The US-China battle over Coronavirus in the news media: Metaphor transfer as a representation of stance mediation

Abstract

Drawing upon a corpus approach to metaphor analysis, stance analysis, and Critical Discourse Analysis, the study analyzes different stances taken by the Chinese news outlet Global Times (GT) and the American The New York Times (NYT) in 2020 Coronavirus narratives to Chinese and English readers. The database includes all Coronavirus-related GT and NYT bilingual opinion articles in 2020, that is, 97 pairs from GT and 73 pairs from NYT which are comparable in Chinese and English tokens. Results show that the differences between GT and NYT in narrating the pandemic and the involved parties, that is, China and the US, are statistically significant with a moderate to strong effect size. The Lambda test of association demonstrates that the knowledge of metaphor transfer methods can significantly increase the correctness of attitudinal intensity prediction, which bears out metaphor transfer as a representation of stance mediation.

Keywords

CDA Coronavirus corpus international politics metaphor translation news translation stance mediation

Introduction

This study compares the positioning of stance on COVID-19 in opinion pieces in Chinese and American news outlets, through exploring metaphor transfer as a representation of stance mediation. The ubiquity of stancetaking in languages has received wide-ranging scholarly recognition, as Englebretson (2007) argues, ‘stancetaking is a pervasive activity which speakers engage in through the use of language’ (2007: 69). Nonetheless, the definition of stance is problematic due to its theoretical and methodological cross-pollination across domains as diverse as Anthropology, Social Psychology, Education, Sociology, and Discourse Analysis. This has resulted in similar, but slightly differently nuanced concepts, such as Affect (Du Bois and Kärkkäinen, 2012), Attitude (Crosthwaite et al., 2017), Appraisal (Martin and White, 2005), Evaluation (Chiluwa and Ifukor, 2015), and Voice (Thompson, 1996), inter alia. In spite of the lack of an agreed-upon definition of stance across these disciplines, some scholars advocate an inclusive framework which ‘recognizes the heterogenous and multifaceted nature of stancetaking’ (Englebretson, 2007: 2) and treat all opinion-expressing discursive practices as stancetaking (Hunston and Thompson, 2000). On the whole, stancetaking is by no means in a state of stasis, inasmuch as ‘discourse positions in society are also in flux’ (Baker, 2006: 14). Importantly, the very fluidity of stance bears out the inevitability of its mediation in discourse. Mediation is a term in translation studies that is correlated with ‘manipulation’ and ‘rewriting’ and is more precisely defined as ‘the ways the translator intervenes, rewrites or manipulates in the [news] transediting process, with an effort to accommodate in the target text stances dissenting from those in the original text’ (Zhang, 2013: 398). Thus, translation is itself a process of altering the original author’s stance into one that may be different in the target context.

Relevant literature on stance mediation in multilingual reports has predominantly been focused on contexts of a highly politicized nature. This includes positioning of east and west in relation to policy in the Middle East (Jaber and Baumann, 2011), race relations impacting international embargos in South Africa (Frassinelli, 2019), political positioning in the dysfunct Weimar Republic (Lovett, 2019). Conflicting stance mediation seems particularly evident in Sino-European bilingual reporting in Asian political contexts, such as the unrest in Tibet, which has been labeled as ‘uprising’, ‘unrest’, and ‘rioting’ among others, depending on the source language of a report (Pan and Liao, 2020). Other examples include the South China Sea disputes (Wu and Zhang, 2015) and the China-US trade war (Qin and Zhang, 2020). Nevertheless, evidence for stance mediation can be found beyond political events, which in effect politicize them. As an ongoing global challenge, the COVID-19 pandemic, or Coronavirus pandemic, becomes part of a political conflict between China and the US, as news media with contrasting ideologies may take distinct narratives in reports. To the best of our knowledge, the inquiries noted above have not touched upon metaphor as a representation of stance and thereby the analysis of stance mediation through metaphor transfer is a dusty corner in academia. Since metaphor choice ‘is determined in part by the writer or speaker’s stance’, rather than a fair reflection of subject matter (Deignan, 2005a: 299), to analyze metaphor as a stance representation engenders a scholarly necessity.

Hence, the current study attempts to fill this gap by analyzing the stance mediation of Global Times and The New York Times via metaphor transfer in news headlines. In the sections that follow, this study clarifies how writers and editors for the Global Times and The New York Times both share commonalities and are disparate, highlighting the significance of comparing metaphor in their Coronavirus news reports. Our analytical model that combines corpus, stance, and Critical Discourse Analysis approaches, together with the dataset compiled in this study are elaborated in detail, followed by a discussion of the analysis results and concluding remarks.

Global Times versus The New York Times: Common ground and differences

Global Times (GT) is a daily tabloid newspaper owned by the Chinese Communist Party’s People’s Daily. It is ‘China’s Fox News’ (Brady, 2015) and a ‘foreign mission’ listed by the United States Department of State (Ruwitch and Kelemen, 2020). Some Western scholars rebuke it for spreading unfounded Coronavirus conspiracy theories and misinformation (Molter and DiResta, 2020). Following the emergence of Coronavirus terminology identifying it as Chinese in American mass media, such as the ‘China virus’, ‘Wuhan virus’, or ‘Kung flu’, the Global Times began in March 2020 to release bilingual opinion articles to counter the negative associations and misinterpretations that might arise from such positioning of the global pandemic in the Western media. GT has an English official website (www.globaltimes.cn) where most opinion articles are directly translated from its Chinese official website (www.huanqiu.com).

The New York Times (NYT) is an American daily newspaper owned by the listed The New York Times Company. It has won more than 100 Pulitzer Prizes, notable for its excellence in journalism. In June 2012, NYT introduced its official Chinese variant, cn.nytimes.com, where readers can view Chinese, English, or bilingual opinion articles. Its Chinese versions are translations of the English reports on www.nytimes.com. NYT has been blocked in Mainland China intermittently and it has a Democratic partisanship (Puglisi, 2011). In mid-March 2020, the Chinese government announced to expel NYT reporters from Mainland, China, Hong Kong, and Macau as a countermeasure to the United States Department of State’s ‘foreign mission’ of censorship, targeting five Chinese news agencies including Global Times (The Information Department, 2020).

