BIASED REPRESENTATION OF HILLARY CLINTON S SCANDAL IN THE MEDIA

Transcription

1 BIASED REPRESENTATION OF HILLARY CLINTON S SCANDAL IN THE MEDIA EVALUATIVE LANGUAGE ON FOX NEWS AND NEW YORK MAGAZINE Written by: Ákos Csernák Supervisor: Dr. Nicole Baumgarten Date of submission: 1 June 2016 Total number of typed characters: 157,346 SYDDANSK UNIVERSITET UNIVERSITY OF SOUTHERN DENMARK

2 Table of Contents Table of Contents... i List of Figures... iv Plagiarism Statement... vi Summary... vii 1. Introduction Hillary Clinton s Controversy Literature and Theoretical Framework Measuring Media Bias The Appraisal Framework: Evaluation in Verbal Language The Appraisal Framework in Linguistics The Appraisal Framework: Attitude Overview The Appraisal Framework: Engagement Overview The Appraisal Framework: Graduation Overview Analytical Approach Data Fox News and New York Magazine: Overview foxnews.com and nymag.com: Overview Situational Characteristics of Fox News and New York Magazine News Articles Data Collection Timeframe Defining Written News Articles Selecting News Articles Using Predefined Criteria Selecting News Articles Using Extended Criteria Data Overview Further Commentary on Data: Article # Further Commentary on Data: Article # Data Measurements Design of the Analytical Framework Design of the Appraisal Framework: Attitude Attitude Affect Attitude Judgement i

3 Attitude Appreciation Design of the Appraisal Framework: Engagement Engagement Heteroglossia Expand Engagement Heteroglossia Contract Engagement Heteroglossia Borderline Design of the Appraisal Framework: Graduation Graduation Reckoning Graduation Scaling Graduation Isolation Design of the Appraisal Framework: Summary And Out-of-Scope Lexical Items Results Frequencies of Attitude, Engagement, and Graduation Normed Rates of Occurrence Occurrences in Total Frequencies of Attitude Categories Frequencies of Engagement Categories Frequencies of Graduation Categories Attitude: A Closer Look Attitude Affect Attitude Judgement Attitude Appreciation Engagement: A Closer Look Engagement Attribute Engagement Entertain Engagement Proclaim Engagement Disclaim Graduation: A Closer Look Graduation Reckoning Graduation Scaling Graduation Isolation Implicit Meaning Making Implicit Meaning Making on Fox News ii

4 Implicit Meaning Making on New York Magazine Discussion Conclusion References... xiii Primary Sources... xiii Secondary Sources... xvii Appendices... xxiv Appendix 1. Part of the New York Magazine News Archive... xxiv Appendix 2. Part of the Fox News News Archive... xxv Appendix 3. Attitude Affect in gate News Articles... xxvii Appendix 4. Attitude Judgement in gate News Articles... xxviii Appendix 5. Attitude Appreciation in gate News Articles... xxix Appendix 6. Engagement Attribute in gate News Articles... xxx Appendix 7. Engagement Entertain in gate News Articles... xxxi Appendix 8. Engagement Proclaim in gate News Articles... xxxii Appendix 9. Engagement Disclaim in gate News Articles... xxxiii Appendix 10. Graduation in gate News Articles: Fox News... xxxiv Appendix 11. Graduation in gate News Articles: New York Magazine... xxxv iii

5 List of Figures Chart 1. Instances of Attitude, Engagement, and Graduation per dataset Chart 2. Frequencies of Attitude Categories Chart 3. Monoglossic vs. Heteroglossic Utterances Chart 4. Frequencies of Engagement Categories Chart 5. Frequencies of Graduation Categories Figure 1. The Appraisal Framework in Systemic Functional Linguistics, Its Three Domains and Their Categories... 9 Figure 2. Analytical Approach: Repeated Cycles of Data Analysis and Framework Consultation Figure 3. Attitude Categories in Fox News and New York Magazine gate News Articles Figure 4. Affect Subcategories in Fox News and New York Magazine gate News Articles Figure 5. Judgement Subcategories in Fox News and New York Magazine gate News Articles Figure 6. Appreciation Subcategories in Fox News and New York Magazine gate News Articles Figure 7. Engagement Categories in Fox News and New York Magazine gate News Articles Figure 8. Heteroglossia Expand Subcategories in Fox News and New York Magazine gate News Articles Figure 9. Heteroglossia Contract Subcategories in Fox News and New York Magazine gate News Articles Figure 10. Heteroglossia Borderline Subcategories in Fox News and New York Magazine gate News Articles Figure 11. Graduation Categories in Fox News and New York Magazine gate News Articles Figure 12. Graduation Reckoning Subcategories in Fox News and New York Magazine gate News Articles Figure 13. Graduation Scaling Subcategories in Fox News and iv

6 New York Magazine gate News Articles Figure 14. Graduation Isolation Subcategories in Fox News and New York Magazine gate News Articles Figure 15. The Full Layout of the Appraisal Framework Figure 16. Voices and Issues at Stake in Clinton's Controversy Table 1. Key Developments in Clinton's Controversy... 5 Table 2. Examples of Attitude Affect Table 3. Examples of Attitude Judgement Table 4. Examples of Attitude Appreciation Table 5. Examples of Engagement Monoglossia Table 6. Examples of Engagement Heteroglossia Table 7. Examples of Graduation Force Table 8. Alexa Internet Metrics for foxnews.com and nymag.com Table 9. Situational Characteristics of Fox News and New York Magazine Articles Table 10. The Dates of Publication and Titles of the 2x14 gate News Articles Table 11. Word Count of Articles on gate Table 12. Mentions of Hillary Clinton in gate Articles Table 13. Out-of-Scope Lexical Items in Articles on gate Table 14. The Calculation of Normed Rates of Occurrence Table 15. Ten Fox News Sentences with Implicitly Evaluative Language Table New York Magazine Sentences with Implicitly Evaluative Language v

