Perceived group threats and right-wing political party membership as driving forces of negative descriptions in Turkish Parliamentary debates (1983–2018)

ABSTRACT

To what extent do perceived threats explain negative descriptions of minorities in politics? This study addresses this ever-significant inquiry by scrutinising all mentions of Armenians in Turkish parliamentary proceedings between 1983 and 2018. Building on group position theory, it proposes that perceived threats influence the perception of Armenians in Turkish politics. The study relies on a quantitative content analysis to test this premise. The findings demonstrate that being a member of a right-wing political party and discussing security and diplomatic threats predict politicians’ negative descriptions of Armenians. The study concludes that perceived threats to Turkey’s national security and international standing contribute to the negative perception of Armenians in Turkish mainstream politics.

KEYWORDS:

• Group position theory
• intergroup hostility
• quantitative content analysis
• parliamentary records
• Turkish-Armenian relations

Introduction

Political parties shape official policies about minorities all around the world. It is imperative to scrutinise the factors associated with political parties’ negative descriptions of minorities. One substantial sociological perspective that could help is group position theory (Blumer 1958; Bobo 1999; Olzak 1993). The approach posits that dominant groups’ perceptions of threats to their prerogatives underlie their hostility towards minority groups. Most of the studies confirming this theoretical approach explore socio-political reasons for prejudice in developed Western countries (Biggs and Knauss 2012; Carter and Lippard 2015; Denis 2015; Kusow and DeLisi 2020; Schneider 2008; Weitzer and Tuch 2005). Considering the academic literature on non-Muslim minorities in Turkey in the light of group position theory, this study suggests that perceived threats contributed to political parties’ negative descriptions of Armenians in Turkish politics between 1983 and 2018. In so doing, it addresses the lack of analyses from non-Western contexts in group position theory and the concurrent lack of systematic analysis of the perception of Armenians in contemporary Turkish politics.

The history of Armenian–Turkish relations provides a background to the discussion. In the Ottoman Empire, the Armenian Gregorian community along with the Jewish and Greek Orthodox minorities lived under the millet system (Barkey 2008). They had religious, educational, juridical and fiscal autonomy (Icduygu and Soner 2006) and were governed by their religious leaderships (Melson 1982). In return, they paid higher taxes than the Muslim Ottomans and could not serve as state officers or in the military (Zürcher 2004). This system gradually collapsed with the fall of the Ottoman Empire between the late eighteenth and early twentieth centuries. The period of collapse witnessed intercommunal violence with particularly horrendous results for the Ottoman minorities (Bloxham 2011; Dyer 1976; Levene 1998). The most notorious was the 1915–1917 deportations referred to as the Armenian Genocide (Göçek 2014). There is a highly charged international debate about recognition of the Armenian Genocide which is beyond the scope of this paper.

The Ottoman Empire collapsed at the end of the First World War (1914–1918). The subsequent Treaty of Sèvres in 1920 imposed heavy conditions on the Ottoman Empire, such as the allocation of most of the imperial territory to the Allies. This triggered the Turkish War of Independence (1919–1923), which ended with the signing of the Treaty of Lausanne. The second treaty led to the establishment of the Republic of Turkey in 1923. The modern republic granted equal rights to all Turkish citizens. However, various events and policies in modern Turkey imply that this was an unfulfilled promise for non-Muslim citizens. For example, Üngör (2011) argues that the Turkish population policies aimed at ethnically homogenising the Eastern region of Turkey between 1913 and 1950 using genocide, deportation, spatial planning, forced assimilation and memory politics. Moreover, the Wealth Tax of 1942 forced non-Muslims to pay very high taxes or sent them to ‘work camps’ in Eastern Anatolia under tough labour conditions (Neyzi 2002). On September 6–7, 1955, there was xenophobic violence against all non-Muslim minorities in Istanbul in which minorities’ shops were looted, for which the Turkish state did not provide enough compensation afterwards (Kuyucu 2005). Additionally, since its establishment, the modern Turkish state has strived to prevent international recognition of the 1915–1917 deportation of Armenians as genocide (Gürpınar 2016; Zürcher 2011); such recognition is viewed as an offensive allegation in Turkey both at an official level and by the general public. The murders of Turkish diplomats and their families by the Armenian Secret Army for the Liberation of Armenia (ASALA) in the 1970s and 1980s also increased the prevalence of negative perceptions of Armenians in Turkey (Gunter 2007). Furthermore, in the ongoing Nagorno–Karabakh military conflict between Armenia and Azerbaijan, Turkey stands against Armenia (Dixon 2010). Despite the tumultuous history between Turks and Armenians, the 85th Patriarch of Turkey’s Armenians, Mr. Sahak Mashalian, estimated that Armenians are currently the largest non-Muslim community in Turkey, with a population of around 60,000 (Özbey 2020).

