used and suggest professionalism, where being trans (media to the audience the ‘need to protect them- assumption) is presented as reason enough to selves and their families’ from the oncoming fear a person and action amounting to public ‘others’ (Reporting on Migration & Minorities, lynching is encouraged against the ’weirdo’ Approach and Guidelines, 2017:10). This type and ‘abnormal woman’. In some EU coun- of reporting creates the impression among the tries, hate mongering and discriminatory calls host population that immigrants/refugees are against the LGBTIQ community are neutralized a destructive force and that their communi- by the media, where they have the right to opt ty is ‘under threat’. Some media outlets have out of reporting on events or calls that affect spread fake news about the alleged security this community.35 problems caused by refugees in Kosovo and Albania, where no facts or evidence are giv- Immigrants/Refugees are covered depending en that the refugees have endangered public on the movements of immigrants/refugees safety. Often there is discrepancy between in the Balkans and Albanian lands. In report- headline and content, and additional ‘facts’, ing about their movements towards Albania such as videos that allegedly prove that the and Kosovo, the prevalent topics are on the perpetrators are refugees, which if examined number of arrivals, their movements within carefully turn out to be fabrications and fake the territory, and their goals to leave or stay. news. It is also important to note here the pro- Concerning their stay in Kosovo and Albania, paganda reporting that presents the refugees immigrants/refugees feature heavily in the as ‘parasites’, who are gaining material bene- crime chronicles reporting of their criminal of- fit from the state of Kosovo. Such news notes fenses in Kosovo and the problem they pose to the assistance schemes for asylum seekers the order and security in host countries. In the and implies that it behooves these asylum reporting of criminal offenses allegedly com- seekers to stay in Kosovo because of the ben- mitted by immigrants/refugees, the ethnicity efits they receive. Some of these news pieces of the protagonists is constantly noted, while boast headlines such as (headlines are para- in the vast majority of cases noting ethnicity phrased): “This is the Number of Syrians in Our does not have any relevance.36 Another prob- Country, Though Supported by the State, They

34 These headlines are paraphrased for the sake of ensuring confidentiality in relation to the media outlets that published the news. The news headline, though paraphrased, is quite close to the original headline. 35 For more, see: https://www.article19.org/wp-content/uploads/2018/06/Responding-to-%E2%80%98hate-speech%E2%80%99-with-positive-measures-A- case-study-from-six-EU-countries-.pdf. It seems significant to note that the media have the right to review religious interpretations, which may promote gender inequality or reinforce stereotypes of various types (page 10) 36 There are also many cases where the media outlets find an immigrant/refugee guilty even though a trial is yet to be held on the alleged offense. In many European countries, the principle of ‘innocent until proven guilty’ is integrated into the Codes of Ethics in Media. For more, see: https://www.respectwords.org/ wp-content/uploads/2017/10/Reporting-on-Migration-and-Minorities..pdf 46 IV. Putting Findings in Context and Giving Them Meaning