7 Plagiarism Statement I hereby solemnly declare that I have personally and independently prepared this paper. All quotations in the text have been marked as such, and the paper or considerable parts of it have not previously been subject to any examination or assessment. Ákos Csernák vi

8 Summary The aim of this thesis is to investigate and identify evaluative language in news articles that discuss Hillary Clinton s controversy. Hillary Clinton s scandal, often labeled as gate, arose in March 2015, one month before she announced her run for President of the United States of America. Since announcing her presidential bid, Clinton and the development of her campaign have been under close scrutiny in the press. Reports on her controversial practices, however, have received frequent and heavy media coverage since March 2015, shifting the focus of the media from her campaign efforts. The online edition of Fox News (foxnews.com) was identified as explicitly opposing the Democratic aspirant Clinton s presidential bid, whereas the online edition of New York Magazine (nymag.com) was identified as explicitly supporting it. The question arose as to whether Fox News and New York Magazine present gate news articles in a language that reflects partisan bias, as well. Thus, this thesis aims to answer the following research question. Is there evaluative language in Fox News and New York Magazine news articles discussing Hillary Clinton s controversy? If so, how is evaluative language expressed and how does it build media bias? In answering the research question, the thesis first offers an overview of how the key events in Hillary Clinton s controversy have unfolded and what repercussions they have had in the media. Clinton s practices gained media attention as The New York Times revealed in March 2015 that the former secretary of state of the United States exclusively used a personal account on her private home server to conduct official business. As secretary of state and head of the State Department, Clinton was responsible for the foreign policy of the U.S. between 2009 and A 2012 terrorist attack on U.S. diplomatic facilities in Benghazi, Libya, had resulted in the death of four American officials and had prompted an official investigation into the accountability of U.S. officials in the events. As part of the investigation, requests were filed for the correspondences of U.S. officials including those of Hillary Clinton. The fact that only eight Benghazi-related Clinton s were submitted in response to the record requests prompted a Vice News investigative reporter to file a Freedom of Information Act in search of other Clinton correspondences. vii

9 Upon The New York Times breaking the news in March 2015 that Clinton s official correspondences passed through her private address on a private server, Clinton claimed to have deemed 30,000 s work-related, which she delivered to the authorities, and she claimed to have deemed 30,000 pages of s private, which were deleted without turning them over. Clinton also said that her s contained no classified information. In response to the news, a federal judge ordered the 30,000 work-related s to be released to the public on a monthly basis. As two inspectors general found that potentially hundreds of Clinton s s could have classified information, the Federal Bureau of Investigation launched its own investigation into the Clinton s. The FBI soon managed to recover the 30,000 pages of deleted private s from Clinton s server. In the media, Clinton s practices generated debate about information security issues and Clinton s responsibilities to defend the nation s security as secretary of state. gate became a frequently discussed topic in the presidential candidates political debates, as well. The issue will likely remain a revisited topic for the rest of 2016 if Clinton wins the presidential nomination of the Democratic Party, especially because the FBI s probe into Clinton s practices has not yet concluded. The thesis next reviews literature on the measuring of media bias. A common means of assessing partisan bias in the media is to investigate language with regard to evaluative/opinionated language use. Martin and White s Appraisal framework is one way to identify evaluation in verbal language. The Appraisal framework relies on Systemic Functional Linguistics, a paradigm of Michael Halliday that approaches language from a social semiotics perspective. The Appraisal framework comprises three domains: Attitude, Engagement, and Graduation. The Appraisal framework is introduced in chapter 3 and its domains exemplified via sentences from written news articles on Fox News and New York Magazine that discuss Hillary Clinton s controversy. Researchers who apply the Appraisal framework to field-specific texts often make adjustments to the framework. In this thesis, some modifications were made to the framework. As a result, the full layout of the system follows in a later chapter. Chapter 3 also introduces the analytical approach used in this thesis. To identify evaluative language on Fox News and New York Magazine, induction was used. viii

10 Following the inductive approach, the data, written news articles published on the two media outlets, were taken as the starting point. First, the data were observed for recurring linguistic elements that construe potentially evaluative meanings. Next, the Appraisal framework was consulted to systematically interpret the found linguistic elements. Afterwards, the data were observed again for a deeper scrutiny, after which the Appraisal framework was consulted again to interpret the new findings. This two-step analytical approach developed into a repeated cycle of analysis and framework consultation until the data were fully exposed for evaluative language and the found linguistic patterns systematically interpreted in the Appraisal framework. The thesis next introduces the data in greater detail, including an overview of Fox News and New York Magazine. Fox News is the only outlet on a national scale to propagate Republican views. New York Magazine is a lesser-known media outlet commonly not included in media bias studies. It is unofficially viewed as having liberal bias. The online edition of Fox News has a bigger reader base than New York Magazine. Foxnews.com is among the 50 most visited American websites, while nymag.com ranks in the top 500. The situational characteristics of Fox News and New York Magazine articles are detailed next. A marked difference between the two media outlets is that the author of many Fox News news articles remain unnamed, decreasing the interactiveness between writers and readers. Next, data collection is detailed. The data investigated in this thesis are 14 gate news articles published on Fox News and 14 news articles published on New York Magazine between 12 April 2015 and 1 February The start date marks Clinton s official announcement of her presidential bid. The end date marks the first Democratic electoral contest in 2016, the first in a series of contests that ends in June 2016, and is designed to help the Democratic Party select its presidential nominee. The 2x14 news articles were then filtered for sentences that reference Hillary Clinton. 410 such sentences were found, which were selected as the core of the data analysis. Narrowing the data was important in order for the focus of the research to shift to how Clinton as a public figure and Democratic presidential candidate is depicted on Fox News and New York Magazine against the backdrop of her controversy. ix