Academic studies on non-Muslim minorities in modern Turkey underline the prevalence and significance of exclusionary practices that they suffered (e.g. Bali 2001, 2004, 2011; Göl 2005; Toktas 2005). While some afford valuable insights about the experience of Armenians in modern Turkey (Ekmekçioğlu 2014, 2016; Suciyan 2018), the scholarship tends to emphasise specific events, predominantly the debates about the international recognition of the Armenian Genocide (Açar and Rüma 2007; Avedian 2013; Bilali 2013; Gürpınar 2016). Thus, the academic literature could benefit from a theory that provides a more generic explanation of the perception of Armenians in contemporary Turkish politics and society. To that end, this study employs group position theory and focuses on the period between 1983 and 2018.

Following the military coup on 12 September 1980, Turkish politics entered a new era, called the Third Republic (Zürcher 2004). The coup dissolved the Turkish parliament, created a new constitution in 1982 and forbade the activities of the old political parties and their politicians. Subsequently, three new political parties competed in general elections, and the country returned to a multi-party parliament on 24 November 1983. The Third Republic is a very interesting period to study, as it involves various milestones of modern Turkish-Armenian relations: the genocide debate, different phases of the Nagorno-Karabakh conflict, the ASALA attacks and the 2009 Zurich Protocols signed by Armenia and Turkey to normalise their relations. This paper analyses Turkish political parties’ descriptions of Armenians in parliamentary debates about these and other topics. Below, while developing its hypotheses, the paper introduces group position theory and its relevance to the Turkish political parties’ perceptions of Armenians. Subsequently, it outlines the research methods used in the study. Finally, the paper presents the findings and discusses their implications.

Academic literature and hypotheses

The main premise of group position theory is that prejudice emerges from dominant group members’ perceptions of outgroups as a threat. According to Blumer (1958), people have a sense of where their social group ought to belong in relation to others. This includes opinions that lay the basis for people’s hostile attitudes regarding the deserved standing of other groups: whenever people feel that their group’s position is challenged by other groups, they react through prejudice. In other words, prejudice is a defensive response to perceived threats from others to maintain the dominance of one’s own group. In this view, four feelings lead dominant group members to prejudice: (1) the dominant group’s sentiment of natural superiority, which involves attributing negative characteristics to others, such as laziness and dishonesty; (2) inherent differences between the dominant group and other groups that cannot be overcome; (3) the dominant group’s proprietary claims of privilege, such as the right to high-paying occupations; and (4) the dominant group’s apprehension of other groups’ desires for and designs to obtain the same advantages. These sentiments provide dominant group members with a cognitive map of society that assumes a ‘natural’ hierarchy that requires protection. Bobo and Hutchings (1996) extend the group position perspective to subordinate groups, arguing that subordinate group members would also be likely to feel alienated and threatened due to their positions, and recommending that the scholarship look equally into the perceived threats of the disadvantaged.

Blumer (1958) explains that people’s beliefs about their group’s deserved dominance are collectively and historically rooted. Therefore, reasons for prejudice cannot be reduced to the contemporary material conditions of individuals. Dominant groups define and redefine their relationships with subordinate groups in various ways, including political speeches, everyday conversations and historical accounts and tales, which function to normalise their feelings of the rightful hierarchical order between themselves and subordinate groups. In parallel, Blumer (1958) states that historically constructed ideas about intergroup relationships would decline if untouched. When members of the dominant group do not define intergroup relationships as threats to the natural order of society, the sense of deserved dominance gradually withdraws. Despite the importance of history, Quillian (1995, 1996) notes a significant lack of studies on the historical and cultural components of dominant groups’ negative attitudes.

