Producing knowledge about Eastern Europe in times of war: the case of Dutch media and the Second Karabakh war

ABSTRACT

This article evaluates the presence of framing mechanisms in Dutch media reporting on the Second Karabakh war. The paper is led by the following questions: To what extent, and why, does the reporting of the Dutch press favour/undermine certain actors in the conflict? What kind of framing patterns are involved in generating such partiality? And did the frames change over the course of the war? In order to evaluate the presence of framing mechanisms in Dutch media reporting on the second Karabakh war, this research conducted a qualitative data analysis of 188 articles on the topic in nine major national Dutch news media. The paper finds that Dutch newspapers created a rather stereotypical, simplified picture of the Second Karabakh war. There are instances where the reporting gave the impression of a possible bias or overemphasis on certain dimensions.

KEYWORDS:

• Media framing
• Dutch newspapers
• Karabakh conflict
• Azerbaijan
• Armenia

Introduction

Media and conflict have often been studied in relation to one another. Journalists can shape the way we see conflicts and crises, by constructing, framing and maintaining the realities of conflicts (Wahl-Jorgensen and Hanitzsch 2009). For instance, news frames can make the stories of conflicting parties more salient or silent in a communicating text.

One of the recent conflicts that has received significant international media attention is the 44-day war between Armenia and Azerbaijan over Nagorno-Karabakh and surrounding regions, which began on 27 September 2020. The war ended on 9 November 2020, when the Presidents of Azerbaijan and Russia and the Prime Minister of Armenia signed a trilateral agreement. The war took the lives of both Armenian and Azerbaijani civilians; as well as around 4000 Armenian soldiers and 3000 soldiers on the Azerbaijani side (TASS 2021; Amnesty International 2021).

The amount of media attention for the war in Dutch news was quite unusual. Armenia and Azerbaijan had already tested their militaries before, with strong clashes in 2008, 2014, 2016, and July 2020. Yet, in those cases, the conflict had received little attention in the Netherlands. One could argue that until September 2020, the conflict had remained underreported in Dutch media. This changed in the Autumn of 2020 when a significant number of articles were published on the situation in the region. Through this reporting, journalists of Dutch newspapers aimed to interpret and explain what is a very complex conflict with contested causes and history, for their audiences. This research found, however, that the studied media did not present a diversity of views and opinions while reporting on the Second Karabakh war. Moreover, there were instances where the reporting gave the impression of a possible bias or overemphasis on certain dimensions.

Dutch news media are generally considered to be trustworthy and unbiased by the Dutch public. The perception of national media offering “a diversity of views and opinions” is highest in the Netherlands among all the EU member states, namely according to 86% of respondents of the 2019 Eurobarometer on media (Publications Office of the European Union 2019, p. 40, 80). Similarly, a majority of Dutch respondents consider national media to provide “trustworthy information”, namely, 84%, with only Finland scoring higher with 86% (Publications Office of the European Union 2019). However, literature on Dutch media reporting on conflicts has found that there often does tend to be a bias (see e.g. Obermann & Dijkink 2008, Ruigrok 2008, Fengler et al. 2020).

This has led us to ask whether or not this was also the case for Dutch media reporting on the Second Karabakh War of 2020. Generally, not much has been published on Dutch media and conflict, in contrast to studies on Anglo-Saxon media coverage of conflicts (Van der Hoeven & Kester 2020, p. 2). Furthermore, there have been very few (Geybullayeva 2012, Goltz 2012, Atanesyan 2020) scholarly assessments of the media coverage of the N-K war thus far. Beyond the empirical contribution, this case study can also help us to gain further insight into conflict reporting in countries beyond a nation's immediate involvement or interest, making the chances of a bias less likely.

This paper is led by the following questions: to what extent, and why, does the reporting of the Dutch press favour/undermine certain actors in the conflict? What kind of framing patterns are involved in generating such partiality? Did these newspapers rely on information from the Armenian or Azerbaijani governments? And did the frames change over the course of the war? The paper builds on a detailed content and framing analysis of all articles on the conflict in nine major Dutch news outlets published for the duration of the conflict.

The article is divided into six more parts. The following section briefly provides background information on the Karabakh conflict. The subsequent part introduces the conceptual framework on which the analysis will be based. The third section sets out the chosen methodology. The fourth and main section of the article discusses the results of the newspaper analysis. The conclusion presents the article’'s findings and also identifies its limitations; and in the final section, it briefly reflects on further steps and aspects for consideration.¹

Context of the conflict

This paper will not go into the historical details of the Karabakh conflict itself, as there are already a large number of works in existence that comprehensively address the historical details from different perspectives (e.g. Abasov and Khachatrian 2006, Abushov 2010, Geukjian 2012, De Waal 2013, Broers 2019, Gasparyan 2019, Makili-Aliyev 2019, Guliyev and Gawrich 2021). This paper will, however, briefly address the background to the Second Karabakh war which erupted in September 2020, as this context is likely to have had an effect on the way national as well as international media have reported on the war.

The Armenian-Azerbaijani conflict over Nagorno-Karabakh (N-K) and seven surrounding districts, which dates back to 1988 in its modern form, has been one of the most complicated and longest-running disputes in wider Europe. It involves an array of political, economic, legal, social and historical factors. “In May of 1994, the first Karabakh war between Azerbaijan and Armenia was ended by the Bishkek Protocol, which was signed by both countries under the supervision of Russia” (Bayramov 2016, p. 117). Between 1991 and 1994, Armenian forces occupied Nagorno-Karabakh as well as seven surrounding regions (Lachin, Kelbajar, Agdam, Jabrayil, Fuzuli, Qubatly and Zangelan) (Rossi 2017). This was condemned by several United Nations Resolutions 822, 853, 874 and 884, which mentioned the deterioration of the relations between Armenia and Azerbaijan and called on all occupying forces to release the regions. The conflict produced an economic, humanitarian and political catastrophe for both countries. According to the United Nations High Commissioner for Refugees (UNHCR 2009), at the time of the ceasefire in 1994, Azerbaijan hosted an estimated 250,000 Azerbaijani refugees from Armenia and more than 600,000 internally displaced persons (IDP). According to the Armenian government, there are more than 360,000 Armenian refugees as a result of the first Karabakh war (as cited by International Crisis Group 2009, p.1).

