2.2.1 Expressions

Expressions in the pragmatic study of language and politics include lexical items which are politically relevant, such as ‘nationalism’ and ‘democracy’, as well as expressions which become important in a particular political context, such as ones through which speech acts like apology are realised in political interaction.Footnote ¹ A related category of politically relevant expressions includes political metaphors, which have been widely studied in the field (e.g. Lakoff Reference Lakoff1995; Fauconnier and Turner Reference Fauconnier, Turner and Gibbs2008; Musolff Reference Musolff2021). In our model, the phenomenon of ‘expression’ also covers conventionalised multi-word sequences. Many such expressions indicate a specific speech act or an interactional move. In the realm of language and politics, many expressions are formulaic, with strongly conventionalised pragmatic use (see Coulmas Reference Coulmas1979). Such expressions often indicate the participants’ awareness of the broader ritual frame holding for a particular political situation. Because of this, in House and Kádár (Reference House and Kádár2021a) we defined them as ‘ritual frame indicating expressions’ (RFIEs). For example, as Bull et al. (Reference Bull, Fetzer and Kádár2020) pointed out, the expression ‘Mr Speaker’ addressed towards the Speaker of the House in the British Parliament indicates awareness of ritual rights and obligations in a situation in which a politician can only make an attack on another Member of the House by addressing the Speaker of the House (see more in Chapter 9).

Among the expression types outlined above, the study of the category of politically relevant expressions such as ‘nationalism’ can be particularly challenging. This is because, in the study of such expressions, one or more of the pitfalls outlined above are prone to influence one’s analysis. For example, in her study of the notion of ‘Bulgarian-ness’, Sotirova (Reference Sotirova2021) used an East–West dichotomy-like analysis to study issues such as why ‘Bulgarian-ness’ triggers negative emotions like anger and frustration. A key problem with an analysis like Sotirova’s relates to the previously mentioned issue that the outcome of the analysis is decided before the analysis itself is conducted: while Sotirova uses interesting data to prove her point, her study ignores the possibility that, for many individuals, the politically relevant notion of ‘Bulgarian-ness’ may actually be positive.

2.2.2 Speech Acts

When it comes to the study of speech acts in politically relevant data, we rely on a finite and minimalist speech act typology (Edmondson and House Reference Edmondson and House1981; Edmondson et al. Reference Edmondson, House and Kádár2023), illustrated by Figure 2.2.

Figure 2.2 Our speech act typology

Source: Edmondson and House Reference Edmondson and House1981; Edmondson et al. Reference Edmondson, House and Kádár2023

Having a finite typology of speech acts allows us to rigorously analyse and compare speech acts through which politicians and other actors in politics interact with one another. It also allows us to capture the interactional patterns and broader pragmatic conventions holding for any pragmatically relevant situation under investigation. Further, in written political language use, the above typology also allows us to break down written messages into speech acts. An advantage of a finite speech act typology like the present one is that it precludes inventing ‘political speech acts’ ad libitum (see more below). Another key advantage of relying on such a finite system of speech acts is that it also allows us to break down discourse into components.

Ever since Austin and Searle, the idea that speech act categories need to be finite has been present in pragmatics (Habermas Reference Habermas and McCarthy1979; Vanderveken Reference Vanderveken1990; Kissine Reference Kissine2013; Levinson Reference Levinson and Huang2017). The idea of finiteness precludes ‘discovering’ new and context-specific speech acts, such as ‘the speech act of persuasion’ in politics (e.g. Natsheh and Atawneh Reference Natsheh and Atawneh2021), and the more general notion of so-called ‘political speech acts’ (see e.g. Hepple Reference Hepple2003; Bolsover Reference Bolsover2018). Operating with speech act categories embedded in the realm of politics shuts the door on the replicability of our research (see House and Kádár Reference House and Kádár2021b). For example, while scholars like Saito (Reference Saito2015) distinguished the speech act of ‘political apology’, we believe that applying such a category is not useful because it precludes studying Apologise realisations in the political arena on a par with other apologies – a train of thought which we will elaborate in Chapter 6.

The speech acts we propose are such simple and basic constituents of linguistic behaviour that they can easily be replicated in the study of interaction across languages and data types (see a detailed discussion in House and Kádár Reference Juliane and Kádár2023). In annotating our data in terms of speech acts, we rely on the standard speech act–analytic procedure of inter-rater coding.

Following our bottom-up approach, we naturally do not assume that a particular political phenomenon is realised by one particular speech act. For example, in Chapter 4 we examine the political ritual of diplomatic mediation. While one could assume that – in the context of a diplomatic conflict which requires the involvement of a mediator – future-related Attitudinal speech acts like Request are frequented, we do not set out to investigate such an assumption. Instead of zeroing in on any speech act, we first examine frequent speech acts in our data and then interpret the outcomes of the analysis. When it comes to quantifying our data in speech act analyses, we annotate our corpus or corpora manually.

In the study of speech acts in freely co-constructed interaction, like in Chapter 8 where we analyse ideologically loaded clashes in public spaces, we break down interaction into moves and examine how speech acts in moves relate to one another, as Initiating, Satisfying, Countering and Contra-ing moves (see Edmondson Reference Edmondson1981).Footnote ² ‘Initiating’ refers to speech acts through which an exchange is started; ‘Satisfying’ includes speech acts through which an Initiating speech act is satisfied; ‘Countering’ points to speech acts through which an initiation is countered – that is, objected to but not entirely rejected; whereas if an utterance is turned down entirely it would be ‘contra-ed’ in our terminology. Let us explain how these categories can be operationalised with a simple example. In a politically relevant conflict like a clash between a protester and a public figure, a speech act through which a morally loaded verbal attack is initiated by the protester may be either ‘contra-ed’ with another speech act, or responded in an unrelated manner, which we will define as the lack of alignment below. In such a case, the responding speech act operates as a ‘re-Initiation’.

With the aid of our interactionally interpreted set of minimal speech act categories, which are formulated on the basis of corpus-based interaction projects (see Edmondson et al. Reference Edmondson, House and Kádár2023), we are able to distinguish interactional moves from speech acts. For example, we do not interpret ‘refusal’ as a speech act because we combine speech acts with interactional moves like ‘counter’, and so we are able to break down more complex interactional phenomena such as refusal into speech acts through which a Countering or Contra-ing move is realised.

