







































Humanities and Social Science Research; Vol. 8, No. 2; 2025 

ISSN 2576-3024   E-ISSN 2576-3032 

https://doi.org/10.30560/hssr.v8n2p75 

 75 Published by IDEAS SPREAD 

 

Metadiscourse Analysis on CEO-Speak in Public Interviews to 

Weather Company’s Crisis 

Zheng Yongmei1 

1 Hubei Preschool Teachers College, China 

Correspondence: Zheng Yongmei, Hubei Preschool Teachers College, Ezhou, China. E-mail: 1453751746  

qq.com 

 

Received: April 3, 2025; Accepted: April 16, 2025; Published: April 18, 2025 

 

Abstract 

Public interview can serve as an effective means for companies to deal with crisis by answering the doubts and 

inquiries from the media and the public. With the United States increasing tariffs or other attacks on Chinese goods, 

more and more companies are facing a crisis. This paper focuses on the deployment of metadiscourse resources 

and their rhetorical features in almost 40 media’s interviews with a company’s CEO. Major findings are 

summarized as follows: (1) The CEO’s utterances are highly persuasive, in which self-mentions and transitions 

occur most frequently. (2) TheCEO pays more attention to the construction of credible appeals and 

affectiveappeals than rational sppeals. Rational appeals can be realized through transitions, code glosses, frame 

markers and endophoric markers; ethos through hedges, boosters, self-mentions and evidentials; affective appeals 

through engagement markers, attitude markers, hedges and self-mentions. (3) Metadiscourse items play an 

important role in The CEO’s efforts to weather the crisis. 

Keywords: metadiscourse, CEO-speak, crisis response, persuasion, rhetorical appeals 

1. Introduction 

It is widely known that U.S. authorities’ allegations of criminality into Huawei had been going on for years, on 

January 28, 2019, the US Justice Department released an indictment, including allegations on technology theft, 

sanctions breaches, lies to banks and conspiracy; in June 2018, the US refused to renew the export licence of 

Futurewei, Huawei’s Silicon Valley research and development unit; but they began to come to a head in December, 

2018 when Meng Wanzhou, the chief financial officer (CFO) of Huawei and daughter of Mr. Ren, was arrested in 

Vancouver on US charges. These series actions have inevitably posed a great threat to Huawei’s business and 

reputation. In face of this, Mr. Ren, who has always kept a low profile since the foundation of Huawei, presented 

frequently in media in an uncharacteristically way to respond to the doubts and inquiries of the media and the 

public.  

The phrase “CEO-speak”, cited from Amernic and Craig (2006, 2007), is the use of discourse by CEOs in speeches, 

statements during press conferences, interviews with the media and CEO messages in corporate annual reports. 

Many scholars agree on the argument that the words of CEO have “clout” (Amernic & Craig, 2006, p.4). Vignone 

(2012) believes that “leaders’ words sculpt reality”, which echoes the idea of Arsenault and Castells (2008) that 

CEOs play a strong role in helping “to shape the social world by exerting control over issue-framing and 

information gatekeeping”. Particularly, the influence of CEOs is important in times of crisis when they are 

expected to assume rhetorical leadership (Amernic & Craig, 2006), and CEOs’ role in weathering a crisis has 

attracted scholars’ attention in organization, communication as well as linguistics studies. Liu and Wang (2006) 

point out that the CEO of a company is the best spokesman and leader of crisis management team because they 

believe the practice can effectively express a company’s sincerity in the eyes of the public. Similarly, Seeger et al. 

(2005) believe that the CEO could become the face of the tragedy and a personification of the losses, and his high-

profile monologic response is instrumental in changing and saving the company.  

Therefore, it is important to investigate how the CEO-speak makes a difference during a crisis. However, the topic 

mostly receives attention from such disciplines as communication and organization, and few of them focused on 

the role of language used by the speaker of a company to respond to a crisis and restore its business (Riaz et al., 

2016). Moreover, a majority of prior studies focus on written data of CEO letters to stockholders in a company’s 

annual report (e.g. Hyland, 1998b), while oral discourse, such as interviews or speeches, receives less attention. 



hssr.ideasspread.org   Humanities and Social Science Research Vol. 8, No. 2; 2025 

 76 Published by IDEAS SPREAD 

 

As we know, oral activities are pervasive in business settings, like business negotiation and business meetings. 