Although GT is a Chinese state-run news agency, whereas NYT is an American private company, both face similar (intermittent) censorship in the target culture due to an ‘ideological affinity’ to the party (Valdeón and Calafat, 2020).¹ They take a sharp tone in opinion articles and regard the Coronavirus pandemic as a political instrument. Moreover, they respond to each other’s claims in news articles, thus constructing a communicative relationship between them.² More importantly, both use metaphorical expressions in news writings. According to Olza et al. (2020), WAR metaphors on Coronavirus are pervasive and problematic, leading to an emergency of promoting non-WAR metaphors. The authors found astonishing metaphors beyond WAR metaphors when reading through GT and NYT reports that may shape the world’s understanding of the pandemic and the roles played by China and the US. Thus, to compare GT and NYT’s stance mediation for Chinese and English target readers in the storytelling surrounding Coronavirus is of vital significance.

Analytical framework

The current study adopts a tool-kit approach, thus drawing upon the complementary strengths from corpus approach to metaphor analysis, stance analysis, and Critical Discourse Analysis.

Corpus approach to metaphor analysis

In Conceptual Metaphor Theory (CMT), metaphors are ways of thinking that involve conceptual mapping with the formula A IS B, where the target domain (A) is understood through a source domain (B). Hence, through this direct identification of one thing as another, metaphor lays bare ideology (Boulanger, 2016), and with ideological revelation so too the politics are revealed (Bazzi, 2014; Khwaileh and Khuwaileh, 2012). Kövecses’s (2020) Extended Conceptual Metaphor Theory (ECMT) goes a step further to advise a multilevel view of metaphor, by assigning a corpus approach to the analysis of the less schematic and more contextualized mental spaces (p. 88). Although scholars in antiquity regard [linguistic] metaphor as ‘a device of the poetic imagination’ or ‘the rhetorical flourish’ (Lakoff and Johnson, 2003: 3), modern metaphor researchers posit that linguistic metaphors are the linguistic realizations of conceptual mapping, establishing the connection between conceptual and linguistic metaphors as ‘one of top-down instantiation from thought to language’ (Cameron et al., 2009: 68).

Metaphor identification typically follows three steps (Deignan, 2005b; Shuttleworth, 2017; Stefanowitsch and Gries, 2007): first, to identify ‘metaphor keywords’ manually in samples; second, to search ‘metaphor keywords’ automatically based on concordance, collocate or semantic prosody in large corpora; third, to decide metaphoricity based on MIP (the Metaphor Identification Procedure, Pragglejaz Group, 2007) or the improved version MIPVU (the Metaphor Identification Procedure Vrije Universitei, Steen et al., 2010). The steps involved in MIPVU are detailed as follows:

First, read the source text (ST) and target text (TT) on a word-by-word basis to establish a general understanding of the meaning.

Second, determine the lexical units. The author adopts the ICTCLAS-NLPIR system (Zhang, 2014), a natural language processing system, to segment Chinese lexical units and manually checks them. Lexical units in English are words separated by spaces.

Third, establish the meaning of each lexical unit in context, through the use of dictionaries. This means that for each lexical unit, its contemporary or colloquial relevance to particular contexts needs to be determined, which contributes to highlighting its uniqueness within the data.

Fourth, if its meaning is contemporary, decide whether the contextual meaning contrasts with the basic meaning but can be understood in relation to it.

Fifth, if its new contextual interpretation can be understood, mark the lexical unit as metaphorical.

After metaphor identification, the authors follow the first standardized corpus-based Source Domain Verification Procedure (SDVP, Ahrens and Jiang, 2020: 47) to identify the corresponding source domains. The authors made final decisions on metaphor recognition after going through Suggested Upper Merged Ontology (SUMO, www.ontologyportal.org), WordNet (wordnetweb.princeton.edu/perl/webwn), Handian Dictionary (www.zdic.net)/Macmillian Dictionary (www.macmillandictionary.com), and the Word Sketch function in Sketch Engine (www.sketchengine.eu).

Once important metaphors in the ST and TT had been identified, it was convenient to tease out the contours of metaphor transfer, namely how ST metaphors are retained, replaced, paraphrased, and omitted in TT. A further step in following the contours of metaphor transfer included tracing how TT metaphors had been created from scratch or from non-metaphorical expressions in their STs.

Stance analysis

As noted in the Introduction section, stance is an umbrella term for any representation of the speaker’s positioning, thus including ‘attitude’, ‘affect’, ‘voice’, and ‘appraisal’ whose research scope is relatively narrow. Although Bednarek (2006) proposes to use ‘evaluation’ as the cover term, this study still follows Englebretson’s (2007) suggestion of using ‘stance’ because the latter is less polysemous and less ambiguous. Meanwhile, in line with Hunston and Thompson (2000), the study also uses the phrase ‘stance evaluation’ alternatively to keep the syntactic and morphological flexibility of the propositions when needed. Despite the variation in terminology, the literature on stance largely distills into two questions: first, the categorization of stance; and second, the grammatical or lexical representation of stance. Martin and White (2005) adopt Appraisal Theory to categorize stance into three dimensions, namely, whether they convey positive or negative textual meanings (i.e. attitude), the intensity or directness of attitudinal utterances (i.e. graduation), and the resources of dialogistic positioning (i.e. engagement). By contrast, Thompson and Hunston (2000: 25) employ a fourfold categorization system, which includes good-bad (positive, negative, neutral), certainty (certain, uncertain), expectedness (expected, unexpected), and importance (important, unimportant). Nevertheless, the most striking features of the datasets for the current study are the direct judgments of good or bad, and the intensity of such judgments when transferred into TT. Hence, neither Martin and White’s (2005) three-dimensional Appraisal Theory nor Thompson and Hunston’s (2000) four-fold categorization suits the current research. The analytical parameters adopted here for the evaluation of stance recognize a matrix of interaction between Attitude and Intensity with negativity-neutrality-positivity (see Table 1). These dimensions have some overlaps with Appraisal Theory (e.g. affect in Attitude and force in Engagement) and Thompson and Hunston’s (2000) good-bad dimension but in a simplified manner. The simplified parameters are chosen because, first, they are mostly related to the research purposes and, second, they are less subjective making them more operationalizable as ‘the more fine-grained the analysis is, the more problematic and subjective classification choices become’ (Fuoli, 2018: 237).

