語言與政治立場：臺灣電視新聞之分析

全文

(1)國立台灣師範大學英語學系碩士論. 文. Master Thesis Graduate Institute of English National Taiwan Normal University. 語言與政治立場：臺灣電視新聞之分析. Language and Political Stances: An Analysis of the Taiwanese Television News. 指導教授：蘇席瑤 Advisor: Dr. Hsi-yao Su 研究生：蔡貴如 Student: Kuei-ju Tsai. 中華民國九十七年 8 月 August, 2008.

(2)

(3) 摘要. 本研究探討媒體報導中語言與政治立場之關係。以臺灣電視新聞為例，我們分析了東森新聞、民視新聞、三立新聞及中天新聞所報導的政治新聞。採用方法包括量化的內容分析(content analysis)及以語言為重的批判言談分析(critical discourse analysis)，並將此兩分析方法所得之成果用於驗證范‧迪克(van Dijk)於 1998 年所提出之意識型態方陣(ideological square)。我們以內容分析量化研究 2007 年 4 月 15 日到 5 月 9 日間，四家新聞台總共 100 集的新聞節目。並從中選出三則重大政治新聞，以批判言談分析作為個案研究。分析結果一致顯示不管在報導數量或報導語言方面，東森及中天皆傾向國民黨，而民視與三立則傾向民進黨。在內容分析中，我們發現新聞台對於政治新聞的處理深受其意識型態及政治立場所影響，而這些影響反映在新聞則數、新聞長度及新聞排序上。也就是說，新聞台對於所喜愛的政黨會給予其較多且較長的報導，並且這些報導會安排在節目一開始做為重點新聞，來凸顯其重要性。相反地，對於敵對的政黨，新聞台則給予較少且偏短的報導，並將其排序在節目後段。在批判言談分析中，我們分析了新聞報導語言中的宏觀以及微觀結構。在新聞的宏觀分析中，我們驗證了新聞台在呈現報導一個新聞事件時所提出的各項資訊，並非隨機出現，而是依據意識型態方陣經過縝密的安排。因此，對我們有利或對他們不利之資訊會獲得傳播，相反地，對我們不利而對他們有利之資訊則會受到相對的封鎖。不只宏觀架構如此，微觀的運用亦如是。根據我們對於新聞報導中語言銜接之分析，我們也發現了不論是連接詞(conjunction)或是字彙銜接 (lexical cohesion)，皆被媒體人操弄用來建構一個符合他們意識型態及政治立場的報導。因此，本研究證實了不論是在報導數量上或是報導語言上，一個新聞台的政. I.

(4) 治立場皆顯著地影響了其政治新聞，亦即驗證了范‧迪克之意識型態方陣：對自己有利而對他人不利之訊息會被呈現，相對應的是對自己不利而對他人有利之訊息會被壓抑。. II.

(5) ABSTRACT. This study aims to investigate the relations between language and political stances in the press. Four TV news stations in Taiwan were selected: ETTV News, FTV News, SET News, and CTi News. Both quantitative content analysis and critical discourse analysis were conducted to testify the ideological square proposed by van Dijk (1998). A total of 100 programs, from April 15 to May 9, 2007, were videotaped and quantitatively measured using content analysis. Three political events were further under critical discourse analysis as case studies. Results of the analyses revealed that ETTV News and CTi News were prone to the KMT, while FTV News and SET News were prone to the DPP. In the content analysis, it is found that a station’s political inclination biases its arrangement of political news in terms of news item, news duration, and news appearing order. Therefore, events congruent with the stations’ stances receive more reports of longer duration and appear earlier in a program. Contrarily, events that go against the stations’ ideologies receive fewer and more sketchy reports which come later in a program. In the critical discourse analysis, both macro- and micro-structures of the reports were examined. The analysis of news schema has manifested that the presentation of information by news stations is not random, but tactically organized according to the ideological square. Hence, information positive about us or negative about them would be expressed, while information positive about them or negative about us is suppressed. In the micro-analysis, it is shown that cohesive devices, including conjunction and lexical cohesion, are also manipulated by journalists to construct a world that best serves their ideologies and political stances. In sum, the study has demonstrated that preferred stances and ideologies are III.

(6) advocated by news stations in terms of quantity of news coverage as well as language of news reporting, whereas unwelcome stances and ideologies are suppressed both quantitatively and linguistically.. IV.

(7) ACKNOWLEDGEMENTS. It takes more than one man’s effort to complete this thesis. With gratitude and pleasure, I would like to acknowledge many generous people’s assistance and support. First and foremost, I want to thank my advisor, Dr. Hsi-yao Su, for her expertise, inspiration, and patience. Her understanding and encouragement led me through all the troubled time in the last year. This paper would not have been possible without her generous guidance. My gratitude also goes to the committee members, Dr. Tammy Miao-Shia Chang and Dr. Sai-hua Kuo, for their insightful comments and valuable suggestions undoubtedly aided me in improving my thesis. I also want to express my gratitude to all the professors who have instructed me throughout my student years at the Department of English, National Taiwan Normal University. It is due to their tireless instructions and enthusiasm that I come to appreciate the beauty and profundities of linguistics. Also I would like to thank all my classmates—Angeline, Cindy, Claire, Harvey, Kyle, Laney, Mike and Natasha—not only for the cheerful and precious moments that we have shared, but also for their readiness to help. These brilliant young linguists have truly lighted up my graduate life. In addition, special thanks go to my supportive roommate, Jill, who has lived a joyful dorm life with me in the past two years. I am definitely indebted to my beloved family. Their unconditional encouragement and warmth, as well as their trust, offer me the strength to move on. Last, I am heartily grateful to my boyfriend, Matt Ho, who not only stood beside me throughout the whole writing process, but, more importantly, gently tolerated my temper in my time of distress. Thank you all. It is you that make this thesis possible.. V.

(8) VI.

(9) TABLE OF CONTENTS. CHINESE ABSTRACT…………………………………………………………….... ENGLISH ABSTRACT…………………………………………………………….... ACKNOWLEDGEMENTS………………………………………………………….... TABLE OF CONTENTS……………………………………………………………... LIST OF TABLES…………………………………………………………………... LIST OF FIGURES…………………………………………………………………... I III V VII IX XI. CHAPTER ONE INTRODUCTION…………………………………………………. 1.1 Motivation………………………………………………………………….... 1.2 Ideology……………………………………………………………………… 1.3 Political Environment in Taiwan…………………………………………...... 1.4 Scope and Goal………………………………………………………………. 1.5 Organization of the Present Study…………………………………………..... 1 1 2 3 4 6. CHAPTER TWO LITERATURE REVIEW…………………………………………... 2.1 Content Analysis……………………………………………………………... 2.2 Critical Discourse Analysis………………………………………………….. 2.2.1 Theoretical Background……………………………………………… 2.2.1.1 Flower and His Colleagues…………………………………..... 2.2.1.2 Fairclough……………………………………………………... 2.2.1.3 van Dijk……………………………………………………….. 2.2.1.4 A Summary of Critical Discourse Analysis…………………… 2.2.2 Empirical Studies……………………………………………………... 7 7 11 11 11 15 18 23 24. CHAPTER THREE METHODOLOGY………………………………………………. 3.1 Data Collection………………………………………………. ……………... 3.2 TV News……………………………………………………………………... 3.3 Data Analysis………………………………………………………………… 3.3.1 Content Analysis……………………………………………………… 3.3.2 Critical Discourse Analysis…………………………………………... 3.3.3 A Sample Construction of Macrostructure……………………………. 31 31 33 34 34 36 36. CHAPTER FOUR CONTENT ANALYSIS…………………………………………… 4.1 Political News in a TV News Program………………………………………. 4.2 Content Analysis of Political News…………………………………………... 41 41 43. VII.