Beg”, “Euro 19.00 Per Asylum Seeker, Syri- discrimination also appears in police reports, ans and Iraqis Plan to Live in Kosovo”, “Middle where the Albanian perpetrator or victim is Eastern Refugees Pose a Problem to Order not noted by ethnicity, rather referred to as and Security in Albania”, “River of Syrians in ‘Kosovan’, while the ethnicity of members of Albania, No Information from the Police”, “Mo- other communities is constantly noted. An- roccans Physically Attack Security Personnel other characteristic of media reporting on the at Magure Asylum Center”. Roma, Ashkali, and Egyptian communities is their framing as ‘victims’ or ‘aggressors’. News In examining reporting on com-munities about them falls in either one of these two ex- Roma, Ashkali, and Egyptian, it is observed tremes, consequently depriving the audience that news about these communi-ties overlook of information that would represent members the context in which a particular phenomenon of these communities as citizens. Closely re- occurs. In other words, news reports tend to lated to this is the practice of not including the present the phenomena relat-ed to these perspectives of these communities on events communities detached from the socio- that concern them. Members of these commu- economic factors, which have enabled these nities can be talked about, researched, judged, phenomena to appear in the first place. As and have news produced about them, but they noted in Reporting on Migration & Minori-ties, themselves as news protagonists are deprived Approach and Guidelines (2017), the problems of talking themselves about their experiences, of ethnic minorities are reported su-perficially goals, and problems they face. Some (para- in terms of not addressing the root causes of phrased) news headlines that provide an these problems, consequently caus-ing the overview of media reporting trends are: “Roma ‘normalization’ of the occurrence of such from Kosovo Suspected of Spreading Coro- problems, as if they were given facts. The navirus in the German City While Celebrat- same type of reporting is present in the media ing Eid”, “In the Week of Vaccination, Doctors subject to examination, where the problems of Say That RAE Community Children Remain a the Roma, Ashkali, and Egyptian communities Problem”, “This Is Why Roma Are Not at Risk are reported as if they appeared in a social from Coronavirus”, “Kosovan Sexually Abus- and economic ‘vacuum’. Sexual abuse, early es Ashkali Minor in Pristina”, “Woman Who marriages, domestic violence, non-vaccination Caused Panic in Kosovo - What Children and of children, and other phenomena are repre- Adults Say”. sented as issues related to the mentality, tra- ditions, unworldliness of these minorities, and Regarding the Serb community, the media not as manifestations of an oppressive system discourse revolves especially around war against members of these communities. As crimes and the obstructions by Serbia in re- with immigrants/refugees, the vast majority of lation to the state of Kosovo through the Serb news reports of illegal acts, allegedly commit- population living in Kosovo. Reporting tends to ted by members of these communities, make emphasize the rivalry/enmity between Alba- note of the perpetrator’s ethnicity either in the nians and Serbs, and the actions of one side headline or in the body of text, although it is towards the other are seen through this prism, irrelevant to the information. In this regard, it where the ‘race’ between them has winners is important to underline the fact that there and losers. Accordingly, sports or music events are some cases where the news highlight the in which Albanians and Serbs participate are perpetrator’s ethnicity when the perpetrator interpreted as competitions where the parties belongs to one of these communities, but not must defeat each other, not only for the sake in cases where the perpetrator is from the of the competition, but also for the ethnicity majority community. This manifestly evident they represent. Contenders enter the compe- IV. Putting Findings in Context and 47 Giving Them Meaning tition not only in their capacity as athletes or performers, but as

Albanians and Serbs who, beyond their own, have a historic mission to fulfill. In addition to generalizing war crimes and using disparaging terms for Serbs, media articles tend to note the ethnicity of the news protagonist merely to associate ethnicity with the protagonist’s conduct. For example, the Serbian ethnicity of the illegal beneficiaries of agricultural grants is emphasized, but not the ethnicity of the beneficiaries from the majority community. The approach is similar in reporting of criminal conduct where the perpetrator’s Serb ethnicity is emphasized, while this is not the case when the perpetrator is a member of the majority community. While reporting on political events, the ethnicity of MPs or ministers is underlined, especially in cases where the news relates to Serbia’s destructive actions and agendas towards Kosovo statehood. Some selected (paraphrased) headlines from media cov- erage of the Serb community are:” Albanian Boxer Before the Match Against Arkan’s Son: I Pledge My Besa - I Shall Not Disgrace You”, “Police Arrest Serbian Who Beat His Alba- nian Wife”, “In Bujanovac, Albanians Buy the Last Serb Owned Shop”, “Fisticuffs Between a Kosovan and a Serb in Pristina”, “At Match End, Xhaka Brawls with the Serb Dubajic”. 48 5. RECOMMENDATIONS