Table 1.

Analytical parameters for stance evaluation (adapted from Thompson and Hunston, 2000; Martin and White, 2005).

Good-bad attitude	–	Negative (Ng)	Neutral (Nt)	Positive (P)
		Ng-Nt	Nt-Ng	P-Ng
		Ng-P	Nt-P	P-Nt
	–	Ng-Ng	Nt-Nt	P-P
Intensity	strengthening (St)	Ng-St	Nt-St	P-St
	weakening (Wk)	Ng-Wk	Nt-Wk	P-Wk
	Unchanged (Un)	Ng-Un	Nt-Un	P-Un

The grammatical and lexical representation of stance is another important line of inquiry. The majority of scholars follow Conrad and Biber’s (2000) benchmark work of investigating adverbial markers as indicative of stance and positioning. Yet, the sub-language of news headlines are also a site for stance evaluation (Hunston and Sinclair, 2000), albeit receiving scant attention. The present study calls for a transition from grammatical stance marker research to exploring the lexical representation of stance mediation, specifically the analysis of metaphors in news headlines. It is true that metaphors imbibe a simultaneous capacity to both hide and highlight meanings. Yet this potency for hidden and overt positioning through metaphor has not yet been researched in scholarship on stance representation or stance mediation (McEntee-Atalianis, 2013).

Such an underrepresentation is even more obvious when analyzing metaphor transfer, as occurs in translation. In Translation Studies, metaphor transfer methods are multifold, and the most complete categorization is by Toury (2012). His approach exhausts the following six possibilities: retaining the same metaphor (M-M), substituting the original metaphor with a new metaphor (M1-M2), paraphrasing the original metaphor (M-P), omitting the original metaphor (M-O), creating a new metaphor (O-M), and from non-metaphorical to metaphorical expressions (P-M) in TT. This sixfold metaphor transfer approach lays bare any stance mediation that might need further clarification.

The Critical Discourse Analysis

Critical Discourse Analysis (CDA) is a branch of linguistics concerned with power inequality, identity, and ideology (Fairclough, 1995). It is debatable whether CDA is a research method. In siding with Van Dijk (2008) who regards CDA as a sociopolitical paradigm, Marchi and Taylor (2018: 2) argue that CDA is more like a paradigm which draws upon diverse methods of analysis. A CDA-based analysis construes meanings from three levels, that is, text, discourse practice, and sociocultural practice (Fairclough, 1995). Thus, it takes a holistic approach to describing, interpreting, and explaining textual features from linguistic, social, political, and ideological realities and perspectives. Research has shown the close-knit relation between stance and CDA (Ehineni, 2014; Ho and Crosthwaite, 2018) and the mutually beneficial collaboration between CDA and corpora (Baker, 2006; Taylor and Marchi, 2018).

The analytical framework adopted in the current study combines the insights of CDA with the corpus and the stance evaluation. Figure 1 illustrates our analytical flowchart.

Figure 1.

Flowchart of the analytical framework.

Research data

The present study has collected all COVID-related bilingual opinion articles published on GT and NYT in 2020. GT published 325 English opinion articles on its official English website (www.globaltimes.cn/opinion/editorial) in 2020. Among them, 133 contain the keywords ‘COVID’, ‘coronavirus’, ‘pandemic’, ‘epidemic’, or ‘pneumonia’ in title or content. By referring back to GT’s official Chinese website (opinion.huanqiu.com/editorial), the authors found that 97 of them have corresponding Chinese versions. Thus, this study has collected 97 bilingual Coronavirus opinion articles from GT. NYT published 221 opinion articles on cn.nytimes.com/opinion in 2020. The authors manually sifted through them and found 121 Coronavirus-related Chinese opinion articles. However, only 73 of them have the corresponding English versions. Hence, this study has collected 73 bilingual Coronavirus opinion articles from NYT. Only news headlines are considered in the study for three reasons: first, the sub-language of news headlines are an important site for stance evaluation (Hunston and Sinclair, 2000); second, they are an excellent resource for metaphor analysis (see Shie, 2011); and third, they usually summarize the main propositions of the news texts.³ Table 2 illustrates the compiled datasets.

Table 2.

Research data from GT and NYT’s bilingual opinion articles in 2020.

	The number of bilingual opinion articles	Total word count of English headlines	Total word count of Chinese headlines
GT	97	775	1622
NYT	73	768	1449

Results

Conceptual metaphors in source language construal in GT and NYT news headlines

Based on the proposed analytical framework, this study analyzes conceptual metaphors that GT and NYT have used in news headlines to describe the Coronavirus pandemic, China, and the US in their Chinese and English opinion articles. Table 3 summarizes the top ten source domains to map the target domain ‘Coronavirus’ based on frequency count. Table 4 displays the corresponding χ² test of independence with α = 0.05 as the criterion for significance.

Table 3.

Top 10 source domains for ‘Coronavirus’ based on frequency in the datasets.

Rank	(GT) Coronavirus is__	ST (Chinese)	TT (English)	Rank	(NYT) Coronavirus is__	ST (English)	TT (Chinese)
1	War	20	19	1	War	5	8
2	Crime	3	2	2	Problem	3	4
3	Marathon	2	0	3	Killer	3	1
4	Shock wave	1	1	4	Opportunity	2	2
5	Sea wave	1	0	5	Risk	2	2
6	Bomb explosion	1	0	6	Fire	0	3
7	Flood	1	0	7	Test	1	2
8	Test	1	0	8	Chernobyl crisis	1	2
9	Competition	1	0	9	Chinese product	1	1
10	Bad weather	0	1	10	Political scheme	1	1

Table 4.