11 The thesis then returns to the Appraisal framework and lays out the system in its entirety based on the data. The Appraisal framework needed modifications in all of its three domains (Attitude, Engagement, and Graduation) in order for it to better accommodate the field-specific language that is used to describe Hillary Clinton and gate. First, the Attitude domain is opened up, with its categories and their subcategories explained in detail and exemplified through language taken from the data. Attitudinal meanings are words that signal emotional responses (e.g. love, excited, fear, anxiety, sympathy), evaluate people s behaviour (e.g. controversial, unusual, shady, sorry), and assess things, objects, and performances (e.g. unsecured server, boring s). Second, Engagement, the second domain of the Appraisal framework, is explained in detail with examples taken from the data. Engagement deals with whether writers and speakers express only their own take on an issue, or their language represents other voices, other opinions, and other takes on an issue, as well. The former is called Monoglossia, where writers present their own ideas, feelings, and thoughts only. The latter is called Heteroglossia. Heteroglossia takes place when writers express that the idea presented in their utterances is only one position out of a range of other positions. The existence of other voices with different ideas, feelings, and thoughts, is recognised. For instance, when a journalist quotes a spokesman or Clinton, he or she introduces a new voice into the text and expresses that the quoted material does not necessarily represent his or her take on the issue. Another example of Heteroglossia is to use negation. When a journalist writes that Clinton never sent classified information, he or she introduces another voice into the text, according to which Clinton sent classified information, a position that is shown not to hold via negation. Third, Graduation, the third domain of the Appraisal framework, is explained in detail. Martin and White, the linguists who created the Appraisal framework, suggest that Graduation is not a separate third domain, but instead it overarches both Attitude and Engagement. Graduation is language that down-scales or up-scales meanings. For instance, highly unusual carries an attitudinal meaning (unusual) up-scaled by the word highly. Some instances of Graduation are less clear-cut. To refer to Clinton s scandal as saga, stories, and an Internet meme is to down-scale gate. x

12 The thesis then presents the research results. New York Magazine news articles are found to contain more evaluative language than those on Fox News. Evaluations are used differently by the two outlets, and they build different dispositions toward Clinton and gate. As for Attitude, Fox News removes the emotional aspect of gate, negatively assesses Clinton s truthfulness and the appropriateness of her usage, and often questions the confidentiality of the Clinton s. New York Magazine describes Clinton in a human context surrounded by friends and everyday matters and hobbies, evaluates the media s attack on gate negatively, and describes the Clinton s as containing uninteresting details on Clinton s everyday life. As for Engagement, Fox News journalists often distance themselves from what Clinton has said in relation to gate, introduce denial into Clinton s statements, and hypothesise that top-secret information leaked from Clinton s use of s, thus jeopardizing the country. New York Magazine journalists often refer to groups of people who intend to find incriminating evidence against Clinton, introduce commentary to dismiss these groups of people, stress that Clinton did not break any laws by using a personal server, and negate statements that otherwise use incriminating language about Clinton. As for Graduation, Fox News up-scales the seriousness of gate, describes it as a long-lasting scandal, blows up the amount of s involved, uses adjectives and adverbs that increase the severity of gate, and lays emphasis on select s that imply that top-secret information passed through Clinton s private server. New York Magazine lays emphasis on select s that contain insignificant details of Clinton s life at the office, describes the media s attention as overblown, and expresses that certain groups of people have failed to use the scandal to hinder Clinton s presidential run. The thesis then interprets the findings. Fox News journalists overplay gate and express disbelief of Clinton s claims in the issue. New York Magazine journalists underplay gate, explicitly dismiss Republicans as the people who wish to find incriminating evidence against Clinton, and present Clinton in a relatable way. These over- and underplaying tones remained unchanged on the two media outlets between 12 April 2015 and 1 February This is interesting because Clinton s usage evolved from a political xi

13 scandal into a morally questioned practice followed by legal consequences. In total, both outlets were found to exhibit bias in their presentation of gate news. Fox News exhibits negative bias toward Clinton. New York Magazine uses evaluative language more often than Fox News, exhibits a positive bias toward Clinton, and makes its Democratic partisan bias explicit. The thesis concludes by stating the interesting aspects of the research. The research sheds light on the responsibility of political journalists in presenting news stories using evaluative language and, thus, shaping how these stories reach the public. This is relevant because gate has been a highly divisive political issue since 2015, possibly contributing to Clinton s lessened chances of winning the presidential nomination. The research also identifies what linguistic resources can be used to defend public figures and their deeds in front of a wide audience. Using these linguistic resources can help protect people and other entities when facing controversies and negative publicity. From the perspective of crisis communication, the research shows language that has been adjusted to aggravate or mitigate issues such as scandals. This type of language can be used to over- and underplay key actors and entities responsibilities in controversies. Lastly, the thesis identifies formatting conventions (parentheses, scare quotes, dashes) used in journalistic discourse as bearing evaluative potential. As these formatting options can be used to infuse written texts with authorial opinions and ambiguity, their further research and possible addition to the Appraisal framework are recommended. xii

