1. Introduction
Cultural studies show that the expression of emotions is guided by display rules that define what emotions it is appropriate to express in a given situation (Matsumoto et al., Reference Matsumoto, Yoo and Nakagawa2008). Accordingly, cultures may vary in the cultural norms regulating whether and how emotions are expressed and how these expressions are perceived by other members of that cultural group (Smith et al., Reference Smith, Easterbrook, Celikkol, Chen, Ping and Rizwan2016). Significant cross-cultural differences have been found concerning the basic emotion of anger, particularly between individualistic and collective cultures, but also between cultures belonging to different religions.
Considering that anger display rules are shaped by cultural norms and values, we would expect cross-cultural differences in how political leaders perform anger across different countries. However, we know very little about how far and in which ways the cultural display rules of anger influence public expressions of anger by high-status individuals in a given society. This paper investigates anger in Turkish culture and presents a case study of anger in the speeches of political leader Recep Erdoğan, tracing how he navigates the cultural possibilities of anger in political rhetoric across the different phases of his time in government.
In the case of Turkey, we might expect a complex interplay between cultural factors. While Turkish culture is generally categorized as collectivist, which encourages the suppression of anger, there are cases such as honour issues, where the expression of anger is expected in certain circumstances (Boiger et al., Reference Boiger, Güngör, Karasawa and Mesquita2014; Okur & Çorapcı, Reference Okur and Çorapcı2016; Yılmaz, Reference Yılmaz2018). Furthermore, Islamic norms advocating patience and the suppression of anger but which also deem anger necessary under certain conditions may also influence political rhetoric in Turkey, given the significant influence of conservative religious groups.
On the other hand, Erdoğan’s socio-political status as a powerful populist leader may also influence his public performances of anger. The expression and incitement of anger is associated with populist leaders in previous research (Breeze, Reference Breeze2020; Wagner, Reference Wagner2014), at least in Western contexts. Furthermore, research on power, emotions, and leadership suggests that individuals in positions of power have more freedom and will be more inclined to express anger, while those with less power are more constrained by cultural norms.
Drawing on extended conceptual metaphor theory (ECMT) (Kövecses, Reference Kövecses2020a) and through corpus-assisted discourse analysis, this study explores to what extent the discursive performance of anger by Recep Tayyip Erdoğan, the Turkish leader since 2003, is influenced by the cultural display rules, religious context, and his role as a long-term populist political leader. It also aims to reveal how Erdoğan’s conceptualization of anger is shaped by the discursive challenges arising from these conflicting cultural, political, and social factors.
2. Theoretical framework
ECMT presents a unified theoretical framework to approach the concept of anger in political discourse by providing categorical tools for analysing concrete usages of metaphors in political discourse and by including contextual components when analysing and interpreting results (Kövecses, Reference Kövecses2020b, Reference Kövecses2020c). In the extended version of his model, Kövecses convincingly argues that ECMT is capable of explaining the conceptual structure of metaphors, as well as the socio-pragmatic and rhetorical functions of metaphorical expressions in discourse.
Contextual meaning can be found at the level of mental spaces which is the ‘individual level’ of metaphorical conceptualization. Cognitive and conceptual components work together dynamically in the actual use of metaphors at the level of mental spaces (or ‘scenarios’ in the sense that Musolff (Reference Musolff2006) uses the term), which function in contextFootnote 1 (Kövecses, Reference Kövecses2020b). ECMT distinguishes four types of context, situational, discourse, bodily, and conceptual-cognitive (Table 1), which each have empirically established contextual factors (Kövecses, Reference Kövecses2021).
Four context types and their contextual factors
Source: Kövecses (Reference Kövecses2021, p. 137).
Context plays a crucial role determining which metaphor is used in which situation. Specifically, the actual use of metaphor results from the priming effect of one or several contextual factors. More importantly, even in the case of established forms of figurative language such as idioms and proverbs, utilizing one metaphorical expression rather than another may serve to achieve pragmatic effects such as the justification of one’s actions, or blame attribution (Kövecses, Reference Kövecses2020b, 127).
ECMT thus allows us to see why and how particular linguistic and conceptual metaphors are used in discourse. In this study, we use the contextual component of ECMT to understand the concept of anger in a specific discourse, so as not only to interpret the metaphorical conceptualizations of anger but also to interpret the results of corpus-assisted discourse analysis of direct expressions of anger.
2.1. Contextual factors specific to Turkish case study
We would expect that several contextual factors would affect the anger performance of Erdoğan, possibly in conflicting ways. Within the situational context, collectivist Turkish culture (cultural situation) encourages the suppression of anger, while Erdoğan’s position as the political leader of Turkey (social situation) would be expected to have the opposite influence allowing easier and more frequent expression of anger. Furthermore, Erdoğan’s conservative religious electoral base (both conceptual-cognitive and discourse context, that is, concerns and interests, and knowledge about hearer) would prime another cultural factor, namely Islamic rules concerning anger control (situational context), and also prioritize the religious discourse on anger (discourse context). Erdoğan as the conceptualizer could be affected by his Islamic-conservative ideological position (conceptual-cognitive context) and by the dominant metaphorical conceptualizations of the anger within Islamic discourse (discourse context). On the other hand, especially after 2013, increasing populist tendencies in Erdoğan’s discourse and practice (surrounding discourse) might well exert an influence in the direction of more frequent performances of anger. Under these theoretically re-formulated conditions, this paper aims to tease out which contextual factors prime and shape the anger expressions (both direct and metaphorical) in the political discourse of President Erdoğan, and in which ways.
