A Corpus-Based Critical Discourse Analysis of News Reports on the COVID-19 Pandemic in China and the UK

Media, as important windows for the public to get to know timely information, play a vital role in influencing citizens’ attitudes as well as behaviors. From 2019, the Coronavirus Disease 2019 (COVID-19) pandemic, a global health emergency, has aroused great concern of the international community, including media. Varied in cultural context, political stand, and people’s ideology, however, media in different countries reported the COVID-19 dissimilarly. According to Fairclough’s critical discourse analysis (CDA) model, it is posited that the discrepancies in the reports of the COVID-19 can reflect ideological differences and have explanatory power in the development of the COVID-19 in distinct countries. Based on this premise, by utilizing the database analysis software AntConc 3.2.4w on self-built corpora, this study analyzed the news reports of different stages on the COVID-19 in China and the UK, i.e., in China Daily and The Guardian, respectively, and attempted to reveal the discourse characteristics in the two media, together with the discussion on their possible relations to the pandemic-controlling practices. The corpus-based analysis showed that China Daily used more objective and neutral words in the descriptions of the COVID-19 and expressed more active attitudes in fighting against the epidemic, whereas The Guardian used more negative words in describing the pandemic and words with weak restricting force when reporting policies concerning the control and prevention of the COVID-19 pandemic. Moreover, the comparison between the discourse before and after the lockdown demonstrated that the descriptions of the COVID-19 in the UK media transformed into a more objective and neutral one than before with an increased use of expressions of restriction and social conflicts. The same comparison in the discourse of China Daily found that words about sharing experience and promoting cooperation augmented noticeably. The above-mentioned findings were also discussed together with these two countries’ domestic epidemic situations and ideological differences, respectively.


Introduction
Ever since the outbreak of Coronavirus Disease 2019  pandemic, countries around the globe have adopted various measures to fight against the pandemic, among which sealing cities from the outside was a common practice, but the corresponding effect varied. Taking China and the UK as examples, ever since the outbreak of the COVID-19 pandemic, China and the UK have implemented mandatory lockdowns on April 8, 2020, and on March 23, 2020, respectively, there were significant differences in the effectiveness of epidemic prevention, as evidenced by the fact that China has quickly contained the spread of the epidemic within a few months, while the number of confirmed coronavirus cases in the UK has continued to increase. Such differences are related to many reasons, such as the early policies and administrative orders, social-economic disparity and vulnerability, mobility and social distancing, the availability of healthcare facilities, climate, economic issues, balancing the open data and privacy protection policies (e.g., , the difference in the levels of promulgated social distancing measures, as well as the difference at the time of promulgation among countries (e.g., Thu, Ngoc, Hai, & Tuan, 2020). Since the discrepancies between countries are noticeable Chen, Li, & Hu, 2015;Chen, Zhang, & Hu, 2020;Hu, 2014;Jing, Li, Wei, Yang, Chen, & Hu, 2018;Xiao, Liu, & Hu, 2019), which calls for multifaceted analysis. However, there are still insufficient studies that have analyzed how news, as an important way for citizens to obtain information and policies about the epidemic,