(10) 4.2.1 Analysis of News Item……………………………………………….. 4.2.2 Analysis of News Duration…………………………………………… 4.2.3 Analysis of News Appearing Order…………………………………... 4.3 Results and Discussion……………………………………………………….. 43 45 48 51. CHAPTER FIVE CRITICAL DISCOURSE ANALYSIS: CASE STUDIES……………….. 53 5.1 Background………………………………………………………………….. 53 5.1.1 Ma Ying-jeou’s Special Expenses Indictment………………………... 53 5.1.2 The DPP Joint Conference………………………………………….... 54 5.1.3 Taiwan’s Rejection of the Beijing Olympic Torch Relay…………….. 54 5.2 Ideological Square………………………………………………………….... 55 5.3 Content Analysis……………………………………………………………... 56 5.4 Critical Discourse Analysis: Macrostructure………………………………… 60 5.4.1 Analysis of Ma Ying-jeou’s Special Expenses Indictment…………… 60 5.4.2 Analysis of the DPP Joint Conference………………………………... 67 5.4.3 Analysis of Taiwan’s Rejection of the Beijing Olympic Torch Relay... 75 5.4.4 Discussion on Macrostructure………………………………………... 83 5.5 Critical Discourse Analysis: Microstructure……………………………….... 87 5.6 Results and Discussion………………………………………………………. 102 CHAPTER SIX CONCLUSION…………………………………………………….. 105 6.1 Summary and Implications…………………………………………………... 105 6.2 Limitations of the Present Study and Suggestions for Future Research……... 108 REFERENCES……………………………………………………………………… 109 APPENDIX A Full content of the FTV News report on Ma’s special expenses indictment………………………………………………………... 117 APPENDIX B Political Headline Stories in ETTV News, FTV News, SET News, and CTi News from April 15, 2007 to May 9, 2007............ VIII. 121.

(11) LIST OF TABLES. Table 3-1. Ranking of the 10 most viewed TV stations in Taiwan (adapted from Table 3-4-5 in Chi and Tsai 2005)……………………………………... Table 3-2. Ranking of the 4 most viewed TV news stations in the four areas of Taiwan (adapted from Table 3-4-10 in Chi and Tsai 2005)……………. Table 4-1. News items and news duration per program of ETTV News, FTV News, SET News, and CTi News from April 15, 2007 to May 9, 2007…..................................................................................................... Table 4-2. Political news items and political news duration per program of ETTV News, FTV News, SET News, and CTi News from April 15, 2007 to May 9, 2007……………………………………………………………. Table 4-3. Political news items per program of ETTV News, FTV News, SET News, and CTi News from April 15, 2007 to May 9, 2007……………. Table 4-4. Percentage of political news items of ETTV News, FTV News, SET News, and CTi News from April 15, 2007 to May 9, 2007……………. Table 4-5. Political news duration per program of ETTV News, FTV News, SET News, and CTi News from April 15, 2007 to May 9, 2007……………. Table 4-6. Percentage of political news duration of ETTV News, FTV News, SET News, and CTi News from April 15, 2007 to May 9, 2007………. Table 4-7. D/K ratio of ETTV News, FTV News, SET News, and CTi News from April 15, 2007 to May 9, 2007………………………………………… Table 4-8. Frequency of political headline stories in ETTV News, FTV News, SET News, and CTi News from April 15, 2007 to May 9, 2007………. Table 4-9. Categorization of top five stories of ETTV News, FTV News, SET News, and CTi News from April 15, 2007 to May 9, 2007……………. Table 5-1 Reports about Ma’s special expenses indictment by ETTV News, FTV News, SET News, and CTi News on April 17, 2007…………............... Table 5-2. Reports about the DPP joint conference by ETTV News, FTV News, SET News, and CTi News on May 7, 2007……………………………. Table 5-3. Reports about Taiwan’s rejection of the Beijing Olympic torch relay by ETTV News, FTV News, SET News, and CTi News on April 26, 2007……………………………………………………………………. Table 5-4. Reports about Taiwan’s rejection of the Beijing Olympic torch relay by ETTV News, FTV News, SET News, and CTi News on April 27, 2007…………………………………………………………………….. IX. 31 32. 41. 42 43 44 46 46 47 49 50 57 57. 58. 59.

(12) Table 5-5. Headlines of the reports about Ma Ying-jeou’s special expenses indictment on April 17, 2007………………………………………....... 61 Table 5-6. Headlines of the reports about the DPP joint conference on May 7, 2007……………………………………………………………………. 68 Table 5-7. Headlines of the reports about Taiwan’s rejection of the Beijing Olympic torch relay on April 26, 2007………………………………… 76. X.

(13) LIST OF FIGURES. Figure 2-1. A framework for critical discourse analysis of a communicative event (Fairclough 1995:59)…………………………………………………. Figure 2-2. Hypothetical structure of a news schema (van Dijk 1988:55)……....... Figure 3-1. Categorization of political news in the present study………………… Figure 5-1. The schematic structure of the ETTV News report about Ma’s special expenses indictment on April 17, 2007……………………………….. Figure 5-2. The schematic structure of the FTV News report about Ma’s special expenses indictment on April 17, 2007……………………………….. Figure 5-3. The schematic structure of the SET News report about Ma’s special expenses indictment on April 17, 2007……………………………….. Figure 5-4. The schematic structure of the CTi News report about Ma’s special expenses indictment on April 17, 2007……………………………….. Figure 5-5. The schematic structure of the ETTV News report about The DPP joint conference on May 7, 2007…………………….. ……………… Figure 5-6. The schematic structure of the FTV News report about the DPP joint conference on May 7, 2007…………………………………………… Figure 5-7. The schematic structure of the SET News report about the DPP joint conference on May 7, 2007…………………………………………… Figure 5-8. The schematic structure of the CTi News report about the DPP joint conference on May 7, 2007…………………….. ……………………. Figure 5-9. The schematic structure of the ETTV News report about Taiwan’s rejection of the Beijing Olympic torch relay on April 26, 2007……… Figure 5-10. The schematic structure of the FTV News report about Taiwan’s rejection of the Beijing Olympic torch relay on April 26, 2007…….. Figure 5-11. The schematic structure of the SET News report about Taiwan’s rejection of the Beijing Olympic torch relay on April 26, 2007…….. Figure 5-12. The schematic structure of the CTi News report about Taiwan’s rejection of the Beijing Olympic torch relay on April 26, 2007……... XI. 16 20 35 62 63 64 65 70 71 72 73 78 79 80 81.