Mediation over a longer-term solution commenced in 1992, led by the Organization for Security and Cooperation in Europe (OSCE) Minsk Group. These mediation efforts are, however, considered to have been futile. On the one hand, the OSCE Minsk Group proved itself an inefficient mediator (Rossi 2017) among others caused by the fact that the Azerbaijani government considered some of the Co-Chairs to be biased in favour of Armenia (Bölükbaşı 2011). Since the signing of the Bishkek-Protocol, the hostility between the two countries has risen significantly. The Azerbaijani government made large investments into its military, using energy revenues and enhanced its (political, economic and military) relations with Turkey (Altstadt 2017), while the Armenian government has received security guarantees and military support from Russia (De Waal 2013). As a result, violent clashes between the two armies occurred at several points in time. Despite the scale and cost of these periodical clashes, the international community turned a blind eye to the conflict resolution process, with both the West and Russia having taken a minimalist approach.

Therefore, when the Second Karabakh war erupted between Armenia and Azerbaijan on 27 September 2020, this did not come as a surprise to observers familiar with the developments in recent years. Azerbaijan liberated the seven regions surrounding N-K, as well as a substantial part of N-K itself, including the strategically-located and hugely symbolic city of Shusha (Shushi in Armenian sources) (De Waal 2021). The trilateral agreement that ended the war on 9 November constitutes more than a mere ceasefire but is much less than an actual peace agreement (Broers 2020). In the aftermath, there still is uncertainty about the long-term prospects for the part of N-K that is still under Armenian control, about the role and presence of Russian peacekeepers, and the future of relations between Armenia and Azerbaijan. Additionally, the near absence of the United States (US) and European Union (EU) during and after the war strengthened the Russian power position in the region further, and confirmed Turkey's role as a regional security actor. Euro-Atlantic neglect and passivity throughout the war are reflected by the deal's sidelining of France, the US, the Minsk Group, and OSCE. All have lost immeasurable standing during the conflict (Broers 2020). Since then, the Minsk Group's prospects have looked ever more uncertain.

Conceptual framework

Media coverage of conflicts has been studied extensively, and there is vast literature on the topic. Although significant research has been done on the partial reporting of Anglo-Saxon war journalists, Dutch newspapers’ war reporting has received limited academic attention. By using the journalism of attachment concept, Ruigrok (2006, 2008) illustrated that Dutch media played an important role in creating a rather stereotypical, simplified picture of the Bosnian conflict. My conceptual framework is inspired by the works of Bell (1995), Ruigrok et al. (2005), Ruigrok (2008), van Oppen (2009). In light of this, the journalism of attachment concept will be the main lens through which the paper will analyse the Dutch media reporting on the Second Karabakh conflict. Adding to the war reporting and conflict literature, I pursue two objectives: First, to show whether the reporting of the Dutch press favours/undermines certain actors in the Second Karabakh conflict. What kind of framing patterns are involved in generating such partiality? In other words, I show whether the Dutch newspapers followed similar journalism of attachment patterns in the Karabakh conflict. Secondly, I argue that the behaviour of partial reporting is not only restricted to Anglo-Saxon war journalists, but it can also be seen in different European/Western newspapers.

Journalism of attachment

Ruigrok (2005, 2006) identifies three characteristics of journalism of attachment. First, Ruigrok et al. (2005, p. 160) state that “conflict-related news focuses not just on causes of the conflict, but also on its victims: good people to whom bad things happen”. According to Bell (1995, 99)

speaking from his own experience in former Yugoslavia, media personalized the conflict, so that people elsewhere could relate to it more easily as if it were their homes and families being targeted, and not some foreign conflict of no consequence.

Ruigrok (2008) further claims that attached journalists take sides with what they consider to be the main victims of war when reporting on it. In doing so, the conflict is portrayed in such a way that “good guys” and “bad guys”, “good” versus “evil” are clearly distinguished. When practicing journalism of attachment, they are actively getting involved in the discussion, distinguishing “right” and “wrong” according to their own views and proposing their own chosen solution to the conflict at hand (Ruigrok 2008). In this regard, attached journalists can play a dominant role in creating and promoting a specific frame. For advocates of this style of journalism, objectivity and accuracy are still “sacred” (Vulliamy 1999). However, in conflicts where there are no “good guys” and “bad guys”, the focus on individual suffering forms an alternative for journalists. These reports produce human-interest stories, which draw the audience attention to a conflict that might otherwise be regarded as irrelevant (Ruigrok 2006).

Ruigrok et al. (2005, p. 158) have applied this notion to their studies on Dutch media reporting on the Bosnian war “finding that Dutch media played an important role in creating a rather stereotypical, simplified picture of the Bosnian conflict”. Obermann and Dijkink (2008) found similar results in their article on the Israeli-Palestinian conflict; and Fengler et al. (2020) found that the Dutch newspapers paid very limited coverage to the Ukrainian conflict in 2014.