They are worthwhile to pay more attention to. 

To contribute to filling these research gaps, metadiscourse can serve as a linguistic theoretical foundation to 

analyze how CEO-speak functions during a corporate crisis. The term is first coined by Zellig Harris in 1959 and 

it focuses on the non-propositional part of the content. Hyland (2008) argues that “metadiscourse is a key linguistic 

device that can help to show how language choices reflect different purposes of writers, the different assumptions 

they make about their audiences, and the different kinds of interactions they create with their readers”. Therefore, 

the present paper will adopt the interpersonal model of metadiscourse (Hyland, 2008) to investigate the linguistic 

features of Mr. Ren’s interviews.  

To summarize, the role of CEOs in dealing with a crisis is so essential for an organization, while previous studies 

on CEO-speak from the perspective of linguistics are still lacking. Therefore, the present study aims to identify 

the deployment of metadiscourse resources in Mr. Ren’s interviews with media based on the interpersonal model 

of metadiscourse. Moreover, Hyland (2008) maintains that metadiscourse is not only used to support the remaining 

propositional part. Rather, it plays an important role in making the latter more coherent, intelligible and persuasive 

to a particular audience since it can help writers to engage their audience, signal relationships, apprise readers of 

varying certainty and guide their understanding of a text. 

2. Literature Review 

2.1 Theoretical studies on metadiscourse 

Metadiscourse now is a popular area in discourse analysis, and according to Hyland (2008), generally it involves 

both cohesive and interpersonal features, combining a text and its context. The leading modern linguists Halliday 

(1994) argues that intentions to express certain meanings in specific situations play an important role in our 

interactions with others, and metadiscourse helps us understand the relationships between the unconscious 

language choices and the social contexts in which we make them. 

Due to diverse meanings realized by metadiscourse markers, many analysts (Crismore, 1989; Mauranen, 1993; 

Vande Kopple, 1985) have attempted and proposed different classification frameworks according to specific 

functions of metadiscourse. The paper will adopt the interpersonal model of metadiscourse developed by Hyland 

because it not only involves the functions of metadiscourse, but also embodies three key principles of 

metadiscourse (Hyland & Tse, 2004): “1) Metadiscourse is distinct from propositional aspects of discourse; 2) 

Metadiscourse refers to aspects of the text that embody writer-reader interactions; 3) Metadiscourse refers only to 

relations which are internal to the discourse”. At a finer degree of delicacy, the model incorporates the insights 

from Thompson and Thetela’s (1995) separation between interactive and interactional resources. The classification 

scheme is summarized in Table 1. 

 

Table 1. Interpersonal Model of Medadiscourse(Hyland,2008, p49) 

Category Function Examples 

Interactive  Help to guide the reader through the text Resources  

Code glosses Elaborate propositional meanings Namely; e.g. in other words 

Endophoric Markers Refer to information from other parts of the 

texts 

Noted above; see Fig.; in section 

2 

Evidentials  Refer to information from other texts According to X, Y states 

Frame Markers Refer to discourse acts, sequences or stages Finally; to conclude; my purpose 

is 

Transition Markers Express relations between main clauses In addition, but; thu, and 

Interactional 

Metadiscourse 

Involve the reader in the text Resources  

Attitude Marker Express writer’s attitude to proposition Unfortunately; surprisingly; I 

agree 

Booster Emphasize certainty or close dialogue In fact; definitely; it is clear that 

Self Mentions Explicit reference to author(s) I; we; my; our 

Engagement Markers Explicitly build relationship with readers Consider; you can see that 

Hedges Withhold commitment and open dialolue Might; perhaps; possible 

 

 



hssr.ideasspread.org   Humanities and Social Science Research Vol. 8, No. 2; 2025 

 77 Published by IDEAS SPREAD 

 

2.2 Aristotle’s Model of Rhetoric 

Apart from Interpersonal Model of Medadiscourse, the study will adopt Aristotle’s model of rhetoric to analyze 

how persuasion is achieved through metadiscourse resources. 

