1.Introduction

Speaking a language is an individual act that may be performed in different ways depending on the intentions of the speaker. This can be observed in politics (Albalat-Mascarell and Carrió-Pastor 2019Albalat-Mascarell, Ana, and María Luisa Carrió-Pastor 2019 “Self-representation in Political Campaign Talk: A Functional Metadiscourse Approach to Self-mentions in Televised Presidential Debates.” Journal of Pragmatics 147: 86–99. ), in academic English (Carrió-Pastor 2014Carrió-Pastor, María Luisa 2014 “Cross-Cultural Variation in the Use of Modal Verbs in Academic English.” SKY Journal of Linguistics 27: 153–166.; Alonso-Almeida and Carrió-Pastor 2017Alonso-Almeida, Francisco, and María Luisa Carrió-Pastor 2017 “Variation and Function of Modals in Linguistics and Engineering Research Papers in English.” In Evidentiality and Modality in European Languages. Discourse-Pragmatic Perspectives, ed. by Juana I. Marín-Arrese, Julia Lavid-López, Marta Carratero, Elena Domínguez Romero, M^a Victoria Martín de la Rosa, and María Pérez Blanco, 242–277. Bern: Peter Lang.), in digital comments on news (Moya Muñoz and Carrió-Pastor 2018Moya-Muñoz, Patricio, and María Luisa Carrió-Pastor 2018 “Estrategias de intensificación en los comentarios digitales sobre noticias en español: un análisis de la variación entre España y Chile.” Spanish in Context 15: 369–391. ), in newspapers (Dafouz 2008Dafouz, Emma 2008 “The Pragmatic Role of Textual and Interpersonal Metadiscourse Markers in the Construction and Attainment of Persuasion: A Cross-Linguistic Study of Newspaper Discourse.” Journal of Pragmatics 40 (1): 95–113. ; Alonso-Almeida and Carrió-Pastor 2019Alonso-Almeida, Francisco, and María Luisa Carrió-Pastor 2019 “Constructing Legitimation in Scottish Newspapers: The Case of the Independence Referendum.” Discourse Studies 21: 621–635. ), etc. All these studies reflect on the variation of language and the dissimilarities in the way communication is carried out by different language users in specific contexts. This can be witnessed when speakers choose one term over another, use specific words to express their thoughts or overuse assertive phrases. This results from the fact that we conceptualize ideas in different ways, and this is reflected in speech. We believe that every speaker processes reality in their own way, and the transmission of this reality is, in turn, bound by a degree of subjectivity. This practice is quite common in political talk (Friedman and Kampf 2014Friedman, Elie, and Zohar Kampf 2014 “Politically Speaking at Home and Abroad: A Typology of Message Gap Strategies.” Discourse & Society 25 (6): 706–724. ; Kampf 2016Kampf, Zohar 2016 “All the Best! Performing Solidarity in Political Discourse.” Journal of Pragmatics 93: 47–60. ).

One linguistic feature that may vary according to the intentions of the speaker is the use of rhetorical devices, e.g. metadiscourse. Abdollahzadeh (2011Abdollahzadeh, Esmaeel 2011 “Poring Over the Findings: Interpersonal Authorial Engagement in Applied Linguistics Papers.” Journal of Pragmatics 43: 288–297. , 288) defines metadiscourse as the way in which “writers tend to convey their personality, credibility, consideration of the reader and the relationship to the subject matter and to readers by using certain devices in their texts”. Metadiscourse devices have been of interest to such researchers as Dahl (2004)Dahl, Trine 2004 “Textual Metadiscourse in Research Articles: A Marker of National Culture or of Academic Discipline?” Journal of Pragmatics 36: 1807–1825. ; Koutsantoni (2004)Koutsantoni, Dimitra 2004 “Attitude, Certainty and Allusions to Common Knowledge in Scientific Research Articles.” Journal of English for Academic Purposes 3: 163–182. ; Hyland (2005)Hyland, Ken 2005 Metadiscourse. London: Continuum.; Andrus (2009)Andrus, Jennifer 2009 “The Development of an Artefactual Language Ideology: Utterance, Event, and Agency in the Metadiscourse of the Excited Utterance Exception to Hearsay.” Language & Communication 29: 312–327. ; Abdollahzadeh (2011)Abdollahzadeh, Esmaeel 2011 “Poring Over the Findings: Interpersonal Authorial Engagement in Applied Linguistics Papers.” Journal of Pragmatics 43: 288–297. ; Mur-Dueñas (2011)Mur-Dueñas, Pilar 2011 “An Intercultural Analysis of Metadiscourse Features in Research Articles Written in English and in Spanish.” Journal of Pragmatics 43: 3068–3079. , Carrió-Pastor (2016aCarrió-Pastor, María Luisa 2016a “Mitigation of Claims in Medical Research Papers: A Comparative Study of English and Spanish Writers.” Communication & Medicine 13: 1–25., 2016bCarrió-Pastor, María Luisa 2016b “A Contrastive Study of the Hedges Used by English, Spanish and Chinese Researchers in Academic Papers.” In Input a Word, Analyze the World: Selected Approaches to Corpus Linguistics, ed. by Francisco Alonso Almeida, Ivalla Ortega Barrera, Elena Quintana Toledo, and Margarita E. Sánchez Cuervo, 477–492. Cambridge: Cambridge Scholars., 2016cCarrió-Pastor, María Luisa 2016c “A Contrastive Study of Interactive Metadiscourse in Academic Papers Written in English and in Spanish.” In Corpus-Based Studies on Language Varieties, ed. by Francisco Alonso Almeida, Laura Cruz García, and Víctor González Ruiz, 89–114. Bern: Linguistic Insights.), and Albalat-Mascarell and Carrió-Pastor (2019)Albalat-Mascarell, Ana, and María Luisa Carrió-Pastor 2019 “Self-representation in Political Campaign Talk: A Functional Metadiscourse Approach to Self-mentions in Televised Presidential Debates.” Journal of Pragmatics 147: 86–99. , among others. In this paper, we focus on the analysis of evidentials (a sub-category of textual or interactive devices, e.g. according to, following, etc.) and boosters (a sub-category of interpersonal or interactional devices, e.g. obvious, fact, extremely, etc.), following Hyland’s (2005)Hyland, Ken 2005 Metadiscourse. London: Continuum. taxonomy. We believe the analysis of the use of these devices in campaign debates may shed new light on the use of rhetorical strategies in politics. Some researchers such as Sclafani (2017)Sclafani, Jennifer 2017 Talking Donald Trump: A Sociolinguistic Study of Style, Metadiscourse and Political Identity. London: Routledge. , Zhang, Sun, Peng, Gan, and Yu (2017)Zhang, Man, Weiwei Sun, Huan Peng, Qiong Gan, and Bo Yu 2017 “A Multidimensional Analysis of Metadiscourse Markers across Spoken Registers.” Journal of Pragmatics 117: 106–118. , and Liu and Lei (2018)Liu, Dilin, and Lei Lei 2018 “The Appeal to Political Sentiment: An Analysis of Donald Trump’s and Hillary Clinton’s Speech Themes and Discourse Strategies in the 2016 US Presidential Election.” Discourse, Context & Media 25: 143–152. have been investigating the use of metadiscourse in spoken English, but we argue that further studies should be dedicated to the analysis of metadiscourse devices used in political debates. We have chosen political debates since they have been defined as “zero-sum games” (García-Pastor 2008García-Pastor, María Dolores 2008 “Political Campaign Debates as Zero-Sum Games: Impoliteness and Power in Candidates’ Exchanges.” In Impoliteness in Language, ed. by Derek Bousfield, and Miriam A. Locher, 101–126. Berlin, New York: De Gruyter Mouton., 101) whose antagonistic character shapes the relationship between the politicians involved and their audiences. The analysis of the presence of evidentials and boosters in these debates may be useful to discover rhetorical strategies that damage the opponent and manage speakers to win over voters.