History is also a decisive factor in contemporary Turkish-Armenian relations. The Ottoman period’s ethno-religious conflicts seem to have contributed to drawing boundaries between the Turkish and Armenian communities as the dominant and subordinate groups in modern Turkey. Suciyan (2018) argues that the Armenian minority was a silenced and repressed community forced to echo the Turkish state’s official position with regards to the Ottoman Empire’s persecution of Armenians, and concludes that the remaining members of the Armenian community continue to experience this persecution in modern Turkey. Similarly, Ekmekçioğlu (2016) states that despite modern Turkey’s promise of equal citizenship to the Armenian minority, it maintained boundaries between the latter and the majority. She describes three layers of the Armenian community: (1) families and homes as the in-side; (2) community institutions, such as schools and churches; as the mid-side through which the Armenian minority had contact with the Turkish state by law, and; (3) the general public space in Turkey as the out-side. These observations are in line with the group position perspective, as they point to the significance of the historical constructions that help to define social boundaries between a dominant and subordinate group.

The academic literature on Turkey confirms group position theory’s insight into the construction and maintenance of intergroup relations in the Turkish-Armenian context. Various studies on the Turkish perception of the non-Muslim minorities illustrate exclusionary policies and practices against these communities (Aktürk 2009; Aviv 2017; Bali 1998, 2001, 2004; Brink-Danan 2012; Haker 2003; Toktas 2005). They highlight Turkish nationalism’s role in justifying the exclusionary attitude through marking the boundaries between dominant and subordinate groups in modern Turkey. Indeed, Ziya Gökalp (1968), one of the founding fathers of Turkish sociology and nationalism, posited that the dismemberment of the Ottoman Empire was a result of cultural dissonance among different ethnic and religious communities. Accordingly, relying on Emile Durkheim’s sociological approach, he formulated a programme of Turkish nationalism that portrayed Hanafi Muslim culture in Turkey as the foundation of Turkish culture (Gökalp 1968). That is to say, his nationalism excluded non-Muslim minorities as outgroups in modern Turkey. Likewise, Ünlü (2014) argues that Turkishness is constructed as the dominant culture and associated with self-righteousness, selective knowledge and apathy towards minority cultures and religions. This positioning of the dominant group includes avoidance of subjects related to the Turkish maltreatment of and violence against the Armenians in the Ottoman period. Members of the dominant group may continue to enjoy privileges unless they breach the Turkishness contract (Ünlü 2014).

Accordingly, this study suggests that Turkish political parties’ threat perception is an important factor influencing their negative descriptions of Armenians in politics. These descriptions are likely to be more common among political actors from the Turkish right-wing, whose ideological orientations are Turkish Nationalism and Political Islamism. Since their political views are relatively more congruent with the Turkish nationalist description of Hanafi Turks as the dominant group in Turkey than are those of parties with other ideological orientations, they would be more likely to perceive Armenians as threats and portray them negatively. Furthermore, this study expects that debates related to perceived security and diplomatic threats would incline parliamentarians to describe Armenians unfavourably by making it more likely to associate them with such threats. Indeed, the relevant academic literature underlines the importance of perceived threats. According to Göl (2005), the state’s perceived threats of Armenian claims on its territory feed its exclusionary policies regarding the minority. With regards to international relations, Zarakol (2010) argues that the perceived threats of recognition of the Armenian Genocide to the global image of the Turks are an important factor in Turkish denial. Taken together, these insights converge on the following interrelated hypotheses:

Hypothesis 1: Right-wing politicians are more likely to describe Armenians negatively.

Hypothesis 2: Discussing threat-related topics makes politicians more likely to describe Armenians negatively.

Data and method

This study relies on content analysis, ‘a research technique for making replicable and valid inferences from data to their context’ (Krippendorff 1989, 403). In particular, it uses quantitative content analysis, a deductive method that codes data into variables and then statistically examines the relationships among those variables to test hypotheses (White and Marsh 2006). To that end, the study analyses all Turkish parliamentary speeches that mentioned the word ‘Armenian’ (Ermeni) between 1983 and 2018. It draws on total population sampling, which enables a thorough perspective on the descriptions of Armenians in Turkish parliamentary politics. The research scrutinises the parliamentary records in the following steps: (1) collecting the speeches that contain the word ‘Armenian’; (2) reading the speeches and generating a coding scheme; (3) two researchers coding the data and calculating the intercoder reliability; and (4) analysing the data quantitatively.