14 1. Introduction As President Barack Obama concludes his second term as President of the United States, America prepares to elect its 45 th president in November saw a number of Democratic and Republican politicians announce their run for the 2016 presidential nomination, launch campaigns and fundraisers, and make clear their political stance on issues such as immigration, taxes, health care, and other social matters. Among the Democratic candidates is Hillary Rodham Clinton, secretary of state of the United States from 2009 to 2013, who is currently considered to be the Democratic front-runner for president. Since Hillary Clinton announced her candidacy for president in April 2015, the national media have covered her political stance on economic and social issues and her campaign-related news extensively. However, a matter of a different kind has been catching the headlines, often overshadowing the politician s campaign. Clinton's controversy, labeled as gate, emerged in March gate soon evolved into a scandal and became one of the most heavily publicised controversies that an American politician was involved in in The controversy arose as it was revealed that Clinton used a personal account, hosted on a privately run server in her New York home, to conduct official correspondence during her tenure as secretary of state. Shortly after the news broke in March 2015, Clinton announced that she had turned over around 30,000 pages of work-related exchanges for the authorities to investigate whether the s contained classified information that could jeopardize the nation s security. Clinton had considered personal another set of approximately 30,000 pages worth of , which her campaign team consequently deleted from the private server without turning them over. Clinton s campaign news, including her controversy, has received wide media coverage in the past year. I found two media outlets to either oppose her presidential bid or support it. Fox News, a prominent conservative media outlet leaning toward Republican values, has commented negatively on Clinton s run for presidency, saying, nothing better defines the hypocrisy of the left than Hillary Clinton s likely coronation as the 2016 Democratic presidential candidate (Peek, 2015, para. 1). New York Magazine, a New Yorkbased online magazine covering politics and culture, has shown a more positive disposition 1

15 toward Clinton s candidacy, saying, the argument for Clinton in 2016 is that she is the candidate of the only major American political party not run by lunatics. There is only one choice for voters who want a president who accepts climate science and rejects voodoo economics, and whose domestic platform would not engineer the largest upward redistribution of resources in American history (Chait, 2015, para. 25). Based on this explicit side-taking by Fox News and New York Magazine, I decided to research whether these two media outlets display media bias in their propagation of gate news. In general, news articles are expected to preserve neutrality and factuality, and minimise commentary. The question arises as to whether the news pertaining to Clinton s controversy contains evaluative language owing to the media outlets negative or positive disposition toward Clinton s presidential bid, i.e. whether journalists exploit gate in their political side-taking. Thus, my research question can be formulated in the following way. Is there evaluative language in Fox News and New York Magazine news articles discussing Hillary Clinton s controversy? If so, how is evaluative language expressed and how does it build media bias? This thesis is structured as follows. I first give a detailed account of Hillary Clinton s controversy in chapter 2. In chapter 3, I first review literature pertaining to the measuring of media bias. Next, I present the Appraisal framework, a system that aims to reveal evaluative language in discourse. Further, I introduce the analytical approach used to conduct the research. In chapter 4, I introduce the contextual properties and production circumstances of foxnews.com and nymag.com, the two sources of my investigation. Then I discuss my data collection process, i.e. the criteria used to select Fox News and New York Magazine gate news articles. In chapter 5, I lay out the Appraisal framework along with the adjustments made to it for the investigation of gate news articles. In chapter 6, I present my findings by showing explicit evaluative language use in my data. In addition, I show examples of the interaction of evaluative meanings that result in implicit evaluation. In chapter 7, I discuss my findings, including how evaluative language contributes to media bias in the two media outlets, and present limitations to my research. In chapter 8, I summarise my thesis. 2

16 2. Hillary Clinton s Controversy In this chapter, I offer an overview of the main events of Hillary Clinton s controversy including their repercussions in the media. In January 2009, Hillary Rodham Clinton became secretary of state of the United States and head of the State Department, primarily responsible for the country s foreign policy. On 11 September 2012, a half a year before the end of Clinton s tenure, Islamic militants attacked two U.S. diplomatic facilities in Benghazi, Libya, killing four American officials. Clinton took responsibility for the Benghazi attack (New York Daily News, 2015). A House Select Committee was appointed in May 2014 to investigate what led to the Benghazi attack and how it could have been prevented, and to assess the accountability of U.S. officials in the attack (About the Select Committee, n.d.). The Committee sent record requests to the State Department to obtain Benghazi-related documents including correspondences. In response, the State Department delivered 15,000 documents to the Committee. The batch included only eight Hillary Clinton s, harvested from state.gov accounts (Kiely, 2015). This raised questions given that Clinton was the President s principal foreign policy adviser at the time of the Benghazi attack and leader of the State Department. In search of other Clinton s about Benghazi, Jason Leopold, a Vice News investigative reporter, filed a Freedom of Information Act in November 2014, when it became clear that [Hillary Clinton] would announce her candidacy for US president (Leopold, 2016, para. 11). In response, Clinton s office delivered 55,000 pages of workrelated s to the State Department the following month. The threads were harvested from her private server. On March , The New York Times revealed that Clinton exclusively used a personal account to conduct government business as secretary of state, which is a possible breach to the Federal Records Act (Schmidt, 2015, para. 3). The New York Times broke the news the following day that several record requests sent from the House Select Committee to the State Department came up empty in the past because the State Department had no access to Clinton s privately stored s (Schmidt & Chozick, 2015, para. 8). 3