Literature Review
Since the onslaught of the COVID-19, news reports on the epidemic have increased exponentially. Linguists from various countries have been collecting COVID-19-related discourse and have built epidemic corpora to interpret the relationship between the progress of the epidemic and related discourse from a linguistic perspective (Debnath & Bardhan, 2020). In general, studies on COVID-19-related discourse can be divided into two strands. One of them is the discourse analysis of COVID-19 topics on various social platforms. Studies have found that there were significant differences in the discourse about the epidemic of different groups on social platforms. For instance, one study found that among the contents related to the COVID-19 posted by Arab twitters, topics related to religion and health were dominant and generally passive (Essam & Abdo, 2020), which was also the case in Malaysia that the majority of the online letters associated the COVID-19 with negative expressions (Joharry & Turiman, 2020). On the contrary, another similar study carried on African Americans found they held a positive attitude towards fighting against the epidemic (Odlum et al., 2020). The other strand is the analysis of news discourse of mainstream media in various countries. The consensus was reached that news discourse concerning the epidemic in various countries can reveal their distinctive ideological and cultural backgrounds. For example, a study on the headlines in eight newspapers from four countries found that the differences in the naming of the epidemic were related to ideological differences (Prieto-Ramos, Pei, & Cheng, 2020). It was proven that people's understanding of the concept of "influencer" during the epidemic was largely affected by regional and socio-cultural backgrounds (Abidin, Lee, Barbetta, & Miao, 2020). Meanwhile, analogous conclusions can be drawn by taking Chinese mainstream media as the research object. For instance, through a discourse-historical approach, it was found that Chinese media closely intertwined the arguments of globalism and nationalism (Yang & Chen, 2020). In addition, research on the use of hedge in the news about the epidemic in China revealed that the way that Chinese media reported on the epidemic reflected the determination and courage of Chinese citizens, as well as the great efforts that the Chinese government took for the control of the epidemic (Chen & Xie, 2020). From the above review, a conclusion could be drawn that there were significant differences in discourse related to the epidemic in the news media of different countries. As the pioneer country fighting against the epidemic, China has achieved remarkable success in epidemic prevention and control. Currently, when the number of infected people is still surging, learning lessons from Chinese epidemic-related discourse and making cross-country comparisons are of great significance to all the countries that are currently being affected by the epidemic. However, through the literature review, there is a dearth of studies comparing epidemic discourse between China and other countries (Zeng & Xie, 2020). Based on this, this study attempted to fill in this research gap by comparing the news discourse corpora of diverse epidemic development stages in China and the UK. By exploring the characteristics of news discourse related to the COVID-19 in two countries, this study aimed to explore the distinct focus and stances of the media in the two countries, as well as the ideological differences behind them.
Regarding the theoretical framework, critical discourse analysis was selected, which reveals how discourse is affected by ideology and power relations, and emphasizes the effects of discourse Chen, Yan, & Hu, 2019;Ding & Liao, 2001;Xiao, Li, & Hu, 2019). Among all the theories of critical discourse analysis, this study adopted the three-dimensional model proposed by Fairclough (1989Fairclough ( , 1995. This model regards discourse as a three-dimensional concept involving text, discourse practice, and social-cultural practice (Fairclough, 1989). In accordance, three steps need to be followed in the discourse analysis, i.e., "description", the "Keyword List" function in AntConc, which could perform a statistical comparison between the target corpus and the reference corpus, where the significantly high-frequency words in the target corpus relative to the reference corpus were listed from top to low according to the degree of difference, also referred as "Keyness" in AntConc. Because the comparative result of keyword list foregrounds the features of the target corpus different from the reference corpus, the statistical results of a single corpus in each pair of corpus groups serving as the target corpus and reference corpus have completely different research meaning. Consequently, three pairs of corpus groups in this study have been statistically processed twice. After the "description" of the keyword list, it was necessary to "interpret" the formation process of the utterance, specifically, the keyword list obtained in the "description" stage needed to be analyzed for its contextual meaning. For this purpose, this research conducted a qualitative analysis of each word in the keyword list through the function of "Concordance" and "File View" to examine the role of the words in the formation of the discourse. Based on this qualitative analysis, some keywords had limited semantic meaning in the research, such as functional words (e.g., "a", "the", and "it"), words on news and publication information (e.g., "Guardian", "com", and "Daily"), and words not related to the epidemic (e.g., "say" and "think"). These words were deleted leaving only 30 keywords for further "interpretation". Subsequent to the above two steps of "description" and "interpretation", this research stepped into the third stage of "explanation", which means to have an interpretation of the different discourse characteristics related to the social backgrounds of China and the UK, to reveal the hidden power, ideology, social and cultural factors in the social context.

Comparison Between the CMC and the UMC
Quantitative analysis results of the keyword list between the UMC and the CMC were presented (see Table 1), followed by qualitative analyses using the function of "Concordance" and "File View". uardian used th nment to call o did not express nment for vari 2020, "Howev an't afford to n t to go out, peo at the epidemi ds with more at, as one of t and the gover , unlike China one goal, such word "crisis" a it was found t "coronavirus c " in the UMC e eighth amon n the other han eason for The uld inevitably i e the society. nia" for referen between the media and the " reveal the cult es more balanc

Comparison Between the UMC1 and the UMC2
Quantitative analysis results of the keyword list between the UMC1 and the UMC2 were presented (see Table 2), followed by qualitative analyses using the function of "Concordance" and "File View". (1) Before the UK blockade on March 23, 2020, there were very few words closely related to the medical care in the top 30 rankings in terms of keyness, namely, "isolate", "confirmed" and "sick"; whereas, there were 11 words after the blockade, namely, "Floyd", "app", "PPE", "scheme", "hydroxychloroquine", "distancing", "care", "tracing", "study", "furlough" and "recovery". This result demonstrated that the UK government did not take the pandemic seriously before the lockdown as much as what the government did afterwards. Besides, words such as "police", "restriction", "distancing", "rules" and "furlough scheme" appeared more frequently after the lockdown while the number of times using the word "advice" had decreased. This result could imply that the UK government has implemented stricter pandemic control policies with increased supervision than before. Additionally, from keywords such as "tracing" and "app", it could be concluded that the UK has begun to use high-tech positioning technology to monitor the development of the epidemic. Keywords such as "PPE" (personal protective equipment), "hydroxychloroquine", and "study" also showed that the UK has put more emphasis on spurring the research on medical treatment of the COVID-19.
(2) Through the comparison before and after the blockade, significant differences in the use of "panic" in The Guardian were identified. Before the blockade, the keyness of "panic" was 1,190.650, ranking the eleventh, being the adjective with the highest keyness; whereas, after the blockade, the keyness of "panic" was significantly reduced. The rationale of this overuse of "panic" in the initial stage was similar to the above analysis on "crisis" which would lead to undesirable consequences. However, the transformation from high-frequency use of "panic" to low-frequency use proved that The Guardian's reports on the pandemic have been gradually becoming more objective and neutral. To illustrate the difference in the use of "panic" before and ijel.ccsenet.
after the b frequency second plo (3) By ana "BAME" through th number o discovery