Based on the findings presented above and 8. Avoid mentioning ethnicity in reporting good international practices in relation to pre- where this is irrelevant to the news being venting and combating hate speech, the Re- reported. port provides the following recommendations: 9. Show sensitivity towards refugee stories 1. Unify definitions of hate speech in laws, and avoid framing refugees/immigrants guidelines, and codes of ethics to avoid in a criminal light. ambiguity and gaps in interpretation. 10. Report professionally on the phenomena 2. Expand the list of characteristics against concerning the Roma, Ashkali, and Egyp- which hate speech is prohibited. Include tian communities, and avoid noting the sexual orientation and refugee/immi- perpetrator’s ethnicity where irrelevant to grant status as grounds on which the the news. incitement or promotion of hate speech is prohibited. 11. Avoid sensational headlines while report- ing on vulnerable groups and communi- 3. Update the Codes of Ethics and monitor ties. their implementation on a regular basis. 12. Censor disparaging references to mi- 4. Provide sanctions against media violating norities in Kosovo and disassociate from professional reporting rules and estab- news containing calls for violence or lish oversight mechanisms for compliance discrimination against a certain social with such rules. group.

5. Strengthen the self-regulatory capacities 13. Avoid generalizing crimes and underline of the media and ensure the inclusion of every time that crimes in Kosovo were the media in setting the rules for self-reg- committed by the state structures of Ser- ulatory mechanisms. bia and not by all Serbs.

6. Re-interpret the principle of ‘objectivity’ in 14. Observe the principle of innocence where reporting and establish a balance be- reporting crimes allegedly committed by tween ‘neutral’ reporting of the event and discriminated groups in Kosovo. preserving human dignity. 15. Break the inter-ethnic relations myth 7. Apply editorial filters to speeches and while covering non-political events, such calls by religious or political authorities as in arts and sports. that incite hatred against certain com- munities. 49 Bibliography

African Union.1981. ‘African Charter on Human Organization of American States. 1969. ‘Amer- and People’s rights’ ican Convention on Human Rights’

Article 11, Prohibiting incitement to discrimina- Palinkas L., Horwitz S., Green C., Wisdom J., tion, hostility or violence, 2012 Naihua D, Hoagwood K (2015), Purposeful sampling for qualitative data collection and Assembly of the Republic of Kosovo. 2008. analysis in mixed method implementation re- ‘Constitution of the Republic of Kosovo’ search, US National Library of Meidicine

Assembly of the Republic of Kosovo. 2012. Press Council of Kosovo. 2015. Code of Print ‘Law No. 04/L-44 on the Independent Media Media. Commission’ The European Council. 1950. ‘European Con- Assembly of the Republic of Kosovo. 2015. vention on Human Rights’ ’Law No. 05/L-021 on Protection Against Dis- crimination’ The United Nations General Assembly. 1948. ‘Universal Declaration of Human Rights’ Assembly of the Republic of Kosovo. 2019. Criminal Code of the Republic of Kosovo. The United Nations General Assembly. 1966. ‘The International Covenant on Civil and Politi- Committee of Minister of the Council of Eu- cal Rights’ rope. 1997. ‘Recommendation No. R (97) 20’ The United Nations General Assembly. 1969. Committee of ministers of the Council of Eu- ‘International Convention on the Elimination of rope. 2010. ‘Recommendation CM/Rec(2010)5’ all Forms of Racial Discrimination’

European Commission. 2016. ‘The EU code Tuchman, G (1972), Objectivity as Strategic of conduct on countering illegal hate speech Ritual: An examination of Newsmen’s notion of online’ objectivity, The American Journal of Sociology vol.77, No.4. European Parliament (LIBE), Hate speech and hate crime in the EU and the evaluation of on- line content regulation approaches, 2020

Fowler, R (1991), Discourse and Ideology in the Press, Routlege, New York

Independent Media Commission. 2016. Code of Ethics for Media Service Providers in the Republic of Kosovo.

National Union of Journalists (UK), NUJ race reporting guidelines, 2016

Views and interpretations expressed in this document do not necessarily reflect the views of the Embassy of the Netherlands in Kosovo.