17 Clinton herself addressed the reports for the first time the following week and claimed to have already delivered her work-related s (referring to the 55,000 pages/30,000 s turned over in November 2014) to the State Department and to not have kept s within the scope of my personal privacy (amounting to 30,000 pages of s). Clinton also pointed out that she did not any classified material to anyone from her private server (The New York Times, 2015). At the end of May 2015, the State Department released a batch of Clinton s in relation to the Benghazi attack (Virtual Reading Room Documents Search Results, 2015). A week later, a federal judge ordered the State Department to release some of the 55,000 Clinton s on a monthly basis starting June 30 and ending Jan. 29, 2016 (Ferrechio, 2015). After the monthly releases began, The New York Times revealed that two inspectors general requested the Federal Bureau of Investigation to conduct a criminal investigation into Clinton s usage after announcing in a joint memo that Clinton s private account contained hundreds of potentially classified s (Schmidt & Apuzzo, 2015). In August, the FBI claimed Clinton s private server and began investigating its contents. Later that month, the question arose as to whether Clinton had deleted her private chains or wiped her server. When a reporter asked her about wiping her server, she responded, What, like with a cloth or something?, furthering the controversy around her use of s (The Wall Street Journal, 2015). In September 2015, IT specialist Bryan Pagliano, who had helped set up Clinton s private server in her New York home, was announced to invoke his Fifth Amendment right, meaning that he would not have to testify in front of the House Select Committee and answer questions about Clinton s server. Later that month, Clinton was interviewed by MSNBC s Andrea Mitchell and said that she was sorry that [her usage] has been confusing to people (MSNBC, 2015). Some media outlets interpreted this as an act of apologizing for the confusion but not for her usage. According to Maureen Dowd of The New York Times, [Clinton] leaves the impression that she is merely sorry to be facing criticism, not that she miscalculated in the first place (Dowd, 2016, para. 17). 4

18 At the end of September, the FBI recovered the 30,000 Clinton s from her private server that had been deemed private and deleted/wiped from the server previously. The State Department also received work-related Clinton s that had not been turned over by Clinton despite her claims that she and her team had provided all of them (CBS News, 2015). Later, the FBI requested Datto, Inc., a tech company that had been unknowingly backing up Clinton s private server, to hand over all the data that the company had on the Clinton s. A month later, the intelligence community announced that two Clinton s previously thought to contain top-secret information were no longer believed to have classified content (Gerstein, 2015). In the meantime, the State Department continued with its monthly release of Clinton s. The Department published over 3,000 s on the last day of 2015, but failed to meet its target of releasing at least 82% of all Clinton s by the end of To meet the goal, it released the remaining s a week later. In January 2016, the intelligence community announced that several dozen s found on Clinton s server contained intelligence known as Special Access Programs. SAP information exceeds regular top-secret classifications (Dilanian, 2016). At the end of the month, the State Department confirmed that 22 Clinton s found on her private server contained top-secret information and withheld their release. For later reference, Table 1 gives a summary of the key events in Clinton s controversy. KEY DEVELOPMENTS IN CLINTON S CONTROVERSY Date Development January 2009 February 2013 Hillary Clinton s tenure as secretary of state September 2012 Benghazi attack on the American consulate May 2014 U.S. House Select Committee established August 2014 Official server data sent to House Select Committee November 2014 Request for Benghazi-related Clinton s December 2014 Private server data sent to State Department February 2015 Private server data sent to House Select Committee 2 March 2015 The New York Times breaks the news about Clinton s use of a private server 5

19 3 March 2015 The New York Times breaks the news about failed Clinton requests March 2015 Clinton holds press conference, speaks up in the issue 22 May 2015 Benghazi-related Clinton s released to the public 27 May 2015 Court order issued requiring regulated release of Clinton s 30 June nd batch of Clinton s released 24 July 2015 FBI requests to conduct investigation into Clinton s 12 August 2015 FBI begins investigation into Clinton s 18 August 2015 Clinton gives controversial reply about deletion 3 September 2015 IT specialist working on Clinton s private server said to take the 5th Amendment 4 September 2015 Clinton interviewed by Andrea Mitchell, says sorry for confusion 23 September 2015 FBI recovers erased s from Clinton s private server 25 September 2015 Clinton s previously not turned over received by State Department 7 October 2015 FBI requests backed up data on Clinton s private server 6 November 2015 Two, previously classified Clinton s no longer thought classified 31 December January 2016 End-of-the-year batch of Clinton s released late 20 January 2016 Several dozen Clinton s found to have intelligence beyond top secret 29 January Clinton s contain top secret information, release withheld Table 1. Key Developments in Clinton's Controversy Although the State Department s release process concluded at the end of February, the FBI s own investigation into Clinton s usage has continued to generate publicity in the media throughout the spring of gate has also been frequently covered in the candidates political debates. Additionally, because The State Department has not ruled out conducting its own review after the FBI investigation is complete, Clinton s controversy is bound to receive media coverage for the rest of her campaign work (Fishel, 2016, para. 12). 3. Literature and Theoretical Framework In this chapter, I first review literature on how media bias has been measured in the past. Afterwards, I introduce the Appraisal framework, an analytical model for the investigation of evaluative language in discourse. Lastly, I introduce the data-driven analytical approach that I used to investigate evaluative language in gate news articles on Fox News and New York Magazine. 6