As briefly mentioned previously, the studies conducted on anger in Turkish culture and language suggest that the suppression or masking of anger is an important cultural rule regulating display of anger (Aksan, Reference Aksan2011; Aksan & Aksan, Reference Aksan, Aksan, MacArthur, Oncins-Martínez, Sánchez-García and Piquer-Píriz2012; Arıca-Akkök, Reference Arıca-Akkök2017; Çorapcı et al., Reference Çorapcı, Aksan and Yagmurlu2012; Diener & Lucas, Reference Diener and Lucas2004; Friedlmeier et al., Reference Friedlmeier, Çorapcı and Cole2011; Matsumoto et al., Reference Matsumoto, Yoo and Nakagawa2008; Yılmaz, Reference Yılmaz2018). On the other hand, given his powerful position, Erdoğan could be expected to express more anger compared to others in society since people associate the emotion of anger with the powerful and the emotion of guilt with the powerless (Tiedens et al., Reference Tiedens, Ellsworth and Mesquita2000; Brescoll & Uhlmann, Reference Brescoll and Uhlmann2008.). Furthermore, some studies in other collectivist cultures such as Asian societies found that anger expression is facilitated by high social status, which serves as a cultural permit or authorization (Park et al., Reference Park, Kitayama, Markus, Coe, Miyamoto, Karasawa, Curhan, Love, Kawakami, Boylan and Ryff2013; Taylor & Risman, Reference Taylor and Risman2006). In general, powerful, higher-ranking people appear to be freer to express anger (van Kleef & Lange, Reference van Kleef and Lange2020), whereas those with less power are more bound by cultural expressive norms (LaFrance et al., Reference LaFrance, Hecht and Paluk2003).
Another contextual factor affecting Erdoğan’s anger performance is populism as a style and practice. Previous research has suggested that at least in the Western context, populist leaders are associated with the expression or incitement of anger (Wagner, Reference Wagner2014; Wahl-Jorgensen, Reference Wahl-Jorgensen2019). While it has been a truism in academic research on Turkish politics that Erdoğan has adopted an increasingly antagonistic populist approach since the 2013 Gezi protests, which was just the first of a series of major events challenging his political authority in Turkey (Ekşi & Wood, Reference Ekşi and Wood2019; Taş, Reference Taş2022), to our knowledge no research has yet examined the role of anger in his populist discourse.
Among the contextual factors with potential to condition anger expression in Erdoğan’s discourse, religion has a special place as a cross-contextual factor. The Islamic understanding of anger may be expected to influence Erdoğan’s political rhetoric, since he draws considerable political support from conservative Islamic groups, playing a role in several types of context (cultural situation, ideology, concerns and interests, dominant forms of discourse, knowledge about hearer).
According to Islamic morality, anger is an emotion to be disciplined, to be civilized and to be kept within the bounds of Islamic rules. It should be expressed only when necessary (Gördük, Reference Gördük2014). While suppression of anger is advised by Islamic authorities for believers, a morally and religiously superior character overcomes anger by using his will instead of suppressing it (Gördük, Reference Gördük2014). The faculty of ‘sabır’ (patience)Footnote 2 is seen as a virtuous antidote to anger and as an internal power needed to control anger (Önal, Reference Önal2008). Sabır is one of the most frequently occurring lexemes in the Quran, indicating the moral strength that a believer should possess to withstand both external and internal destabilizing forces. Anger is an evil force which challenges sabır and occurs as overflow of sabır (Aksan & Aksan, Reference Aksan, Aksan, MacArthur, Oncins-Martínez, Sánchez-García and Piquer-Píriz2012). ‘Sabır’ and ‘çile’ (suffering) are culturally salient concepts that serve to structure Turkish speakers’ understanding of life, morality, and emotion, and the cultural significance of the metaphors moral strength is sabir and anger is an overflow of sabir can scarcely be overstated. The person is understood to be under pressure from offending events and is enduring them all patiently, as morally expected. The virtuous person endures patiently up to the very last moment humanly possible. The final event or confrontation adds the very last drop (Aksan & Aksan, Reference Aksan, Aksan, MacArthur, Oncins-Martínez, Sánchez-García and Piquer-Píriz2012).
Consequently, anger is seen as loss of control and as moral weakness. However, anger may be justified in Islam as righteous indignation and may be even necessary or required in some situations harmful to the community of believers and to religious/moral order, such as disobedience to God’s commandments, injustice to the people, humiliation, threats to the nation, honour, or dignity (Gördük, Reference Gördük2014). Yet even then, anger should be controlled not to give rise to disproportionate or unjust reactions (Gördük, Reference Gördük2014). The concept of ‘had’ (boundary) plays an important role in justifications of anger in Islamic teachings. In fact, anger, and the resulting retribution and punishment against phenomena that exceed the boundaries of Islamic morality, may be seen as morally good. Trespassing the moral bounds laid down by Islamic rules (‘haddini aşmak’) can be seen as an instance where anger is justified and even necessary. Thus, conceptual metaphors of ‘had’ in Islamic discourse can be expected to occur in other discourses, just as Biblical metaphors are often recycled in later discourses over ages (Kövecses, Reference Kövecses2015, pp. 55–Reference Tiedens, Ellsworth and Mesquita56).