In this study, we analyse political debates on Brexit, as such speeches may reveal how politicians persuade voters. Our first objective is to analyse the evidentials and boosters most frequently used among pro- and anti-Brexit politicians in debates and relate these to the speakers’ goals. Our second objective is to study the use of evidential devices and boosters among pro-Brexit politicians considering the results of the referendum. Thus, the research questions of this study are as follows:

1. What evidential and boosting devices are the most used in the campaign debates held before the Brexit referendum? Do pro- and anti-Brexit politicians use evidentials and boosters with the same frequency and with the same function?

2. What devices are used more frequently and/or with the same frequency by pro-Brexit politicians: evidentials or boosters? What may be the associated causes?

This article is organised as follows. Section 2 discusses previous studies on evidentials, boosters, and political speech. Then, in Section 3, we describe the corpus compiled and the method followed in this study. Section 4 shows the results obtained from our analysis. Finally, our conclusions are drawn in Section 5.

2.Evidential devices, boosters, and political discourse

Metadiscourse plays a key role in knowledge construction by managing the interactions between speakers and audiences, who often come from the same discourse community and engage in shared social and cultural practices (Albalat-Mascarell 2015Albalat-Mascarell, Ana 2015 “El metadiscurso como marco de análisis comparativo funcional entre el inglés y el español en los discursos de especialidad.” Revista Académica Liletrad 1: 87–96.; Carrió-Pastor and Muñiz-Calderón 2015Carrió-Pastor, María Luisa, and Ruth Muñiz-Calderón 2015 “Identification and Causes of Lexical Variation in Chinese Business English.” English Today 31: 10–15. ; Carrió-Pastor 2016aCarrió-Pastor, María Luisa 2016a “Mitigation of Claims in Medical Research Papers: A Comparative Study of English and Spanish Writers.” Communication & Medicine 13: 1–25., 2016bCarrió-Pastor, María Luisa 2016b “A Contrastive Study of the Hedges Used by English, Spanish and Chinese Researchers in Academic Papers.” In Input a Word, Analyze the World: Selected Approaches to Corpus Linguistics, ed. by Francisco Alonso Almeida, Ivalla Ortega Barrera, Elena Quintana Toledo, and Margarita E. Sánchez Cuervo, 477–492. Cambridge: Cambridge Scholars., 2016cCarrió-Pastor, María Luisa 2016c “A Contrastive Study of Interactive Metadiscourse in Academic Papers Written in English and in Spanish.” In Corpus-Based Studies on Language Varieties, ed. by Francisco Alonso Almeida, Laura Cruz García, and Víctor González Ruiz, 89–114. Bern: Linguistic Insights., 2019aCarrió-Pastor, María Luisa 2019a “Different Ways to Express Personal Attitudes in Spanish and English Engineering Papers: An Analysis of Metadiscourse Devices, Affective Evaluation and Sentiment Analysis.” Lodz Papers in Pragmatics 15 (1): 45–67. , 2019bCarrió-Pastor, María Luisa 2019b “Do Writers Express the Same Attitude in Historical Genres?” In Writing History in Late Modern English. Explorations of The Coruña Corpus, ed. by Isabel Moskowich, Begoña Crespo, Luis Puente-Castelo, and Leida Maria Monaco, 237–259. Amsterdam: John Benjamins. ). However, while metadiscourse is a useful tool in explaining the rhetorical features of languages in different domains and genres, it has mostly been examined in relation to academic writing (Hyland 1998Hyland, Ken 1998 “Persuasion and Context: The Pragmatics of Academic Metadiscourse.” Journal of Pragmatics 30 (4): 437–455. , 2010Hyland, Ken 2010 “Metadiscourse: Mapping Interactions in Academic Writing.” Nordic Journal of English Studies 9 (2): 125–143. ; Hyland and Tse 2004Hyland, Ken, and Polly Tse 2004 “Metadiscourse in Academic Writing: A Reappraisal.” Applied Linguistics 25 (2): 156–177. ). Little attention has been given to the role of metadiscourse devices in non-academic discourses with an overtly persuasive component such as political discourse.