(14) CHAPTER ONE INTRODUCTION. 1.1 Motivation A news report is perceived by laymen and assumed by journalists to be objective and unbiased. It is to reproduce the one and the only truth, functioning as a source of information for the public. Yet, as a social institution and an economic entity, the press is by no means independent and impartial. Rather, it is influenced and impeded by various forces, for example, the search for advertising revenue, the favor of a party or an ism over another, or the prevailing socio-cultural atmosphere (Bell 1991, Fairclough 1995, Fowler 1991, Herman and Chomsky 1988, van Dijk 1988). In the process of making news, news organizations construct ‘reality’ in a manner congruent with their underlying ideologies and interests. The press has taken over the ideology formation role of the family, the Church, and the school in contemporary information societies. It ‘not only influences what we think about, and not only what we form opinions about, but also among which opinions we must choose’ (van Dijk 1991:241). It is through language that their ideologies are realized and presented to the public. Therefore, news language is never value-free, but ideologically-laden (Fairclough 1989, 1992, 1995, Fowler 1991, Fowler et al. 1979, Kress and Hodge 1979, van Dijk 1988, 1991). Yet, without direct access to news events and alternative sources of information, few ever question or even challenge what has been presented in news. Hence, approaches like content analysis and critical discourse analysis are designated to expose the possible media bias to the public through a quantitative measurement of news coverage and a critical examination of news language. The present study investigates the relations between language and ideology in Taiwan TV news stations 1.

(15) adopting both quantitative content analysis and linguistics-oriented critical discourse analysis.. 1.2 Ideology Various definitions and positions of ideology have been identified in the literature. In the present study, ideology is not taken as dominant ideology and false consciousness, which involves distortion in the service of power, as discussed in the Marxist tradition; nor is it taken as hegemony through which the working class comes to identify with the bourgeois’ values and thus develop a consensus culture. Rather here ideology is perceived to be ‘a systematic body of ideas, organized from a particular point of view’ (Kress and Hodge 1979:6), which leads people to make sense of the world in a goal-directed and interest-related way. Thus, ideology is not limited as an instrument of domination, for ‘there are also ideologies of opposition or resistance, or ideologies of competition between equally powerful groups, or ideologies that only promote the internal cohesion of a group, or ideologies about the survival of humankind (van Dijk 1998:11).’ As cognitive phenomena, ideologies are nevertheless social systems shared by social members. They are dynamic instead of static, that is, they can be acquired, used, and, most of all, changed in social situations. Ideologies operate alongside and inside the material, political, and institutional environments (Blommaert 2005). It is through language, as well as other practices, that ideologies become observable. The present study observes the underlying ideologies of TV news stations in Taiwan through their language. To be more specific, it aims to determine how their political stances influence related political reports in terms of the amount of news stories and linguistic structures of news reporting.. 2.

(16) 1.3 Political Environment in Taiwan The long-term tension between Taiwan and China since 1949 has given rise to two conflicting ideologies and powers in Taiwan. One is pro-unification, led by the Kuomintang (KMT, the Chinese Nationalist Party), which forms the pan-blue coalition with the People First Party and the New Party. The other is pro-independence, represented by the Democratic Progressive Party (DPP), and together with the Taiwan Solidarity Union, they comprise the pan-green coalition. The KMT established its regime over Taiwan from the end of World War II in 1949. In the first ever direct elections for the 9th-term President and Vice President in 1996, incumbent President Lee Teng-hui and Vice President Lien Chan of the ruling KMT won a majority of 54% of the votes.1 However, in the following 2000 election, DPP Chen Shui-bian beat KMT nominee Lien Chan, and the popular independent candidate, James Soong, which put an end to half a century of KMT rule on Taiwan. It is commonly believed that the split of the KMT between Lien and Soong, as well as the secret support from former President Lee, contributed to Chen’s unexpected victory. In the aftermath, Soong announced the formation of the People First Party (PFP), which belonged to the pan-blue coalition, while Lee formed the pro-independence Taiwan Solidarity Union and cooperated with the DPP. In the subsequent 2004 election, incumbent President Chen Shui-bian and Vice President Annette Lu of the DPP were re-elected by a margin of 0.22% of valid votes over a combined opposition ticket of KMT Chair Lien Chan and PFP Chair James Soong. Lien and Soong refused to concede, but the challenges turned out to be unsuccessful.. 1. The election information mentioned in the section is based on the database of Central Election Commission. 3.

(17) During the time of data collection, i.e. 2007, the DPP was the ruling party, whereas the KMT remained the biggest opposition party. For the 2008 presidential election to come, the two parties had numerous disputes and reconciliations both within and between them, making this period an opportune moment to study news organizations’ political stances hidden in supposedly neutral reports. The KMT ticket of Ma Ying-jeou and Vincent Siew was officially formed in June, 2007. Later in August, 2007, Frank Hsieh of the DPP chose Su Tseng-chang as his running mate. On March 22, 2008, the Ma-Siew ticket, with 58.45% of the popular vote, won landslide victory over the Hsieh-Su ticket, bringing the KMT back to power in Taiwan.. 1.4 Scope and Goal The goal of the present study is to reveal the relations between news language and political stances in the Taiwanese press. Four most popular TV news stations were selected, i.e. Eastern Television (ETTV, 東森電視), Formosa Television (FTV, 民視), Sanlih Entertainment Television (SET, 三立電視), and Chung T’ien Television (CTi TV, 中天電視). The midnight programs were videotaped for a span of 25 days, from April 15, 2007 to May 9, 2007. ETTV was established in 1990, and later developed into the Eastern Multimedia Group, which runs a total of eight TV stations and also sets foot in internet, newspaper, and radio broadcast. From 2002, ETTV began it overseas broadcasting, and its service areas cover America, South East Asia, Australia and New Zealand, Hong Kong and Macao. FTV, established in 1996 in Kaohsiung, is the fourth wireless TV station in Taiwan, and the only privately-owned one among the five. It was founded by Trong-rong Chai, a party member of the DPP, and thus was wholly operated by the DPP. To live up to the DPP’s perspective, FTV is also the first station to use 4.