Using our typology implies that, for us, speech acts represent utterance-level phenomena which do not span different turns in face-to-face interaction. Further, in written or scripted data such as the case of Apologise in political settings (see Chapter 6), we break down the phenomena under investigation into different speech acts.

An advantage of defining speech acts on the utterance level and approaching discourse by examining the relationship between speech acts is that this procedure allows us also to examine how an interactional goal is achieved through speech acts in subsequent moves. For example, in Vladimirou et al. (Reference Vladimirou, House and Kádár2021) we examined a social protest on Twitter through the lens of realisations of speech acts, in particular Complain. Examining the relationship between speech acts in Tweets allowed us to consider how the interactional behaviour of complaining escalated as the participants in the protest reflected on others’ Complains in their own Complain realisations.

As noted in Chapter 1, in the present book we capitalise all speech acts; for example, we use ‘Apologise’ instead of ‘apology’. When we talk about an interactional phenomenon like ‘complaining’ (versus the speech act Complain) we do not use capitals.

2.2.3 Discourse

Discourse is the highest unit in our analytic framework. The importance of discourse in the study of language and politics relates to the fact that our ultimate goal is to understand political language use as a discourse process, and we often apply smaller units of analysis like speech acts and expressions to break down this process into replicable components. Our definition of discourse is linguistic; that is, we interpret discourse as a unit larger than a sentence or an utterance. This is different from how ‘discourse’ is used in social theory and in many CDA inquiries, in which discourses – often in the plural – describe social engagement and language use unified by a certain goal or set of characteristics (e.g. Wodak Reference Wodak, Zienkowski, Östman and Verschueren2011). As Jaworski and Coupland (Reference Jaworski and Coupland1999: 6) argue,

Following the renowned debate between Widdowson (Reference Widdowson1995) and Fairclough (Reference Fairclough1996), we do not argue that discourse is simply synonymous with ‘text’ because it is a product of interaction often involving a complex participation framework in language and politics. For us, discourse is therefore a unit of analysis which can be best studied if we break it down into contextually embedded components and first analyse the contextually embedded use of these components, and then consider whether such uses reflect broader pragmatic conventions or not. It is also possible to have a systematic look at discourse by categorising context and role types in a corpus of discourse data (see e.g. Chapter 9).

It is also necessary here to consider types of discourse through which one can reflect on the nature of a politically relevant data type under investigation. Figure 2.3 overviews our system of types of discourse.

Figure 2.3 Types of discourse

Source: taken from House Reference House, House and Blum-Kulka1986: 179

Discourse in pragmatics always interrelates with a specific medium. In language and politics, medium is particularly relevant to the way in which a message is formulated. As an example, let us here refer to Tweets, which have played a significant role in recent political events worldwide (see, among many others, Breeze Reference Breeze2020; Scott Reference Scott2021). Furthermore, many key topics in the field of language and politics – which are to be covered in Part Two of this book – need to be considered through the medium of discourse (see the right-hand side of Figure 2.3). A particular issue which one is advised always to consider is whether a piece of political data is spontaneous or not. ‘[–marked]’ in Figure 2.3 refers to the fact that spoken data are non-premeditated due to their real-time synchronous nature, whereas ‘[+marked]’ indicates that many types of asynchronous written interaction are premeditated. The line pointing from ‘spoken’ to ‘[+marked]’ shows that some types of spoken data, such as institutional ritual interaction, often lack ad hoc character, at least to some degree, and are of a premeditated nature. Computer-mediated communication (CMC) represents an ‘in-between’ category, as some types of CMC are of a synchronous nature whereas others are of an asynchronous nature (an excellent overview is Susan Herring’s work, e.g. the edited volume Herring Reference Herring1996). As studies such as Smith et al. (Reference Smith, John and Sturgis2013) have shown, the (a)synchronous nature of politically relevant data has a significant influence on the dynamics of such data. What we define as ‘examples’ in Figure 2.3 represents discourse types.

As we have shown thus far, Figure 2.3 presents a basic typology of discourse types based on the pragmatic dynamics of the (a)synchronous nature and the related degree of premeditatedness of a data type. Another distinction we need to consider when it comes to the study of discourse in language and politics is whether a particular instance of discourse is created by a single producer or multiple producers (multiple producers of a single-source discourse like an anonymous news item may be invisible). This is different from the standard terminology in various areas of the humanities and social sciences like literature where a frequently used binary pair through which types of discourse are analysed is ‘dialogue’ and ‘monologue’ (see also studies like Musolff Reference Musolff2011 in the field of language and politics). We propose to use another terminology, to align with the strictly linguistic scope of our inquiry, namely single-source and multiple-source discourse. As House (Reference House, House and Blum-Kulka1986: 179, 180, references elided) argues,

In using the three analytic units outlined above, it is important to bear in mind what we have already emphasised: they are distinct but always interrelated. For example, as already noted, in order to analyse discourse from both qualitative and quantitative points of view, it is preferable for the pragmatician to break manifestations of discourse down into expressions and speech acts, and also to consider expressions through which, for example, the broader ritual frame of an institutionalised political interaction is indicated. Expressions often indicate speech acts as well, they are discursively embedded, and they also have an intrinsic relationship with the context of the interaction. Finally, speech acts are never ‘stand-alone’ or isolated phenomena: similar to expressions, they are discursively embedded, and they are often realised through conventionalised formulae.

2.2.4 Context

As Figure 2.1 shows, we interpret our units of analysis in context, which often imposes a ritual frame onto the participants in the realm of language and politics.

In the tradition of pragmatics, conceptualisations of context have played such an important role that the very definition of pragmatics is often bound up with the notion of context. Thus Stalnaker (Reference Stalnaker1999: 43) writes, ‘Syntax studies sentences, semantics studies propositions. Pragmatics is the study of linguistic acts and the contexts in which they are performed.’ And we might even say, with Levinson (Reference Levinson1983: 32), that pragmatics is ‘a theory of language understanding that takes context into account’. The underlying assumption here is that in order to arrive at an adequate theory of the relation between linguistic units and what they express, one must consider the context in which these units are realised. In pragmatics, attention is given to how the interaction of context and content can be represented, and how the linguistic realisations used relate to context. The relationship between content and context is, however, never a one-way street: content expressed also influences context; that is, linguistic actions influence the context in which they are performed. The effects of this dependency are omnipresent and decisive for the construction and recovery of meaning. But context also plays a role in the overall organisation of language, affecting its syntactic, semantic, lexical and phonological structure to the point that, as Ochs (Reference Ochs, Ochs and Schieffelin1979: 5) puts it, ‘we could say that a universal design feature of language is that it is context-sensitive’. Any pragmatic framework of language and politics, then, needs to include a general representation of contextual features that determine the values of our pragmatic units, with context being represented by a body of information presumed to be available to the participants in the interactional situation.