In Aristotle’s (2010, p.6) book Rhetoric, he illustrate that rhetoric is “the art of discovering in any giving case, the 

available means of persuasion” (Hyland, 2008, p.64). It involves demonstrating how something is true or how it 

can be shown to be true. Aristotle developed the first and most influential theory of persuasion, which includes 

three means of persuasion, that is, rational appeals, affective appeals, the credible appeals.  

As for the rational appeals, the writer or speaker needs to conduct logical reasoning or give examples to reach a 

common ground between himself/herself and the audience. It concerns the speech itself, its arrangement, length, 

complexity, types of evidence and arguments, etc. According to Hyland (1998b), by providing explanations, 

orientations and guidance for the audience, interactive metadiscourse can help the speaker or writer achieve the 

rational appeals. Hyland (2008) probes into rational appeals of CEO’s letters to shareholders, and proposed that 

transitions, code glosses, frame markers and endophoric markers were the metasdiscourse subcategories that 

contribute most to the CEO’s rational appeals. This discovery is also applied to many other studies in 

metadiscourse and persuasion (e.g. Jiang, 2019; Wang, 2013). In this way, the present study will take transitions, 

code glosses, frame markers and endophoric markers as rhetorical means to achieve logos. 

The credible appeals means the personal appeal of one’s character. Whether the persuasion is successful or not is 

closely connected with the speaker’s or writer’s capability to create an effective ethos. To this end, the perceived 

integrity and authority of the writer or speaker are crucial elements. In some sense, if the speaker is believed to be 

trustworthy, reliable, and competent, the audience are likely to accept what he or she says. According to Hyland 

(2008, p.78), the subcategories of engagement markers, boosters, hedges and evidentials can indicate authors’ 

assessments on truth and their convictions in the opinions, thus facilitating most to the realization of CEOs’ 

credibility appeals. Another strategy is the combination of self-mentions and boosters, which can help the speaker 

or writer build a positive, determined and confident image. The study will take the subcategories of hedges, 

boosters, self-mentions and evidentials as rhetorical means to realize credible appeals. 

The affective appeals means the appeal to emotions, which can be achieved by exerting emotional influence on 

the audience, whose education background, gender, age, interest, previous knowledge as well as communities need 

to be taken into consideration. Therefore, this means is more reader-centered than the two previous. With regard 

to the connection between pathos and specific metadiscourse categories, Hyland (2008) said that engagement 

markers, attitude markers and hedges, along with the manipulation of pronoun reference can contribute to the 

establishment of relationship of the two sides, thus facilitating the realization of affective appeals. Through an 

extensive reading on the data, it is found that the first-person pronoun I and the second-person pronoun you are 

frequently used. In this end, the paper will take engagement markers, attitude markers, hedges as well as self-

mentions to achieve affective appeals, as Jiang (2019) has done in investigating persuasive effects of metadiscourse 

in Chinese and American company profiles. 

3. Materials and Methods 

The research data are collected from the book, In His Own Words, Dialogues with Ren, consisting of three 

roundtable discussions and eight interviews with media of international profile coming from different countries. 

The content is based on transcripts of Huawei’s founder and CEO Mr. Ren’s interviews with 39 media in total 

from January to April, 2019. Its Chinese and the translated English version are public accessible from Xinsheng 

Community in Huawei’s official website. This study keeps Mr. Ren’s answers and excludes the journalists’ 

questions in the interviews of translated English version to get the research data. The practice can enable the 

present study to focus on the discourse Mr. Ren uttered despite the fact that it, to some extent, neglects the effect 

of context made by journalists’ questions on his answers. The translated version is selected for the reason that it is 

highly consistent with the original one in utterance interpretation and is officially admitted by Huawei. What’s 

more, the practical reason is that it is more convenient to process English words with corpus analysis tool compared 

with Chinese characters. All his answers constitute the small corpus of 56,807 English words. 

The current study conducted a quantitative analysis to identify the frequency and distribution of metadiscourse 

subcategories under the framework of Hyland’s (2008) interpersonal model of metadiscourse. He labelled 

comprehensive words under various metadiscourse subcategories and reminds that many items can express either 

interactive or interpersonal meanings. The present study will also take the list of words as primary searching items. 