As Table 1 below demonstrates, the study creates a binary dependent variable: politicians’ negative and non-negative comments about Armenians. The unit of analysis is a member of parliament (MP)’s speech. Each coder assessed whether the general tone of a talk was negative or not by paying attention to how central the negative and non-negative comments were to the speech and considering their latent and manifest meanings. Indeed, few speeches (34 out of 2,024) contained both negative and non-negative comments. As the speeches had been prepared and edited by the politicians or their staff prior to having been spoken in the parliament, it was easy to understand their overall tone. This might have contributed to the high level of intercoder reliability discussed below. An example of a negatively coded speech can be given from a talk by Nidai Seven from the Nationalist Action Party in which he described Armenian presence as a potential threat: ‘Some news in the media mentions that Protestant Armenians plan to immigrate to our city, Diyarbakır. What kind of precautions is the government taking?’¹ An example of a non-negatively coded description comes from Selina Dogan of the Republican People’s Party. She celebrated the contribution of Armenian intellectuals to the feminist movement in Turkey: ‘I would like to remind you of various Armenian feminists, authors and activists who contributed to the struggle for women’s rights in this country … ’²

Table 1. Codebook.

Perception	NEGATIVE: Description of Armenians’ identity, existence or actions in unfavourable terms, which includes defending an unfavourable treatment or perception of Armenians.
	NON-NEGATIVE: Description of Armenians’ identity, existence or actions using value-free or favourable terms, which includes criticising an unfavourable treatment or perception of Armenians.
Debate topic	International affairs	Threat-related: Armenian genocide recognition; Karabakh conflict.
		Others: International relations with Armenia; International relations with other countries; the Armenian diaspora.
	National affairs	Threat-related: National security threats; the Turkish War of Independence.
		Others: Minority life in Turkey; Economy; Culture; Turkish citizenship law; Local politics; Geography.
Ideological orientation	Right: Speaker is a member of a Turkish nationalist or political Islamist party.
	Centre-right: Speaker is a member of a liberal-democratic party.
	Centre-left: Speaker is a member of a social-democratic party.
	Left: Speaker is a member of a socialist or Kurdish rights movement party.

To ensure reliability, the author and a research assistant (RA) compared their codes and calculated the percentage of coding agreement for the perception of Armenians. The RA only coded a sample of the speeches, which contained every twentieth speech starting from the first, totalling 102 among 2024 speeches. The simple agreement between two coders was 96% (.96), and Cohen's Kappa was .88, which could be interpreted as a strong intercoder agreement (McHugh 2012). The high agreement rate might be related to having a simple, binary variable (Lacy et al. 2015). The independent variables, i.e. the ideological orientation of the political party and the debated topics, are based on information available from official parliamentary records and political party documents. For example, there is an index for each parliamentary sitting that briefly describes the topic of debate. Therefore, the study did not require intercoding measures for these variables.

The study analyses how politicians portray Armenians during different types of debates to test the relationship between perceived threats and negative perceptions. It distinguishes between debates that concern international and national politics to provide a comprehensive analysis, as political parties’ approaches and perceived threats of Armenians might differ between national and international issues. Indeed, there is a statistically significant association between the perception of Armenians (negative or non-negative) and whether a political debate was nationally or internationally oriented (Pearson’s chi-square χ2(1) = 157.037, p < .000). International affairs debates are concerned with Turkey’s international presence and relations. The speeches on the international recognition of the Armenian Genocide and the Karabakh conflict between Azerbaijan and Armenia are related to perceived threats. The genocide debate is predominantly seen by the majority of the political spectrum, in varying ways, as a diplomatic threat to Turkey’s global image (Zarakol 2010) and as unjust foreign accusation (Gürpınar 2016). The debates on the genocide in the Turkish parliamentary records between 1983 and 2018 also present the dominant tendency to supress it as a threat. For example, in 591 speeches on the topic, none of the speaking politicians openly accepted that the genocide took place. The Karabakh conflict (1988-2020) involves an ethnically Turkic community, Azerbaijan, seen as a natural and eternal ally; therefore, the conflict is a perceived threat. Indeed, some reports declare that Turkey took an active part in the latest military confrontation in the Karabakh region (Gabuev 2020). The rest of the internationally-oriented debate topics are about relations with various countries and the Armenian diaspora (see Table 1 above). Discussions of the national security threats posed by the PKK (Partiya Karkerên Kurdistanê/Kurdistan Workers’ Party) and ASALA (Armenian Secret Army for the Liberation of Armenia), and the Turkish War of Independence period (1919-1923), are classified as threat-related. The former reflects on different phases of the Turkey-PKK conflict and the ASALA terror attacks between 1975 and 1990. The speeches on the Turkish War of Independence period refer to a past national security threat. The rest of the discussions are on the living conditions and problems of minorities; local municipal matters; Turkish citizenship law; and Turkish geography, economy and culture, including artefacts and social norms and regulations, as shown in Table 1 above.