3. Emotion concepts in ECMT
In the cognitive linguistic view, metaphor is a means of conceptualizing one domain of experience in terms of another (Kövecses, Reference Kövecses1986; Lakoff & Johnson, Reference Lakoff and Johnson1980). The domain of experience used to comprehend another domain is typically more physical, more directly experienced, and better known than the domain we wish to comprehend, which is typically more abstract, less directly experienced, and less known. The more concrete domain is called the source domain and the more abstract one is called the target domain.
As vast literature on CMT shows, emotion concepts are largely metaphorically and metonymically constituted and defined. Anger is one of the most widely analysed emotion concepts within CMT (Kövecses et al., Reference Kövecses, Szelid, Nucz, Blanco-Carrión, Akkök, Szabó, Heredia and Cieślicka2015; Kövecses, Reference Kövecses1986, Reference Kövecses2009, Reference Kövecses2015; Lakoff & Kövecses, Reference Lakoff, Kövecses, Holland and Quinn1987). Previous CMT research on anger shows that many unrelated languages and cultures share the generic-level metaphor: the angry person is a pressurized container. This metaphor is motivated by universal experiences of the embodiment of anger: the pressurized container metaphor underlies the widespread conception that anger is a force that makes the angry person perform aggressive or violent actions, and the actual physiology of anger provides considerable support for this conceptualization (Kövecses, Reference Kövecses2009). In Turkish, too, some previous studies found that anger is often conceptualized as an entity in a container (Kövecses et al., Reference Kövecses, Szelid, Nucz, Blanco-Carrión, Akkök, Szabó, Heredia and Cieślicka2015). The container is the human body, and anger is a substance, which can be either a fluid or a solid, inside it.
CMT also offers through metaphor analysis a prototypical scenario of anger in American English (Table 2; Kövecses, Reference Kövecses1986; Lakoff & Kövecses, Reference Lakoff, Kövecses, Holland and Quinn1987):
Stages of anger
Source: Lakoff and Kövecses (Reference Lakoff, Kövecses, Holland and Quinn1987).
S is the person who gets angry, short for the self.
Lakoff and Kövecses (Reference Lakoff, Kövecses, Holland and Quinn1987) point out that the various conceptual metaphors all map onto a part of the prototypical scenario and jointly converge on that scenario. This is a culture-specific folk theory of anger and may differ from culture to culture, while even within the same culture, many deviations are possible (‘non-prototypical cases’). Research on the Turkish prototypical anger scenario is rare, but what there is argues that the foregoing scenario is valid also for Turkish culture, with some minor variations (Arıca-Akkök, Reference Arıca-Akkök2017). Kövecses gives a non-exhaustive list of non-typical scenarios of anger, as detailed in Table 3.
Although some non-typical scenarios – specifically the ones where retribution (Stage 5) is prioritized, such as ‘don’t get mad, get even’ and ‘wrath’, and those where the offending event (Stage 1) is prioritized, such as ‘righteous indignation’ – clearly omit various stages of the prototypical scenario, most CMT research on anger, including Kövecses’s original works, has focused on conceptual metaphors that can be mapped into Stages 2–4, that is, the stages where embodied effects of anger can be observed. These stages are also those where intensity and control, the two most important aspects of the concept of anger, are clearly present (Kövecses et al., Reference Kövecses, Szelid, Nucz, Blanco-Carrión, Akkök, Szabó, Heredia and Cieślicka2015).
On the other hand, Stages 1 and 5 may be considered the areas where morality intersects with the concept of anger. Angry behaviour at Stage 5 is, in itself, viewed as a form of retribution. However, the metaphors which could belong to Stage 5 have rarely been studied in previous CMT research. Even in Kövecses’s original work, this dimension of anger is discussed very little. He mentions that warnings and threats of retribution can count as angry behaviour such as I’ll pay you back, −with interest! And are thus metonymies for anger. Arguably, Stage 5 is the dimension where the metaphorical conceptualization of anger by Kövecses (Reference Kövecses1986) and the metaphorical conceptualization of moral metaphors by Lakoff (Reference Lakoff2010) overlap.
Taking this further, Kövecses openly states that emotions are commonly based on moral ideas so that the concept of emotion evokes the notion of social norms, of right or wrong, appropriateness of response, and the appropriate measure of feeling (Kövecses, Reference Kövecses2020b), constituting a ‘domain matrix’. ECMT thus extends its scope to contextual factors influencing emotion concepts. We can relate this to Lakoff’s conceptualization of moral metaphors in politics, particularly with metaphors that conceptually cut across anger and retribution on the one hand, and anger and offending event on the other. For example, the non-typical ‘don’t get mad, get even’ scenario is a moral scenario of anger, metaphors for which would be positioned at the intersection of moral accounting and anger. Moral accounting simply refers to the metaphorical understanding of moral action as a financial transaction whereby financial morality is carried over to morality in general. There is a moral imperative not only to pay one’s financial debts, but also one’s moral debts (Lakoff, Reference Lakoff2010). In Lakoff’s (Reference Lakoff2010, pp. 45–46) words: ‘Just as literal bookkeeping is vital to economic functioning, so moral bookkeeping is vital to social functioning’. The ‘getting even’ scenario implies balancing the scales of justice in the sense that retribution can alleviate or prevent anger.