(18) Taiwanese in a majority of its programs. SET rose as a local station in Kaohsiung in 1990, but was later relocated to Taipei. After the 2000 presidential election, it produced a soap opera based on the romance between then President Chen Shui-bian and then First Lady Wu Shu-chen, which attracted high audience ratings. CTi TV was founded in 1994 by a Hong Kong businessman. Due to financial difficulty, it later merged into the China Times group, which also owns China Television. The China Times group is widely considered to be pan-blue. Some even suggest that CTi TV is especially prone to Communist Party of China, instead of the KMT in Taiwan. To determine how the stations’ underlying political inclinations, either pan-blue or pan-green, influence the relevant news reports, two approaches are adopted in the study. To begin with, a quantitative content analysis is employed. Stories dedicated to the DPP and those to the KMT are measured and compared in terms of the number of news items, the accumulated news duration, and the news appearing orders. The other approach is critical discourse analysis. Reports on three political events are selected as case studies. Both macrostructures and more local linguistic strategies are analyzed to verify the influence of the ideological square (van Dijk 1998) on news reports. To sum up, the present study aims to answer the following three research questions: First, according to content analysis, how do the four news stations differ in their coverage of political events? Do their reports dedicated to the DPP and the KMT differ in terms of the number of news items, the accumulated news duration, and the news appearing orders in the rundown? What does the quantitative data tell us about the news stations’ political stances? 5.

(19) Second, according to critical discourse analysis, how do the four news stations differ in their news language? Do their reports on the three political events differ in terms of macro- and micro-structure? Does their language conform to the ideological square proposed by van Dijk (1998)? Third, how do the results of content analysis and those of critical discourse analysis correspond to each other? What do the two analyses tell us about the relations between political stances and coverage language in the press?. 1.5 Organization of the Present Study The organization of the thesis is as follows. Chapter 2 begins with an introduction of content analysis, and then elaborates the theoretical background and empirical studies of critical discourse analysis. Chapter 3 states the procedures of data collection and two analyses conducted in the present study: content analysis of news coverage, and critical discourse analysis of news language. In Chapter 4, content analysis is employed to measure the news stations’ political news in terms of news item, news duration, and news appearing order. The prominence of the DPP and the KMT in the news reveals the stations’ political preferences. Chapter 5 presents a critical discourse analysis of the news language in three political events. Both macroand micro-structures of the news stories are under critical examination to disclose the stations’ political stances. The reciprocal influences between language and political stances further verify the ideological square by van Dijk. Finally, Chapter 6 concludes the major findings of the present study by addressing the relations between language and ideology in the press, discussing limitations of the thesis, and putting forth suggestions for future research.. 6.

(20) CHAPTER TWO LITERATURE REVIEW. The chapter reviews relevant literature. Section 2.1 introduces the content analysis approach and relevant media studies in Taiwan. Section 2.2 focuses on critical discourse analysis and its precursor, critical linguistic. The theoretical background is first presented in Section 2.2.1, followed by empirical studies in Section 2.2.2.. 2.1 Content Analysis News is often viewed as a source of information, and a reproduction of reality. However, as a social product, it is not a faithful copy of the world, but an interpreted reconstruction of discourses. News is not pre-existing but manufactured. What event is qualified as a news story is determined by a set of ideologically-driven news value criteria, and the formation process includes selection, representation, and, most importantly, transformation (Fairclough 1995, Fowler 1991, Herman and Chomsky 1988, van Dijk 1988). Therefore, a news report is anything but neutral, with ideologies imprinted in its language (Fairclough 1989, 1992, 1995, Fowler 1991, Fowler et al. 1979, Kress and Hodge 1979, van Dijk 1988, 1991). To disclose the media bias, content analysis is extensively used in media studies (D’Alessio and Allen 2000). It is ‘a research technique for the objective, systematic, and quantitative description of manifest content of communications’ (Berelson 1952:74). It codes texts and breaks them down into a number of discrete and manageable categories. The frequencies are counted and interpreted as meaningful information. With the help of content analysis, Chen and Chen (1992) found that after the 7.

(21) liberation of journalism in Taiwan in 1988, reports about the 1989 legislative election were much more balanced than before. Reports about the KMT candidates accounted for 35.3%, and those about the DPP candidates rose to 21.7%. Nevertheless, reports in the United Daily News, the Central Daily, and the Taiwan News were highly positive about the KMT, but negative about the DPP. On the contrary, the Capital Daily and the Independent Morning spoke well of the DPP, but ill of the KMT. The China Times and the Taiwan Times gave positive and more neutral reports to both parties without obvious partisanship. On the 2000 presidential election, Lo and Huang (2001) compared campaign reporting in state-owned newspapers and that in privately-owned newspapers. After the content analysis of news sources, amounts of news stories, and overall party images, the research showed that state-owned newspapers unanimously supported then ruling KMT, while privately-owned newspapers showed diversities. The Liberty Times supported the KMT Lien-Siew and the DPP Chen-Lu tickets, but disliked the Soong-Chang ticket. In contrast, the United Daily News sided with the Soong-Chang ticket, but disapproved of the Chen-Lu ticket. The China Times seemed to be more balanced than the other three. A similar pattern was drawn in Yeh (2001) as well. Also on the 2000 election, Chen (2001) found that the amount of attack news on hopefuls significantly increased with the approach of the election. And the negative coverage of candidates, especially by newspapers that held opposite stances, had negative effects on their approval ratings. Thus, the attack news on Lien and Soong from the Liberty Times was threatening to their polls, while Chen suffered popularity loss from negative reports in the United Daily News. Following Lo and Huang (2001), Lo et al. (2007) identified structural bias in newspapers coverage of the 2004 presidential election. That is, incumbent President Chen received far more reports than candidates of the opposition party. In addition, 8.

(22) partisan bias was found in the Liberty Times, which was in favor of the DPP Chen-Lu ticket, but in opposition to the KMT Lien-Soong ticket. A similar political inclination was drawn up in the Apple Daily, whereas the China Times showed its support for the Lien-Soong ticket. The United Daily News was considered to be more neutral among the four. With regard to the most recent 2008 presidential election, Li (2008) examined both newspapers and news stations in Taiwan. The content analysis showed that the United Daily News and the Apple Daily assigned more positive reports to the KMT Ma-Siew ticket, but more negative reports to the DPP Hsieh-Su ticket. Yet the Liberty Times preferred the Hsieh-Su ticket to the Ma-Siew ticket. As for the news stations, reports in the Formosa Television and the Sanlih Television advantaged the Hsieh-Su ticket but disadvantaged the Ma-Siew. On the other hand, the China Television Company, the Public Television Service, and TVBS-N favored the Ma-Siew ticket, though no significant bias was detected in reports on the Hsieh-Su ticket. The study concluded that the United Daily News, the Liberty Times, the Apple Daily, the Formosa Television and the Sanlih Television showed stronger political biases than the China Television Company, the Public Television Service, and TVBS-N. Among them, the Liberty Times, the Formosa Television and the Sanlih Television were pro-DPP, whereas the United Daily News, the Apple Daily, the China Television Company, the Public Television Service, and TVBS-N were pro-KMT. In addition to campaign reporting, content analysis is often applied with frame analysis to studies on political controversies. For example, concerning the 319 shooting in 2004, in which incumbent President Chen Shui-bian and Vice-President Annette Lu of the DPP were both shot while campaigning in Tainan, Chang (2006) found that the China Times, the United Daily News, and the Apple Daily quoted mainly from the pan-blue coalition, while the Liberty Times preferred statements by 9.