Chi-Square test of independence between source domain for ‘Coronavirus’ and news outlets in ST.

	Value	df	Asymptotic significance (two-sided)
Pearson Chi-Square	186.231^a	54	0.000
Likelihood Ratio	133.085	54	0.000
Cramér’s V	0.942		0.000
Number of valid cases	70

Seventy cells (92.1%) have expected count less than 5. The minimum expected count is 0.01.

The χ² test of independence (χ² = 186.231, df = 54, α < 0.001) illustrates that there is a significant difference between GT and NYT in terms of the source domains for ‘Coronavirus’ with a strong effect size⁴ (V = 0.942, α < 0.001) in ST. Both have depicted the pandemic as a global ‘war’ and a ‘test’ for governments, thereby highlighting the magnitude of the pandemic’s severity. On the one hand, GT’s ST headlines compared the pandemic to natural disasters, such as, ‘sea waves’, ‘floods’, and ‘bad weather’. These natural disaster metaphors are not carried over into the English translations.

On the other hand, they also represent Coronavirus as a ‘crime’, implying the existence of a ‘criminal’ (see Example 1 below) and a ‘marathon’ or ‘competition’ between world powers, highlighting the politicization of the pandemic.

Example 1⁵. (CM formula: CORONAVIRUS IS CRIME; US IS SUSPECT)

ST:⁶ 接受国际督查, 美国的实验室应是第一拨

jiēshòu guójì dūchá měiguó de shíyànshì yīng shì dìyī bō

receive international investigation US AUX laboratory should be first patch

TT: US should make bio-labs more transparent (GT, May 14, 2020)

In stark contrast, headlines of the opinion pieces in NYT conceptualize the pandemic as a ‘killer’, a ‘Chernobyl crisis’ (see Example 2 below), a ‘Chinese product’ and a ‘political scheme’ in the STs. Hence, NYT headlines take a sharp tone in staging the Coronavirus narratives, barely hiding their anti-China sentiment for Chinese readers.

Example 2. (CM formula: CORONAVIRUS IS CHERNOBYL CRISIS)

ST: Trump’s Chernobyl (NYT, March 13, 2020)

TT: 特朗普的切尔诺贝利时刻

tèlǎngpǔ de qiēěrnuòbèilì shíkè

Trump AUX Chernobyl time

GT and NYT hold evidently contrasting positions to parties involved in the pandemic (see Table 5). In GT metaphor choices, China is a ‘hero’ (see Example 3), ‘champion’, ‘winner’, and ‘humble man’ in the virus fight, whereas the US is a ‘terrorist’, ‘actor’, ‘sinner’, ‘liar’, ‘criminal’, and ‘war maker’. Nevertheless, the NTT English headlines paint both China and the US negatively, with China as ‘dictator’, ‘insurgent’, ‘troublemaker’, and ‘tyrant’ (see Example 4, where out-China China is from the notion to out-Herod Herod, meaning to out-do, or exceed, even the icon of tyranny), and America as the land of ‘death’, ‘denial’, and ‘chaos’ (see Table 5).

Example 3. (CM formula: CHINA IS WAR HERO)

ST: 白皮书记录了可歌可泣的中国抗疫

báipíshū jìlù le kěgēkěqì de zhōngguó kàng yì

white paper record AUX deeply moving AUX China fight pandmeic

TT: Paper documents China’s heroic response (GT, June 7, 2020)

Example 4. (CM formula: CHINA IS TYRANT)

ST: America, Don’t Try to Out-China China (NYT, September 2, 2020)

TT: 美国, 不要学中国打民族主义牌

měiguó búyào xué zhōngguó dǎ mínzúzhǔyì pái

America don’t learn China play nationalism card

There were a range of political positions being mediated by GT and NYT source texts such as both regarding China as a ‘political tool’ for Trump’s administration to win the 2020 Presidential election and NYT commenting on Sino-US relations as ‘allies’, ‘war enemies’, and ‘spouses’ (see Table 5). Yet overall, GT was less favorable toward US than NYT was toward China. The next section will flesh out metaphor transfer in detail.

Table 5.

Source domains for ‘China’ and ‘the US’ based on frequency in GT and NYT.

Rank	(GT) China is__	ST	TT	(NYT) China is__	ST	TT	Rank	(GT) US is__	ST	TT	(NYT) US is__	ST	TT
1	Hero	2	2	US’s political tool	2	2	1	Terrorist	5	4	Land of death	1	1
2	Champion	1	0	Dictator	1	2	2	Actor	4	1	Land of denial	1	1
3	US’s excuse	1	0	US’s war enemy	2	1	3	Sinner	3	2	Land of chaos	1	1
4	US’s stepping stone	1	0	US’s ally	1	1	4	Burden	2	2	Child	1	1
5	Humble man	1	0	US’s spouse	1	1	5	Militant	2	1	Dying person	1	1
6	Sharpshooter	1	0	Prison	1	1	6	Liar	2	1	–
7	Suspect	1	0	Insurgent	0	1	7	Maniac	2	1
8	US’s political tool	0	1	Trouble maker	1	0	8	Tricker	0	3
9	Winner	0	1	Tyrant	1	0	9	Criminal	1	1
10	–			–			10	War maker	2	0

Metaphor transfer methods in GT and NYT news headlines

The current study adopts the sixfold metaphor transfer methods, that is, M-M, M1-M2, M-O, O-M, M-P, and P-M. However, the datasets barely adopt M-P or P-M methods, so only four methods are counted in the study. Figure 2 illustrates that GT and NYT processed metaphorical expressions using mixed methods. In other words, they kept (M-M), changed (M1-M2), omitted (M-O), and created metaphors (O-M) interchangeably in Chinese and English translated news headlines. Overall, GT and NYT adopt M-M method most frequently (35.66% vs 34.95%), thereby retaining the original images in most cases. However, while GT has a strong propensity to replace the ST metaphor with a new one in the TT (i.e. M1-M2, 27.91%), NYT tends to create new metaphors in the TT (i.e. O-M, 26.21%). Meanwhile, GT and NYT also omit metaphors in some cases (i.e. M-O). Nonetheless, Table 6 illustrates that their differences in metaphor transfer methods are not statistically significant (χ² = 5.436, df = 3, α > 0.005).