Another stage which potentially entails an intersection of morality and anger is Stage 1 (cause of anger) in the prototypical scenario of anger. For example, cause of anger is trespassing (the limits) is one of the examples given by Kövecses (Reference Kövecses1986) for conceptualizing anger metaphors (also see Zlatev et. al. Reference Zlatev, Jacobsson and Paju2021). In fact, in some cultures, such as China, morality as bounded space is a very common conceptual metaphor (Slingerland, Reference Slingerland2004, Reference Slingerland2007). Thus, in Stage 1, the concept of morality and concept of anger may intersect as cause of anger is trespassing the moral bounds.
4. Methodology
4.1. Corpus
The data were collected from Turkish Presidency digital books containing all public speeches given by President Erdoğan.Footnote 3 We selected only political speeches given at mass meetings during his electoral campaigns for local, national, and presidential elections (in total 7 elections) from 2004 to 2018. Between 2004 and 2018, nine elections were held in Turkey (Table 4). Elections in 2015 were not included in the corpus as Erdoğan did not organize any electoral campaign.Footnote 4
Elections held in Turkey between 2004 and 2018
The resulting corpus comprised 326 political speeches, consisting of 700,575 words. To follow the diachronic change in expression of anger in discourse, we created two subcorpora by taking the antigovernment Gezi protestsFootnote 5 in 2013 in Turkey as a turning point after which Erdoğan’s populism was intensified through the adoption of an increasingly antagonistic, authoritarian, conspiratorial, and nativist discursive style (Destradi et al., Reference Destradi, Plagemann and Taş2022; Ekşi & Wood, Reference Ekşi and Wood2019; Taş, Reference Taş2022). Table 5 shows the breakdown for the Pre-2013 and Post-2013 subcorpora.
Corpus: Pre-Gezi and Post-Gezi subcorpora
4.2. Analytical approach
To explore how anger is conceptualized and verbally expressedFootnote 6 in Turkish political discourse, this research uses a mixed-methods approach combining quantitative and qualitative analysis and loosely drawing on corpus-assisted cognitive discourse analytical methodology (Fabiszak & Hebda, Reference Fabiszak, Hebda, Winters, Tissari and Allan2010).
For corpus-assisted discourse analysis of anger, a culturally sensitive lexicon of Turkish anger-related terms was constructed for qualitative and quantitative processing. The anger lexicon was built by manually scanning the terms, idioms, and metaphors used in the Turkish dictionaries for anger using a snowball method following synonyms and semiotic closeness in meaning, complemented by terms gathered from previous research in Turkish linguistics on anger expressions (Akın, Reference Akın2016; Aksan, Reference Aksan2006; Arıca-Akkök, Reference Arıca-Akkök2017; Atay, Reference Atay2022; Baş, Reference Baş2015; Çet, Reference Çet2006; Darıcı, Reference Darıcı2012; Şaş, Reference Şaş2023). To expand the lexicon, lists of anger terms in English were compiled from previous research on anger in disciplines from linguistics to social-psychology and sociology (Bednarek, Reference Bednarek2008; Parrott, Reference Parrott2001; Plutchik, Reference Plutchik2001; Russell, Reference Russell1980; Turner, Reference Turner2007) and translated to Turkish with the help of English–Turkish dictionaries. The lexicon was subdivided into direct and indirect expressions of anger (Breeze & Casado-Velarde, Reference Breeze and Casado-Velarde2019). The resulting lexicon comprised 13 direct anger termsFootnote 7 (anger nouns, verbs, and adjectives) and 139 indirect anger terms (figurative speech, i.e., metaphors, metonymies, idioms). The lexiconFootnote 8 was uploaded to MaxQDA for analysis using its Dictionary feature. Since MaxQDA has no lemmatization tool for the Turkish language, each dictionary term was lemmatized manually.
The results of the automatic analysis were manually cleaned for all terms. Then, since it is obviously not the same to express one’s own anger and to describe other people’s anger, we classified all the instances of anger into two different categories: (a) ‘extended’ inscribed anger (including words such as I, We, Turkey, People) and (b) ascribed anger (anger attributed to others) (Bednarek, Reference Bednarek2008; White, Reference White, Tracy, Sandel and Ilie2015), according to who feels that particular emotion (i.e., the emoter) (Bednarek, Reference Bednarek2008; Breeze, Reference Breeze2020). We revised the definitions of ‘ascribed’ and ‘inscribed’ to adapt these categories to the genre of political discourse. In its classical definition, ‘inscribed’ refers to the cases where the emoter is the first person, in other words the author of the sentence and the one who feels (or not) the emotion are the same person. In order to understand the use of anger in an antagonistic political discourse as I/we versus them or the self/selves versus others (Kövecses & Douthwaite, Reference Kövecses, Douthwaite, Douthwaite and Tabbert2023), we manually coded all the instances where the anger was expressed as felt (or not) by Erdoğan, his party members, the nation and first-person plural (‘we’), as ‘inscribed’ anger. All other instances where the emoter was others (other parties, out-groups, other countries, other nations, etc.) were coded as ‘ascribed anger’.