Second, according to Kepplinger et al. (1991, p. 250), “attached journalists, who already have a framework in mind about the conflict to cover, use ‘instrumental actualisation’, up- or downplaying certain events or statements of experts in order to support their own opinions”. In doing so, attached journalists select different events and present them in a different way without giving their explicit opinion about the events at hand. Kepplinger et al. (1991) argue that news selection is an intentional act to achieve the individual aims of the journalist involved. Their research concludes that journalists’ subjective beliefs account for about one-third of the variance in decisions during the news selection process. This is in line with Entman’s (2007) research on the selection and reporting of news. According to Entman (2007, p. 163), journalists present “a few elements of perceived reality and assembling a narrative that highlights connections among them to promote a particular interpretation”. As a result, they cover the events in such a way as to motivate public action to do something about the situation. According to Ruigrok (2008), attached journalists in Bosnia had a clear goal in mind with their news coverage: triggering a military intervention to set free the victims of the war. With a specific goal in mind, attached journalists exhibit a functional model of journalistic practice. As explained above, this research is interested in the way in which Dutch journalist constructed their frames about the Second Karabakh war.

Another aspect of attachment can be “the selection of sources by using opportune witnesses” (Hagen, 1993, p. 320). Hagen (1993, p. 320) argues that “journalists prefer to quote sources that support the journalist's political stance or their editorial line. By selecting sources that put forward a certain opinion, journalists can convey the impression that experts share their personal views”. According to Van Gorp (2005), this process includes identified metaphors, catch phrases, lexical choices, graphics, stereotypes, dramatic characters and visual images. These sources contribute greatly to “the (political) work that creates and defines the public's beliefs about the causes of social issues, and they thereby justify the action of certain actors” (Van Gorp, 2005, p. 485). By selecting specific quotes, journalists make a piece of information more noticeable, meaningful, or memorable to audiences (Entman 1993, p. 52). Additionally, by selecting and reporting, news media influence not so much “what we think, but it tells us what to think about” (Ruigrok et al. 2005). Therefore, “much of what readers learn may have little to do with the facts, names, and figures that journalists try to present so accurately” (Bird and Dardenne 1988, p. 190).

It is important to consider the different forms of news coverage, namely factual news coverage and op-eds to trace journalism of attachment in news coverage. The framework that attached journalists have in mind is most clearly presented in editorials and op-ed pieces by journalists on the forum pages. However, Ruigrok (2008, p. 45) argues that “attached journalists do not confine their opinions to the op-ed pieces and editorials, but their opinions will also pervade the straight news coverage”. In this regard, the coding groups will include both op-ed articles and straight news coverage.

In addition to selection and salience, this paper argues that journalism of attachment can also imply the act of silencing a group's voice in practice or in discourse. While highlighting a particular aspect of the conflict, attached journalists silence other parts of the conflict. According to Dingli (2015, p. 160)

silence in its strict definition, is the complete absence of voice or sound and a state in which one abstains entirely from speech. She argues that the silent issues in global politics are not physically voiceless; rather, they are politically silenced by external actors.

Similarly, Li (2021, p. 360) notes that

silences are not only absences; silences are exclusions. More specifically, any successful framing implies as a consequence the exclusion of other possible constructions of meaning. This means the inability of those who have been silenced to “speak” because their voices are either volume-less or unintelligible.

As a result, receivers’ reactions are clearly influenced if they perceive and process information about one interpretation and possess little or incommensurable data about alternatives. To evaluate the narrative qualities of news, it is necessary “to look critically at whose values are encoded in the news and whose stories are being told” (Colby 1975, p. 230). Silence can be intentionally or unintentionally used to manipulate information or to keep information away from the audience or from the political discourse (Li 2021). In this regard, framing does not just exclude the experiences of others but, rather, it makes them inadmissible for consideration. Silenced views could not gain adherents and generate an actual effect on public opinion. This is why the silence of interpretations by frames is as significant to outcomes as salience. Both will receive a prominent place in the analysis.

The code groups that will be assessed in light of the journalism of attachment and silence literature are those concerning historical facts, the use of city names, the use of sources, perceptions, and references to international decisions.

Methodology

In order to evaluate the presence of framing mechanisms in Dutch media reporting on the Second Karabakh war, this research conducted a qualitative data analysis of 188 articles (8 op-eds and 180 news articles) on the topic in nine major national Dutch news media. To trace journalism of attachment in news coverage, this article will consider the different forms of news coverage: the news in which journalists put forward their own opinions; the news in which they bring facts; and the news in which journalists allow stakeholders to speak. Journalism of attachment can be researched on an individual and collective level (surface phenomenon). When a group of newspapers becomes attached, the effects of their combined coverage can be far more significant. In such a situation, journalism of attachment can be seen as a surface phenomenon occurring throughout the total news coverage of a newspaper (Ruigrok 2006). In this study, I will concentrate on journalism of attachment as a surface phenomenon.

The news media were selected based on the size of their readership (Ministerie van Onderwijs, Cultuur en Wetenschap S.D., latest data from 2017, FDMG 2021). They include (in order of readership): De Telegraaf, Algemeen Dagblad, de Volkskrant, NRC Handelsblad, Trouw, het Financieele Dagblad, and the online news media NU, RTL Nieuws, and NOS.² A distinction can still be made between more left-wing, central, and right-leaning media, and in the analysis I will check whether or not any obvious patterns can be discerned. More specifically, De Volkskrant and Trouw would be more left-wing; NRC liberal-conservative; and Telegraaf and AD are considered right-wing (Bakker & Vasterman in Terzis 2007, p. 147).

During the period between 27 September and 10 November 2020, 188 articles were published across these newspapers. When looking at the number of articles published by each medium, I can already observe some differences between them, with NRC Handelsblad publishing 45 news articles during the analysed period, and RTL and Financieele Dagblad only 8 each.