Figure 2.

Frequency count of metaphor transfer methods in GT and NYT TTs (Unit: %).

Table 6.

Chi-Square test of independence between metaphor transfer methods and news outlets.

	Value	df	Asymptotic significance (two-sided)
Pearson Chi-Square	5.436^a	3	0.143
Likelihood Ratio	5.429	3	0.143
Cramér’s V	0.153		0.143
Number of valid cases	232

Zero cells (0.0%) have expected count less than 5. The minimum expected count is 20.42.

Consider Example 5 and 6 (below), which illustrate headlines from both source and target GT texts. In Example 5, the US is ‘a morbid maniac’ and ‘a hegemonic power’ that ‘suppresses’ China in the Chinese original version. However, the corresponding English translated text omits these metaphors (i.e. M-O). The TT of Example 5 only maintains the metaphor, ‘the Coronavirus is war’ with the word ‘fight’. Therefore, this is an example of M-M, metaphor retention. Likewise, in Example 6, the US is a man who keeps the same pattern of behaviors in the Chinese ST but becomes a trickster in the English TT, thereby illustrating M1-M2, metaphor substitution. Moreover, the TT also reconceptualizes the Coronavirus epidemic as a ‘disaster’, which creates a new metaphor in TT, adopting a O-M method.

Example 5. (M-O, M-M)

ST: 病态, 华盛顿打压中国比抗疫还起劲

bìngtài huáshèngdùn dǎyā zhōngguó bǐ kàng yì hái qǐ jìn

morbid Washington suppress China compare fight pandemic more with gusto

TT: US fights harder against China than Coronavirus (GT, August 9, 2020)

Example 6. (M1-M2, O-M)

ST: 确诊 200 万, 华盛顿仍一套一套的

quèzhěn wàn huáshèngdùn réng yītàoyītào de

confirmed cases 2million Washington still keep same pattern AUX

TT: US tricks prevail despite epidemic disaster (GT, June 11, 2020)

On the other hand, Example 7 and 8 (below) show metaphor transfer from the NYT English originals into Chinese. In Example 7, the NYT regards Coronavirus as ‘a Chinese story’ in the English source version, implying China’s manipulation and dishonesty. However, in the corresponding Chinese target version, Coronavirus becomes China’s ‘pain’, ‘anger’, and ‘reflection’, thus playing down its anti-China sentiment. In Example 8, NYT Source Text does not directly comment on China and America, only praising Germany as a ‘role model’. Nonetheless, the TT reconceptualizes the US as ‘disordered’ and China as ‘hegemonic’, adopting an O-M strategy. Meanwhile, the TT does not mention Germany as a role model, illustrating a M-O strategy.

Example 7. (M1-M2)

ST: The Coronavirus story is too big for China to spin (NYT, February 18, 2020)

TT: 中国政府无法删除的痛苦、愤怒和反思

zhōngguó zhèngfǔ wúfǎ shānchú de tòngkǔ fènnù hé fǎnsī

China government cannot delete AUX pain anger and reflection

Example 8. (M-O, O-M)

ST: Germany’s Lessons for China and America (NYT, May 26, 2020)

TT: 除了失调的美国和强权的中国, 世界还有选择

chúle shīdiàode měiguó hé qiángquánde zhōngguó shìjiè hái yǒu xuǎnzé

besides disordered US and hegemonic China World still have choice

Hence, Example 6–8 demonstrate that both GT and NYT adopt mixed methods to mediate their stances in texts. In other words, they are equally aware of the necessity to modify metaphors with a view to expressing opinions to ST and TT readers.

Stance mediation in GT and NYT news headlines

Based on Table 1 produced above, the authors have codified good-bad attitude and intensity in news headline metaphors in both GT and NYT translations.

Figure 3 illustrates the percentages of good-bad attitudinal transfer from ST to TT toward the pandemic, China, the US, President Donald Trump, President Xi Jinping, and China-US relations in GT and NYT. The results show that over a half of GT news headlines express a negative-to-negative (Ng-Ng) opinion toward the US, whereas NYT holds relatively balanced negative attitudes toward the Coronavirus pandemic (20.27%), China (10.81%), the US (10.81%), and President Trump (17.57%).

Figure 3.

Frequency count of good-bad attitudinal transfer in GT and NYT news headlines (Unit: %).

In contrast to the anti-US and pro-China stance taken in GT translations, NYT does not avoid negative comments on its own government and President, when translated into Chinese. Nonetheless, NYT also openly criticizes Chinese President Xi Jinping and is pessimistic about China-US relations (4.05% of Ng-Ng attitudes), whereas GT holds a neutral-to-neutral stance toward the bilateral relation (2.20% of Nt-Nt attitudes). Table 7 demonstrates the χ² test between news outlets and their attitudes toward Covid-19, China, and the US. Results show that GT and NYT differ significantly in their attitudes toward Covid-19 (χ² = 20.719, df = 4, α < 0.001), China (χ² = 29.568, df = 6, α < 0.001), and the US (χ² = 11.019, df = 3, α < 0.001), with a strong to moderate effect size (V = 0.893, 0.947, 0.422 respectively, α < 0.001).

Table 7.

Chi-Square test of independence between good-bad attitudinal transfer toward Covid, China and US, and news outlets.

	Value (Covid, China, US)	df	Asymptotic significance (two-sided)
Pearson Chi-Square	20.719^a	4	0.000
	29.568^b	6	0.000
	11.019^c	3	0.012
Likelihood Ratio	20.609	4	0.000
	39.976	6	0.000
	8.246	3	0.041
Cramér’s V	0.893		0.000
	0.947		0.000
	0.422		0.012
Number of valid cases	26
	33
	62

Nine cells (90.0%) have expected count less than 5. The minimum expected count is 0.23.