To begin with, the author needs to prepare the data for further processing in corpus analysis tool AntConc 3.4, 

including downloading the transcripts of Mr. Ren from the internet, and revising them into “txt” for further 

calculation. Then, the data file will be input into AntConc 3.4., in which Hyland’s list of words will be input as 



hssr.ideasspread.org   Humanities and Social Science Research Vol. 8, No. 2; 2025 

 78 Published by IDEAS SPREAD 

 

searching items one by one to produce concordance lines. These lines will be checked manually to determine 

whether specific items in their contexts can be counted as metadiscourse or not. Moreover, it is obvious that the 

interviews of Mr. Ren took place in business setting and the raw data extracted in the present study is oral rather 

than written, which is quite different from the basis on which Hyland’s list of metadiscourse is formed. Therefore, 

the author will make a second check of metadiscourse devices through close reading of the data to figure out the 

words or phrases what are specific to business discourse and oral genres to generate a new list of metadiscourse 

devices. Followed by this, these items will be complemented to the list provided by Hyland to form a new list that 

is appropriate to the data of the present study. The author will also will employ AntConc 3.4 to process the newly 

identified items for their frequency. e only accept papers written in English and without orthographic errors. 

4. Results and Discussion 

4.1 Overall Distribution of Metadiscourse 

The quantitative analysis shows the importance of metadiscourse in Mr. Ren’s interviews with media in the context 

of crisis, with 12,407 occurrences in total and about one metadiscourse signaling every 4.6 words. The occurrence 

is higher than that of most previous researches, demonstrating the pervasiveness of metadiscourse in Mr. Ren’s 

discourse. Table 2 provides the frequency and percentage of the metadiscourse categories. 

 

Table 2. Distribution of metadiscourse categories 

 Frequency Percentage(%) 

Interactive metadiscourse 3923 31.6% 

Code glosses 295 2.4 

Endophoric Markers 40 0.3 

Evidentials  646 5.2 

Frame Markers 777 6.3 

Transition Markers 2165 17.4 

Interactional Metadiscourse 8484 68.4 

Attitude Marker 692 5.6 

Booster 1443 11.6 

Self Mentions 4235 34.1 

Engagement Markers 777 6.3 

Hedges 1337 10.8 

total 12407 100 

 

On the whole, Mr. Ren uses more interactional resources (68.4% of the total) than interactive resources (31.6% of 

the total). Interactional items occur more than two times as much as interactive ones. With regard to interactive 

metadiscourse, transition markers rank first in frequency, followed by frame markers (6.3%), and then evidentials 

(5.2%), code glosses, endophoric markers. In terms of interactional metadiscourse, self-mentions take up the lion’s 

share in the corpus, followed by boosters and hedges, and then engagement markers (6.3%) and attitude markers 

(5.6%). 

Self-mentions, transition markers, boosters and hedges rank the four most frequently-used metadiscourse 

subcategories in the corpus. Self-mentions significantly account for 34.1% of the total metadiscourse and half of 

all interactional forms. In addition, transition markers comprise more than half of the interactive counts. As Hinds 

(1987, p.147) maintained that the writer needs to use proper transition markers so as to help the audience 

understand his/her logic in organizing the composition. Next to transition markers are boosters and hedges, with a 

frequency of  11.6% and 10.8% respectively.  

Code glosses (2.4% of the total) and endophoric markers (0.3% of the total) of interactive resources are two devices 

that are most rarely used. Genre of public interviews needs to be taken into consideration. Compared with 

argumentative essays, Mr. Ren, in face of journalists’ questions, did not have enough time for preparation and may 

pay less attention to the organization of his information, thus leading to the lower frequency of interactive 

metadiscourse. 

4.2 Using Metadiscourse to Achieve Persuasion 

Metadiscourse can be closely connected with the purposes of speakers or writers, and persuasion can be completed 

through metadiscourse by contributing to rational appeals, credible appeals and affective appeals. 



hssr.ideasspread.org   Humanities and Social Science Research Vol. 8, No. 2; 2025 

 79 Published by IDEAS SPREAD 

 

Rational appeal is assumed to be established mostly through transitions, frame markers, endophoric markers as 

well as code glosses. Credible appeal is assumed to be mainly achieved by the deployment of evidentials, hedges, 

boosters and self-mentions. At last, self-mentions, hedges, attitude markers and engagement markers serve to 

realize affective appeal. 