Studying context in the realm of language and politics implies that the researcher needs to consider as much information about the historical events and the participants as possible. For example, in Chapter 5 we examine a corpus of colonial-era documents. For us, these documents represent the unit of discourse, and in order to analyse their contents by breaking them down to sequences of speech acts, we consider as much contextual information as possible, including information on the participants, the diplomatic event which brought these documents alive, and so on. Such information not only helps us to consider patterns of language use in our data, but also allows us to draw inferences regarding the replicable pragmatic features of our results, for example by considering how the pragmatic patterns emerging from our data represent the aggressive language use of the coloniser. We do not, however, assume at the very outset that the data we study constitute a particular historically embedded genre or discourse(s). This latter approach is preferred in CDA, in particular in its discourse-historical approach (for a summary see Reisigl and Wodak Reference Reisigl, Wodak, Wodak and Meyer2001). The discourse-historical approach argues that a text always belongs to something much broader, and this broader phenomenon defines the features of discourse(s) in an intertextual and interdiscursive way. As Flowerdew (Reference Flowerdew, Flowerdew and Richardson2018: 166, original emphasis) argues,

Studying intertextuality and interdiscursivity allows the CDA scholar to capture broader social discourses, which influence the formation of a particular text. For instance, Flowerdew (Reference Flowerdew, Flowerdew and Richardson2018) illustrates the use of context by interpreting particular features of a data set he studies through abstract intertextual and interdiscursive notions, such as ‘Western conceptions of democracy’ (Flowerdew Reference Flowerdew, Flowerdew and Richardson2018: 172). However, in a language-anchored pragmatic investigation like ours it is problematic to interpret features of a particular data set through a pre-existing extratextual notion like ‘Western conceptions of democracy’ without conducting a data-driven contrastive analysis of relevant discourse corpora. Intertextuality therefore may only be used in our analysis if it evidently emerges from individual discourse units in our corpus, while we do not consider interdiscursivity at all because it assumes that an inquiry operates with corpus-external phenomena. Context in our pragmatic framework therefore only includes historical information relating to the particular interaction under investigation, which helps us to interpret what is said in our data in the here and now.

2.2.5 Ritual Frame

Along with context, the frame in Figure 2.1 also includes the notion ‘ritual frame’. ‘Ritual frame’ refers to the fact that many instances of political language use, in particular institutionalised ones, are ritual in nature (see more in Chapter 9). Following Goffman (Reference Goffman1967) and Kádár’s (Reference Kádár2017; Reference Kádár2024) pragmatic adaptation of Goffman’s view, we argue that political language tends to represent a form of interaction ritual where pragmatic conventions follow underlying rights and obligations, indicate an invisible ritual frame, are often ostensible in nature and embody a moral order. Furthermore, political language use is also ritual in a more ceremonial sense in settings such as parliaments, where ritual in the very literal sense of the word is frequent. The more institutionalised a political interaction is, the more visible the ritual frame becomes; for example, in a televised presidential debate the pragmatic rights and obligations of the participants are regulated by rules and by a mediator who can enforce such rules. While in less institutionalised political settings the ritual rights and obligations are not set in stone, and political actors often have means to follow strategic agendas, any form of strategic behaviour often needs to be hidden under a veneer of ritual conventions (see e.g. Chapter 5).

Unlike intertextuality and interdiscursivity, the features of a ritual frame cannot be taken for granted but should rather be examined in a bottom-up way across data types. Let us take diplomatic mediation as an example, which is to be studied in Chapter 4. While we assume that a frame is in place in diplomatic mediatory situations right at the outset of our research, we have no way to know how it actually influences discursive behaviour before conducting a bottom-up examination of our data.

The term ‘frame’, which we use in the context of the ritual nature of political language use, is also frequently used in linguistics, as witnessed by high-impact cognitive research such as Schank and Abelson (Reference Schank and Abelson1977), Tannen (Reference Tannen and Friedl1979), Fillmore (Reference Fillmore1982), Barsalou (Reference Barsalou, Lehrer and Feder Kittay1992), Chafe (Reference Chafe1994), Bednarek (Reference Bednarek2005) and so on. What distinguishes our use of the term ‘ritual frame’ from other interpretations of ‘frame’ is our predominantly bottom-up focus, which traces a path from linguaculturally embedded units of analysis to the underlying cognitive frame. This path is rarely discussed outside pragmatic inquiries. Using the concept of frame implies that we look at conventionalised and ritual aspects of language use and we exclude idiosyncratic behaviour from the scope of our analysis.

2.2.6 The Right-Hand Side of Figure 2.1

In Figure 2.1, the right-hand side includes contrastive approaches (see also Chapter 1). As already noted, these contrastive views are rooted in cross-cultural pragmatic research. Contrasting can be done in many different ways, and what we advocate in this book only represents certain types of contrasting, without the claim of being comprehensive.

3.3.1 ‘Nationalism’ in CDA Research

In order to position our case study, let us briefly summarise previous CDA research involving the concept of nationalism. In a body of such inquiries, cross-cultural research has been used to validate dominant ‘Western’ conceptualisations of ‘nationalism’ as an exclusively negative phenomenon. For instance, KhosraviNik and Zia (Reference KhosraviNik and Zia2014) adopted a CDA approach to study how a form of Iranian national(ist) identity is constructed and presented on the popular Facebook page ‘Persian Gulf’. While KhosraviNik and Zia used Persian data, they interpreted Iranian ‘nationalism’ without considering native conceptualisations of this notion. Others, such as Setiyadi et al. (Reference Setiyadi, Hersulastuti and Widayanti2018), examined linguacultural equivalents of the term ‘nationalism’, and they approached these equivalents in a top-down manner, assuming that ‘nationalism’ is an essentially negative phenomenon in the linguacultures studied. By so doing, they unavoidably fell into Pitfall 1: notwithstanding whether or not a scholar is a linguacultural insider of Persian or other linguacultures involved in the inquiries outlined above, the uncritical use of a Western definition of ‘nationalism’ shuts the door on any objective study of this subject.