Each of these articles was coded using the codes established on the basis of the abovementioned literature. The coding process was divided into two stages. First, the draft coding book was produced with codes divided in coding groups. The code book was also open to creating new codes where needed during the coding process. Second, the content of the articles was coded, using Atlas.ti software. Coding was done at the paragraph level, which means a code or coding group may have been applied multiple times per article. This research is based on an independent coding of all articles and each article and coding were double-checked by another Dutch native speaker. The final dataset consisted of 1444 quotations across 83 codes, divided in 13 code groups. In order to explain and visualise the data and coding groups I used graphs in the results part.

Results

Frames on the nature of the conflict

The research started its investigation of the textual framework by examining how Dutch newspapers frame the Karabakh conflict. As mentioned above, it is “one of the most complex conflicts in the post-Soviet space as it reflects political, legal, ethnic and historical factors” (Abushov 2010). Therefore, it is important to investigate whether the Dutch media understands its complexity and how results explain it to their audience (see below Figure 1).

Figure 1. Frames: nature of the conflict.

As mentioned, attached journalists play an important role in creating and promoting a specific frame and tone. The paper found four main frames that were used to different degrees to explain the conflict: interstate (a conflict between Armenia and Azerbaijan); intrastate (a conflict between N-K as a separatist region and Azerbaijan), religious (Christian Armenia versus Muslim Azerbaijan) and geopolitical (the conflict being closely linked to regional or global power dynamics). First, the newspapers referred to the interstate frame 64% (217 times) to explain the conflicting parties. One reason why this was so often used is that the ceasefire and peace negotiations were conducted between the Armenian and Azerbaijani governments.

In contrast, the intrastate frame was only used 19% (61 times). This mostly happened when they provided background information or if they referred to the statements of the de facto authorities in N-K. However, the news media used this framework in a confusing way by mixing intrastate and interstate aspects within the same article. For example, NRC (October 28, 2020b) mentions that “Nagorno-Karabakh is recognized as part of Azerbaijan but in practice, it is an independent region where mainly Armenians live. Both Azerbaijan and Armenia believe that Nagorno-Karabakh belongs to them”. Additionally, the regions surrounding N-K were only mentioned in 10% (directly 11 times, and another 19 times indirectly). This is a simplification of the conflict situation because the conflict includes both Nagorno-Karabakh and seven surrounding districts.

While the studied newspapers made the interstate, intrastate and religious frameworks more salient, the ethnic aspect of the conflict was mentioned only in 2% (7 times). This was only done by the AD (1 time), NRC (2), Telegraaf (2) and Trouw (2). According to Allen and Seaton, conflicts presented as being concerned with ethnicity and difference have an immediate distancing effect on media audiences, for ethnicity is seen as something that those strange and wild people have, not “us”.

The religious framework was present much more prominently and was mentioned both directly and implicitly, by all outlets 15% (50 times). For example, according to the RTL Middle East correspondent “[ … ] the conflict is not just about land. More importantly, it's also about religion”. Or: “It is also a conflict between (Armenian) Christians and (Azerbaijani) Muslims. That makes this so incredibly sensitive and painful” (RTL, 27 September 2020). NRC referred to the conflict as “the most recent exchange of violence in a decades-old conflict over the territory between mainly Orthodox Christian Armenia and predominantly Muslim Azerbaijan” (NRC, September 27, 2020a). The newspapers also used the religious frame to refer to the competition between Turkey and Russia. For example, De Volkskrant mentioned that “the war puts global powers against each other. Islamic Azerbaijan is provided with modern weapons by NATO member Turkey. Russia supports Christian Armenia” (Volkskrant, October 26, 2020a). The conflict is not considered a religious one, however, by most experts and even government officials (see e.g. De Waal 2013, p. 9, Armenpress, December 24, 2021). According to Armen Sarkissian, the former President of Armenia, “the conflict was never, never a religious war. Armenia has wonderful relations with a lot of states where Islam is a major religion” (Armenpress, December 24, 2021). Similarly, the Azerbaijani government and local Muslim leaders mentioned that the Second Karabakh war had no religious dimension (The State Committee on Religious Associations of the Republic of Azerbaijan, October 15, 2020). Broers (2019) argues that framing the conflict in religious terms serves more to locate an obscure conflict than to accurately convey its nature. The religious framework is one of the important misinterpretations of the conflict resulting from journalistic oversimplification, ignorance and stereotypical assumptions.

NOS (11 times), De Volkskrant (7), and NRC (7) most frequently highlighted the religious frame the most among all the studied media. This framework is present in the op-ed and news articles. One of the possible reasons for this religious framing is that some newspapers assigned non-expert, and mainly the Middle East, journalists to the South Caucasus, who may not actually be familiar with the regional context of the South Caucasus, which is substantially different from that in the Middle East. Therefore, these journalists already had a (religious) framework in mind about the conflict to cover. For example, while reading the articles in De Volkskrant, I found that one of the journalists was based in the Middle East and had never reported about the Karabakh conflict before. When asked by a fellow respondent in De Volkskrant “How did you, our Middle East correspondent, end up in Nagorno-Karabakh?” the journalist replied that:

Due to the covid restrictions, Tom [the journalist responsible for the South Caucasus] cannot leave Moscow at the moment. [ …] These are crazy times, which force you to make strange choices. But I talked a lot about the conflict with Tom and coincidentally, I also have a lot of Armenian contacts. In Beirut, where I live, you have a large Armenian community. (Bahara, 2020)

As mentioned, attached journalists, on the other hand, actively and collectively take a certain side in a conflict and have a more systematic intention in their coverage. Considering this, there is a possibility that the journalist may have received one-sided information.