Metadiscourse is a fuzzy concept. It is generally characterized as “discourse about discourse” (Hyland 2015Hyland, Ken 2015 “Metadiscourse.” In The International Encyclopedia of Language and Social Interaction 3 Volume Set, ed. by Karen Tracy, Cornelia Ilie, and Todd Sandel, 997–1006. London: John Wiley & Sons. , 1), but while some analysts restrict this to features of textual organization, others include in their analyses how speakers react to what they are saying and create rapport with audiences. In this paper, we adopt the second position, which sees metadiscourse as something related to the interpersonal character of communication. Proponents of this broader idea of metadiscourse tend to establish a distinction between interactive and interactional metadiscoursal categories (e.g. Thompson 2001Thompson, Geoff 2001 “Interaction in Academic Writing: Learning to Argue with the Reader.” Applied Linguistics 22 (1): 58–78. ; Hyland 2005Hyland, Ken 2005 Metadiscourse. London: Continuum.; Mur-Dueñas 2011Mur-Dueñas, Pilar 2011 “An Intercultural Analysis of Metadiscourse Features in Research Articles Written in English and in Spanish.” Journal of Pragmatics 43: 3068–3079. ; Carrió-Pastor 2016aCarrió-Pastor, María Luisa 2016a “Mitigation of Claims in Medical Research Papers: A Comparative Study of English and Spanish Writers.” Communication & Medicine 13: 1–25., 2016bCarrió-Pastor, María Luisa 2016b “A Contrastive Study of the Hedges Used by English, Spanish and Chinese Researchers in Academic Papers.” In Input a Word, Analyze the World: Selected Approaches to Corpus Linguistics, ed. by Francisco Alonso Almeida, Ivalla Ortega Barrera, Elena Quintana Toledo, and Margarita E. Sánchez Cuervo, 477–492. Cambridge: Cambridge Scholars., 2016cCarrió-Pastor, María Luisa 2016c “A Contrastive Study of Interactive Metadiscourse in Academic Papers Written in English and in Spanish.” In Corpus-Based Studies on Language Varieties, ed. by Francisco Alonso Almeida, Laura Cruz García, and Víctor González Ruiz, 89–114. Bern: Linguistic Insights.), depending on the speaker-listener relationship built into discourse. Items of interactive categories aim to organize information depending on the expectations of the audience, as opposed to interactional metadiscoursal features designed to stimulate interaction between the addresser and the addressee.

In this study, we analyse evidential devices, which are one sub-category of interactive devices, and boosters, which are one sub-category of interactional devices. Both are common rhetorical strategies that lend credibility to arguments either by drawing on external sources of information or by emphasizing certainty about a proposition.

On the one hand, evidentials “indicate the source of textual information which originates outside the current text” (Hyland 2004Hyland, Ken 2004 “Disciplinary Interactions: Metadiscourse in L2 Postgraduate Writing.” Journal of Second Language Writing 13: 133–151. , 139). In this paper, we adopt a dynamic view of evidentials based on the idea that they consist of the source of information rather than the assessment of the epistemic status of the proposition, a perspective followed by other researchers (e.g. Aikhenvald 2005Aikhenvald, Alexandra Y. 2005 Evidentiality. Oxford: Oxford University Press.; Marín-Arrese 2011Marín-Arrese, Juana I. 2011 “Epistemic Legitimizing Strategies, Commitment and Accountability in Discourse.” Discourse Studies 13 (6): 789–797. ; Marín-Arrese et al. 2013Marín-Arrese, Juana I., Marta Carretero, Jorge A. Hita, and Johan Van der Auwera (eds) 2013 English Modality: Core, Periphery and Evidentiality. Berlin: de Gruyter. ; Berlin and Prieto-Mendoza 2014Berlin, Lawrence N., and Alejandra Prieto-Mendoza 2014 “Evidential Embellishment in Political Debates during US Campaigns.” Intercultural Pragmatics 11 (3): 389–409. ; Alonso-Almeida 2015Alonso-Almeida, Francisco 2015 “On the Mitigation Function of Modality and Evidentiality. Evidence from English and Spanish Medical Research Papers.” Intercultural Pragmatics 12 (1): 33–57. ; Marín-Arrese, Hassler and Carretero 2017Marín-Arrese, Juana I., Gerda Hassler, and Marta Carretero (eds) 2017 Evidentiality Revisited: Cognitive Grammar, Functional and Discourse-Pragmatic Perspectives. Amsterdam: John Benjamins. ; Estellés 2019Estellés, María 2019 “The Evolution of Parliamentary Debates in Light of the Evolution of Evidentials: Al Parecer and Por Lo Visto in 40 Years of Parliamentary Proceedings from Spain.” Corpus Pragmatics 4: 59–82. ). In this study, the function of an evidential device is to introduce discourse-external references that endorse a proposition. The focus of this paper is on the analysis of evidential markers that provide evidence to make a factual claim (Marín-Arrese, Hassler and Carretero 2017Marín-Arrese, Juana I., Gerda Hassler, and Marta Carretero (eds) 2017 Evidentiality Revisited: Cognitive Grammar, Functional and Discourse-Pragmatic Perspectives. Amsterdam: John Benjamins. ). In the political debates analysed, we searched for devices that indicate the “structural dimension of grammar that codifies the source of information” (Bussmann 2006Bussmann, Hadumod 2006 Routledge Dictionary of Language and Linguistics. London: Routledge. , 390). Aikhenvald also makes it clear that evidentiality is “a grammatical category that has source of information as its primary meaning” (Brown 2005Brown, Keith 2005 Encyclopedia of Language and Linguistics. Amsterdam: Elsevier., 320). In the political debates analysed, we decided to focus on the study of the devices (evidentials and boosters) used to emphasize the force of politicians’ propositions to convince voters.