Another key variable is ideological orientation. Table 2 below ideologically taxonomises the political parties in the dataset. It is based on an approximation of positions of the political parties informed by the relevant scholarship (Carkoglu 2007). In so doing, the study does not account for minor shifts in time or differences in the profiles of individual politicians. In the taxonomy, right-wing political parties are followers of Turkish nationalist and political Islamist ideologies. The latter stems from the Millî Görüş (National Vision) movement. The centre-right political parties are the liberals, and social-democratic political parties constitute the centre-left. The left-wing political parties in the Turkish parliament in this period are the socialists and the pro-Kurdish movement. Four members of the parliament were independent at the time of their speeches. The study ideologically categorised the independents by considering their biographies, previous and subsequent political party affiliations and speeches: three on the centre-right (Mr. Mesut Yilmaz, Ms. Ummet Kandogan, Mr. Mehmet Kemal Agar) and one on the centre-left (Mr. Kamer Genc). It should be noted that the left-right distinction in Turkey does not exactly match the political spectrum in Western democracies. Aydogan and Slapin (2015) argue that the centre-right and centre-left order is reversed in Turkish politics compared to the political parties in the Western European system. They find that the centre-left employs more populist rhetoric than the centre-right political parties in Turkish politics, which presents a contrast to the West. In parallel, Özbudun (2006) suggests that while centre-left and left-wing political parties in the West show more sensitivity to lower income groups, this is the case for centre-right and right-wing political parties in Turkey.

Table 2. Ideological orientations of political parties.

Right-wing	Centre-right	Centre-left	Left-wing
Welfare Party (RP)	Motherland Party (ANAP)	Populist Party (HP)	People’s Labour Party (HEP)
Virtue Party (FP)	True Path Party (DYP)	Social Democratic People's Party (SHP)	Democratic Society Party (DTP)
Felicity Party (SP)	Justice and Development Party (AKP)	Republican People’s Party (CHP)	Peace and Democracy Party (BDP)
Nationalist Democracy Party (MDP)	Nation Party (MP)	Democratic Left Party (DSP)	People’s Democratic Party (HDP)
National Action Party (MHP)	Democrat Turkey Party (DTP)		Freedom and Solidarity Party (ODP)
Great Unity Party (BBP)

Findings

Descriptions of Armenians in discussions about international relations

The dataset contains 1281 talks about international affairs. Table 3 below illustrates that discussions of international relations that mentioned Armenians were more often on threat-related topics: the Armenian Genocide recognition debate and the Karabakh conflict. These were followed by international relations with Armenia and other countries and the Armenian diaspora in the Western countries.

Table 3. Descriptions of Armenians in debates about international relations.

		Non-negative description	Negative description	Total
Debate	Genocide recognition	43	548	591
	Karabakh conflict	35	429	464
	Relations with Armenia	44	131	175
	Relations with other countries	22	27	49
	The Armenian diaspora	2	0	2
Total		146	1135	1281

The frequency of references to Armenians is similar among the political parties of the right, centre-right and centre-left, as seen in Table 4 below. The left-wing-leaning political parties mentioned the word ‘Armenian’ relatively less than the others in foreign affairs debates.

Table 4. Descriptions of Armenians and ideological orientation in debates about international relations.