 

Table 3. Persuasive functions of metadiscourse subcategories 

Functions  Metadiscourse  Frequency(n) Percentage(%) 

Rational appeal Code glosses 295 / 

Endophoric markers 40 / 

Frame markers 777 / 

Transtions  2165 / 

Total  3277 18.2% 

Credible appeal Evidentials  646 / 

Boosters  1443 / 

Self-mentions 4235 / 

Hedges  1337 / 

Total  7661 42.6% 

Affective appeal Attitude markers 692 / 

Self-mentions 4235 / 

Engagement markers 777 / 

Hedges  1337 / 

Total  7041 39.2% 

Total 17979 100% 

 

The distribution and frequency of the three means of persuasion is presented in Table 2. When code glosses, 

endophoric markers, frame markers and transitions are added together, the author gets the number of metadiscourse 

items contributing to rational appeal, which is 3,277. It is found that transition is the most frequently-used 

metadiscourse subcategory that contributes to the realization of rational appeal. In the same way, we can get the 

figure of credible and affective appeal, which is 7661 and 7041 respectively. Self-mentions play an important role 

in achieving credible and affective appeal. On the whole, the frequency of rational appeal is much lower than that 

of credible and affective appeal, which implies that in Mr. Ren’s interviews, he pays more attention to the 

construction of a credible persona and a close relationship with his audience. 

Furthermore, from the distribution presented in table 3, it is found that interactive metadiscoursal resources mainly 

contribute to the establishment of rational appeal, while interactional resources play a more important role in 

building credible and affective appeal. Meanwhile, it is notable that one metadiscoursal device is likely to perform 

over one function in practice. For example, it can be seen from table 3 that self-mentions and hedges often appeal 

to both credible appeals and affective appeals, which also explains why their frequency have been calculated twice 

in the analysis. 

5. Conclusions 

The present study probes into the use of metadiscourse in Mr. Ren’s answers of the public interviews, and explore 

how persuasion is realized through its use. Here are the major findings and limtations. 

Firstly, there are 12,563 metadiscourse items in the small self-built corpus of 56,778 words, and it is found that 

the proportion of metadiscourse items in the self-built corpus is larger than that of other researches focusing on 

the area, with one metadiscourse item every 4.5 words. Therefore, Mr. Ren’s utterances in the interviews are highly 

persuasive. Besides that, Mr. Ren uses more interactional resources (67.1% of the total) than interactive resources 

(32.9% of the total). Interactional items occur more than two times as much as interactive ones, indicating Mr. 

Ren’s attempts to be more evaluative and his concern for the audience. However, argumentative essays are quite 

different from public interviews, so the oral genre of the interviews should be taken into consideration. 

Furthermore, self-mentions and transitions are two most frequently used in the corpus, followed by boosters, and 

then hedges, while code glosses and endophoric markers occur the least frequently. 

Secondly, Aristotle’s three means of persuasion: rational, credible and affective appeal, are employed to analyze 

the rhetorical effects of the metadiscourse. Overall, the frequency of logos is much lower than that of credible 

appeals and affective appeals, demonstrating that in Mr. Ren’s interviews, he pays more attention to the 



hssr.ideasspread.org   Humanities and Social Science Research Vol. 8, No. 2; 2025 

 80 Published by IDEAS SPREAD 

 

construction of a credible persona and a close relationship with his audience. In Mr. Ren’s interviews investigated, 

rational appeals can be established mostly through transitions, frame markers, endophoric markers as well as code 

glosses. Credible appeal is mainly achieved by the deployment of evidentials, hedges, boosters and self-mentions. 

Attitude markers, self-mentions, engagement markers and hedges serve to realize affective appeals. Furthermore, 

it is found that the frequency of rational appeals are much lower than that of credible and affective appeals, which 

implies that in Mr. Ren’s interviews, he pays more attention to building a credible persona and a close relationship 

with his audience. Meanwhile, we conclude that in the present study, interactive metadiscoursal resources mainly 

contribute to the establishment of rational appeals, while interactional resources play a more important role in 

building credible and affective appeals. 