Following Mur-Dueñas (2011Mur-Dueñas, Pilar 2011 “An Intercultural Analysis of Metadiscourse Features in Research Articles Written in English and in Spanish.” Journal of Pragmatics 43: 3068–3079. , 3070), evidentials comprise “both references to textual information as originated in a scholar’s or several scholars’ work (personal evidentials) and references to what the writers assume to be common, shared knowledge within the discipline (impersonal evidentials)”. Thus, evidential markers are divided into two types (Hyland 2005Hyland, Ken 2005 Metadiscourse. London: Continuum.; Mur-Dueñas 2011Mur-Dueñas, Pilar 2011 “An Intercultural Analysis of Metadiscourse Features in Research Articles Written in English and in Spanish.” Journal of Pragmatics 43: 3068–3079. ; Carrió-Pastor 2016cCarrió-Pastor, María Luisa 2016c “A Contrastive Study of Interactive Metadiscourse in Academic Papers Written in English and in Spanish.” In Corpus-Based Studies on Language Varieties, ed. by Francisco Alonso Almeida, Laura Cruz García, and Víctor González Ruiz, 89–114. Bern: Linguistic Insights.):

• Personal, which includes according to X, as X argued, in X’s study.

• Impersonal, which comprises previous, past research, previously.

On the other hand, boosters are linguistic devices that emphasise certainty about a proposition or highlight confidence in an assertion, increasing the illocutionary force of the speech act and closing off potential alternative viewpoints (Mur-Dueñas 2011Mur-Dueñas, Pilar 2011 “An Intercultural Analysis of Metadiscourse Features in Research Articles Written in English and in Spanish.” Journal of Pragmatics 43: 3068–3079. ). Several taxonomies of the words or phrases that can be classified as boosters have been made by researchers. For example, Hyland (2005Hyland, Ken 2005 Metadiscourse. London: Continuum., 2010Hyland, Ken 2010 “Metadiscourse: Mapping Interactions in Academic Writing.” Nordic Journal of English Studies 9 (2): 125–143. ) labelled the following words as boosters: obvious, obviously, very, extremely, far, full, never, certain, certainly, sure, find, must, realize, really, surely, think, true, without doubt, etc. More recently, Mur-Dueñas (2011)Mur-Dueñas, Pilar 2011 “An Intercultural Analysis of Metadiscourse Features in Research Articles Written in English and in Spanish.” Journal of Pragmatics 43: 3068–3079. compiled a longer list of words identified from a corpus of academic papers on business management, such as determine, show, demonstrate, reveal, highlight, confirm, emphasize, conclude, hold, underscore, establish, assert, prove, know, clearly, significantly, generally, largely, particularly, indeed, widely, highly, primarily, consistently, strongly, actually, mostly, especially, extensively, entirely, essentially, dramatically, substantially, always, fully, considerable, clear, vast, evident, substantial, evidence, fact, majority, assertion, conclusion, in fact, for the most part, of course, to a large extent, etc. As can be observed, boosters demonstrate confidence and convey the right amount of self-assurance to listeners, although their meaning may vary depending on the context or on the specific field of the discourse analysed.

As indicated above, despite the increasing number of studies devoted to the use of interpersonal metadiscourse devices such as evidentials and boosters in academic written genres, limited research (Albalat-Mascarell and Carrió-Pastor 2019Albalat-Mascarell, Ana, and María Luisa Carrió-Pastor 2019 “Self-representation in Political Campaign Talk: A Functional Metadiscourse Approach to Self-mentions in Televised Presidential Debates.” Journal of Pragmatics 147: 86–99. ) has been conducted on the deployment of interactive and interactional metadiscourse in political discourse. Generally, political talk has been analysed from a pragmatic perspective, as its persuasive nature has often been of interest. Researchers have focused on specific topics such as irony and cognitive aspects (Cap 2017Cap, Piotr 2017 “Studying Ideological Worldviews in Political Discourse Space: Critical-Cognitive Advances in the Analysis of Conflict and Coercion.” Journal of Pragmatics 108: 17–27. ; Mussolff 2017Musolff, Andreas 2017 “Metaphor, Irony and Sarcasm in Public Discourse.” Journal of Pragmatics 109: 95–104. ), interaction (Frachiolla 2011Fracchiolla, Béatrice 2011 “Politeness as a Strategy of Attack in a Gendered Political Debate. The Royal–Sarkozy Debate.” Journal of Pragmatics 43 (10): 2480–2488. ; Boyd 2014Boyd, Michael S. 2014 “(New) Participatory Framework on Youtube? Commenter Interaction in US Political Speeches.” Journal of Pragmatics 72: 46–58. ), interview styles (Eriksson 2011Eriksson, Göran 2011 “Follow-Up Questions in Political Press Conferences.” Journal of Pragmatics 43: 3331–3344. ; Proctor and Su 2011Proctor, Katarzyna, I. Lily, and Wen Su 2011 “The 1st Person Plural in Political Discourse. American Politicians in Interviews and in a Debate.” Journal of Pragmatics 43 (43): 3251–3266. ; Gnisci, Zollo, Perugini and Di Conza 2013Gnisci, Augusto, Pierpaolo Zollo, Marco Perugini, and Angiola Di Conza 2013 “A Comparative Study of Toughness and Neutrality in Italian and English Political Interviews.” Journal of Pragmatics 50 (1): 152–167. ), address terms (Rendle-Short 2007Rendle-Short, Johanna 2007 “ ‘Catherine, You’re Wasting Your Time’: Address Terms within the Australian Political Interview.” Journal of Pragmatics 39 (9): 1503–1525. ), political genres (Cap and Okulska 2013Cap, Piotr, and Ursula Okulska (eds.) 2013 Analyzing Genres in Political Communication: Theory and Practice. Amsterdam: John Benjamins. ), political speech acts (Remer 2008Remer, Gary 2008 “Genres of Political Speech: Oratory and Conversation, Today and in Antiquity.” Language & Communication 28 (2): 182–196. ; Kampf 2016Kampf, Zohar 2016 “All the Best! Performing Solidarity in Political Discourse.” Journal of Pragmatics 93: 47–60. ), media communication from a cross-cultural perspective (Fetzer and Lauerbach 2007Fetzer, Anita, and Gerda Lauerbach (eds.) 2007 Political Discourse in the Media: Cross-Cultural Perspectives. Amsterdam: John Benjamins. ), and argumentation (Lauerbach 2007Lauerbach, Gerda 2007 “Argumentation in Political Talk Show Interviews.” Journal of Pragmatics 39: 1388–1419. ). Additionally, previous research on evidentials and boosters in political genres such as parliamentary and campaign debates can be found (e.g. Ilie 2003Ilie, Cornelia 2003 “Discourse and Metadiscourse in Parliamentary Debates.” Journal of Language and Politics 2 (1): 71–92. ; Berlin and Prieto Mendoza 2014Berlin, Lawrence N., and Alejandra Prieto-Mendoza 2014 “Evidential Embellishment in Political Debates during US Campaigns.” Intercultural Pragmatics 11 (3): 389–409. ; Buckledee 2018Buckledee, Steve 2018 The Language of Brexit: How Britain Talked Its Way Out of the European Union. London: Bloomsbury Publishing.; Estellés 2019Estellés, María 2019 “The Evolution of Parliamentary Debates in Light of the Evolution of Evidentials: Al Parecer and Por Lo Visto in 40 Years of Parliamentary Proceedings from Spain.” Corpus Pragmatics 4: 59–82. ). Nevertheless, no studies so far have considered the use of evidentials and boosters in campaign debates from the point of view of interpersonal metadiscourse.