The work of Ruth Wodak and her colleagues deserves particular attention here. Wodak (Reference Wodak, Flowerdew and Richardson2017) defined ‘nationalism’ using cultural and ethnic criteria, by proposing a discourse-historical approach to this notion (see also Chapter 2). Further, Wodak and Boukala (Reference Wodak and Boukala2015) explored the diachronic patterns in the emergence of ‘nationalism’ across various European countries, providing a model for the historical development of nationalism. While we agree with Wodak that examining the historical development of nationalism (and its equivalents) provides a key to understanding ‘nationalism’ as a concept, our approach is very different from it because we do not start from the assumption that equivalents of nationalism always have a negative default connotation across different linguacultures.

3.3.2 Methodology and Data

Following the tripartite design proposed in Figure 3.1, in our case study we proceed as follows:

1. 1 Contrastive analysis 1. We study the diachronic development of M/minzu-zhuyi and nationalism in China and the US respectively, by examining the ways in which politicians used these expressions over time. This research revealed major cross-cultural contrastive pragmatic differences between the use of M/minzu-zhuyi and nationalism. The results also revealed a certain rupture which followed the end of early modern Chinese politics, coinciding with the founding of the People’s Republic of China in 1949.

2. 2 Contrastive analysis 2. In order to investigate whether this rupture holds for language use beyond speeches made by politicians, we explore the ways in which M/minzu-zhuyi and nationalism have been used in the Chinese and US print media between 1949 and the present day.

3. 3 Testing the results of the contrastive analyses. We conduct a set of semi-structured interviews with Chinese and US respondents of two different age groups in order to measure – at least in a modest way – how present-day language users in China and the US perceive the meanings of M/minzu-zhuyi and nationalism.

Adopting this approach, we examined the following three corpora:

1. 1 Corpus for contrastive analysis 1. Transcripts of speeches made by politicians between the 1920s and 2019. The Chinese corpus consists of 951,887 Chinese characters, while the US corpus has 978,768 words. The Chinese corpus was collected from twelve Chinese websites. The US corpus was retrieved from millercenter.org/the-presidency/presidential-speeches.

2. 2 Corpus for contrastive analysis 2. Newspaper articles published between 1949 and 2019. The Chinese corpus of newspaper articles consists of approximately 630,000 characters, while the US corpus has 205,500 words. The Chinese newspapers studied include China Daily and Global Times, while the US newspapers include the Washington Post and the New York Times. We studied data drawn from these newspapers because, internationally, they are among the best-known newspapers published in the two countries involved in our case study.

3. 3 Testing the results of the contrastive analyses – interviews. Our interview corpus includes two lots of thirty interviews with Chinese and US respondents from two different age groups: under and over thirty-five years. We chose thirty-five years in a relatively arbitrary manner, arguing that by reaching this age people have usually settled down, and so after this age interviewees may have a different view on politics than younger respondents. We assumed that the age factor is a crucial determinant when it comes to evaluations of the expressions M/minzu-zhuyi and nationalism. Due to the effect of globalisation and related exposure to international media, younger people have a different view of M/minzu-zhuyi and nationalism than members of the older generation. This assumption was based on previous research on the pragmatics of globalisation, such as Sifianou (Reference Sifianou2013) and House and Kádár (Reference House and Kádár2020).

When constructing our corpora, we used the computer software AntConc v. 3.5.8 to search for occurrences of M/minzu-zhuyi and nationalism. We examined not only occurrences of these expressions, but also their collocations with adjectives: up to three words before and after the occurrence of nationalism and, similarly, up to three words before but not after the occurrence of M/minzu-zhuyi because Chinese adjectives only occur before nouns. We examined adjectives to determine whether they tended to have a positive, negative or neutral meaning in the linguacultures being contrasted. As part of this analysis we also examined the broader contextual use of each example, to determine whether the expressions under investigation were used positively or negatively in a particular text. Typical examples of positive and negative adjectives modifying M/minzu-zhuyi in our Chinese corpus are the following: fandiguo-de-Minzu-zhuyi 反帝国主义的民族主义(‘anti-imperialist Minzu-zhuyi’) [positive meaning] and youyi-de-minzu-zhuyi 右翼的民族主义 (‘right-wing minzu-zhuyi’) [negative meaning]. In the case of our US corpus, while we could not identify positive adjectives modifying nationalism, we found various adjectives such as economic, which we initially classified as ‘neutral’. However, as our more detailed analysis later revealed, in many cases such semantically neutral adjectives also gain a negative meaning in US political language use. A typical adjective negatively modifying nationalism is nastiest.

In our semi-structured interviews, we asked three questions in English and Chinese (see more below). When selecting our respondents, we included people with a university education background who said ‘yes’ to our question whether they are interested in politics and frequently read or watch different media. The interviews were conducted via email and Skype or WeChat in two steps. First, we provided our respondents with the interview questions by email or WeChat and then arranged the Skype or WeChat interviews. The average interview lasted approximately ten minutes. The interview data were stored in compliance with the standard ethical-research requirements of linguistic inquiries. The participants’ names were anonymised and any sensitive information was removed from the interview transcripts.

3.3.3 Analysis

The following Tables 3.1, 3.2 and 3.3 summarise the quantitative analysis of parts 1 and 2 of our research. They include the number of occurrences of M/minzu-zhuyi and nationalism in our diachronic corpora:

The figures given in Tables 3.1, 3.2 and 3.3 clearly show major cross-cultural contrastive differences between the Chinese and US expressions under investigation – we will revisit this point in more detail later. More importantly here, they also point to the existence of a diachronic rupture in the Chinese data. While Table 3.1 indicates that M/minzu-zhuyi was almost always used positively in pre-1949 political speeches, it also shows that, after 1949, Chinese politicians used it only negatively. This is in accordance with the fact that, in pre-1949 Chinese political speeches, M/minzu-zhuyi had a dual use: it was used as a proper noun in a positive sense, and (rarely) as a common noun in a negative sense.