Compa
Quantitativ followed b  Vol. 11, No. 2; A noticeable increase in the keyness of "experience" and "cooperation" was found comparing the CMC 1 and the CMC 2. A qualitative contextual study of these two words showed that "experience" mostly referred to China's experience in fighting the epidemic, aiming to share China's experience in effectively controlling the pandemic with other countries. Just as reported in this excerpt from China Daily on February 24, 2020: "lessons can be learned from China's experience to help all countries defeat this common enemy through awareness, responsibility, and prompt action". The word "cooperation" was often used to call on other countries to combine their efforts in fighting against the pandemic, which was also reported from the following excerpt on March 25, 2020, from China Daily: "Xi reiterated China's advocacy of boosting global cooperation in fighting the pandemic based on the vision of a community with a shared future for mankind." The frequent use of "experience" and "cooperation" is consistent with the Chinese government's stance that the international community should work together in response to the COVID-19. Cooperation is of utmost importance in today's community with shared destiny for mankind, because no matter which country it is, it cannot stand without the help and support from the international community (Chen & Hu, 2021).

Conclusion
Based on the coverage of the epidemic in different stages of the news media in China and the UK, this study utilized corpus linguistics software AntConc 3.2.4w to have a critical discourse analysis of self-built corpora by using quantitative and qualitative methods. By utilizing "Keyword List", "Concordance", "Collocates", "Concordance plot" and "File View" and other functions, this research had an in-depth analysis of the keywords with significant differences, thereby revealing the different language strategies adopted by the Chinese and UK media in reporting the epidemic and thus unveiling the ideological differences behind. The main findings of this study are listed as follows. First, by comparing the UMC and the CMC, it was found that Chinese media used more objective and neutral words to describe the epidemic and words used in how to deal with the epidemic showed strong morale; whereas the UK media described the epidemic in a comparatively negative manner. This finding reflected the differences between the "harmonious" and "conflict" discourse modes of the Chinese and UK media regarding emergencies, as well as the ideological differences between the two countries. Second, discourse characteristics of The Guardian and China Daily before and after the lockdown reflected the changing attitudes and actions toward the epidemic control between the two countries, which also indirectly revealed two countries' ideological differences. Through the comparison between two stages in China and the UK, an increasing use of "experience" and "cooperation" was observed, which indicated that China has been consistently sharing experience with other countries and calling for more international cooperation. Lessons can also be drawn from the comparison of the UK media discourse in two stages. More specifically, although the UK did not control the spread of the pandemic as effectively as China did after the blockade, a positive change of news discourse before and after the blockade in the UK media could still be found, with more attention paid to the medical prevention and physical restriction. This finding could play a positive role in inspiring other countries struggling with the pandemic to reconstruct their news discourse following the positive transformation that the UK media has undergone. An undesirable outcome of the pandemic, however, was also inferred in the discourse of UK media in the second stage, that words related to social conflict increased noticeably, so more attention and solid measures should be taken to solve these problems.
By comparing the news corpus of the two major news media, China Daily and The Guardian on the reports of the COVID-19 at different stages, the results of this research could inspire readers to interpret one specific issue in different stages from a critical perspective, thereby making readers aware of the influence of national ideology on the discourse of news reports. In addition, by comparing the methods of pandemic news reports in different countries, this study further advised the media to make full use of the positive effects of the discourse, to optimize the discourse strategy use in reporting news on COVID-19. Experience sharing and international cooperation were also suggested in concordance with the stance of China to triumph in the battle with the pandemic for mankind.
Limitations of this study lie in the following perspectives. First, since the development of the COVID-19 in different countries was influenced by a multitude number of factors, and it is impractical to control all the other factors, so the findings of this study are highly reliant on the correlation rather than strict causation. Second, until now, the COVID-19 has not reached an end, so the discourse collection of this study is not the complete picture of the development of pandemic, and thus cannot disclose the discourse features to the fullest in this regard. To address the listed limitations, further investigations are correspondingly suggested. First, to mitigate the intervention of other factors, comparative studies could be initiated among different countries holding backgrounds similar. Second, an expanded time-line is suggested to generate a more in-depth analysis of the development of the COVID-19 in different stages.