In this paper, we address this gap by analysing the presence and function of evidentials and boosters in the 2016 campaign debates on the United Kingdom European Union membership referendum (also known as the Brexit referendum). The designated official slogan leading the Remain campaigning group was “Britain Stronger in Europe”, and the designated official slogan leading the Leave campaigning group was “Vote Leave”. We focused on the political speeches delivered before the Brexit referendum given that their focus was persuasive. Politicians tried to provide arguments for and against Britain leaving the European Union. In a previous study, Alonso-Almeida and Carrió-Pastor (2019)Alonso-Almeida, Francisco, and María Luisa Carrió-Pastor 2019 “Constructing Legitimation in Scottish Newspapers: The Case of the Independence Referendum.” Discourse Studies 21: 621–635. paid attention to the Scottish independence referendum. Specifically, the epistemic stance strategies deployed by journalists to legitimise their ideas concerning the referendum were analysed. In this study, the persuasive function of the political campaign debates held before the referendum is at play. Our intention is to analyse how politicians persuade readers of the veracity of their propositions and the possible effect of their use of evidential devices and boosters on the results of the poll.

3.Materials and methods

The corpus compiled to meet the objectives of this article includes the transcripts of speeches given by eleven different British politicians and one trade union representative at the BBC’s “Great Debate” held at Wembley Arena on 21st June 2016 and hosted by David Dimbleby in front of an audience of 6,000 people.

We chose this from the numerous TV debates, question sessions, and interviews that took place during the referendum campaign given its significance and format, as it was marketed as the most significant event of the referendum campaign. The debate featured three main speakers nominated by each official leading a campaigning group and a second stage involving a further ten guests with five representatives from each side. Each panellist on the main stage was given an opportunity to answer questions from audience members (split evenly between Leave and Remain supporters) and to debate subjects related to three key referendum issues, i.e. the economy, immigration, and Great Britain’s place in the world. In the second stage, politicians, businesspeople, and other prominent speakers who had also played a part in the campaign provided additional opinions and commentary on these key issues. The debate came only two days before polling day at a moment when both sides were looking for a clear victory that could possibly determine the outcome of the referendum.

As our material, we used the video of the debate uploaded on YouTube by different channels. The debate was originally broadcasted on BBC One between 8pm and 10 pm (120 minutes). We prepared a thorough transcript of the whole debate, including details on miss-starts, repetitions and overlaps relevant to a metadiscoursal analysis approach. As we only aimed to examine speeches delivered by politicians of different ideological backgrounds adopting a particular side of the referendum campaign, we only used those parts of the transcript that corresponded to politicians’ turns and to those belonging to O’Grady’s speeches, who was also included in the analysis as a main speaker of the debate. Speeches selected for our research were those delivered by:

• Conservative MP and former Mayor of London Boris Johnson, Labour MP Gisela Stuart, and Conservative MP Andrea Leadsom representing Leave on the main stage.

• Leader of the Scottish Conservative and Unionist Party Ruth Davidson MSP, London Mayor Sadiq Khan, and General Secretary of the Trades Union Congress Frances O’Grady representing Remain on the main stage.

• Minister of State for Employment Priti Patel and Deputy Chairman of UKIP Diane James representing Leave on the second stage.

• MSP Humza Yousaf, Conservative MP Sarah Wollaston, Liberal Democrat leader Tim Farron, and Green Party MP Caroline Lucas representing Remain on the second stage.

Our procedure first involved compiling the corpus by classifying transcripts of the speeches selected into two categories: Leave campaigners and Remain campaigners. See the data compiled for the analysis in Table 1.

It can be observed that the size of the files and tokens are different in the two sub-corpora, as the Leave campaigners’ sub-corpus was composed of 5,798 tokens while the Remain campaigners’ sub-corpus contained 6,411 tokens. Additionally, more sentences were found in the remain campaigners’ corpus. To compare the results obtained after the analysis, raw occurrences were normalised to 1,000 to calculate the frequencies for each per the same number of words. A total of 12,209 tokens composed the corpus of this analysis.

Second, evidential devices and boosters were searched for electronically through the whole corpus using WordSmith Tools 5.0 to extract frequency lists. After that, METOOL was used to check the context and verify the occurrences that functioned as metadiscourse devices, eliminating false-positive results.