Twelve cells (85.7%) have expected count less than 5. The minimum expected count is 0.48.

Six cells (75.0%) have expected count less than 5. The minimum expected count is 0.16.

To analyze how these metaphor-based propositions were strengthened, weakened, or kept unchanged in translation, the authors codified the datasets and plotted the results in Figure 4. Comparatively, GT has a strong propensity to soften its bitter tone toward the US (24.18% of Ng-Wk) and to play down its positivity toward China for its English readers (6.59% of P-Wk) while NYT both hides and highlights its negativity toward China for Chinese readers in different cases (6.76% of Ng-Wk and 6.76% of Ng-St). NYT headlines also mitigate the adverse influence of the pandemic via weakening its negativity toward Covid-19 (8.11% of Ng-Wk) and strengthening the positivity (1.35% of P-St). Nonetheless, statistical analysis in Table 8 shows that there are no significant differences between GT and NYT in the intensity of attitudes toward Covid-19 (χ² = 1.155, V = 0.235, df = 4, α > 0.005) and the US (χ² = 1.705, V = 0.170, df = 2, α > 0.005). On the contrary, the two news outlets hold contrasting attitudinal intensity toward China (χ² = 134.000, V = 1.000, df = 12, α < 0.001), which is partly attributed to the contesting good-bad attitudinal transfer that they hold.

Table 8.

Chi-Square test of independence between intensity of attitudes toward Covid, China and US, and news outlets.

	Value (Covid, China, US)	df	Asymptotic significance (two-sided)
Pearson Chi-Square	1.155^a	4	0.885
	134.000^b	12	0.000
	1.705^c	2	0.426
Likelihood Ratio	1.539	4	0.820
	118.029	12	0.000
	1.628	2	0.443
Cramér’s V	0.235		0.885
	1.000		0.000
	0.170		0.426
Number of valid cases	21
	67
	59

Eight cells (80.0%) have expected count less than 5. The minimum expected count is 0.05.

Eighteen cells (85.7%) have expected count less than 5. The minimum expected count is 0.16.

Three cells (50.0%) have expected count less than 5. The minimum expected count is 1.36.

Figure 4.

Frequency count of attitudinal intensity in GT and NYT news headlines (Unit: %).

Example 9. (M-O; P-P, P-Wk toward China)

ST: 无中国疫苗参战, 全胜病毒很难

wú zhōngguó yìmiáo cān zhàn quánshèng bìngdú hěn nán

no China vaccine join war completely win virus very difficult

TT: Hard to beat COVID-19 without China’s vaccine (GT, December 24, 2020)

Example 10. (M-M; Ng-Ng, Ng-Un toward China)

ST: Coronavirus spreads, and the world pays for China’s dictatorship (NYT, January 30, 2020)

TT: 疫情蔓延, 世界为中国的独裁统治付出代价

yìqíng mànyán shìjiè wéi zhōngguó de dúcáitǒngzhì fùchū dàijià

epidemic spread world for China AUX dictatorship pay cost

Example 11. (M1-M2; Ng-Ng, Ng-Wk toward the US)

ST: 疫情如此紧张, 华盛顿还在操弄政治！

yìqíng rúcǐ jǐnzhāng huáshèngdùn hái zài cāonòng zhèngzhì

epidemic so tense Washington still be manipulate politics

TT: Virus-plagued US still playing politics (GT, March 27, 2020)

Example 12. (M-M; Ng-Ng, Ng-Un toward the US)

ST: This land of denial and death (NYT, March 30, 2020)

TT: 美国, 否认与死亡之国

měiguó fǒurèn yǔ sǐwáng zhī guó

US denial and death AUX nation

GT and NYT hold contrasting attitudes toward China (see Example 9 and 10), though they hold the US in negative esteem (see Example 11 and 12). In Example 10 and 12, the negative ST metaphor of China as ‘dictator’ and the US as ‘a land of denial and death’ are retained in TT (M-M), thereby consistently keeping NYT’s negativity toward China and the US (Ng-Un). Hence, it stands to reason that attitudes are retained with the direct transfer method, M-M.

By contrast, the transfer method of M-O usually suggests a weakening of attitude intensity. This is evident in Example 9, where the ST metaphor of China as a ‘warrior’ is omitted in TT (M-O), thus playing down its positivity toward China (P-Wk attitude). The alteration of metaphor from one type to another in the M1-M2 strategy, can also soften negative attitudes expressed in the original. In Example 11, GT ST metaphor depicting the US as a ‘manipulator’ is transferred into the US as a ‘player’ in its TT. Thereby, the transfer methods M-O and M1-M2 hint at a weakening of attitudes.

To provide statistical evidence to these observations, Lambda, a simple measure of association between variables was utilized here, with metaphor transfer methods as the independent variable and the attitudinal intensity as the dependent variable (see Table 9). The results show that there is a strong⁷ association between metaphor transfer methods and the attitudinal intensity where the knowledge of metaphor transfer methods can increase 53.4% of correctness of attitudinal intensity prediction (λ = 0.534, α < 0.001). This strong association further corroborates that metaphor transfer is an important representation of stance mediation since attitudinal intensity of an utterance can be better predicted from metaphor transfer methods.

Table 9.

Lambda test of association between metaphor transfer methods and attitudinal intensity.

		Attitudinal intensity (frequency count)							Total
		Ng-St	Ng-Un	Ng-Wk	Nt-Un	P-St	P-Un	P-Wk	Total
Metaphor transfer method	M-M	2	45	1	11	0	4	0	63
	M-O	1	0	27	1	1	0	4	34
	M1-M2	20	2	13	3	3	1	2	44
	O-M	18	0	2	1	2	1	0	24
Total		41	47	43	16	6	6	6	165
		Value	Asymptotic standardized error^a	Approximate T^b	Approximate significance
Lambda	Attitudinal intensity as dependent variable	0.534	0.047	9.614	0.000

Not assuming the null hypothesis.

Using the asymptotic standard error assuming the null hypothesis.