Framing: geopolitical context

The geopolitical frame was a very dominant one. Rather than focusing on the war as a conflict between Armenia and Azerbaijan, or between Azerbaijan and N-K, the newspapers additionally framed the conflict as a geopolitical struggle between regional powers, particularly Turkey and Russia. Turkey was presented as the main regional power supporting Azerbaijan. For example, Trouw (November 8, 2020) mentioned that “Azerbaijan can count on the support of ‘brother state’ Turkey in the conflict”. NOS (October 17, 2020a) reported that “Ankara supports Islamic Azerbaijan”, thereby including an indirect religious frame as well (Figure 2).

Figure 2. Geopolitical framework: regional powers.

The type of support provided by either Russia or Turkey was sometimes specified, and sometimes kept vague. While reporting on Turkey's support for Azerbaijan, the newspapers (50%/85 times) did not specify the type of support. The newspapers also did not always specify Russia's support (12.7%/29 times).

Turkey's military support for Azerbaijan was mentioned quite frequently (41%/71 times), specifically by Telegraaf (19 times), Trouw (12) and NRC (11). In addition to reporting on the delivery of weapons, it was incorrectly mentioned by the Volkskrant (23 October 2020b) that Turkey had also sent money to Azerbaijan. Azerbaijan has military trade relations with several countries, namely Russia, Israel, Ukraine, Belarus, and Turkey. According to the Stockholm Peace Research Institute (SPRI) (Kuimova et al. 2021), “Turkey accounted for 2.9 per cent of Azerbaijan's imports of major arms over the decade 2011–20”.

While Turkey's military support to Azerbaijan was emphasised, the newspapers gave less attention to the Russian military support for Armenia (14.9%/34 times). There was also relatively little attention to the fact that Russia has provided military support to both Armenia and Azerbaijan. Yet, Kuimova et al. (2021) show that “despite Russia acting as a leading mediator in the conflict between the two countries, in 2011–20, it accounted for 94% of Armenia's imports of major arms and 60% of Azerbaijan's”.

In contrast to military support, the research found that both Turkey’s diplomatic support to Azerbaijan, and Russia’s diplomatic support to Armenia specifically, received less attention from the newspapers (namely 6.5%/11 times regarding Turkey’s diplomatic support; 1.7%/3 times regarding Turkey’s mediating role; and 3.9%/9 times of the coded quotations concerning Russian diplomatic support).

While Turkey's role in the conflict has been described in mostly negative terms, the newspapers mainly described Russia in a more positive manner, particularly as a mediator (51%/125 times) in the conflict. Russia did play a very active and diplomatic role, and since Russia is a Co-Chair of the OSCE Minsk group, one may argue that it is only natural to highlight this role. However, the research found that media often did not mention Russia in light of the Minsk group at all.

According to Vladisavljević (2015), by selecting and reporting news, the media highlights some problems, different sets of facts and people, and not others. One of the main disturbances I found is that the role of Iran in the region received very limited attention (namely, 2.8%/11 times) of the coded quotations. Yet, Iran did play a role in the conflict by offering to mediate, and it also has relations with Armenia and, to a lesser extent, with Azerbaijan (Broers 2019). More specifically, Iran has direct borders with both parties to the conflict and has sought to prevent any expansion of the conflict into its territories since the outset of the conflict. Iran is also interested in prolonging the status quo established by the Bishkek agreement, and blocking the intervention of the non-regional powers (e.g. Israel and the US). Therefore, Iran unsuccessfully attempted to achieve a ceasefire between Armenia and Azerbaijan in the early 1990s. Furthermore, the world's largest single community of ethnic Azerbaijanis lives in Iran. However, Iran's relations with Armenia are rhetorically warmer than with Azerbaijan because of Baku's close relations with Israel, and its energy agreements with the West (Mahammadi and Huseynov 2022). While Turkey closed its land borders with Armenia in the 1990s because of the first Karabakh war, Iran provided Armenia with an open southern border and an important resource in overcoming the blockade (Mahammadi and Huseynov 2022). Additionally, Armenia has solved its acute energy problems thanks to the gas pipeline from Iran (De Waal 2013). Considering these factors, one can argue that Iran's role in the conflict should have been highlighted.

Nevertheless, AD, Trouw, RTL, FD, NU did not mention Iran at all. This might indicate a form of silencing. This is what Wolfsfeld (1997, p. 34) calls the “narrative fit between incoming information and existing media frames”. Nevertheless, this geopolitical framing downplays the significance and independent agency of local actors, which resulted in an inaccurate representation of developments in the conflict.

Salience and silence

Looking at the use of sources, one can see that government sources were cited quite frequently, but nearly always in an indirect manner, for example, by reporting on decisions announced by either side or by mentioning that the different authorities had made accusations or denied those. Parts of press conferences were translated and selected and put into context and were therefore considered as “indirect use of government sources”, in the analysis. However, the overall number of coded quotations referring to government information was not balanced as the newspapers referred to the Azerbaijani government 41.5%/139 times but they referred to the Armenian government and the de facto Karabakh authorities combined 58.5%/196 times. In line with Hagen’s (1993) argument, one could argue that sources were selected on an unequal and imbalanced basis (Figure 3).

Figure 3. Use of sources.

According to Wolfsfeld (1997, p.4), antagonists in conflicts “compete over access to the news media and they compete over media frames”. There were only very few instances where information from either government was directly reproduced by the news media without editing (8 in total, 5 photos and 3 Tweets). Most of the reproduced information could be found in De Telegraaf (5), which would for instance re-tweet President Aliyev's tweets, or display pictures of military action, released by the Armenian or Azerbaijani Ministries of Defence. This does suggest a critical stance towards government-produced visuals and information by most Dutch media.