(23) the law enforcement agency. Kochurova (2008) analyzed how newspapers in Mainland China and in Taiwan report former President Lee Teng-hui’s visit to Japan in 2007. During the trip, Lee paid a tribute at the Yasukuni Shrine, where Japan honors her soldiers killed during World War Two, to commemorate his brother who was killed while serving in the Japanese army. The tribute angered the China’s press, and Lee was described as a betrayer and a problem-maker in the news. Content analysis showed the visit received more attention in Taiwan than in China. In addition to the betrayer and problem-maker frames, the Taiwanese press brought out nostalgia and humanitarian as motives for Lee’s visit. Yet, the China Times was comparatively harsh and critical toward Lee, while the Liberty Times reports were more positive and tolerant. Also on Lee Teng-hui, Chien (2002) researched how his images were presented and changed in the United Daily News and the Liberty Times from January, 1988 to December, 2001. It was found that both focused mainly on political subjects and political images. For the United Daily News, most straight news remained positive toward Lee, but editorials were more critical. Evaluations of his personality character and political preference worsened year by year. As for the Liberty Times, both straight news and editorials were positive, and so were evaluations of personality character. Yet evaluations of political preference went downhill. Overall, the Liberty Times presented a more positive image of Lee than the United Daily News. This section has demonstrated how the quantitative data of content analysis can provides an objective and reliable basis to research on media bias. Yet content analysis is at its best when supplemented by such a qualitative insight as critical discourse analysis employed in this study.. 10.

(24) 2.2 Critical Discourse Analysis The previous section has reviewed the comprehensive employment of quantitative content analysis in exploring media bias in Taiwan. This section turns to critical discourse analysis and its precursor, critical linguistics, and is aimed at the relations between language and ideology in the press.. 2.2.1 Theoretical Background To investigate the dialectic relation between language and ideology, Fowler and his associates from the late 1970s began to critically examine news texts, using the critical linguistics approach (Fowler 1991, Fowler et al. 1979, Kress and Hodge 1979), which was later expanded into critical discourse analysis by Fairclough (1989, 1992, 1995) and van Dijk (1988, 1991). Despite with different names and to some extent different perspectives, both approaches aim to display to the public the underlying ideologies the press embraces by critically studying structures and features of news language.. 2.2.1.1 Fowler and His Colleagues Inspired by Halliday’s (1978) systemic-functional grammar, which asserts that language is as it is because of its functions in the social structure, Fowler and his colleagues at the University of East Angelia tried to read off social structures by critically analyzing linguistic structures (Fowler 1991, Fowler et al. 1979, Kress and Hodge 1979). Language use is not arbitrary for options are available and ‘each particular form of linguistic expression in a text—wording, syntactic option, etc.—has its reason’ (Fowler 1991:4). For example, any social disturbance can be termed as either a ‘demonstration’ or a ‘riot’; the former suggests people’s exercising their right, while the latter condemns their behavior. Ideologies exist in every discourse, and 11.

(25) news is no exception. To unveil the hidden ideologies of news, Fowler and his colleagues adopted analytic tools in Halliday’s discussion of the language functions1, and developed critical linguistics, in which lexical choices used in news media, such as referring expressions and their collocated predicates, and syntactic patterns, including transitivity and transformation processes, are under critical examination. Vocabulary of a language, to critical linguists, is in the form of a taxonomic map instead of a list, which sorts boundless concepts into strictly defined categorial relationships and helps language users stabilize and converse about their life experiences. The categorization of vocabulary is, hence, of significant ideological importance. The words chosen in the press reflect not only how an event is perceived by journalists, but how the audience is directed and expected to construct the event in the same way. Referring expressions in news categorize people and events into groups, and place discriminatory values on them. For example, any social unrest can be either a ‘demonstration’ or a ‘riot’, as has been mentioned above, and a suicide bomber is likely to be portrayed as a ‘terrorist’ by western media, but a ‘freedom fighter’ by Jihad supporters. All the expressions used in news, in other discourses as well, are not a random selection but embody categorial and ideological function (Caldas-Coulthard 1993, Fang 1994, 2001, Flowerdew et al. 2002, Kuo 2001, Kuo and Nakamura 2005, Teo 2000, Wang 1993). The significance of referring expressions cannot be ignored, and so is that of their accompanying predicates (Flowerdew et al. 2002, Kuo 2001). In Flowerdew et al.. 1. Halliday (1978) proposed three meta-functions of language: ideational, interpersonal and textual. The ideational function allows people to represent their experiences of the world and their beliefs in language. The interpersonal function helps construct social identities and social relationships between people. Last, the textual function relates to how a text is constructed out of sentences. 12.

(26) (2002), it was found that Hong Kong’s leading English newspaper, the South China Morning Post, habitually ascribed negative attributes to Mainland immigrants—poor, unemployable, uneducated, unhygienic, to name a few. These predicates clearly unfolded the newspaper’s hostility toward Mainland immigrants. At the sentence level, critical linguists focus on transitivity and relevant transformations. The concepts of transitivity and transformation employed in critical linguistics differ from the sense of the terms in Chomsky’s generative grammar, but conform to Hallidayan grammar. The transitivity system of language, according to Halliday (1985), is the foundation of representation, for it construes the world of experience into a manageable set of process types. Each experience can be represented by clauses with processes, participants, and circumstances, mainly who does what to whom in when and where. The processes involved and the roles participants take allow journalists to encode the institutional ideologies into news texts (Fang 1994, 2001, Teo 2000). Trew (1979a) studied transitivity used by two British newspapers, The Times and the Guardian, in a series of 1975-Salisbury-riot reports, with the first day headlines as followed. (1) a. Rioting blacks shot dead by police as ANC leaders meets (The Times, June 2, 1975) b. Police shoot 11 dead in Salisbury riot (Guardian, June 2, 1975) The content is similar in these two cases, but the processes and the participants presented are distinct. The Guardian headline is in the active form, and The Times one is in the passive. The Guardian stance was less detectable, but with ‘rioting blacks’ in the focal position instead of ‘police’, The Times suggested that ‘blacks’ be responsible for the shooting, and thus legitimize police’s killing. The Times position was 13.

(27) reinforced by its characterization of the dead as ‘blacks’, whose information was unspecified in the Guardian. Also while ‘riot’ is the circumstance in which the tragedy occurred in the Guardian, ‘rioting’ is the quality of the dead in The Times. This examination of transitivity was shown to help unravel ideological positions the newspapers held. Transformation processes related to transitivity—mainly passivization and nominalization—are also under examination (Fang 1994, Kuo and Nakamura 2005). Passivization is a process of demoting the agent, and simultaneously promoting the patient. The patient is put in the focal position, but the agent is marginalized or omitted. This transformed presentation may be an accommodation of textual coherence or an indication of a specific interpretation, just like The Time headline above, in which the journalist’s interpretation that the responsibility was on the shot instead of the shooter was signaled by the employment of the passive construction. The other transformation process is nominalization, which is a radical transformation of a clause into a nominal. For example, a complete proposition ‘X has alleged against Y that Y did A and that Y did B [etc.]’ may be trimmed into a nominal ‘allegation’, with participants and indication of time and modality deleted. Thus, the nominalization process is potentially mystification, which permits concealment. The critical linguistics approach, which explores lexical choices and syntactic patterns of text, as well as modality and speech acts, to disclose the connection between linguistic and ideological processes in news, has proven to be fruitful and, what’s more, it has laid a sound foundation for future theory building. In the following sections, we will see respectively how Fairclough and van Dijk expanded the critical linguistics approach into critical discourse analysis.. 14.