However, this type of categorization leaves out instances where Erdoğan denies, rejects, disapproves, or negates anger as an emotion. Therefore, we also manually coded cases where anger was negated grammatically (not, no, never, etc.) (Bednarek, Reference Bednarek2008), that is, sentences without ‘emoter’ but with direct anger terms, as examples of appraisal (negative or positive appraisal of anger).
In addition to quantitative lexical analysis, figurative expressions of anger were qualitatively analysed to further explore the cultural conceptualizations of anger, drawing on ECMT (Kövecses, Reference Kövecses2020a). We listed the metaphorical idioms and proverbs in the corpus using our Turkish anger lexicon for indirect anger terms.Footnote 9 We did not include metaphorical expressions containing direct anger terms,Footnote 10 in order to see the implied, disguised, or implicit metaphors of anger. After grouping conceptual metaphors in the corpus and categorizing them according to their use as ‘ascribed’ or ‘inscribed’ anger, we mapped the anger metaphor groups onto the prototypical stages in the anger scenario and compared metaphorical conceptualizations of ascribed and inscribed anger in terms of their pragmatic effect in different anger scenarios.
5. Corpus-assisted quantitative analysis
In general, the expression of anger, whether through direct anger terms or through figurative speech, was rare. Only 496 anger terms were detected (relative frequency 0.00070799, or 707 per million). Additionally, there were no substantial differences between the two subcorpora in terms of relative frequencies (Table 6). Log-likelihood (LL) tests (Rayson & Garside, Reference Rayson and Garside2000)Footnote 11 (Table 7) showed that anger terms were not significantly more frequent for Post-Gezi than Pre-Gezi (LL was 2.70, lower than the 3.84 needed to be significant at p < 0.05).
Distribution of anger words, idioms, and metaphors
Log-likelihood results
Note: O1 is observed frequency in Corpus 1. O2 is observed frequency in Corpus 2. The values %1 and %2 show relative frequencies in the texts. A positive value indicates overuse in O1 relative to O2. A negative value indicates underuse in O1 relative to O2.
We then investigated whether these anger expressions were ascribed or inscribed. The results were unexpected concerning the anger display rules in collectivist cultures: inscribed anger was almost four times more frequent than ascribed anger in terms of absolute frequencies (Table 8). Furthermore, while only 14% of the inscribed anger was negated or negatively appraised, all ascribed anger was in positive form.
Inscribed anger versus ascribed anger
As a third step, we compared the distribution of direct and indirect anger terms to explore differences between the style of ascribed and inscribed anger expressions. We found that 88% of inscribed anger was expressed using figurative speech, compared to 31% of ascribed anger (Table 9). Ascribed anger was more frequently expressed by using direct anger terms (69%). Among the inscribed anger expressed using direct anger terms, only 5% was in the affirmative, meaning that direct anger terms were generally used to express ‘not feeling angry’ when they were used to express the mood of Erdoğan/nation/Turkey or ‘Us’ (Table 10).
Cross-comparative: inscribed versus ascribed – direct versus figurative
Direct anger terms cross-comparative: inscribed versus ascribed – positive versus negated
Additionally, when direct anger terms were used to express others’ anger, they were never negated in the corpus. In fact, this was also true for figurative language (Table 11). Ascribed anger was never negated even in the instances where indirect expression of anger was attributed to others. Overall, in the whole corpus, others were never evaluated as ‘not angry’.
Cross-comparative figurative language: inscribed versus ascribed – positive versus negated
Although the difference between the two corpora in terms of frequency of anger terms (Table 7) was not statistically significant (LL), there was a statistically significant difference between the two corpora in terms of figurative language use in expressing anger. While in the Pre-Gezi Corpus direct and indirect expressions were almost equal in percentage, in the Post-Gezi corpus, figurative expressions of anger constituted 90% of all anger expressions (Table 12).
Cross-comparative figurative language: direct versus figurative – pre-Gezi versus post-Gezi
In Pre-Gezi, 76% of figurative expressions, but no direct terms, were used to convey inscribed anger in positive form (Tables 13 and 15). In the Post-Gezi corpus, 91% of figurative expressions were used to express inscribed anger (Table 14), with no significant increase in direct anger terms used for inscribed anger (Table 16).
Pre-Gezi figurative language
Post-Gezi figurative language
Pre-Gezi Direct Anger Terms
Post-Gezi direct anger terms
Overall, if we compare the frequency of inscribed anger expressed positively between the two subcorpora (Table 17), we find 62 Pre-Gezi versus 268 Post-Gezi, which gives us relative frequencies of 0.00022896 (Pre-Gezi) and 0.00062357 (Post-Gezi) respectively (LL score 60.90). We may safely conclude that Erdoğan’s election campaigns became angrier after 2013 with a major increase in the use of figurative speech for conveying anger.
Log-likelihood results (for inscribed anger expressed positively)
6. Figurative language in the corpus
The preference for figurative language over direct terms to express anger, especially for inscribed anger, made us look closely at the metaphors, metonyms, and idioms used in the corpus.Footnote 12 Using ECMT, we grouped the metaphors in use in the corpus (Tables 18 and 19).