METOOL was designed through a joint project involving the Research Institute for Information and Language Processing (University of Wolverhampton, GB) and Universitat Politècnica de Valencia (Spain) to compile, tag, identify, and analyse metadiscoursal devices. METOOL has been specifically developed to detect metadiscoursal strategies and is part of the research project FFI2016-77941-P (funded by Ministerio de Economía y Competitividad, Spain). It is not still available in open access as the tagging of the base corpus that will allow the automatic identification of metadiscoursal devices is still in progress. This tool is in an extensive phase of testing at the University of Wolverhampton and the University of Coventry. The tool has proven very useful for analysing and identifying metadiscoursal elements in context, but, as the testing phase has not finished, we preferred to use both tools, METOOL and WordSmith. METOOL includes different kinds of analysis. It extracts frequencies and collocations, and allows us to study markers in context. We could study boosters and evidentials in context with a link that shows the marker in the text analysed. In fact, almost all the boosters and evidentials identified were checked in order to be sure the data retrieved was accurate and thus false positives were not included. While the tool was initially designed to analyse academic discourse, it has also proven to be useful for the analysis of political speech in this study.

The boosters identified from the Brexit speeches are as follows:

• Adjectives: clear, extraordinary, great, major, greatest, secure, complete, leading, vital, credible, crucial, enhanced, enormous, fundamental, huge, permanent, securer, strong, stronger.

• Adverbials: all, really, actually, right, absolutely, rightly, clearly, completely, always, indeed, certainly, mainly, constantly, continually, enormously, frankly, incredibly, inevitably, perfectly, steadily, successful, vividly.

• Verbs: know, believe, knows, knew, admit, admitting, believed.

• Nouns: security, truth, evidence, trust, agreement, credit, fact, leadership, stability, optimists, primacy, success.

The evidentials found in the Brexit speeches are as follows:

• Personal: say, ask, saying, agree, asked, according, quote, report, argument, show, shows, asks, asserted, claim, founder, quotes, reporting.

• Impersonal: recently, former, past, earlier, early.

Occurrences of boosters and evidentials were counted, analysed, and contextualized. At that point, the two sub-corpora of Leave campaigners and Remain campaigners were analysed separately, and occurrences of evidentials and boosters were counted by calculating frequencies. After obtaining the quantitative results of evidential devices and boosters, examples were carefully analysed in context to ensure that they performed a rhetorical function in the debate and could be incorporated into the counts. Raw occurrences were then normed to occurrences per 1,000 words to facilitate comparisons across the debates, as explained above.

Then, the statistical analysis was carried out using http://​vassarstats​.net​/index​.html, a user-friendly tool for performing statistical computation. The tool has been created and maintained by Richard Lowry, PhD, Professor of Psychology Emeritus at Vassar College. In this study, the significance of the difference between two independent proportions was calculated using ANOVA tests. We established the p-value at < 0.05.

4.Results and discussion

The analysis of the corpus showed that the politicians who participated in the Brexit campaign debate generally used more boosters than evidentials. We believe that this may be attributable to the intentions of political discourse used in the Brexit campaign, i.e. to convince voters to leave or remain, appealing to emotions rather than to facts or evidence. In total, we obtained 108 occurrences (8.84 normalised to 1,000) in the category of evidential devices and 281 occurrences in the category of boosters (23.01 normalised to 1,000).

Table 2 illustrates the occurrences and normalised frequencies (NF) to 1,000 words of evidential devices and boosters used by the pro-Brexit and anti-Brexit politicians:

After calculating the difference between two independent proportions, we obtained that the proportion of boosters (0.023%) was higher than evidentials (0.008%) with a p-value of < .0002. Thus, the normalised results showed that the first hypothesis of this paper was confirmed, i.e. politicians preferred boosters to motivate voters, increasing their illocutionary force.

A significant difference was not found in the use of boosters by pro-Brexit (0.022%) and anti-Brexit politicians (0.024%), with a p-value of 0.368. The findings are significant (p-value < .0001), but the difference was only 3.0 frequencies per 1,000 words. On the contrary, the data obtained concerning evidential devices (pro-Brexit speakers 0.011% and anti-Brexit 0.006%) displayed a significant difference (p = 0.004). Pro-Brexit politicians provided more evidence about their claims than anti-Brexit politicians. Considering the outcome of the referendum, which resulted in 51.9% of the population being in favour of leaving the European Union, it seems voters trusted politicians who used more devices with source of information as its primary meaning (i.e. evidentials).

4.1Results and discussion of boosters used in the BBC’s “Great Debate”

The most frequent boosters used by both pro- and anti-Brexit politicians in the campaign were all, know, and really. The overall results concerning the use of boosters in the Brexit campaign debate and their normalised frequencies can be seen in Table 3:

Table 3.Most frequent boosters in pro- and anti-Brexit political speeches

Boosters	NF pro-Brexit	Boosters	NF anti-Brexit
All	3.58	Know	5.51
Actually	2.33	All	4.48
Know	2.18	Really	1.89
Clear	1.55	Believe	0.68
Right	1.4	Evidence	0.68
Really	1.24	Right	0.68

We have divided this section into two sub-sections; one is devoted to the analysis and discussion of examples of the boosters found in the Remain EU (anti-Brexit) sub-corpus, and the other to the boosters used in the Leave EU (pro-Brexit) sub-corpus.

4.1.1Boosters in the anti-Brexit sub-corpus

Figure 1 illustrates the results for the boosters used by politicians attempting to convince voters to remain in the European Union.

Figure 1.Comparison of normalised frequencies (per 1,000 words) of boosters used by anti-Brexit politicians

Anti-Brexit politicians used a limited range of frequent boosters to convince voters of their beliefs (know, all, really). One of the most frequent was all, which was used to refer to a big group of people (e.g. all the experts) or to emphasise the speakers address all the voters (e.g. all you). Some examples of the boosters extracted from the remain EU corpus are shown below in (1), (2), and (3).

(1)

“Well, the evidence is undeniable and all the experts agree – we are stronger, safer and better off in Europe”. (REMAIN_KHAN).