(28) 2.2.1.2 Fairclough The previous section has demonstrated how Fowler and his colleagues demystify readings of ideology-laden texts with critical linguistics. However, for Fairclough (1989, 1992, 1995), it is impossible to read off social structures simply from linguistic structures. Instead only in the domain of interpretation, that is, in the way how the audience interprets a text, can social meanings and ideologies be fully explored and discussed. Fowler (1996) himself also acknowledged that ‘the original (critical linguistic) theory … privileges the source of texts, ascribing little power to the reader because the reader simply is not theorized’ (p. 6&7). There is no one-to-one correspondence between a text and its interpretation, for texts are open to interpretations dependent on contexts and interpreters. Each interpreter, as a social member, processes each text based on his position and resources he acquires in the society, and thus arrives at his own specific interpretation. The analysis of text alone is inadequate for an ideological study. The concepts of discourse, i.e. text with context, and society are needed if a more consolidated theory on language and ideology is to be advanced. With respect to the interface between language and society, critical linguists center on how language, as a social practice, reproduces existing ideologies and social values in a unidirectional way, whereas Fairclough takes the position that language is both socially shaped and socially shaping. That is, language can not only reproduce and thus help maintain social norms and conventions, but challenge and further transform social values and beliefs when used creatively. In Fairclough’s conception, critical studies should not be delimited at the textual level, but extended to the discourse level; it should not be pursued purely in the linguistic domain, but rather within a social perspective. He is the key figure to transit studies on language and ideology form text analysis to discourse analysis and a 15.

(29) precursor in the critical discourse analysis framework, in which he proposed a three-dimensional framework to analyze a discourse as text, discourse practice, and sociocultural practice.. text production TEXT text consumption DISCOURSE PRACTICE SOCIOCULTURAL PRACTICE Figure 2-1. A framework for critical discourse analysis of a communicative event (Fairclough 1995:59). Text, the product of discourse practice, may be written (newspaper), spoken (broadcast), or spoken with visual (TV news). It is analyzed under four headings: vocabulary, grammar, cohesion, and text structure. Vocabulary and grammar deal with individual words and how they are combined into clauses and sentences. Cohesion manages the linkage between clauses and sentences (Halliday and Hasan 1976, Halliday 1985) and text structure organizes the global structure of the whole text. Discourse practice is on how a text is produced and consumed. As we have pointed out earlier that news organizations are not self-contained bodies but social institutions, news production is always conditioned by other factors of society, and involves complex institutional routines of a collective nature. In consumption, social member is at the premium in the sense that a text will not mean anything or carry any. 16.

(30) ideological significance until it is interpreted by a social member. Both processes of production and consumption are constrained by sociocultural practice. They are constrained by members’ available resources, which are internalized in social structures, and by the nature of specific communicative events involved. In the discourse practice dimension, three aspects of analysis are brought up: the force of utterances, the coherence of texts, and the intertextuality of texts. The forces of utterances, i.e. speech acts, are actual components of any communicative event (Searle 1969). A coherent text requires that constituents of a text be meaningfully connected so that the text as a whole conveys an intelligible idea. Intertextuality, coping with the interdependence among discourses (Foucault 1972), looks at traces of discourse practice in the target discourse, and strives to interpret social and cultural meanings hidden beneath. News reports, which represent and transform utterances and happenings of other discourses, are thus always intertextual (Waugh 1995). In illustration of intertextuality, Fairclough (1992) studied how a committee report on drug trafficking was transformed into a news report in a British tabloid, the Sun. In the news report, it was found that the informal, colloquial language of private life was used to win readership, and meanwhile part of official discourse remained to preserve the legitimacy and authority of the report. The heterogeneity of the language revealed the newspaper’s contradictory positions and identities. Sociocultural practice is the last dimension, in which ideologies and power negotiation has a material existence in practices. Ideologies built into conventions are more or less naturalized, and most of the people through their whole lives are never aware of the fact that their automatic practices contain ideological functions. However, what Fairclough wants to emphasize is that people’s practices, on the other hand, are socially constitutive, that is, they possess the potential to either reproduce or 17.

(31) restructure social norms. Social struggles can be undertaken and social changes can be achieved through creation and negotiation of discourse practices. To sum up, for Fairclough, text is what is there to be described, but society is what is there to be explained. As for discourse, it functions as a mediator between text and society in the sense that ‘properties of sociocultural practice shape texts, but by way of shaping the nature of the discourse practice, i.e. the ways in which texts are produced and consumed, which is realized in features of texts’ (Fairclough 1995:60). The three dimensions, though focusing on separate facets of a discourse, are indispensable for a more holistic critical discourse analysis framework.. 2.2.1.3 van Dijk Same as Fairclough, van Dijk (1988, 1991) is an advocate of critical discourse analysis, believing that the study of ideologies in language will not be satisfactory if the scope of analysis is limited to text only with discourse and society left unexplored. Yet, to van Dijk, no direct relation can be constructed between social practice and discourse practice, but the influence of society over discourse is exerted by the way social values and beliefs shape models and scripts in people’s cognition, which are realized through people’s language. Therefore, cognition is the missing link to bridge the gap between discourse and society. The three components—cognition, discourse, and society—form van Dijk’s triangle theory of ideology. In his thorough and influential study on the structure of news discourse, van Dijk has differentiated the macrostructures, which concern the overall textual structures, from the microstructures, which focus on actual linguistic expressions used. The global, abstract macrostructures, including the thematic and schematic structures, lead the discourse, but need to be realized at the more local, concrete microstructures of words and sentences. In what follows, both macro- and micro-structures are 18.

(32) discussed. News discourse exhibits a structure that is basically top-down and relevance-controlled. What is considered newsworthy is presented first, regardless of its relation with other elements of the same event. Thus the outcome and impact of an event may be presented first if laden with news values when its cause and circumstances are still left unspecified. The particular writing style constructs an ‘inverted pyramid’, a standardized format for news writing (Franklin et al. 2005, Yopp and McAdams 1999). Not only does the top-down processing control the way journalists organize news stories, but it also influences the way audience interpret news, for given the topic alone, as experienced news consumers, we may predict what might follow afterward with some confidence. In news, topics at the higher level dominate those at the lower level, forming a hierarchical structure. The thematic structure of a discourse organizes its topics and content, and the schematic structure deals with the form it takes. Each discourse type has its more or less fixed schema in a specific culture, and the schematic structure of news proposed by van Dijk as in Figure 2-2, like the thematic one, is composed of a series of hierarchically ordered categories, determining how the topics of a text could or should be ordered and how sequences and sentences should appear in the text. The categories, except Summary, are usually delivered cyclically in installments. Also, it should be noted that, as an underlying abstract structure, under the influence of specific constraints, such as relevance and ideologies, transformations are possible, that is, the news schema maybe be realized in the text in different ways. Within all the categories, Headline and Lead deserve most attention, for they together form the Summary and accommodate most relevant and valuable information in a news report (Fang 2001, Kuo and Nakamura 2005, Lee and Craig 1992, Teo 2000, Wang 1993). Main Events are the core of a News Report with Circumstances 19.