Metaphors and metonymies of ascribed anger
Metaphors and metonymies of inscribed anger
A comparison of the categories of conceptual metaphors for inscribed and ascribed anger shows that they were significantly different. In fact, there was only one common category between two groups, namely violent frustrated behaviour.
6.1. Metaphor and metonymies of ascribed anger
Our results show that Erdoğan’s speeches use significantly more metaphors and metonymies related to behavioural effects of anger (angry/aggressive/violent behaviour), rather than embodiment related expressions (container and heat) to express the anger ascribed to others (CN = 3). This is followed by ascribed anger metaphors which conceptualize anger’s distorting effect on correct perception (CN = 11): anger causes loss of judgement and leads to loss of self-control with insane (CN = 8) and animal-like behaviour (CN = 4), or violent frustrated behaviour (CN = 4).
6.1.1. interference with accurate perception
Eleven metaphors of ascribed anger belong to interference with accurate perception, a subset of the physiological effects of an emotion stand for the emotion. These metaphors are related to distortion of visual ability, such as gözü dönmek (CN = 9) and gözü kararmak (CN = 2). Gözü dönmek (literally, one’s eyes are rolled back) refers to not being able to see straight with anger/being blinded with rage.
Gözü kararmak (lit. one’s eyes are darkened) means being blindfolded by rage.
In their semantic meaning, gözü kararmak and gözü dönmek refer to losing the ability to think sanely, thus engaging the semantic category loss of judgement (Kövecses, Reference Kövecses1986). In fact, the effect of anger on physical and mental abilities refers to a state in which anger starts to obtain control of body and mind. These metaphors are thus related to the stage of loss of control over body and mind, rather than being simply metonymies indicating that the person is angry or furious.
6.1.2. behavioural effects of anger
behavioural effects of anger is the next step in loss of control because anger takes full control of bodily movement and free will, and makes us do things we would not normally do. behavioural effects of anger can be ordered according to gravity (in terms of the degree of behavioural effect of anger) as follows:
-
1. violent frustrated behaviour;
-
2. insane behaviour;
-
3. aggressive animal behaviour.
In the last two, the person’s mental capacity is overruled by anger.
The most frequent metonymic conceptualization for the category of behavioural effects of anger is anger is insanity/anger is insane behaviour (CN = 8). Various metaphorical idioms were used to convey anger felt by others, such as kudurmak (lit. ‘to become rabid, to go mad with rage’), çıldırmak (lit. ‘to go crazy’), çıldırtmak (lit. ‘to drive crazy, to make very furious’), küplere binmek (lit. ‘to mount on the barrels, to get furious’), and köpürmek (lit. ‘to foam due to anger, foam at the mouth’):
Other people’s anger is also ascribed using metonymies about animals (CN = 4) in the source domain of aggressive animal behaviour, such as Kaplan kesilmek (lit. ‘to play the tiger’) which means to behave in an aggressive way.
In the category of violent frustrated behaviour, one figurative expression was used in the form of reported speech (CN = 4). Dişlerini sökmek refers to pulling someone’s teeth out (to defang).
6.1.3. anger is heat in a container
Notably, in Erdoğan’s political discourse, container metaphors for anger are rare and used only to ascribe anger to others (CN = 3, in total 372 metaphors, less than 1%). In container metaphors, the container is the human body, and anger is conceptualized as a hot substance, which can be either a fluid or a solid, in this container (Kövecses et al., Reference Kövecses, Szelid, Nucz, Blanco-Carrión, Akkök, Szabó, Heredia and Cieślicka2015). We detected two subcategories for anger is heat in a container: a) anger is fire and b) anger is explosion. Ateşler salmak (lit. ‘to spread fire’) was categorized as anger is fire:
Çatlayıp patlamak (lit. ‘to crack and explode’), which refers to anger with envy, was categorized as anger is a heat in a container explosion:
6.1.4. anger is a sharp object
The metaphorical idiom bilenmek may be literally translated as ‘to hone oneself’. Here, anger becomes one with the self and sharpened during time and is conceptualized as anger is a sharp object. In Turkish, bilenmek refers to anger mixed with grudge and resentment since this covers a time span in which anger grows inside the person and sharpens feelings, getting prepared for revenge.
6.1.5. loss of self-control
Çileden çıkmak is one of the culture-specific complex metaphors in this discourse. It may literally be translated as to come out of religious seclusion/hermitage. It refers to breaking the religious rules which order people to stay in hermitage for 40 days. It is linked to the end of patience earlier than required by religious rules. Çile is about disciplining the self through suffering by enduring patiently. Çileden çıkarmak is to cause someone to interrupt this endurance with extremely provocative actions, thereby causing the hermit’s loss of self-control.
6.1.6. Prototypical scenario of ascribed anger
When we mapped the use of anger metaphors in our corpus into the prototypical scenario of anger (see Table 20), most of the metaphors of anger ascribed to others were those describing loss of control (Stage 4) in the prototypical anger scenario. The behavioural effects were presented such that anger controls the mind of others until the point that they behave like animals or insane people. Even the metaphors mapped in Stage 2, about bodily effects of anger, were about anger controlling physical abilities such as visual ability and (implicitly) mental ability, implying loss of judgement. Overall, loss of control and being captivated by anger account for most metaphors for ascribed anger (79%) and others’ actions are metaphorically represented as irrational, unjust, excessive, and animal-like. Interestingly, the metaphors categorized within the domains of cause of anger (Stage 1. Offending event), control of anger (Stage 3. Attempt at control), and acts of retribution (Stage 5) are non-existent in the metaphorical conceptualization of others’ anger in the corpus.