Discussion and concluding remarks

This study combines corpus approach to metaphor analysis, stance analysis, and Critical Discourse Analysis to investigate how metaphor transfer can serve as a representation of stance mediation in GT’s and NYT’s Coronavirus narratives. The differences between GT and NYT in the metaphors used to depict Coronavirus, China, and the US in STs are statistically significant with a strong effect size. Where Metaphors are retained from source to target text, there is likewise a retention of attitude and intensity. What is similar about all source and target texts are their acknowledgment of the destructive impact of Coronavirus. In this, the headlines for GT and NYT and their translations are united. Only NYT and its translation represent the pandemic as having new opportunities. Meanwhile, the mediation of stance toward the Chinese and US governments is also transferred directly as ‘troublemaker’ and ‘culprit’, respectively. Despite the fact that both news outlets use mixed metaphor transfer methods, GT favors the replacement strategy in translation whereas NYT creates more new metaphors in its TTs. Nonetheless, these differences do not reflect a statistical significance.

Stance analyses illustrate that there is a strong association between metaphor transfer methods and the attitudinal intensity, which further corroborates metaphor transfer as a representation of stance mediation. GT and NYT, though unable to represent the whole news industry in China and the US, have an obvious partisanship toward the respective Chinese Communist Party and American Democratic Party, as can be seen from their negative attitudes toward each other and President Donald Trump. Taking a closer look, translation brings more social actors into the newspapers’ stancetaking, since stancetakers (news editors and translators) as well as the addressees of stance (Chinese and English readers) undergo some transition. However, no clues are given in the articles as to the linguistic and cultural identities of the translators of GT and NYT news. Yet, the native language and allegiances of both writers and translators will impact their orientation to the source and target cultures and, hence, their stancetaking when writing and translating the news. In this regard, the study proposes that newspapers practice a more transparent procedure of news production, or at least make the readers aware that they are reading a translation rather than an original article written by someone having a similar/different cultural background.

The analyses underpin Fairclough’s (1992: 64) view of discourse, including that of the translated news text, as having a dialectical relationship with social identities, social relationships, and systems of knowledge and belief. Metaphor, part of the vocabulary used by the two newspapers, their news production and distribution practices, and the context of the practices on the situational, institutional or societal level bear out the three dimensions of discourse in the CDA model, that is, text, discursive practices, and sociocultural practices. In effect, the study has exemplified the China-US battle of Coronavirus in news headlines via the tactics of metaphor transfer. Nonetheless, the study only analyzes news headlines and future research can consider the news texts to explore more nuanced metaphor usage. More meaningful conclusions on stance representation might be drawn with follow-up interviews with these newspaper editors, as this will inform more sophisticated insights into the news writers as well as the translators.

Footnotes

Acknowledgements

We would like to thank the anonymous reviewers for reading and commenting on this research. We also want to express our gratitude to Dr. Kathleen Macdonald for proofreading this article.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

Author biographies

Yufeng Liu is a PhD candidate at The Department of Chinese and Bilingual Studies, The Hong Kong Polytechnic University. Her research interests lie in the intersections of metaphor, translation, and communication studies.

Dechao Li is an associate professor at Department of Chinese and Bilingual Studies, The Hong Kong Polytechnic University. His main research areas include corpus-based translation studies, empirical approaches to translation process research, history of translation in the late Qing and early Republican periods and problem-based learning and translator/interpreter training. He has published over 50 articles in journals such as Perspectives: Studies in Translatology, The Translator and Interpreter Trainer, Interpreting, Discourse & Society, Target, Frontiers in Psychology as well as some book chapters published by Routledge, Springer, and Wayne State University Press.

References

Ahrens

Jiang

(2020) Source domain verification using corpus-based tools. Metaphor and Symbol 35(1): 43–55.

Baker

(2006) Using Corpora in Discourse Analysis. London and New York: Continuum.

Bazzi

(2014) Foreign metaphors and Arabic translation: An empirical study in journalistic translation practice. Journal of Language and Politics 13(1): 120–151.

Bednarek

(2006) Evaluation in Media Discourse: Analysis of a Newspaper Corpus. London and New York: Continuum.

Boulanger

P-P

(2016) When the media translate the crisis: The metaphor used by the mainstream press during the subprimes crisis. Meta 61: 144–162.

Brady

A-M

(2015) Authoritarianism goes global (II): China’s foreign propaganda machine. Journal of Democracy 26(4): 51–59.

Cameron

Maslen

Todd

, et al. (2009) The discourse dynamics approach to metaphor and metaphor-led discourse analysis. Metaphor and Symbol 24(2): 63–89.

Chiluwa

Ifukor

(2015) ‘War against our Children’: Stance and evaluation in# BringBackOurGirls campaign discourse on Twitter and Facebook. Discourse & Society 26(3): 267–296.

Conrad

Biber

(2000) Adverbial marking of stance in speech and writing. In. Hunston

Thompson

(eds) Evaluation in Text: Authorial Stance and Construction of Discourse. Oxford: Oxford University Press, pp.56–73.

10.

Crosthwaite

Cheung

Jiang

(2017) Writing with attitude: Stance expression in learner and professional dentistry research reports. English for Specific Purposes 46: 107–123.

11.

Deignan

(2005a) Review of the book Corpus Approaches to Critical Metaphor Analysis. Metaphor and Symbol 20(4): 295–301.

12.

Deignan

(2005b) Metaphor and Corpus Linguistics. Amsterdam: John Benjamins Publishing.

13.

Du Bois

Kärkkäinen

(2012) Taking a stance on emotion: Affect, sequence, and intersubjectivity in dialogic interaction. Text & Talk 32(4): 433–451.

14.

Ehineni

(2014) A critical discourse analysis of modals in Nigerian political manifestos. International Journal of Linguistics 6(3): 109.

15.

Englebretson

(2007) Stancetaking in discourse: An introduction. In. Englebretson

(ed.) Stancetaking in Discourse: Subjectivity, Evaluation, Interaction. Amsterdam: John Benjamins Publishing, pp.1–26.