“You’re going to be asked to vote in two days’ time and all you’ve heard tonight is trust us and it will all be fine”. (REMAIN_DAVIDSON)

“I don’t think it’s unreasonable people worried about their local hospital, people worried about their local school, worried what would happen if our economy went into recession or if they lost their jobs or businesses struggled, because all the experts say, and you know this, that leaving the EU would cause problems for our country”. (REMAIN_KHAN)

Here, Khan tries to motivate voters with the booster all, reinforced with the adjectives stronger, safer, and better, boosting his speech act. Davidson also emphasises the need to vote and participate in the referendum in “all you’ve heard” and “all be fine”. In the third sample, Khan uses know to emphasize that not only all the experts, but also voters, know a fact, boosting the importance of the sentence, consolidating the speaker-voter’s commitment to the proposition.

(2)

“Jane made a really, really important point in her question about our NHS, and let’s not forget the NHS was built by and for working people, and we should all be really proud of it”. (REMAIN_OGRADY)

The illocutionary force of boosters is obvious here, when O’Grady repeats really before the attitude marker important, being used later again before proud.

(3)

“We certainly haven’t had it throughout this campaign. Britain deserves better than people who say they’ve got a quick fix but won’t tell you what it actually means for Britain”. (REMAIN_DAVIDSON)

Thus, anti-Brexit politicians used several boosters in a sentence, with the two functions identified by Jalilifar and Alavi-Nia (2012Jalilifar, Alireza, and Maryam Alavi-Nia 2012 “We Are Surprised; Wasn’t Iran Disgraced There? A Functional Analysis of Hedges and Boosters in Televised Iranian and American Presidential Debates.” Discourse & Communication 6 (2): 135–161. , 140) “making things less fuzzy there by suppressing alternatives and emphasizing or boosting the force of a speech act”. So, anti-Brexit speakers gave emphasis to their speech repeating the same booster followed by attitude markers (e.g. really, really important) or using different boosters in the same sentence to support the claim that British citizens should remain in the European Union.

4.1.2Boosters in the pro-Brexit sub-corpus

Figure 2 shows the results of devices used as boosters by the pro-Brexit politicians to convince voters to leave the European Union.

Figure 2.Normalised frequencies (per 1,000 words) of boosters used by pro-Brexit politicians

Pro-Brexit politicians used a wide range of devices (all, actually, know, clear, right, really, truth, complete, extraordinary, security…) to increase the illocutionary force of their campaign debates. It should be noted that while anti-Brexit politicians used quite frequently three boosters (know, all, and really), pro-Brexit politicians used a wider range of devices. Therefore, although the difference in the frequencies of the two groups of politicians is not significant, we found a greater variety of boosters in the Leave EU sub-corpus.

The function of boosters in this context is evident: to include all British citizens, emphasizing the idea that leaving the European Union is the best option. This can be seen in different parts of the speeches. Some examples of the functions of the boosters used by pro-Brexit politicians are shown below in (4), (5), and (6).

(4)

“You do – you do deserve the truth and if the Prime Minister is refusing to say that he would exercise his veto over Turkey, what do you think he’s going to veto? And let’s just be clear about the United States of Europe, which Douglas addresses”. (LEAVE_STUART)

“I think it was extraordinary to hear that – um, that we would have tariffs imposed on us, because everybody knows that this country receives about a fifth of Germany’s entire car manufacturing output”. (LEAVE_JOHNSON)

In the first sample, the booster clear (or even the whole sentence let’s just be clear) signals honesty and transmits that the speaker is a politician who does not deceive his audience. The speaker explains a fact using bare information and he implicitly indicates that leaving the European Union is the option of honest voters. In the second one, Boris Johnson uses the verb know to consolidate the politician-voter’s commitment to the value advanced in the speech, emphasized with everybody.

(5)

“Now, the problem with free movement, for me as a mum, it’s not just about uncontrollable numbers coming here and putting pressure on public services. It’s all about security for all of our children, for all of us”. (LEAVE_LEADSOM)

In this speech, the speaker uses the booster security, one of the most frequently used in the corpus of pro-Brexit politicians. Right was also commonly used to convince voters that this was the correct option (to leave the European Union) as shown in Example (6).

(6)

“This is right, the EU has also created fifty percent youth unemployment across southern Europe. It’s a total – it has wrecked the prospects for a generation of young people”. (LEAVE_PATEL)

The boosters clear, security, and right are used to support the idea that leaving the European Union is a positive approach and thus that voters should support this move. The semantic charge of these devices boosts the political discourse of the speaker, thereby emphasizing the force or the speaker’s certainty about a fact.

4.2Results and discussion of evidential devices used in the BBC’s “Great Debate”

The evidential devices most frequently used by both pro-Brexit and anti-Brexit politicians include the lemmas say and ask, which belong to personal evidential devices. The overall results concerning the frequency of use of evidentials in the Brexit campaign debate and their normalised frequencies can be observed in Table 4.

Table 4.Most frequent evidentials in pro- and anti-Brexit political speeches

Evidentials	NF pro-Brexit	Evidentials	NF anti-Brexit
Say	7.89	Say	3.41
Ask	1.77	Ask	1.02
Agree	0.46	Quote	0.68
Former	0.31	Show	0.68
Reporting	0.31	According	0.51

The sections below discuss the different evidential devices identified in the corpus and some examples are commented.

4.2.1Evidentials in the anti-Brexit sub-corpus

Figure 3 shows the evidential devices used by the anti-Brexit politicians to convince voters to remain in the European Union.

Figure 3.Normalised frequencies (per 1,000 words) of evidentials used by anti-Brexit politicians

The most frequently used evidential device is the lemma say, which is a quite commonly used evidential verb to cite the opinion of other people (personal evidential device). This is the reason why the results of the evidential devices found in the anti-Brexit and pro-Brexit sub-corpora (see Figure 4) are quite similar. It should be highlighted that anti-Brexit politicians used fewer evidentials than pro-Brexit politicians did. This may indicate that they focused on other rhetorical strategies to convince voters to stay in the European Union. Some examples of evidentials extracted from the anti-Brexit politicians (remain EU corpus) are shown below in (7), (8), and (9).