20 3.1. Measuring Media Bias The American journalistic discourse is often perceived as exhibiting partisan bias, i.e. a conservative or liberal political bias (Covert & Wasburn, 2009). This is held to commonly manifest in a slanted coverage of either conservative or liberal perspectives in the discussion of core social issues. A common approach to measure media bias is to investigate how much coverage these issues receive in a media outlet. For example, Covert and Wasburn (2009) looked at the extent to which four social issues (crime, the environment, gender, and poverty) were covered in two national news magazines and journals. According to the researchers, an imbalanced coverage of these topics leads to the preferential propagation of partisan issue positions. Journalists can attain these slanted representations by way of advancing the political perspectives that they align with and omitting or dismissing positions that belong to the opposite political side. Covert and Wasburn (2007) also remark that investigating language use is another widely used means of detecting bias, because, as communication theorists have argued, bias in news reporting is unavoidable due to the inherently evaluative character of language (Geis, 1987). Bednarek (2006) looked at the expression of opinion through language in her investigation of evaluation in media discourse (Bednarek, 2006, p. 3). Bednarek investigated how tabloid and broadsheet journalists evaluate the world along the clines of, for instance, what is good/bad, important/unimportant, expected/unexpected, as well as the reliability of propositions and other parameters. Bednarek also reviews the Appraisal framework, which this thesis builds on, and identifies issues within it. (Bednarek, 2006, p ). For instance, she points out that categorising words and propositions within the framework system is sometimes problematic. On the other hand, she also states that the Appraisal framework is an ongoing research project and the result of an over 20-year-long research (Bednarek, 2006, p. 34). Given the progressive nature of the Appraisal framework, the system allows for extensions and amendments when field-specific texts are under investigation. For example, Hommerberg and Don (2015) used and amended the Appraisal framework to analyse winespeak (Hommerberg & Don, 2015). Furthermore, Economou (2009) adapted the 7

21 Appraisal framework from the verbal to the visual mode and analysed evaluation in news photos (Economou, 2009) The Appraisal Framework: Evaluation in Verbal Language In this section, I lay out an initial overview of the Appraisal framework as developed by Martin and White (Martin & White, 2005). The data that I applied the Appraisal framework to are two text sets: written news articles on Hillary Clinton s gate scandal published by Fox News and New York Magazine. In order to investigate evaluative language in these text sets, some adjustments to the Appraisal framework were necessary. Therefore, I describe the full Appraisal framework specific to my text sets in chapter 5. Here, I briefly describe the framework as established by Martin and White and exemplify the Appraisal domains with instances from the Fox News and New York Magazine text sets The Appraisal Framework in Linguistics The Appraisal framework was developed by Martin and White with the aim of identifying linguistic resources in texts that authors rely on both to formulate stances and to position their audiences to supply their own assessments (Martin & White, 2005, p. 2). As the definition indicates, the Appraisal framework looks at language from an interpersonal point of view, i.e. the way people interact and attempt to share feelings and thoughts with each other. Interpersonal meaning comprises one of the three modes of meaning within Systemic Functional Linguistics, a paradigm of Michael Halliday and an approach to language from a social semiotics perspective (Martin & White, 2005, p. 8). The Appraisal framework comprises three domains that work together in the creation of evaluation: Attitude, Engagement, and Graduation. Figure 1 1 shows the modes of SFL, the Appraisal framework, and the three framework domains along with their categories. 1 The layout of the Appraisal framework as proposed by Martin and White (2005) is not reproducible. Therefore, I altered the visual layout of the system and use this color-coded version throughout this thesis paper. 8

22 Systemic Functional Linguistics (SFL) Textual mode Ideational mode Interpersonal mode A ATTITUDE AFFECT JUDGEMENT APPRECIATION APPRAISAL FRAMEWORK E ENGAGEMENT MONOGLOSSIA HETEROGLOSSIA G GRADUATION FORCE FOCUS Figure 1. The Appraisal Framework in Systemic Functional Linguistics, Its Three Domains and Their Categories Before introducing each domain, it is important to note that evaluative meaning making can be explicit and implicit. Positive and negative evaluations, i.e. stances, are sometimes explicitly realized by a single lexical item that falls under Attitude, Engagement, or Graduation. Often, however, meanings are not isolated in such clear-cut ways. Instead, the Appraisal domains and their categories co-occur and blend to facilitate more subtle, i.e. nonexplicit meaning making. These types of meanings are constructed in a way that they, when dissected into separate lexical items, lose their evaluative values. As a result, instead of limiting my research to only one domain within the Appraisal framework, I analyzed occurrences across all of the three domains in my data. 9