16.

Fairclough

(1992) Discourse and Social Change. Cambridge: Polity Press.

17.

Fairclough

(1995) Critical Discourse Analysis: The Critical Study of Language. Harlow: Longman.

18.

Frankfort-Nachmias

Leon-Guerrero

(2018) Social Statistics for a Diverse Society, 8th edn. Los Angeles, CA: Sage.

19.

Frassinelli

(2019) Borders, Media Crossings and the Politics of Translation: The Gaze from Southern Africa. London and New York: Routledge.

20.

Fuoli

(2018) A stepwise method for annotating appraisal. Functions of Language 25(2): 229–258.

21.

SYE

Crosthwaite

(2018) Exploring stance in the manifestos of 3 candidates for the Hong Kong Chief Executive election 2017: Combining CDA and corpus-like insights. Discourse & Society 29(6): 629–654.

22.

Hunston

Sinclair

(2000) A local grammar of evaluation. In. Hunston

Thompson

(eds) Evaluation in Text: Authorial Stance and Construction of Discourse. Oxford: Oxford University Press, pp.74–101.

23.

Hunston

Thompson

(2000) Evaluation in Text: Authorial Stance and the Construction of Discourse. Oxford: Oxford University Press.

24.

Jaber

Baumann

(2011) The BBC World Service in the Middle East: Claims to impartiality, or a politics of translation? Journalism 12(2): 171–182.

25.

Khwaileh

Khuwaileh

(2012) The semantic confusing connotation of terrorism: Terminologies generated in Arabic and used in English. Babel 58(1): 19–30.

26.

Kövecses

(2020) Extended Conceptual Metaphor Theory. Cambridge: Cambridge University Press.

27.

Lakoff

Johnson

(2003) Metaphors We Live By. Chicago, IL: University of Chicago Press.

28.

Lovett

(2019) The politics of translation in the press: Siegfried Kracauer and cultural mediation in the periodicals of the Weimar Republic. Translation and Interpreting Studies 14(2): 265–282.

29.

McEntee-Atalianis

(2013) Stance and metaphor: Mapping changing representations of (organizational) identity. Discourse & Communication 7(3): 319–340.

30.

Marchi

Taylor

(2018) Introduction: Partiality and reflexivity. In. Taylor

Marchi

(eds) Corpus Approaches to Discourse: A Critical Review. London and New York: Routledge, pp.1–16.

31.

Martin

White

(2005) The Language of Evaluation. London: Palgrave Macmillan.

32.

Molter

DiResta

(2020) Pandemics & propaganda: How Chinese state media creates and propagates CCP coronavirus narratives. Harvard Kennedy School Misinformation Review 1(3): 1–24.

33.

Olza

Pérez-Sobrino

Koller

, et al. (2020) #ReframeCovid. Available at: https://sites.google.com/view/reframecovid/initiative?authuser=0 (accessed 20 April 2021).

34.

Pan

Liao

(2020) News translation of reported conflicts: A corpus-based account of positioning. Perspectives. Epub ahead of print 17 July 2020. DOI: 10.1080/0907676X.2020.1792519.

35.

Pragglejaz Group (2007) MIP: A method for identifying metaphorically used words in discourse. Metaphor and Symbol 22(1): 1–39.

36.

Puglisi

(2011) Being the New York Times: The political behaviour of a newspaper. The BE Journal of Economic Analysis & Policy 11(1): 1–43.

37.

Qin

Zhang

(2020) Taking mediated stance via news headline transediting: A case study of the China-US trade conflict in 2018. Meta 65(1): 100–122.

38.

Ruwitch

Kelemen

(2020) Trump Administration labels 4 more Chinese news outlets ‘Foreign Missions’. Available at: www.npr.org/2020/06/22/881755421/trump-administration-labels-4-more-chinese-news-outlets-foreign-missions (accessed 10 April 2021).

39.

Shie

(2011) Metaphors and metonymies in New York Times and Times Supplement news headlines. Journal of Pragmatics 43(5): 1318–1334.

40.

Shuttleworth

(2017) Studying Scientific Metaphor in Translation. Oxford: Taylor & Francis.

41.

Steen

Dorst

Herrmann

, et al. (2010) A Method for Linguistic Metaphor Identification: From MIP to MIPVU. Amsterdam: John Benjamins Publishing.

42.

Stefanowitsch

Gries

(2007) Corpus-Based Approaches to Metaphor and Metonymy. Berlin: Walter de Gruyter.

43.

Taylor

Marchi

(2018) Corpus Approaches to Discourse: A Critical Review. London and New York: Routledge.

44.

The Information Department (2020) China takes countermeasures against the US’s suppression of Chinese media agencies in the United States. The Ministry of Foreign Affairs of the People’s Republic of China. Available at: www.new.fmprc.gov.cn/web/fyrbt_673021/t1757128.shtml (accessed 10 April 2021).

45.

Thompson

(1996) Voices in the text: Discourse perspectives on language reports. Applied Linguistics 17(4): 501–530.

46.

Thompson

Hunston

(2000) Evaluation: An introduction. In. Hunston

Thompson

(eds) Evaluation in Text: Authorial Stance and Construction of Discourse. Oxford: Oxford University Press, pp.1–27.

47.

Toury

(2012) Descriptive Translation Studies and Beyond, revised edn. Amsterdam: John Benjamins.

48.

Valdeón

Calafat

(2020) Introduction: The politics of translation and the translation of politics. Translation & Interpreting 12(2): 1–6.

49.

Van Dijk

(2008) Discourse Reader. London: SAGE.

50.

Zhang

(2015) Translating political ideology: A case study of the Chinese translations of the English news headlines concerning South China Sea disputes on the website of www.ftchinese.com. Babel 61(3): 394–410.

51.

Zhang

(2014) ICTCLAS-NLPIR. Available at: http://ictclas.nlpir.org (accessed 10 March 2021).

52.

Zhang

(2013) Stance and mediation in transediting news headlines as paratexts. Perspectives 21(3): 396–411.