(7)

“And perhaps the (inaudible) of the whole campaign is Michael Gove saying we’ve had enough of experts. That’s perhaps an explanation of why he was such a dreadful Education Secretary”. (REMAIN_FARRON)

Farron quotes Gove’s campaign to show evidence of the statement, i.e. a way of criticizing the Education Secretary. Farron manages to mitigate the statement with the repeated use of the hedge “perhaps”, which is an epistemic word used to mitigate the evidence of what is being said by Michael Grove and to indicate that he was ‘dreadful education Secretary’. Here, the evidential device is used to show a negative aspect of one pro-Brexit politician.

(8)

“She said sixty percent of our laws are made in Europe and it’s simply not true. Thirteen percent of our laws, according – according to the independent House of Commons library – according to the independent House of Commons library, that number is thirteen percent”. (REMAIN_DAVIDSON)

Davidson cites Andrea Leadsom, from the conservative party, and repeats an impersonal evidential device (i.e. according) to highlight the source of information and to convince future voters that Leadson included false information in her speech. In this way, Ruth Davidson emphasizes that her ideas are based on real facts, providing evidence from the independent House of Commons Library.

(9)

“How is it – how is it having more control if you have less money in your pocket as Martin Lewis the money expert says?” (REMAIN_KHAN)

Khan specifies the source of information with the use of the verb says referring to Martin Lewis, in this way providing personal evidence of his statement. The speaker refers to Lewis as an expert and so the speaker demonstrates that he is right, reinforcing his statement.

4.2.2Evidentials in the pro-Brexit sub-corpus

Pro-Brexit politicians also used evidentials to support their speech and to persuade voters about leaving the European Union. Frequencies are shown in Figure 4.

Figure 4.Normalised frequencies (per 1,000 words) of evidentials used by pro-Brexit politicians

As in the anti-Brexit speeches, pro-Brexit politicians preferred the lemma say to express personal evidence, citing the statements of other politicians to base their speeches on real facts. Some examples of evidentials extracted from the speeches of leave EU corpus are shown below in (10), (11), and (12).

(10)

“Now, who do you think that was? It was – it was Sidi Khan, again. I agree, I do agree with Sidi. I think you need a grown-up approach to this”. (LEAVE_JOHNSON)

Johnson repeated personal evidentials (e.g. agree) to render the ideas exposed in his speeches more certain; this is also emphasised with the personal pronoun I and the auxiliary verb do.

(11)

“And just to quote the former head of Interpol, Ronald Noble, he said, “The EU border system is like hanging a sign welcoming terrorists to Europe”. It is making us insecure rather than making us securer”. (LEAVE_STUART)

The pro-Brexit speaker refers to the personal opinions of other speakers participating in the debate and in some cases even quotes well-known people (Ronald Noble, head of Interpol) to indicate that remaining in EU is insecure for British.

(12)

“And it would be a fine thing if, as Lord Rose says, if people on low incomes got a pay rise as a result of us taking back control of our country and our system”. “And it is no wonder that they have not been, as Andrea rightly says, they have not been able to do”. (LEAVE_JOHNSON)

In this case, Johnson quotes other politicians with the personal evidential verb says, repeating the previous word or words before and after citing other politicians and then reinforcing the meaning of the personal evidential device.

To sum up, similar boosting and evidential devices are used by pro- and anti-Brexit politicians to increase the number of voters that believed in leaving or remaining in the EU. Politicians aimed at increasing the degree of the truth of a proposition using boosters or citing other politicians to endorse their ideas. It has been proven that there was not a significant difference in the use of boosters, but the use of evidentials was significant in the case of pro-Brexit politicians. The evidential devices identified tend to work in combination towards the representation of a credible self, challenging opposing views on the same issue, i.e. to leave the European Union.

5.Conclusions

The results presented here relate to the use of evidential devices and boosters from speeches given by eleven different British politicians and one trade union representative at the BBC’s “Great Debate” held at Wembley Arena on 21st June 2016. Results extracted from the analysis of the corpus reveal that British politicians used boosters and evidentials to persuade future voters, but it was the use of evidential devices that may have influenced the outcome of the referendum. Evidentials were more frequently used by pro-Brexit politicians to provide evidence about the topic discussed, i.e. leaving the European Union. If we consider the outcome of the referendum (which resulted in most of the voters being in favour of leaving the European Union), it may seem that the public preferred politicians who endorsed their ideas with real facts and data to those using intensifiers.

The first objective of this research was to compare the evidentials and boosters most frequently used in debates and to relate them to the speakers’ goals. On the one hand, the evidentials most frequently used in both sub-corpora were the personal ones, mainly say and ask. Statistical differences were observed in the use of evidentials by pro- and anti-Brexit politicians, but no differences in the functional or qualitative use of these devices were identified. On the other hand, the boosters most frequently used by pro-Brexit speakers were all, know, clear, and actually. The speakers used recurrently the inclusive term all to convince voters and the term clear to transmit confidence. The boosters most used by anti-Brexit politicians were found to be know and really.

The second objective and research question of our analysis was to contrast the use of evidentials and boosters with the results of the referendum. The results of the study showed that pro- and anti-Brexit politicians used similar frequencies of boosters but the use of evidentials may have established a difference in the speeches. At the time, Great Britain faced an uncertain future and may have needed politicians who could convey security and evidence. As a consequence, the use of evidential devices that convey confidence and rely on objective information (impersonal evidentials) could have been crucial to convincing future voters to abandon the European Union. Pro-Brexit politicians showed that leaving the European Union was necessary for Great Britain. In answering our second research question, our view is that, in this debate, pro-Brexit politicians may have used rhetoric to transmit a sense of security with evidential devices and this may have had an impact on the results of the referendum.

We are conscious that the small size of the corpus used may have affected our conclusions, and thus more research should be carried out on this topic. Our aim was to show that evidentials and boosters are rhetorical devices are used by politicians strategically to convince voters of their beliefs. Boosters emphasize the convictions of the politicians and evidentials might influence voters’ opinions. Our future studies will be aimed at examining a larger corpus and identifying metadiscoursal devices mostly used by politicians to persuade voters.