After 1949, Chinese politicians only used minzu-zhuyi as a common noun (see more below), in a negative sense. Table 3.2 (in conjunction with Table 3.1) shows that this rupture is less valid for post-1949 media texts, in which M/minzu-zhuyi can be used in both negative and positive ways, although the negative uses are much more marked than the positive uses. Table 3.2 (in conjunction with Table 3.1) also indicates that it is not always clear in post-1949 media texts whether M/minzu-zhuyi is a proper or common noun when used positively. That is, unlike the situation in pre-1949 political speeches, in post-1949 media texts there is a sense of ambiguity surrounding the modern use of this expression. Table 3.3 confirms this ambiguity: while in our corpus of political texts Minzu-zhuyi as a proper noun is never modified by an adjective, in the media texts one can observe cases where M/minzu-zhuyi has been modified by an adjective and expresses a positive meaning. Table 3.3 also confirms that negative uses of M/minzu-zhuyi are substantially more frequent than positive uses. It is worth noting that the ambiguity regarding whether M/minzu-zhuyi is a proper or common noun stems from the nature of Chinese writing. Unlike with many other writing systems, when using Chinese characters, one cannot indicate whether a word is a proper or common noun, and so the status of a noun can only be discerned from its context. While we attempted to resolve this ambiguity by carefully analysing the contextual use of M/minzu-zhuyi in each example in our corpus, and also by validating our interpretation with the aid of a panel of native speakers, in certain cases the ambiguity could not be completely resolved.

In the following, we first conduct a qualitative interpretation of the above quantitative results and then check the validity of the findings by implementing the third interview-based step in our model.

4.3.1 Background

Mediatory negotiation as a form of preventive diplomacy has been studied from many angles, such as international relations (e.g. Bercovitch and Rubin Reference Bercovitch and Rubin1994; Bercovitch and Gartner Reference Bercovitch and Gartner2007), conflict (e.g. Wallensteen and Svensson Reference Wallensteen and Svensson2014; Beardsley et al. Reference Beardsley, Cunningham and White2018), history (e.g. Gürkan 2015; Van Gelder and Krstić Reference van Gelder and Krstić2015; Ciftci Reference Ciftci2022) and rhetoric and argumentation (e.g. Murau Reference Murau2012). Some scholars also examined the phenomenon of imbalanced and partial mediation: Wehr and Lederach (Reference Wehr and Lederach1991), Tome (Reference Tome1992), Svensson and Lindgen (Reference Svensson and Lindgen2013), Blakemore (Reference Blakemore2019) and others explored the role of partial mediator in conflict management.Footnote ¹ Covertly imbalanced diplomatic mediation – where the conflicting sides did not agree that the mediator can be imbalanced – has been studied by Siniver (Reference Siniver2022) and Awwad (Reference Awwad2023), who considered problems caused by partial mediators in Middle Eastern conflicts. Further, Rosoux (Reference Rosoux2022) argued that certain seemingly ‘democratic’ mediation attempts in postcolonial conflicts are doomed to fail due to the mediator’s inherently biased position. This latter body of inquiries is particularly relevant for both our case studies because we also consider language use in an event where the mediator was supposed to remain ‘neutral’.

Research on back-door diplomatic negotiations is clearly relevant for our study (for an overview see De Lange Reference Lange and Lange2010). In language and politics, negotiation and mediation have been studied in three different areas:

1. 1 various scholars have examined mediation in the context of translation where the interpreter/translator of a diplomatic text is referred to as a ‘mediator’ (e.g. Ayyad Reference Ayyad2012; Pérez-González Reference Pérez-González2012; Bendazzoli Reference Bendazzoli2023);

2. 2 others have considered the role of institutionalised media in conflict resolution (see Gilboa Reference Gilboa2001; Kampf Reference Kampf and Tracy2015; Friedman Reference Friedman2017; Friedman et al. Reference Friedman, Kampf and Balmas2017);

3. 3 a few scholars have examined the actual face-to-face language use of the mediator in mediation and conflict resolution (e.g. Smith Reference Smith2012; Barbé et al. Reference Barbé, Herranz-Surrallés and Natorski2015).

For us, the third area above is of particular relevance. We hope to contribute to previous research in the field by conducting a pragmatic investigation of what happened to the powerless party during a closed-door diplomatic mediation session, and also by considering ex post facto reflections of this powerless side.

Research on mediation in legal contexts (e.g. Maley Reference Maley1995; Jacobs Reference Jacobs2002) and ritual conflict resolution (e.g. Ran and Zhao Reference Ran and Zhao2018) also bears relevance for our study. Such research has focused on how the pragmatic conventions through which mediation is realised are constrained by a ritual frame, which cannot be violated due to the institutionalised nature of mediation. Since in our first case such a violation was kept under the veneer of diplomatic ‘civility’, it is interesting to consider how it was interpreted by the representatives of the powerless side.

Finally, historical research, like the studies conducted by Lyddon (Reference Lyddon1996) and Glaurdic (Reference Glaurdic2013), provides a detailed outline of the diplomatic strategies of the EEC in the historical events discussed in our case studies.

4.3.2 Data

The data in our first case study consist of the transcript of a tape recording of a mediation event between EEC representatives and the newly established Slovenian and Croatian states. Following the Slovenian and Croatian declaration of independence on 25 June 1991, the Yugoslav People’s Army launched an attack on Slovenia. The EEC intervened partly because they had a vested interest in Yugoslavia’s remaining intact and politically and economically stable (Repe Reference Repe2002), and partly out of fear that similar independence movements might occur elsewhere in Europe. On 28 June 1991, after reaching an agreement with Yugoslavian state representatives that they will convince the Slovenian and Croatian representatives to temporarily suspend further steps towards independence, the EEC representatives met with the president of the Presidency of the Republic of Slovenia, Milan Kučan; the Slovenian foreign minister, Dimitrij Rupel; the president of the Republic of Croatia, Franjo Tuđman; and Croatia’s representative in the Presidency of Yugoslavia, Stjepan Mesić. Kučan was persuaded to agree to recommend to the Slovenian assembly suspending the independence process for at least three months in the form of the so-called ‘Brioni declaration’.

The data in our second case study are a transcript of another tape recording taking place on 8 July 1991, between Kučan and other Slovenian representatives, some of whom were not present in the diplomatic mediation session. The participants in this second meeting faced the following dilemma: implementing the agreement by having the Brioni declaration ratified by the Slovenian parliament meant that Slovenia would militarily put itself at the mercy of Yugoslavia, with no guarantee from the EEC that they would intervene if the Yugoslav state backtracked on what they had promised, blocking Slovenian attempts to reach independence. However, were the Slovenians not to implement the agreement and the declaration, the Yugoslav state would have most likely declared war on Slovenia, and Slovenia would consequently have lost all support from the EEC. In this second meeting, the Slovenian leadership regarded what was said during the first meeting as a ‘dictate’. Still, they convinced the Slovenian parliament to ratify the Brioni declaration and suspend the independence process temporarily (for three months) a few days later (Repe Reference Repe2004: 146–148).