		Non-negative description	Negative description	Total
Ideology	Right	7	396	403
	Centre-right	45	415	460
	Centre-left	39	322	361
	Left	55	2	57
Total		146	1135	1281

Table 5 below demonstrates that ideological orientation is a significant predictor of the negative descriptions of Armenians (p<.001). The odds of Armenians being described negatively is approximately 6 (1 to .163) to 1000 (1 to .000) times higher when they are mentioned by a speaker from a right-wing political party compared to a member of a political party with a different ideological orientation. Model 1b includes threat-related topics and presents a statistically significant relationship (p<.001). It shows that the odds of Armenians being described negatively is 4 times higher in a discussion of the Karabakh conflict than in other international relations debates. The odds of Armenians being described negatively is 10 times higher in a discussion of Armenian Genocide recognition than in other foreign affairs debates. Nagelkerke’s R-squared in both models indicates a relationship between prediction and grouping. The analysis confirms the main hypotheses by showing that ideological orientation and topics of security are reliable predictors for negative remarks about Armenians in parliamentary debates about international relations.

Table 5. Logistic regression models of the negative descriptions of Armenians in debates about international relations.

Variable	Model 1a		Model 1b
Ideology	B (SE)	Odds	B (SE)	Odds
Centre-right	−1.814*** (.412)	.163	−2.262*** (.424)	.104
Centre-left	−1.925*** (.417)	.146	−2.184*** (.426)	.113
Left	−7.350*** (.815)	.001	−7.930*** (.848)	.000
Debate topic	-	-	-	-
Genocide recognition	-	-	2.322*** (.299)	10.194
Karabakh conflict	-	-	1.394*** (.270)	4.031
Constant	4.036*** (.381)	56.571	2.984*** (.399)	19.766
Model χ2 (df)	279.044*** (3)		344.745*** (5)
Pseudo R^2 (Nagelkerke)	.385		.464
N	1281		1281

Note: *p < 0.05; **p < 0.01; ***p < 0.001. Reference category for ideology: right-wing. Reference categories for debate topics: discussing other foreign relations topics.

Descriptions of Armenians in discussions about national affairs

The dataset contains 743 speeches concerned with national matters. As seen in Table 6 below, the discussions were more frequently on the general living conditions of minorities in Turkey and the threat-related topics of national security threats and the period of the Turkish War of Independence. This was followed by culture, local matters, economy, Turkish citizenship law and geography.

Table 6. Descriptions of Armenians in debates about national affairs.

		Non-negative description	Negative description	Total
Debate	National security threats	7	233	240
	The Turkish War of Independence	18	148	166
	Minorities in Turkey	206	56	262
	Economy	5	13	18
	Local issues	3	20	23
	Geography	0	3	3
	Citizenship law	1	7	8
	Culture	16	7	23
Total		256	487	743

Table 7 below outlines that with regards to national affairs, the distribution of the frequency of mentions of Armenians is more homogenous among political parties of different ideological orientations, most often by the centre-right, followed by the right, centre-left and left.

Table 7. Descriptions of Armenians and ideological orientation in debates about national affairs.

		Non-negative description	Negative description	Total
Ideology	Right	18	181	199
	Centre-right	62	208	270
	Centre-left	47	97	144
	Left	129	1	130
Total		256	487	743

Table 8 below shows that ideological orientation is a significant predictor of negative descriptions (p<.001). The odds of Armenians being described negatively is approximately 3 (1 to .334) to 1000 (1 to .000) times higher when they are mentioned by a speaker from a right-wing political party compared to a member of a political party with a different ideological orientation. Model 2b includes the threat-related topics and shows a statistically significant relationship (p<.001). It illustrates that the odds of Armenians being described negatively are 16 times higher in a discussion about the Turkish War of Independence than in the other national affairs debate topics. The odds of Armenians being described negatively is 165 times higher in a discussion of national security threats than in other internal politics debates. Nagelkerke’s R-squared in both models indicates a relationship between prediction and grouping. The analysis confirms the hypotheses by showing that ideological orientation and reflections on perceived threats are reliable predictors of negative remarks about Armenians in parliamentary debates about national matters.

Table 8. Logistic regression models of the negative descriptions of Armenians in debates about national affairs.