Prototypical scenario of ascribed anger
6.2. Metaphors of inscribed anger
In the corpus, we found 257 instances of anger metaphors of retribution and punishment, all referring to inscribed anger (Table 21). Metaphors of retribution and punishment constitute 76% of inscribed anger metaphors. This is followed by the metaphors of cause of anger is trespassing (13%).
Metaphors of retribution and punishment
6.2.1. act of vengeance stands for anger
Retribution and punishment metaphors of anger belong to the retribution stage (Stage 5) and differ from the angry behaviour at Stage 4, which marks the angry person as having lost control. Accordingly, we categorized them as retributive act stands for anger and punishing behaviour stands for anger. These metaphors stand at the intersection of morality and anger, as the model of retributive justice is built into our concept of anger and seeking vengeance is part of the prototypical anger scenario (Lakoff & Kövecses, Reference Lakoff, Kövecses, Holland and Quinn1987). In fact, most of these metaphors belong, at the same time, to the category of moral accounting developed by Lakoff (Reference Lakoff2010).
In fact, most of the retribution metaphors (CN = 151) in the corpus directly contain the lexeme hesap (‘account-ing’): hesabını sormak (CN = 110)’, hesabını vermek (CN = 40), and hesaba çekmek (CN = 1). On the other hand, we categorized the figurative expressions haddini bildirmek, dersini vermek, and tokat/şamar/sille as metaphors of punishment, which is another moral scheme within moral accounting. Metaphors of punishment differ from metaphors of retribution by establishing a hierarchical relation between punisher and punished by ascribing the necessary authority to the punisher. This is clear in the metaphor of dersini vermek (‘to teach them a lesson’), which metaphorically depicts the punisher as a teacher.
However, the metaphor haddini bildirmek (to make someone know his line/place) needs more explanation as a metaphor of punishment. The Turkish lexeme ‘had’ in the metaphor openly carries anger into the moral domain. The dictionary defines haddini bildirmek as to punish someone since (s)he threatens the established moral hierarchy by exceeding his/her limits with his/her acts/behaviour. Furthermore, had has strong roots in Islamic thought. Had in Islam refers to the limit which God defines for people’s acts and behaviours in the Quran (Kubbealti Lugati, Reference Lugati2020). The person who punishes, in the metaphor of haddini bildirmek, is implicitly seen as a moral authority over the deviating person; thus, the metaphor creates a hierarchical relation between the punisher and the person who incites the anger with his morally deviant actions.
6.2.2. cause of anger is trespassing the moral bounds
The concept of had is frequently used in Turkish in metaphorical idioms of anger suggesting that the cause of anger is trespassing. We found 33 instances of had idioms of anger and categorized them as conceptual metaphors of cause of anger is trespassing the moral bounds (Table 22).
cause of anger is trespassing the moral bounds
These metaphors of had in the corpus emphasize that the cause of anger (Stage 1) is violation of moral bounds. Anger is not only incited but also expected, since immoral people’s actions transgress the prescribed bounds and threaten the moral and social order. For example, the idiom haddini bilmemek (lit. ‘not to know one’s place/limits’) refers to ‘not behaving as required by one’s level, social position or status’ (12a), whereas haddini aşmak (‘to overstep/exceed one’s line/boundaries’) refers to a similar concept linked to violation of the rules which provide the appropriate bounds. Hadsiz may be translated literally as ‘without (moral) limits, impertinent’ (12b).
Finally, patience metaphors in the corpus were categorized as the cause of anger is trespassing. Here, the offending event exceeds the limits of human patience. Two patience metaphors of anger, sabrın sonu (‘end of patience’) and sabırlar tükenmiş (‘run out of patience’), are found. Although sabır is not frequent among the anger metaphors in the corpus, it is widely and only used to describe ‘us’, that is, as an identity marker and moral trait of ‘we-self’ (Table 23).
Domain of sabır in the corpus
6.2.3. Inscribed anger in the prototypical scenario
When we map the anger metaphors from our corpus onto the prototypical scenario of anger, most of the metaphors of inscribed anger conceptualize retributive acts and punishment (Stage 5. retribution, CN = 287) and the cause of anger as trespassing the moral bounds (Stage 1, CN = 42) in the prototypical scenario of inscribed anger (Table 24). The Stage 3 metaphors of agitation Footnote 15 were always negated (CN = 7), and Stage 4 metaphors occurred only 11 times, 5 of which were negated. In fact, inscribed anger fits neatly into the non-typical scenario of moral anger which is don’t get angry, get even where Stages 2–4 are avoided and instead the emoter passes directly to Stage 5. The moral dimension is emphasized with conceptual metaphors of trespassing moral bounds as the cause at Stage 1. Retribution and punishment for immoral acts by the legitimate authority is located at the centre in this non-prototypical scenario of moral anger.