(33) depicting its settings, Previous Events tracing its causes, and Consequences bringing about its effects. Verbal Reactions feature opinions cited from other sources while Evaluations are made by reporters, who may also put forth Expectations of future development concerning Main Events.. News Report. Summary. Headline. Story. Lead. Episode. Situation. Comments. Background. Main Events Consequences. Context. Verbal Reactions Conclusion. History. Expectations. Evaluations. Circumstances Previous Events Figure 2-2. Hypothetical structure of a news schema (van Dijk 1988:55). The value of the macrostructures of news discourse is evident in Wang’s (1993) study on the New York Time’s and Renmin Ribao’s ‘the People’s Daily’ coverage of the 1991 Soviet coup. Comparing the thematic and schematic structures of the Ribao report with those of the Time report, Wang made it clear that as the mouthpiece of the Chinese Communist Party, the Ribao report on the Soviet coup was conservative and incomplete, lacking all the background and contextual information. The microstructures of news discourse include a variety of linguistic devices, such as implicatures2 (Grundy 2000, Levinson 1983), presuppositions3 (Abbot 2000, 2. Conversational implicature was first brought out in Grice’s theory of the cooperative principle and the four maxims (1967), which Sperber and Wilson later revised as Relevance Theory (1995). As an 20.

(34) Levinson 1983, Saeed 2003), local coherence4 (Halliday and Hasan 1976), style and rhetoric. Like most discourses, news reports leave many things unsaid, which are either presupposed or must be inferred for full comprehension. Presuppositions are information that is taken-for-granted or contextually presumed, constituting the background in a discourse upon which an assertion stands up. Owing to their resistance to detectability and defeasibility, presuppositions, as well as implicatures, come to be effective tools to conceal one’s ideological intention (see also Fairclough 1989, 1992, 1995). The same is true for local coherence, which depend on subjective knowledge and beliefs about the world. Thus propositions may cohere for the speaker but not for the hearer. Learning how propositions are linked together in a discourse enables us to make explicit the latent assumptions and beliefs (see also Fairclough 1995). Style of a discourse bears traces of context in a text, which specifies the kind of occasion, the degree of formality, the relations among participants, and so forth. ‘It is the total set of characteristic, variable structural features of discourse that are an indication of the personal and social context of the speaker’ (van Dijk 1988:73). Lexical and grammatical styles involve selections of features from all the possibilities in a language, which are controlled by the speaker’s opinions and attitudes. Thus the analysis of stylistic choices helps decode the speaker’s underlying ideologies. The last microstructure is rhetoric. Unlike rhetoric featured in literature, here van Dijk discusses linguistic strategies employed to make a text persuasive, including. inferred meaning arising from the hearer’s calculation of the speaker’s intention, conversational implicature is non-detachable and defeasible. In contrast, conventional implicature is closely associated with a particular lexical item or expression by convention, and thus detachable and non-defeasible. 3 While semantic presupposition is discussed in a truth-relation approach and thus is non-defeasible, presupposition in pragmatic theories deals with defeasibility and the projection problem. That is, presuppositions are dependent on the discourse context and the hearer’s encyclopedic knowledge, and thus are defeasible for they may fail to survive in different levels of context. 4 Local coherence at the micro-level of a text is a counterpart of the global macro-coherence of the text. 21.

(35) numbers, news sources, and quotations. Provision of numbers or news sources makes a report authentic and authoritative. Quotations may present information newsworthy in it own right, or it may be a channel to insert opinions of preferred personages who voice the position of the news organization. Hence, the selection of quotation may be propagandist (Fairclough 1995, Fang 2001, Kuo and Nakamura 2005, Kuo 2007, Lee and Craig 1992, Teo 2000). Other than the discourse analysis of macro- and micro-structures, for van Dijk, ‘a complete empirical account of discourse also requires a description of cognitive processes of discourse production and understanding and of social interactions in sociocultural situations’ (1988:30). The framework of van Dijk is similar to that of Fairclough, yet more effort is devoted to the cognition domain to explicate the connection between language use and social cognition, namely knowledge and beliefs shared by group members. The macro-micro distinction also exists in van Dijk’s conceptions of cognition and society. Personal beliefs, knowledge, and practices—such as the use of language—are at the micro-level; social cognition and system principles belong to the macro-level. The macro-micro relations are both top-down and bottom-up. Social cognition and principles govern social members’ beliefs, knowledge, and practices. Yet individuals have free will to deviate from social norms, and if certain deviation becomes systemized and generalized, then a social change is taking place. The interaction between discourse and cognition is manifest through cognitive processes of discourse production and understanding, involving the operations of models and scripts. Models are the fundamental cognitive knowledge structures people build up when comprehending an event (references). Scripts are general knowledge people acquire toward the world of experience through socialization (Schank and Abelson 1977, Schank and Kass 1988). In comprehending a news report, 22.

(36) people construct a specific model for this particular event, with scripts filling in the relevant missing information (Fang 2001, Lee and Craig 1992). Take the Salisbury case in Section 2.2.1.1 for example. In the mental model of this disturbance, we may have detailed information like ‘11 were dead’, ‘Salisbury’, ‘1975’, and ‘the meeting of ANC leaders’. Yet our knowledge about ‘riot’, ‘blacks’, and ‘police’ comes from scripts, accumulated from our previous experiences. Therefore, people always get more than what is linguistically presented in the news. van Dijk contributed to critical discourse analysis the structures of news discourse in terms of content and form. He also enriched the approach by including the concept of cognition, different from Fairclough’s focus on society. The triangle theory of discourse, cognition, and society has provided a comprehensive framework for later studies on the ideological import of language.. 2.2.1.4 A summary of Critical Discourse Analysis As pioneers in the study on language and ideology, Fowler and his associates focused on linguistic characteristics, including lexical choices and syntactic patterns, and formed critical linguistics. Their analysis of the lexical and clausal levels was expanded to the discourse level in critical discourse analysis by Fairclough and van Dijk, who are interested in the context in which a text is produced and consumed. Fairclough has presented a three-dimensional framework with text, discourse, and society, whose concern is on how language may reproduce and restructure ideologies and social structures. The structure facet of news discourse was not fully explored until van Dijk, who builds up macro- and micro-structures of news, and works with a more cognition-oriented triangle theory of discourse, cognition, and society. It should be noted that the terms ‘text’ and ‘discourse’ denote disparate 23.