23 According to Martin and White (2005), evaluative linguistic resources can be explicit, implicit, or provoked. In explicit instances, a single lexical unit carries the meaning. Implicit statements look factual on the surface, apparently stripped of any evaluation, but they have the potential to evoke evaluation in the reader depending on whether the reader shares the emotional, cultural, and/or aesthetic values presented by the writer. Attitudinal positioning can also be provoked. Lying between explicit and implicit evaluations, provocations contain detectable, vaguely evaluative language, where a single lexical unit cannot be pinpointed as the carrier of stance. As a main focus of my research is to show how explicit evaluations cooccur and interact to express subtle authorial stances, this thesis maintains the difference between explicit and non-explicit categories. However, due to the sometimes fuzzy boundaries between provoked and implicit evaluations, I do not make further distinctions within non-explicit meaning making. Instead, I discuss both implicit and provoked categories under implicit evaluative meaning making The Appraisal Framework: Attitude Overview Attitude, the first domain within the Appraisal framework, is the sum of linguistic resources used for the signalling of emotional responses (Affect), the evaluation of people s behaviour (Judgement), and the assessment of things, objects, natural phenomena, and performances (Appreciation). The category Affect attends to registering positive and negative feelings. Table 2 offers exemplary instances of Attitude Affect taken from my data. In the Fox News utterance, Hillary Clinton is quoted by the journalist as she shares her feelings about being called a liar 2 and dishonest (Malloy, Associates, & Relations, 2016, p. 9). Her use of good together with the negative clearly indicates her emotional response to a certain event or occurrence. In the New York Magazine proposition, the journalist assigns an emotional response to Clinton, a strong feeling of sadness and regret (Remorse, n.d.). 2 Examples taken from Fox News and New York Magazine gate news articles are italicised throughout this thesis paper. This is to avoid interference with double quotes, which have evaluative potential. Section details how double quotes can carry evaluation. 10

24 Attitude Affect (emotional responses) Fox News example Well it certainly doesn t make me feel good, Clinton said. New York Magazine example Clinton's remorse was mostly directed at the hell the hath wrought instead of the fact it existed. Table 2. Examples of Attitude Affect Affect positions the author who selects these words to describe a feeling, but also the reader: the audience may show solidarity and align with the statements, or reject them and distance themselves from the emotional perspective of the statement. The category Judgement, the evaluation of behaviour, is concerned with a person s conduct in terms of social esteem, morality, and legality. Table 3 offers exemplary instances of Attitude Judgement taken from my data. In the Fox News assertion, improperly points to the appropriateness of Hillary Clinton s use of a private server to accommodate classified communication, with legal issues being at stake due to the Justice Department s possible involvement in the issue. The journalist proposes a possibility for unethical conduct by using the adjective in its negative form. The writer of the New York Magazine sentence states that Clinton has been described as liar and dishonest, two adjectives of negative judgement, showing that Clinton s truthfulness has been called into question. Attitude Judgement (evaluating people s behaviour) Fox News example key inspectors general asked the Justice Department to open an investigation into whether classified material was improperly shared on the former secretary of state's account. New York Magazine example the top two words that come to voters minds when they hear the words Hillary Clinton are liar and dishonest. Table 3. Examples of Attitude Judgement The category Appreciation deals with the valuation of things and objects surrounding people, natural phenomena, and performances. Table 4 offers exemplary instances of Attitude Appreciation taken from my data. Appreciation occurs explicitly three times in the Fox News sentence. These instances describe the s contained on Hillary Clinton s private server and their release. The speaker of the utterance, directly quoted by the journalist, assigns a level of appropriateness to the Clinton s (while assuming that a percentage of 11

25 them may be deemed inappropriate), and describes their release as timely and transparent. These adjectives evaluate the level of suitability (appropriate) of the s, as well as the level of punctuality (timely) and the level of veracity/truthfulness (transparent) of the release process. The second sentence, written by a New York Magazine journalist, assesses the worthiness of recent releases, speculating that some upcoming releases will also contain boring bureaucratic data. Attitude Appreciation (assessing things, objects, natural phenomena, performances) Fox News example As [Clinton] has said, it is her hope that State and the other agencies involved in the review process will sort out as quickly as possible which s are appropriate to release to the public, and that the release will be as timely and transparent as possible, Merrill said Tuesday. New York Magazine example Based on the last few releases, [the releases] will mostly entail bureaucratic boringness Table 4. Examples of Attitude Appreciation The Appraisal Framework: Engagement Overview Engagement, the second domain of the Appraisal framework, defines attitudinal positioning along the dimension of dialogism, opening up two subcategories: Monoglossia and Heteroglossia. An utterance is monoglossic when there is only one overt voice behind the assertion: the author. If a proposition references other voices apart from that of the author s or opens up alternative positions to the statement, the utterance is heteroglossic. By invoking other voices or positions, the writer of an utterance reduces the absoluteness of the statement, relocates some of the responsibility taken for the statement to an external voice, or expresses that the utterance is not taken for granted but is still up for discussion. The writer also takes stand with respect to what has been said by others: the author stands with or against the statement, can remain undecided, or expresses neutrality. A common way to recognise alternative voices or positions is to quote what another person stated previously, to use negation, or to open up alternative interpretations by marking the statement as hypothetical. From this perspective, heteroglossic utterances can be contractive or expansive. Contraction happens when alternative voices are invoked to then challenge, shut down, or 12

26 refute them, thus closing down the space of dialogism. Alternatively, a voice is singled out as highly warrantable, shutting out any other, possible voices. Expansion, on the other hand, brings in external voices and accepts their existence, and thus opens up the space of dialogism. Table 5 offers exemplary instances of Engagement Monoglossia taken from my data. In all of the three sentences, the writers do not invoke any additional voices that the statement or a part of the statement could be allocated to. The authors, thus, take full responsibility for their statements and assign a sense of absoluteness to them. Potentially evaluative language (growing questions, unusual usage, notion inspiring bad headlines) is formulated by the journalists and remains to be associated with them. As for absoluteness, the sentences introduce ideas that are shown to be not up for discussion. Instead, they seem taken-for-granted assertions. Engagement Monoglossia (single-voiced utterances) Fox News example The revelation adds to the growing questions related to the Democratic presidential front-runner's unusual usage of a private account and server while in government. New York Magazine example The notion that Clinton had willfully jeopardized national security for the sake of convenience inspired months of bad headlines for the Democratic front-runner. The interview took place a day before the 20th anniversary of Hillary Clinton's U.N. speech in Beijing, where she famously said, "It is time for us to say here in Beijing, and the world to hear, that it is no longer acceptable to discuss women's rights as separate from human rights. Table 5. Examples of Engagement Monoglossia A further consideration arises when apparently externally sourced, heteroglossic utterances are overridden by the author, who introduces some attitudinal meaning to express their own alignment/disalignment with the externally sourced proposition. In the last sentence in Table 5, the journalist expresses evaluation of Clinton s speech by assessing it as famous. Thus, the utterance is considered monoglossic despite Clinton s external voice invoked in it, which would otherwise render the assertion heteroglossic. Heteroglossic utterances have external voices and/or positions that are assigned to somebody other than the writer. Table 6 offers exemplary instances of Engagement 13