The transcript of the first meeting under investigation consists of approximately 7,000 words, and the second data set consists of approximately 14,500 words. As far as we are aware, the original tape recordings of these events are not available, but owing to Repe’s (Reference Repe2004) comprehensive historical work the transcripts of the events are available in Slovenian. While the first data set included speakers of other languages as well, what was said by the EEC representatives is only available in a Slovenian translation (with elements of Serbo-Croat), provided by an interpreter who was present in the meeting.

4.3.3 Methodology

We interpret diplomatic mediatory negotiation as a ritual. As outlined in Chapter 2, following Goffman (Reference Goffman1967) and Kádár’s (Reference Kádár2017; Reference Kádár2024) pragmatic adaptation of Goffman’s view, we argue that in ritual situations – which involve both ceremonies and as seemingly more ad hoc interactions – pragmatic conventions follow underlying rights and obligations and a related moral and interactional order. In the context of mediatory negotiation, while such rights and obligations are not set in stone and both the mediator and the conflicting parties have significant leeway to pressurise the others, any form of pressurising needs to be hidden under a veneer of civility (House et al. Reference House, Kádár, Liu and Song2023). Pragmatic constraints are particularly strict for the mediator: provided he has an outsider and non-partial footing (i.e. takes such a participant role in an interaction; see Goffman Reference Goffman1979), he is bound by the ritual frame to sound ‘neutral’, while the other side is not bound by such a pragmatic constraint; that is, the parties in conflict can afford to sound ‘subjective’. Setting out from this basic difference, which is in the DNA of the ritual, the contrastive approach is particularly useful in the analysis of our first case study: we aim to capture recurrent patterns in the language use of the EEC representatives and the Slovenian and Croatian representatives and contrast these patterns. In our second case study, we consider how the Slovenian representatives reflected on what happened during the mediatory session once they were not constrained by its ritual frame.

We break down our data in both our first and second data sets into speech acts, relying on our finite speech act typology. To help the reader to follow our analysis below, here we present our typology again (see Figure 4.2 below).

Figure 4.2 Our speech act typology

Following our bottom-up take, we do not assume that the ritual phenomenon of diplomatic mediation or the ex post facto debriefing is realised by one particular speech act. Instead of zeroing in on any speech act, we first examine frequent speech acts in our data from both qualitative and quantitative points of view, and then interpret our outcomes. In quantifying our data, we annotated our corpus manually, and so broader speech act annotation issues relevant to the study of large corpora are not relevant to the current investigation (see here Weisser Reference Weisser, Aijmer and Rühlemann2014). To keep our manual annotation rigorous, we used sentences rather than utterances as a unit to be categorised as a speech act.

In both our analyses – in particular in the first case study – we refrain from using the concepts of ‘politeness’ and ‘impoliteness’, and instead use the notion of ‘covert aggression’ (e.g. House et al. Reference House, Kádár, Liu and Song2023) to refer to cases where significant pressure is exerted on the other in a diplomatic exchange. In diplomatic exchanges, one can rarely observe fully fledged ‘politeness’ and ‘impoliteness’ in Brown and Levinson’s (Reference Brown and Levinson1987) sense simply because the participants are bound to follow the ritual conventions holding for the situation, so attempts to pressurise the other normally represent a covert form of ‘civil’ aggression rather than impoliteness. The notion of ‘covert aggression’ therefore includes instances of language use which have an aggressive overtone but which are not explicitly aggressive and rude, and which do not afford the other’s making the speaker accountable.

4.3.4 Case Study 1

Table 4.1 summarises recurrent speech acts in our first data set. As Table 4.1 shows, by far the most frequent speech acts in data set 1 are the Informative speech acts Tell and Opine. While this tendency accords with what we previously observed about the language of diplomatic conflicts in general, it is noteworthy that there are more Opines than Tells in the utterances of the Slovenian and Croatian representatives. Other salient tendencies are the following:

• the EEC representatives realise more Requests than the Slovenian and Croatian representatives;

• Willings are nearly exclusively realised by the Slovenian and Croatian representatives.

In the following, we interpret these tendencies from the point of view of the power imbalance which they reflect in the seemingly ‘impartial’ procedure of diplomatic mediation in our case.

4.3.5 Case Study 2

In the following, we focus on those parts of our second data set which feature a meeting between Slovenian decision makers who reflect on what happened during the diplomatic mediatory meeting. Such ex post facto reflections are realised by the speech acts given in Table 4.4 below.

4.3.6 Reflections

In this section, we have considered how to avoid Pitfall 2 with the aid of a bipartite analysis, featuring two closely related case studies. We investigated whether there is a tension between the ideally ‘neutral’ language of diplomatic negotiation during mediation and the presence of a powerful actor who takes up the role of the mediator, with the powerless side facing defeat from the start. We studied this tension by using a speech act-anchored approach, allowing us to precisely capture differences between the behaviour of the powerful and the powerless sides in our first data set, and to interpret ex post facto reflections by the powerless side in our second data set. Figure 4.3 summarises our findings from a contrastive pragmatic point of view.

Figure 4.3 Outcomes of our analysis

As Figure 4.3 shows, while only a small variety of speech acts are frequented in both our data sets, looking at the realisation patterns of these speech acts has still allowed us to tease out the power dynamics underlying the first meeting and reflections about this during the second meeting.

Our findings suggest that although diplomatic language use on the part of a mediator – like the EEC representatives in our case – is ideally ‘neutral’, such ‘neutrality’ is often covertly violated. While various scholars, such as Rosoux (Reference Rosoux2022), have noted that diplomatic negotiation is often biased, here we have been able to capture such a bias from a speech act point of view. Furthermore, we have found that representatives of the powerless side are often clearly aware of such violations. The powerless side may also reinterpret their own behaviour after the diplomatic event, in order to save face. For example, we were able to witness various cases when negatively reflecting on the aggressive behaviour of the powerful side also prompted representatives of the powerless side to reinterpret very basic facts, such as whether the agreement was signed or not.