In addition to government information, several articles also shared views of civilians. Having conversations with “regular citizens” is a popular means for Western media to “contextualise” conflicts but obviously tends to be unrepresentative (Luyendijk 2006, p. 54). A few articles (14%/16 quotations) displayed the views of Azerbaijani citizens, be it in Azerbaijan itself or of Azerbaijanis living in the Netherlands. However, the overall majority (86%/101 quotations) of “citizens views” were from Armenian civilians, again both in Karabakh, Armenia, and the Netherlands. Individuals were, for example, interviewed to express their concerns and share their personal stories. A few news articles told the stories of Armenian soldiers fighting at the front. However, there was much less attention for the “people behind the conflict” on the Azerbaijani side. The news medium that stood out in this regard is Algemeen Dagblad (AD), with 38 instances where Armenian civilians were directly used as a news source; versus 5 Azerbaijani civilians. For the Volkskrant, this was 23 instances citing Armenian civilians versus none citing Azerbaijani citizens. NRC, RTL and Telegraaf also cited only Armenian, and no Azerbaijani civilian sources. In line with the journalism of attachment, these findings illustrate that Dutch newspapers only focused on one part of the events and mainly quoted Armenian people in their coverage. These examples also illustrate the use of opportune witnesses because Volkskrant, NRC, RTL and Telegraaf chose specific stakeholders to cite in their publications. This means that journalists chose participants to direct the public discussion in a certain direction.

Implicit instrumental actualisation: historical context

Most news coverage included very brief references to the historical context of the N-K conflict. They would, for instance, take the first war of 1988–1994 as a historical reference point. The conflict goes back further in time than 1988, however. A handful of articles (6%/17 times) from De Volkskrant, Trouw, NOS and NRC referred to developments in the 1920s, specifically to the argument that is sometimes posed, that Stalin would have given the region of N-K to Azerbaijan, implying that the historical claim to the land could therefore be disputed – but this argument has not been proved (Saparov 2018).

As mentioned, attached journalists can either up- or downplay certain sub-themes and contesting parties in their articles. While describing the historical, ethnic composition of the region, the studied newspapers mostly mentioned ethnic Armenians 24.3% (68 times), while not mentioning the Azerbaijani population that lived and lives in the region. In contrast to this, some of the articles referred several times (2.5%) to Azerbaijani Internally Displaced People (IDP), who were displaced during the first war (de Volkskrant 3 times, Trouw 1 time and RTL 3 times). Considering this, it can be argued that while highlighting a particular aspect of the Karabakh conflict, attached journalists silenced other parts of the conflict (Figure 4).

Figure 4. Historical context.

Regarding the status of N-K, there generally appears to be quite a lot of confusion. On the one hand, 31.5%/88 times stated that the international community recognises Nagorno-Karabakh as part of Azerbaijan. On the other, most media (except for RTL and Telegraaf) simultaneously referred to N-K as an “Armenian enclave” (14.3%/40 times). NRC (18) and FD (9) used the term the “Armenian enclave” most frequently. Besides the contradiction, some have made the point that using the term “enclave” is incorrect when speaking about N-K, in the first place (Goltz 2012). According to the Oxford Public International law,

an enclave in international law is an isolated part of the territory of a State, which is entirely surrounded by the territory of only one foreign State—the surrounding, enclaving, or host State—so that it has no communication with the territory of the State to which it belongs—the mother or home State—other than through the territory of the host State.

One last theme that needs to be covered in the context of historical references is the issue of the Armenian genocide. Although there is no connection to the relations with Azerbaijan, nor to the war over N-K, several newspaper articles AD (4 times) Volkskrant (15), NRC (6), Trouw (4) did mention the genocide and the N-K war in one and the same text. This was done indirectly 11 times (e.g. by discussing Turkey's role in the conflict, and then also referring to Turkey-Armenia relations and mentioning the genocide of 1915); and directly 18 times, thus in total 29 times (10.3%). One of the reasons for this is that an immediate link was implied by Armenian citizens who were interviewed. However, there was very little mentioning of historical reference points that are significant to Azerbaijan, such as the Khojali massacre (0.3%/1 time) or the aforementioned UN resolutions (2.8%/8 times). In line with the implicit instrumental actualisation, it can be argued the studied newspapers downgraded certain sub-themes of Azerbaijan within the conflict. Similar to Kepplinger et al. (1991) and Entman's arguments, these findings illustrate that Dutch newspapers selected different events and different sources depending on the goal they wanted to achieve, without giving their explicit opinion about the events at hand. In doing so, they presented selective elements of perceived reality and created a framework that highlights connections among them to promote a particular interpretation.

Portrayed roles of conflicting parties

As explained in the conceptual part of this paper, “journalism of attachment” predominantly refers to the representation of conflicting parties as “right” and “wrong”, of “victims” or “aggressors” (Ruigrok 2008). In practice, this can for instance result in using certain sources but not others; or in making the suffering of some more visible than that of others. Citizens on both sides of this conflict have suffered greatly. Many civilians have died, have been wounded, have lost their homes or had to flee; and there were many deaths and wounded among the soldiers from both warring parties.

As said, there were civilian casualties on both sides of the conflict. International estimates are that 11 Armenian civilians and 72 Azerbaijani civilians lost their lives during the 2020 war (Amnesty International 2021, p. 7 and p. 13). Reporting on civilian victims on both sides of the conflict was relatively even, with 38% (48 mentions) from Azerbaijan, and 29% (37 mentions) from Karabakh and Armenia. Often, media would also simply refer to “casualties” without specifying on which side mentions.