(37) conceptions in divergent discussions. ‘Text’ in critical linguistics falls upon lexical choices and syntactic patterns; similar concepts are ‘vocabulary’ and ‘grammar’ in Fairclough and ‘style’ in van Dijk. To Fairclough, ‘text’, meaning more than lexicon and sentences, includes cohesion between sentences and the overall textual structure. ‘Text’ in Fairclough, however, is under the ‘discourse’ heading in van Dijk with macrostructures being the overall textual structure and microstructures dealing with vocabulary, grammar and cohesion. ‘Discourse’ in Fairclough focuses on how a ‘text’ is produced or interpreted by social members, which quite resembles the way van Dijk discusses in ‘cognition’ how people’s models and scripts influence their production and interpretation of a ‘discourse’. In spite of those terminological differences, Fairclough’ and van Dijk’s frameworks are much alike in that they both emphasize the overall structure of news, people’s production and interpretation processes involved, and the social factors lying behind. These emphases are what distinguish their frameworks of critical discourse analysis from critical linguistics.. 2.2 Empirical Studies The major goal of critical linguistics and critical discourse analysis is to make explicit the underlying ideologies of seemingly neutral and objective language. The analysis of language and ideology has significantly contributed to various social studies,. including. sexual. discrimination. (Fairclough. 1992,. Fowler. 1991,. Caldas-Coulthard 1993), racism (Achugar 2004, Erjavec 2001, Fairclough 1995, Flowerdew et al. 2002, Fowler et al. 1979, Richardson 2004, Teo 2000, Thetela 2001, van Dijk 1991), and political science (Fairclough 1989, Fang 1994, 2001, Fowler 1991, Fowler et al. 1979, Kuo 2001, 2007, Kuo and Nakamura 2005, Lee and Craig 24.

(38) 1992, Wang 1993). This thesis aims to examine how language signals the underlying political stances of four TV news stations in Taiwan. In what follows, critical studies on political inclinations are introduced, with westerns studies first, followed by China- and Taiwan-related studies. Apart from the analysis of the Salisbury riot discussed in Section 2.2.1.1, Trew (1979b) examined how two British newspapers, the Sun and the Morning Star, reported the 1977 Notting Hill Carnival, in which black youths confronted the police. Through the analysis of transitivity and characterizing expressions used in the news, it was found that as a tabloid which benefited from capitalism, the Sun was hostile to the ‘black’. On the contrary, being an official organ of the Communist Party of Great Britain, the Star sympathized with the ‘youths’, who were deprived by the society and in need of help instead of punishment. Also on British newspapers, Fowler (1991) studied how they responded to the bombing of Libya by the US with the help of the UK on April 15, 1986. The Guardian condemned the action, and remained neutrality by using names or formal titles to refer to principals involved. In contrast, the Sun was ecstatic over the bombing and extensively employed evaluative terms, such as ‘Rambo’ for then US President Reagan, ‘Maggie’ for then UK Premier Thatcher, and ‘mad dog’ to then Libya Colonel Gaddafi. These all served as substantial clues to the Sun’s position in the tension between the US/UK and Libya. On the other side of the Atlantic, language and its ideological properties also attracted considerable attention in the US. Herman and Chomsky (1988) proposed the propaganda model to explain the performance of the US press. As so-called gatekeepers, the press uses a set of filters5 to sift out events that comport with its 5. Herman and Chomsky (1988) brought out five filters. First, size, ownership, and profit orientation of the mass media; second, the advertising license to do business; third, sourcing mass-media news; forth, flak and the enforcers; fifth, anticommunism as a control mechanism. 25.

(39) ideologies as news reports. To defend capitalism, the very root of western values and media, the anti-communism fervor makes victims in a communist country ‘worthy’, while those in friend countries ‘unworthy’. The consequence is that disturbances in an enemy country stand a greater chance of becoming headline stories than those in an ally, which are likely to be marginalized or even concealed. The discriminatory treatments constitute the ‘us versus them’ dichotomy. The press acts in propaganda to pursue and defend its own interests and its language can be best understood if it is analyzed in this propaganda model. In illustrating the propaganda model, Lee and Craig (1992) analyzed the US reports on labor strikes in South Korea and Poland during 1980s. The content analysis showed that more reports, longer length and more prominence were dedicated to strikes in Poland than those in South Korean. In the critical discourse analysis, it was found that the strike in South Korea was considered a wage dispute, in which the government performed as a successful mediator. Nevertheless, the strike in Poland was portrayed as a political dissidence, a challenge to the communism and a demand for freedom, and the communist government was discredited to be weak and indecisive. In addition to research on the western media, studies targeting at the China’s and Taiwanese press are also fruitful. Wang (1993) compared the Chinese Communist Party newspaper, Renmin Ribao ‘the People’s Daily’, with the New York Times in their coverage of the 1991 Soviet coup. In the content analysis, it was found that the Times devoted a lot more news items than Ribao to the coup, yet with similar percentages of front-page stories. In the critical discourse analysis, according to the macrostructures, the Times report was exhaustive in content and multi-angled in structure, while the Ribao report was comparatively incomplete. In the micro-level, the Times gave more prominence 26.

(40) to reformers Yeltsin and Gorbachev, contrary to Ribao’s attention to hardline Yanayev. What’s more interesting is that Ribao never referred to the event as a ‘coup’, but as a ‘power transfer’. The implicit but systematic linguistic differences between the Times and Ribao suggested their opposing positions to the 1991 Soviet ‘coup’. Another study on Renmin Ribao is Fang (1994), in which critical linguistics was adopted to examine how the ‘us versus them’ dichotomy influenced its coverage of civil unrest in foreign countries. Results showed that different labels and accounts were given to different countries according to the dichotomy. Thus, unrest in ‘us’ countries was baoluan/saoluan ‘a riot’, and the rioters should bear the bitter consequences of their irrational behaviors. Nevertheless, unrest in ‘them’ countries was shiwei ‘a demonstration’ and shiwei youxing ‘a march’, in which the people had the rights to voice themselves, and the authority was to blame. Years later, Fang (2001) adopted critical discourse analysis in studying Renmin Ribao (RR) and the Taiwanese newspaper Central Daily News (the CDN). The two newspapers reports on civil unrest in South Africa and Argentina were analyzed. Argentina established amicable relations with both Taiwan and China, but South Africa maintained friendly relations with Taiwan, but not with China. In the South African reports, two opposing scripts were employed by RR and the CDN to account for the outbreak of the conflict. For RR, it was a ‘racial struggle’ in which the black were victims; to the CDN, however, it was a ‘law and order’ crisis in need of riot control actions. Moreover, it was found that events against the newspaper’s perspectives were marginalized or omitted in the reports. Therefore, RR did not cover the incident that black officers were also under attack, whereas the CDN used it as evidence to support its ‘law and order’ frame. Nevertheless, the CDN downplayed the event that the police opened fire on unarmed mourners. On the other hand, the two newspapers reports on the Argentina ‘riots’ were 27.