Through these case studies, we are able to capture how diplomatic mediatory negotiation operates when there is a salient power difference between the participants, which is often the case in such settings. This outcome led to a replicable description of this context, but not so much of the actors: while in our case studies the representatives of the EEC behaved in a controversial – not to say arrogant – way, we do not think that this behaviour should lead to any grandiose value judgement about the ‘democratic values’ of the EEC beyond the actual case.

5.3.1 Key Concepts

We use the collective term diplomatic note to refer to a note deployed in national conflicts of interest. In modern diplomacy, this subgenre of diplomatic notes is often known as démarche (zhaohui 照會 in historical Chinese). In diplomatic notes, aggression is veiled through the operation of a conventionalised ritual frame (Kádár and House Reference Kádár and House2019): as already pointed out, covert aggression in diplomacy always occurs under a veneer of civil diplomatic language. In our present case study, the intensity of veiled aggression correlates with the power of the coloniser over the colonised.

In diplomacy, especially exchanges of written discourse studied here, interaction often takes place in the form of genres with strict conventionalised ritual features (see Kádár Reference Kádár2017; Reference Kádár2024), including

• the intensive use of expressions of deference, such as ceremonial forms of address;

• the operation of complex participation and ratification in Goffman’s (Reference Goffman1967; Reference Goffman1974) sense: while a diplomat may exchange seemingly ‘personal’ remarks with the recipient, ultimately he corresponds as a representative of a country rather than as an individual;

• the operation of a ritual frame, which manifests itself in participants’ rights and obligations (see also our previous case studies).

In diplomatic notes, aggression tends to be realised in the form of veiled threats; that is, aggressive behaviour hidden behind conventional diplomatic civility with the goal of coercing representatives of the other state to do what the aggressor wants. As Culpeper et al. (Reference Culpeper, Bousfield and Wichmann2002: 1572) argue, veiling threats is usually ‘insincere’, provided that the goal of the language user is to realise a real threat. Considering that the veneer of diplomatic ‘civility’ precludes being explicitly impolite while at the same time affording aggression, in the current study we use the term ‘covert aggression’, similar to our previous case study. The aggressive veneer in written diplomatic language studied in our present case study resembles what Watts (Reference Watts1999) defined as the ‘iron fist in a velvet glove’. The ‘velvet glove’ implies that the aggressor uses insincere ‘politeness’ to frame his aggression as ‘civil’, but also allows him to morally legitimise the aggression (see Kádár et al. Reference Kádár, Parvaresh and Ning2019).

The ritual characteristics of diplomatic notes mentioned above manifest themselves in the operation of a ritual frame, imposing a certain sense of ritual ‘constraint’ (Goffman Reference Goffman1967) on aggressive language use: what a diplomat can and cannot do is tightly regulated, and it is practically impossible to cross a certain invisible conventionalised threshold in a ritual diplomatic genre.

5.3.2 Positioning the Case Study

The genre of diplomatic notes has received limited attention in pragmatics, and it has been mostly examined by historians and political scientists (e.g. Fendrick Reference Fendrick2012; Beres Reference Beres2015). A major contribution to this area involves research on ‘coercive diplomacy’, such as George (Reference George1991), Jakobsen (Reference Jakobsen2011) and Schettino (Reference Schettino2011). Our study relates to this latter body of research: we examine a case in which a colonising state coerces a state to be colonised into ‘coping’ with the claimed needs of the so-called ‘international community’. Such claimed needs involve the coloniser’s demand to be allowed to represent himself in the capital of the colonised and establish free trade, only benefitting the coloniser. From a contemporary point of view, enforcing such demands may be more than sheer ‘diplomacy’, in that such coercive behaviour violates what we today understand as normal international diplomacy. However, at the time of colonialism many Western countries officially claimed that they were entitled to colonise ‘underdeveloped’ nations, and when making threats their representatives often argued that they were simply enforcing their countries’ perceived diplomatic ‘rights’ (see Fitzmaurice Reference Fitzmaurice2003).

Chinese linguists have devoted particular attention to diplomatic notes, due to the humiliating role such notes played in the nineteenth-century colonisation of the country. For example, various scholars have used such notes to explore diplomatic issues surrounding translation problems in late imperial China (see e.g. Qu Reference Qu2017). Another interesting line of inquiry in this area is represented by the study of Ding and Mao (Reference Ding and Mao2000), who investigated the diachronic development of diplomatic documents in China and explored the diplomatic issues which the archaic style inherited from ancient China triggered during nineteenth-century Chinese diplomatic encounters with Western nations. Other scholars, such as Guo (Reference Guo2003), explored how diplomatic notes evolved in late imperial China under the influence of exchanges with Westerners. Guan (Reference Guan2017), on the other hand, explored Chinese influence on British diplomatic notes written to the Chinese government. Our present case study fills a knowledge gap by approaching nineteenth-century diplomatic notes between China and a Western power, the US, and by focusing on the ways in which an American diplomat realised aggression through veiled threats. While we mostly focus on covert aggression in our US corpus because we are interested in the behaviour of the coloniser (see Figure 3.4), we also briefly examine Chinese diplomatic responses to the US diplomatic notes to enrich our analysis from a cross-cultural pragmatic point of view.

The context of colonialism provides insights into the relationship between aggressive threats and aggression (see Kádár et al. Reference Kádár, Parvaresh and Ning2019). In our case study, the coloniser who has the military might to cause significant harm and loss of face to the other practically always embeds threats in his realisation of diplomatic ‘politeness’ rather than impoliteness proper. In this respect, the data under investigation are very different from what (im)politeness and impoliteness-anchored research studied in the realm of language and politics. That is, while a small body of studies touched on aggressive language use in diplomacy (e.g. Swain Reference Swain2015) and other political settings (see e.g. Bull Reference Bull2013), such analyses mostly focused on impoliteness rather than on the role of politeness as a veneer of civility – our phenomenon of covert aggression. Also, while ritual has received considerable attention in pragmatic research on language and politics (see e.g. Chilton Reference Chilton1990; Kampf Reference Kampf, Aldar, Danziger and Schreiber2019), little work has been done on ritual as an interactional frame facilitating and condoning aggressive political exchanges under the polite veneer discussed above.