Military casualties were estimated to be 3773 on the Armenian side and 2908 on the Azerbaijani side (Radio Free Europe, August 24, 2021; MoD Azerbaijan, October 21, 2021). Military deaths and wounded were reported much more frequently when it concerned Armenian soldiers 19.8% (25 times), versus 5.5% (7 mentions) regarding Azerbaijani soldiers; alongside 7% (9 instances) where military deaths on both sides were reported, but this could possibly be due to the fact that the Azerbaijani government did not issue official information on military casualties. This aspect is thus also linked to the issue of government control over information.

However, one aspect in which I did observe a clear discrepancy is the reporting on infrastructure damage and refugees. There are more references to damage on the Armenian side, with 70%/47 mentions referring to instances in Karabakh; and 30%/27 mentions regarding instances on the Azerbaijani side. However, this cannot be fully accounted for by the location of combat. The damage in residential parts of the Azerbaijani city Ganja, was mentioned several times, but not discussed in terms of the people affected: the implications for Azerbaijani civilians were only mentioned in terms of damage to infrastructure, whereas in the case of N-K, the focus was more on human suffering. According to the Internal Displacement Monitoring Center (2021, p. 47), the Second Karabakh war led to 84,000 new displacements in Azerbaijan in 2020. Many news media provided rather personalised accounts of Armenian people caught up in the conflict within Karabakh, and that personalised dimension appears to be the biggest difference in the reporting when it comes to victims of the war. This corresponds to the research findings regarding the use of civilian sources, where I also observed that there was more space for the viewpoints of Armenian citizens than for Azerbaijanis. Considering this, one may also argue that took sides in this conflict, by pointing out who is to blame for the conflict and who is the main victim in need of protection (Figure 5).

Figure 5. Instances.

Although not always transparent, the phenomenon of taking sides with the underdog often occurs when journalists frame conflicts (Ruigrok 2006). These are often portrayed in such a way that “attacker” versus “victim” are clearly distinguished. In line with this, the way the warring parties were described, often either as an attacker, or the attacked, was also inconsistent. Azerbaijan was described as attacker much more frequently than was Armenia (75%/103 times and 25%/33 times, respectively), especially by De Volkskrant and NRC. For example, NOS (28 September 2020b) mentions that “the armed forces of oil-rich Azerbaijan are much stronger than those of impoverished Armenia”.

Discussion

As stated by Bird and Dardenne (1988), newspapers retell stories in different ways. The findings illustrate that the Dutch newspapers’ reporting on the Second Karabakh War of 2020 was often oversimplified and lacked nuance. Specifically, the religious framework is one of the important misinterpretations in my opinion; and there was substantial contradiction over the status of Karabakh and the seven surrounding regions. The geopolitical frame overemphasised the role of Russia and Turkey at the expense of understanding the dynamics between Armenia and Azerbaijan. In doing so, the newspapers treated Armenia and Azerbaijan as mere pawns rather than as agents in their own right. Furthermore, the positive portrayal of Russia as a mediator overlooked that Russia is the largest exporter of arms to both Armenia and Azerbaijan. In general, Turkey's role was described in rather negative terms. Based on Obermann and Dijkink's work on the impact of large political events on news reporting of other issues, one explanation to be explored is that relations between Turkey and the Netherlands have deteriorated quite significantly over the past years, and this negative attribution may therefore come forward in news articles on Karabakh, too (Obermann & Dijkink 2008, p. 160). Possibly, an even broader factor may play a role in the reporting on Azerbaijan, namely a more negative view towards Islam, which might explain why religion is consistently mentioned when referring to Azerbaijan. An analysis of the frames on the historical context showed how little context was often provided in news articles, and how selective these were. As part of the frames, salience was given disproportionally to accounts of suffering of Armenian civilians, over those of Azerbaijani citizens. Journalism of attachment, lastly, became apparent when looking at the reporting of casualties, victims, and attackers, with more attention for the “human” dimension in Armenia and N-K, as opposed to in Azerbaijan.

One may ask whether the frames changed over the course of the war, and whether media were able to represent the regional context more accurately as time progressed? The short answer to this question is both yes and no. On the one hand, the religious frame was maintained throughout time (especially by NOS, NRC and De Volkskrant). Other references that, in my opinion, are not conducive to accurate reporting, such as the references to genocide, or the historical argument about Stalin's decisions, were also still used towards the end of the conflict. A cross-check analysis shows that both news and op-eds articles in NRC and reports by De Volkskrant, applied several of these frames simultaneously, using the following references: “religion”, “Stalin”, “genocide”, “Armenian enclave”, “aggressive Azerbaijan”.

In a similar vein, while Telegraph, NRC, NOS, Trouw and de Volkskrant kept highlighting Turkey's (2.6%) military trade with Azerbaijan throughout their stories, Russia's military relations with both countries (60% Azerbaijan, 93% Armenia) were mentioned only either in the first or in the last articles. Trouw paid significant attention to the role of Turkey. As mentioned, the seven surrounding districts and the four UN resolutions received very limited coverage: either in their first or last articles. Although AD, Trouw, De Volkskrant and NU stopped referring to N-K as an Armenian enclave from approximately the middle of the war, NRC and the FD kept this incorrect reference throughout their reporting.

While a significant degree of bias could be found in every news medium that I examined, there appear to be some variations across the media. As was shown in the analysis above, government information was reproduced only very infrequently; but in most instances (5 out of 8) this was done by the Telegraaf. AD, NRC, NOS, Trouw and de Volkskrant did give more space to Armenian (government and civilian) sources in general. As a result, the facts and figures from the Azerbaijani side were silenced, namely Azerbaijani city names in Karabakh and Internally Displaced People. Therefore, in contrast to the 2019 Eurobarometer's findings on Dutch media, the paper found that the studied media did not always present a diversity of views and opinions while reporting on the Second Karabakh war.