Variable	Model 2a		Model 2b
Ideology	B (SE)	Odds	B (SE)	Odds
Centre-right	−1.098*** (.286)	.334	−1.750*** (.342)	.174
Centre-left	−1.584*** (.304)	.205	−1.491*** (.349)	.225
Left	−7.168*** (1.034)	.001	−8.070*** (1.230)	.000
Debate topic	-	-	-	-
National security threats	-	-	5.106*** (.737)	165.030
The Turkish War of Independence	-	-	2.785*** (.346)	16.205
Constant	2.308*** (.247)	10.056	1.093*** (.271)	2.983
Model χ2 (df)	351.575*** (3)		570.875*** (5)
Pseudo R^2 (Nagelkerke)	.521		.740
N	743		743

Note: *p < 0.05; **p < 0.01; ***p < 0.001. Reference category for ideology: right-wing. Reference categories for debate topics: discussing other national affairs topics.

Overall perception

To delineate the full picture of the relationships between the descriptions of Armenians, the ideological orientation of the speakers’ political parties and threat-related debate topics, the study ran a binary logistic regression of the entire dataset, which includes debates about both national and international affairs (N=2024). Table 9 below points to the significance of ideological orientation (p<.001) and the impact of right-wing ideology. The odds of Armenians being described negatively are approximately 4 (1 to .252) to 1000 (1 to .000) times higher by a speaker from a right-wing political party than a politician with a different ideological orientation. Model 3b adds threat-related topics and shows a statistically significant relationship (p<.001). The odds of Armenians being described negatively is 83 times higher in discussions of national security threats and 9 times higher in discussions of the Turkish War of Independence period than in other discussions. Moreover, the odds of Armenians being described negatively is 22 times higher in discussions of genocide recognition and 9 times higher in debates about the Karabakh conflict than in other discussions. Nagelkerke’s R-squared in both models indicates a relationship between prediction and grouping. The analysis supports the hypotheses by demonstrating that ideological orientation and perceived threats are significant predictors of negative remarks about Armenians in Turkish parliamentary debates.

Table 9. Logistic regression models of the negative descriptions of Armenians in all discussions.

Variable	Model 3a		Model 3b
Ideology	B (SE)	Odds	B (SE)	Odds
Centre-right	−1.377*** (.230)	.252	−1.983*** (.251)	.138
Centre-left	−1.555*** (.236)	.211	−1.772*** (.254)	.170
Left	−7.255*** (.617)	.001	−7.858*** (.682)	.000
Debate topic	-	-	-	-
National security threats	-	-	4.425*** (.631)	83.530
The Turkish War of Independence	-	-	2.223*** (.325)	9.231
Genocide recognition	-	-	3.128*** (.256)	22.839
Karabakh conflict	-	-	2.208*** (.221)	9.093
Constant	3.139*** (.204)	23.080	1.875*** (.216)	6.522
Model χ2 (df)	709.744*** (3)		1027.796*** (7)
Pseudo R^2 (Nagelkerke)	.469		.631
N	2024		2024

Note: *p < 0.05; **p < 0.01; ***p < 0.001. Reference category for ideology: right-wing. Reference categories for debate topics: discussing other topics.

Discussion and conclusion

This paper reports two major findings that support the research hypotheses. First, it presents a strong statistical relationship between ideological orientation and describing Armenians negatively in the parliamentary speeches between 1983 and 2018. In particular, being a member of a right-wing political party, defined as Turkish nationalist or political Islamist, exponentially increases the odds of making negative comments. The odds gradually decrease, in hierarchical order, among centre-left, centre-right and left-wing politicians. The finding that the centre-right seems to be more liberal towards Armenians than the centre-left is in line with the argument that they tend to act in reverse order in Turkish politics compared to Western politics (Aydogan and Slapin 2015). Second, the study demonstrates that the debates which are related to perceived national security and diplomatic threats are strong predictors of negative comments. In other words, the odds of politicians describing Armenians negatively are significantly higher when the former speak about past and present threats. These debate topics are national security threats posed by the PKK and ASALA; the international debate on the recognition of the Armenian Genocide; the period of the Turkish War of Independence; and the Karabakh conflict between Armenia and Azerbaijan.