Furthermore, while a number of scholars have studied the role of aggression in institutional language use (see e.g. Harris Reference Harris2001; Grainger Reference Grainger2002; Archer Reference Archer, Bousfield and Locher2008), little attention has been devoted to the institution of diplomacy. Also, in terms of the mode of realising aggression, pragmaticians frequently examine spoken and computer-mediated modes of interaction (see e.g. Vladimirou et al. Reference Vladimirou, House and Kádár2021), while the type of written interaction which we analyse in our case study has been backgrounded.

5.3.3 Methodology

In the present case study, we investigate the following research question: ‘How is aggression realised in the ritual genre of diplomatic notes?’ The rationale behind asking this question is that one may assume that diplomatic communication is blandly professional, avoiding aggression and conflict, considering that diplomats are responsible for upholding international relationships.

As elsewhere, we follow the logic of Karl Popper’s empirical research methodology, which we already discussed in detail in Chapter 2. An advantage of such an empirical procedure is that it often leads us to falsify our own assumptions. For example, when conducting the present case study we had an initial assumption, namely that aggression in diplomatic notes is essentially realised through upgraded Requests (see Blum-Kulka et al. Reference Blum-Kulka, House and Kasper1989). This was a logical assumption because, in a context where a more powerful party wants the powerless side to do something, the speech act Request is a ‘natural’ choice. This is all the more because Request can be realised on a wide scale of (in)directness. Notwithstanding our assumption, we decided to explore our corpus without any pre-categorisation, by basing our work on our typology of speech acts. Our examination has shown that the speech act Tell is by far the most frequent speech act in our corpus; that is, our initial assumption turned out to be completely wrong.

As already noted, in our speech act typology Tell is an Informative illocution, which is defined as follows:

Our category Tell is close to what Searle (Reference Searle1979) in his classic study defines as ‘Representatives (Assertives)’; that is, a speech act category consisting of speech acts committing the speaker to the truth of his proposition. The reason why we use the category of Tell rather than ‘Representatives (Assertives)’ is that we pursue interest in the interactional features of Tell following our typology of speech acts and related analytic system; that is, we do not assume that Tell is always informative. More specifically, we focus on cases when Tell is not about committing the speaker to the truth, but rather serves other pragmatic functions.

Our analysis will reveal that in the US corpus (see below) the coloniser’s aggression is overwhelmingly realised through the Informative speech act of Tell, which not only changes its default Informative function as a means of covert aggression, but also is often intertwined with other speech acts, such as Request and Complain. By the ‘intertwinedness’ of speech acts we mean that the speech acts under investigation in our diplomatic data cannot always be rigorously disentangled. This has a significant implication for the understanding of covert aggression in the ritual genre of diplomatic notes: whenever diplomats have the power to act as aggressors and realise threats, they often ‘package’ aggression in the insincerely ‘harmless’ informative speech act Tell.

5.3.4 The Historical Background of Our Corpus

In 1842, the First Opium War ended with the Treaty of Nanking between Britain and the Qing Dynasty (1644–1911), through which Britain forced China to open its ports. This was not a harmless ‘opening of trade’: the British ruthlessly swamped China with opium. Following this treaty, many Western powers, including the US, decided to follow the example of Britain. The US government appointed Caleb Cushing (1800–1879) as envoy extraordinary and minister plenipotentiary and dispatched him to the south of China to negotiate a treaty similar to the one signed between the British and the Chinese. In February 1844, Cushing arrived at Macao, then a Portuguese colony. He immediately contacted Ching Yucai (Pinyin: Cheng Yucai 程矞採, 1783–1858) – China’s acting governor general of Guangdong and Guangxi Provinces – with a series of diplomatic notes. Cushing demanded that Ching allow him to proceed to Peking (Beijing) to negotiate a ‘treaty of peace’ (i.e. an open-trade arrangement) with the imperial government. Cushing’s demands rang alarm bells in the Chinese government, and Ching responded to Cushing’s notes by attempting to prevent, or at least delay, Cushing’s proceeding to the capital. Cushing, on the other hand, was very well aware that China, having recently suffered a humiliating defeat by the British, had essentially no power to prevent him signing a treaty for his government. Because of this, Cushing’s diplomatic notes became increasingly assertive. Not surprisingly, the exchange ended with the Chinese government giving in to Cushing’s demands.

5.3.5 Data

Our corpus consists of the twenty-five diplomatic notes exchanged between Cushing and Ching, between 27 February and 24 May 1844. Within this corpus, we devote special attention to the twelve notes written by Cushing, with occasional references (for contrastive purposes) to Ching’s diplomatic notes. Originally Cushing’s diplomatic notes were translated into Classical Chinese, and a sinologist in Cushing’s team translated the Chinese diplomatic notes into English. In this case study we only mention translational problems in passing, as they do not relate to the main issue we wish to investigate – how the coloniser behaves. The translations of the examples featured in this case study are the original translations. The size of our corpus is given in Table 5.1.

Diplomatic notes in our corpus are publicly available.Footnote ¹ Note that while our corpus is relatively small, we agree with Sharoff et al. (Reference Sharoff, Rapp, Zweigenbaum and Fung2013) that a certain degree of ‘imperfection’ of corpora does not invalidate pragmatic research, in particular if we look at historical data.

5.3.6 Analysis

5.3.7 Summary of Analysis

Our analysis has shown that Tells are fundamental in our US data for realising aggression under a veneer of civility. We have also examined other speech acts which, being intertwined with Tells, are also used to realise covert aggression. In the following, we summarise the results of our analysis from a quantitative point of view. Note that in this analysis we only count the approximate length of covertly aggressive speech acts in our corpus, bearing in mind all the difficulties which speech act annotation entails, and also that intertwinedness only allows us to quantify our data for illustrative purposes.Footnote ³ Based on our annotation of the data, here we simply refer to the number of words representing speech acts in our annotation, rather than arguing that Cushing used a particular ‘quantity’ of speech acts. In annotating our data, we manually distinguished sections which are relevant for the realisation of aggression from others which we regarded as ‘neutral’; we dismissed the latter from our quantitative analysis. Table 5.2 provides a summary of our quantitative findings.

In the following, we will conduct a comparison of the outcomes of our analysis of Cushing’s notes with the pragmatic features of Ching’s responsive diplomatic notes. We keep this comparison brief because in the procedure provided for resolving Pitfall 3, we propose to analytically prioritise the behaviour of the actor whose behaviour we aim to investigate.