Conclusion

This article has analysed Dutch media reporting on the Second Karabakh War. The article was based on nine newspapers and 188 articles in the Dutch media between 27 September 2020 and 10 November 2020. The paper investigated the tone of the reporting and the framing. In contrast to the 2019 Eurobarometer's findings on Dutch media, this paper found that the studied media did not present a diversity of views and opinions while reporting on the Second Karabakh war. Although the studied newspapers covered civilian victims from both sides, they presented Azerbaijan as an aggressor and Armenia as a weak side in the conflict. Overall, the article illustrated that the studied newspapers gave more space to the Armenian side of stories and less space to the Azerbaijani side of stories. Similar to the previous studies on Dutch media (e.g. Ruigrok et al 2005, Obermann & Dijkink 2008, Ruigrok 2008, Fengler et al. 2020), the article illustrated that Dutch newspapers created a rather stereotypical, simplified picture of the Second Karabakh war. The research also found that the predominantly negative tone towards Azerbaijan was limited to opinion pieces.

Although this research was based on documents from only the Netherlands during a specific time, the findings contribute to a more nuanced discussion about the Western media coverage of Post-Soviet conflicts. The findings suggest that journalists could work harder to highlight both sides on opinion pages, if they are concerned about the objective representation of their work in the media. For a foreign conflict, of which the Dutch public has little knowledge, balanced news coverage of the events is important in achieving a comprehensive understanding of the conflict.

In terms of the limitations of this article, this paper has focused particularly on the Second Karabakh war and nine Dutch newspapers, which has left other newspapers, the First Karabakh War and peace process unexamined. In this regard, further research is necessary to address whether other newspapers also followed similar reporting strategies, and how they reported the First Karabakh war or other Post-Soviet conflicts (e.g. Georgia-Abkhazia).

Table

Concept from literature	Code Groups	Codes	Comments
Journalism of attachment	Victimisation: Textual	● Armenian civilian victim in Karabakh (death + wounded)
		● Armenian soldier victim in Karabakh (death + wounded)
		● Armenian victim (not clear soldier or civilian)
		● Azerbaijani civilian victim (death + wounded)
		● Azerbaijani soldier victim (soldier + wounded)
		● Azerbaijani victim (not clear soldier or civilian)
		● Civilian victim both sides
		● Military death both sides
		● Victim both sides not clear soldier or civilian
		● Refugees and lost home Armenia
		● Refugees and lost home Azerbaijan
		● Unclarity about Azerbaijani soldier victim numbers
Salience + silence	City names	● City Names in Azerbaijani
		● City names in Armenian
	International decisions	● The UN four resolutions (but not clear)
		● Describing Nagorno-Karabakh as Armenian Enclave
		● Legally Recognized as Azerbaijan, de facto Armenia
		● Describing NK as independent entity
	Historical facts	● Seven Surrounding regions
		● Mentioning some of seven districts but not clear
		● Historical: in-depth mentioning history of conflict
		● Historical: briefly mentioning history of conflict
		● Indirectly Armenian genocide in the context
		● Referring Armenian genocide (in context of the conflict)
		● Mentioning Azerbaijani IDPs (checked)
		● Mentioning Armenian IDPs
		● Mentioning only ethnic Armenian population in Karabakh
		● Referring USSR (general/indirect reference)
		● Russia/Stalin gave to Azerbaijan (specifically about Stalin argument)
	Perception	● Describing Armenia weak but Azerbaijan strong checked
		- Describing Azerbaijan as attacker
		- Describing Armenia as attacker
Framing	Framework	● Religious Christians vs Muslims (checked)
		● Mentioning Religion (note for analysis: it creates unconscious implication Muslim vs Christians)
		● Interstate conflict
		● Intrastate
		● Ethnic
Sources/representation (part of silence/salience?)	Government sources	● Sources: report on civilians Armenia
		● Sources: report on civilians Azerbaijan
		● Sources: report on officials Armenia
		● Sources: report on officials Azerbaijan
		● Sources: report on officials NK
		● Sources: reproduce info Armenia
		● Sources: reproduce info Azerbaijan
Geopolitics (part of framing 2)	Regional versus external powers: i) Regional	● Turkey military support Azerbaijan
		● Turkey diplomatic support Azerbaijan
		● Turkey support Azerbaijan unspecified
		● Russian military support Azerbaijan
		● Russia diplomatic support Azerbaijan
		● Russian military support Armenia
		● Russia Diplomatic support Armenia
		● Russia support Armenia unspecified
		● Russian military support both
		● Russia diplomatic support both
		- Russia as mediator
		● Describing Turkey as mediator
		● Regional powers: Russia-Turkey relations
		● Highlighting regional powers Iran
		● Iran neutral
	ii) external	- External powers: EU
		- Highlight external powers (incl. OSCE)
		- External powers Israel military support Azerbaijan

Acknowledgement

For helpful comments on earlier versions of the contribution, the author would like to thank Mustafa Ali Sezal, Seiki Tanaka, and Senka Neuman-Stanivukovic as well as the reviewers.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 Research into any conflict is inherently sensitive. I want to add that as an author, I do not take a stance on the conflict. My view is that there has been terrible suffering on both sides, and that civilians on both sides have been the victims of this conflict.

2 NOS, NU.nl, and RTL are broadcasting media rather than newspapers, and have been included on grounds of them having among the highest readership.