These findings align with group position theory in two main ways. To begin with, right-wing political parties underline the importance of Turkish ethnic identity and Islam. In other words, they place Muslim Turks as the dominant group, and, therefore, they are more inclined to position non-Muslim minorities as outsiders and potential sources of threats to the dominant group. This seems to make them lenient towards negative comments about Armenians. Furthermore, the data shows that left-wing political parties, consisting primarily of pro-Kurdish and pro-minority movement politicians, tend to exhibit a reverse trend relative to the right-wing. They seem to position themselves alongside Armenians as outgroups and act in line with their perceived interests. Second, in accordance with the tenets of the group position perspective, the study highlights that discussing perceived threats increases the probability of making negative comments about Armenians in Turkish politics. That is to say, perceived threats seem to be associated with a prejudiced attitude towards Armenians in both national and international affairs. It should be noted that the paper does not propose any causal relationship between perceived threats and negative perception. That is to say, talking about perceived threats does not directly lead politicians to making negative comments, but, rather, provides more opportunities for relating Armenians to perceived threats.

The significance of perceived threats to negative perceptions confirms various arguments from the relevant scholarship. Most generally, it supports group position theory by highlighting the implications of the historically rooted boundaries between the dominant and subordinate groups and the perceived threats of Armenians in modern Turkish politics. Also, the tendency of pro-Kurdish movement political parties to avoid negatively describing Armenians, compared to other political parties, might be viewed as evidence of empathy and solidarity among non-dominant minority groups. This interpretation aligns with Bobo and Hutchings (1996) emphasis that subordinate group members could also act according to the feelings of alienation and threats caused by their disadvantaged positions. In other words, feelings of subordination may also inform perceptions of minorities. Furthermore, the paper’s findings resonate with the argument that the Turkish state’s perceived threats of Armenian territorial claims contribute to negative state policies (Göl 2005). They also support the academic literature’s view of non-Muslims as an excluded outgroup.

All in all, the pertinence of group position theory to explaining negative descriptions of Armenians in Turkish politics between 1983 and 2018 implies that Armenians are generally seen as an outgroup and as perceived threats, which presents a challenge to democracy and human rights by contributing to intolerance towards minorities. Indeed, a recent report on Turkish media found that Armenians were the group most frequently targeted by hate speech in 2019 and labelled as an enemy community posing a threat to Turkey, both currently and during the period of the Turkish Independence War, and also to Azerbaijan in the Nagorno–Karabakh conflict in Turkish media (Dink Foundation 2020). Deveci and Binbuğa Kınık (2019) claim that Turkish nationalism biases the legal process concerning hate speech, and the state generally does not punish hate speech targeting minorities. Accordingly, Turkish society and politics could benefit from more effective management of perceived security threats and stricter punishment of hate speech in order to become more inclusive. This might help Turkish society to avoid defining intergroup relationships in the framework of perceived threats to and rightful privileges of a dominant group. In so doing, they could create conditions for the decline of intergroup hostility.

Future studies could test the validity of group position theory in explaining perceptions of other minorities in Turkey. In so doing, they would not only contribute to filling gaps in the literature on Turkish politics, but would also test the validity of group position theory in non-Western contexts. This paper also calls attention to the fact that Turkish parliamentary proceedings afford a very rich evidence base. These records reflect, in detail, political parties’ official positions and differences on important topics in Turkish politics. This enables systematic analysis of political party descriptions of minorities in real life rather than of responses prompted by researchers. This level of depth is not easily available to other research methods. For example, if we had conducted a survey among all members of the political parties involved in this analysis, we might not have an equally lucid picture of how these parties would react in actuality; individuals or groups could be decision-makers in a political party whose views might differ from the average of their fellows. Future research could exploit the easy online access to Turkish parliamentary debates on a variety of topics. This would enable a new approach to prejudice and how it informs important policy decisions. Draege (2019) also claims that parliamentary debates are very useful resources to analyse political decision-making, as they contain comprehensive evidence. Such analysis could also allow for international comparisons of ethnoreligious hostility by exploring parliamentary proceedings in different contexts.

Acknowledgements

This research was made possible by a Gulbenkian Foundation research grant. I would like to acknowledge my gratitude to Dr Metin Yüksel for his help in creating the dataset. Last, this article benefited greatly from the valuable comments of two anonymous reviewers.

Disclosure statement

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

Additional information

Funding

This work was supported by Fundação Calouste Gulbenkian: [Grant Number 227637].

Notes

1 TBMM, Tutanak Dergisi, Term 21, Legislative Year 4, Sitting 30 (04.12.2001), 122.

2 TBMM, Tutanak Dergisi, Term 26, Legislative Year 2, Sitting 78 (07.03.2017), 98.