Inscribed anger in prototypical scenario
7. Discussion
Our preliminary quantitative analysis of anger expressions in the corpus gave unexpected results in terms of the emotion display rules concerning the suppression of anger. Anger was expressed mainly as an emotion attributed to the self/selves rather than to others (78% to 22% in all anger expressions). However, a more detailed quantitative comparative analysis showed that rather than direct expressions of anger using emotional lexis (12%), figurative expressions were dominant in the cases of inscribed anger (88%). Ascribed anger expressions used significantly more direct emotion terms compared to the expressions of inscribed anger. Furthermore, for inscribed anger, the direct anger terms were used mainly in negations of feeling (95%), whereas ascribed anger was never negated when it was ascribed using direct anger lexemes.
This clear tendency to negate anger when using direct anger terms (43 out of 47 instances), and the dominant preference for figurative expressions for conveying inscribed anger (87% of all figurative expressions of anger in the corpus) suggest that Erdoğan deliberately abstains from expressing anger directly, probably under the influence of cultural display rules. Furthermore, the frequency of inscribed (non-negated) anger was meaningfully and substantially higher in the Post-Gezi corpus comparing to Pre-Gezi, showing that angry discourse indeed accompanied the rise of populism in Turkey after 2013.
The significant weight of figurative language for inscribed anger led us to analyse the conceptual metaphors of anger in the corpus and compare further possible differences between expressions of inscribed and ascribed anger. One important result of the ECMT analysis is that ascribed anger and inscribed anger were expressed in a way that creates two contrasting anger scenarios for us and others. These scenarios centralize the relationship between the subject and anger. Metaphors of ascribed anger are distributed between the physiological effects of the anger (Stage 2) and loss of control (Stage 4). Even metaphors mapped in Stage 2, which is about the bodily effects of anger, reflected how anger seizes control over physical abilities such as visual ability and – implicitly – mental ability (loss of judgement impedes one from seeing or acting appropriately). Overall, metaphors relating to loss of control and being captivated by anger constitute the majority of the metaphors used to express ascribed anger (79%). The discourse portrays others’ actions as irrational, unjust, excessive, and animal-like through the use of these specific anger metaphors.
The narrative-like portrayal of the other’s relations with anger also serve to depict others as in conflict with the prescribed moral/cultural rules that should regulate this relation. Others lose their self-control due to anger, and anger controls others’ selves, that is, it captivates their body and mind. The others-anger relation is metaphorically framed as a moral weakness in parallel with cultural rules about anger. Others’ anger is ‘frustrated anger’ which shows their incompetence because they cannot get past Stage 4 (i.e., they cannot take retributive action despite the intensity of their anger), a scenario which portrays them as ‘weak’ in power-relations in addition to their moral weakness.
On the other hand, the scenario of inscribed anger is almost the opposite in terms of subject–anger relations. Here, the subject (I/we/nation) never loses its self-control due to anger. As in the non-prototypical scenario of ‘cool anger’, there are no physiological effects on the self and the self remains in control. In fact, the mapping of metaphors of inscribed anger show that ‘we’ as the self, keep our control over the anger so that neither bodily effects, nor loss of judgement nor loss of control is experienced, and we continue to act appropriately as prescribed by cultural rules. All the metaphors for inscribed anger can be conceptualized in the general domain of control over anger is moral strength. The controlled anger of us is complemented by the identity markers of sabır which portray the ‘we-self’ as having the necessary moral strength against anger.
The retribution and punishment by the legitimate authority are located at the centre in this non-prototypical scenario of moral anger, which is similar to the retributive ‘don’t get angry, get even’ scenario where Stages 2–4 are bypassed and we proceed instead directly to Stage 5. In this scenario, moral strength is combined with competence as power to take retributive actions and punishment.
The inscribed anger scenario is not only compatible with the cultural rules which emphasize self-control over anger, but it is also in line with the rules that regulate how and where feeling and expressing anger is justified. With the metaphors of cause of anger is trespassing the moral bounds, we see that punishment and retribution are deserved by those who transgress the limits: wrong-doing threatens society by blurring the clear, prescribed, socially accepted boundaries between right and wrong. Inscribed anger is moralized and justified only in the non-typical scenario of ‘righteous indignation’, since the cause of anger is metaphorically framed as a moral offence against which anger is culturally not only acceptable but also expected. The ‘we-self’, with its cool anger, is expected to take retributive actions against transgressors and deviants who ‘are dangerous to society not only because they can lead others astray, but because they create new paths to traverse, thus blurring the clear, prescribed, socially accepted boundaries between right and wrong’ (Lakoff, Reference Lakoff1995, p. 188).
In conclusion, due to culture-specific differences, Turkish political discourse inclines toward implied, disguised forms of anger and instrumentalizes anger in culture-specific ways to overcome the dilemmatic discursive position of anger arising from the conflict between contradictory cultural display rules. This paper provides the first analysis of anger in contemporary Turkish political discourse, shedding light on the structure of this emotion in a lesser-known cultural context and permitting comparisons with studies on anger from other cultural and political contexts.
Data availability statement
All dataset related to the coding in this study are available at the Open Science Framework: https://osf.io/ep62b/?view_only=57015a8defe34c029fa3804d4e64c57b.
Competing interest
The authors declare none.
Funding statement
The research described in this article has received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No. 89631.
Open access