As to limitations of this study, to begin with, the scale of the self-built corpus consists of only 56,708 English 

words, which is relatively small. Consequently, the results obtained from this study might be limited. Then, the 

data we collected is translated version of the transcripts of Mr. Ren’s answers in the interviews with media. 

Although they are translated word by word exactly, we have to admit that there exist some differences between 

the transcripts and the English translated version, which may disrupt the research results. Meanwhile, despite the 

fact that we have made great efforts to ensure that all metadiscourse resources have been identified and classified 

into the according subcategories through a double check of the author’s supervisor, it is difficult to avoid missing 

and misclassifying some due to the contextual and multifunctional nature of metadiscourse. Also, the manual 

recognition of metadiscoursal items is likely to be influenced by the author’s subjectivity. 

There are following suggestions for future research, firstly, it is necessary to build a larger corpus. Secondly, to 

collect and analyze the reports of Huawei by the media involved after the interviews will, to some extent, help 

gain insights of whether Mr. Ren’s words have made a difference. Last but not least, it will be of great significance 

to conduct a cross-disciplinary research to further explore the connection between metadiscourse and corporate 

crisis responses in Mr. Ren’s discourse. 

Reference 

[1] Jiang, L. Y. (2019). 《中美企业简介中的元话语研究》 (Master’s thesis, Guangdong University of Foreign 

Studies).  

[2] Liu, Q., & Wang, G. (2006). 国外企业危机管理理论研究综述. 科学学研究(S1), 255-260. 

[3] Amernic, J. H., & Craig, R. J. (2006). CEO speak: The language of corporate leadership. Montreal: McGill-

Queen’s University Press.  

[4] Amernic, J. H., & Craig, R. J. (2007). Guidelines for CEO-speak: Editing the language of corporate leadership. 

Strategy and Leadership, 35(3), 25–31. https://doi.org/10.1108/10878570710757966 

[5] Annual report. (n.d.). Retrieved from https://www.huawei.com/cn/annual-report.  

[6] Aristotle. (2010). Rhetoric (W. D. Ross, Trans.). London: Oxford University Press.  

[7] Crismore, A. (1989). Talking with readers: Metadiscourse as rhetorical act. Peter Lang, New York. 

[8] Halliday, M. A. K. (1994). An introduction to functional grammar (2nd ed.). London: Edward Arnold. 

[9] Hyland, K. (1998a). Hedging in scientific research articles. Amsterdam: John Benjamins. 

Hyland, K. (1998b). Exploring corporate rhetoric: Metadiscourse in the CEO’s letter. Journal of Business 

Communication, 35(2), 224–245. https://doi.org/10.1177/002194369803500204 

Hyland, K., & Tse, P. (2004). Metadiscourse in academic writing: A reappraisal. Applied Linguistics, 25(2), 156-

177. https://doi.org/10.1093/applin/25.2.156 

[10] Hyland, K. (2008). Metadiscourse. London: Continuum.  

[11] Mauranen, A. (1993). Cultural differences in academic rhetoric. Frankfurt: Peter Lang. 

[12] Riaz, S., Buchanan, S., & Ruebottom, T. (2016). Rhetoric of epistemic authority: Defending field positions 

during the financial crisis. Human Relations, 69(7), 1533–1561. https://doi.org/10.1177/0018726716633335 

[13] Seeger, M. W., Ulmer, R. R., Novak, J. M., & Sellnow, T. (2005). Post-crisis discourse and organizational 

change, failure, and renewal. Journal of Organizational Change Management. 

https://doi.org/10.1108/09534810510615051 

[14] Vande Kopple, W. J. (1985). Some exploratory discourse on metadiscourse. College Composition and 

Communication, 26, 82-93. https://doi.org/10.2307/357915 

[15] Vignone, M. J. (2012). Family, buildings, and wars: Organizational conceptual metaphors. OD Practitioner, 



hssr.ideasspread.org   Humanities and Social Science Research Vol. 8, No. 2; 2025 

 81 Published by IDEAS SPREAD 

 

44(1), 34–37.  

 

Copyrights 

Copyright for this article is retained by the author(s), with first publication rights granted to the journal. 

This is an open-access article distributed under the terms and conditions of the Creative Commons Attribution 

license (http://creativecommons.org/licenses/by/4.0/). 