27 Heteroglossia taken from my data. Said, comment on, and the quote marks explicitly add other speakers to the discussion. The writers acknowledge the assertions of the external voice. These instances expand dialogism. Another type of Heteroglossia is expressed with negation. It s not clear, no s, and would not comment carry with themselves the positive counterparts, i.e. open up new dialogic alternatives (e.g. it s clear, some s), then reject them. These instances contract dialogism. Any and but represent a different type of disalignment: they express counterexpectations, signalling that a different proposition was expected in place of the current one. These instances also contract dialogism. If and whether are used to open up the dialogic space: the writer acknowledges the proposition and announces it as but one of the possible positions in the issue, expanding dialogism. Lastly, presumably renders the utterance hypothetical and helps the writer explicitly present the proposition as grounded in its own subjectivity, and invokes dialogic alternatives, expanding dialogism again. Engagement Heteroglossia (utterances sourced out to other voices/positions) Fox News example It's not clear if the device will yield any information Clinton's attorney said in March that no s from the main personal address she used while secretary of state still "reside on the server or on back-up systems associated with the server." New York Magazine example A Justice Department spokesman would not comment on whether the feds are looking for another device, but presumably the hunt for Clinton's yoga schedules continues. Table 6. Examples of Engagement Heteroglossia In sum, journalists can use monoglossic and heteroglossic sentences when propagating news contents. Monoglossic utterances represent the voice of the journalist only. Heteroglossic utterances introduce other voices to the statements by means of quotation marks, negation, words that render the statement hypothetical, etc The Appraisal Framework: Graduation Overview The third domain of the Appraisal framework, Graduation, attends to the scalability of both Attitude and Engagement values. Attitudinal meanings as well as Engagement values can be mitigated or aggravated in order to convey lesser or greater degrees of positive or negative 14

28 stances. Graduation involves the down-scaling and up-scaling of e.g. amount (a few, many), mass (small, heavy), and proximity (recently, in times gone by). Martin and White (2005) also propose that Graduation overarches both Attitude and Engagement (Martin & White, 2005, p. 136) by decreasing or increasing the intensity of attitudinal meanings (a bit sad, somewhat staggering, utterly unacceptable, a most memorable performance) and scaling Engagement values (she hesitantly suggested, she stated, she firmly insisted). Further, a single lexical item can carry lower or higher grades of attitudinal meanings (contented happy ecstatic) and Engagement values (suspect believe be convinced). Martin and White (2005) discuss Graduation in terms of its two categories Force and Focus. Table 7 offers exemplary instances of Graduation Force taken from my data. In the Fox News utterance, Graduation attends to the intensity of the statement. The writer upscales Clinton s denial as he or she intensifies the vigour of the denial via the word repeatedly. In the New York Magazine statement, the writer discusses a number of s in question and chooses not to provide the precise amount of it. Instead, the journalist downscales the gravity of the proposition by providing imprecise reckonings of underestimation. Other common examples that create a similar effect in my data include a handful of s, more than 3,600 pages of documents, etc. Graduation Force Fox News example Clinton has repeatedly denied sending or receiving any then-classified information on her personal account. New York Magazine example The s in question, fewer than ten of them, were sent between Clinton both before and shortly after she became secretary of State and now-retired general David Petraeus, who was at that point in charge of the wars in Iraq and Afghanistan. Table 7. Examples of Graduation Force As for Graduation Focus, the second Graduation category as proposed by Martin and White (2005), the data under my investigation contained no linguistic resources that fall under Focus. Martin and White (2005) discuss here resources that sharpen or soften meanings that are otherwise not scalable. These involve hedges and vague language, which are unexpected in written journalistic discourse. The following mock examples, not part of my 15

29 data, feature Focus: Clinton s decision to use a private server borders on the unreasonable and Clinton used kind of/sort of a homebrew server, where borders on, kind of, and sort of exemplify Focus Analytical Approach In order to identify evaluative language in Fox News and New York Magazine news articles, I relied on the inductive analytical approach. In inductive reasoning, we begin with specific observations and measures, begin to detect patterns and regularities, formulate some tentative hypotheses that we can explore, and finally end up developing some general conclusions or theories (Trochim, 2006). Induction takes data as its starting point and the data s organization into an analytical model as its goal. Therefore, I started with an initial investigation of my data for recurring linguistic patterns that construe potentially evaluative meanings. Second, I consulted Martin and White s Appraisal framework (2005) to interpret the linguistic patterns that I had found to be notable with respect to evaluation. Next, I returned to the data for a deeper scrutiny and then to the Appraisal framework again to interpret the found linguistic patterns. This two-step analytical work developed into cycles of data analysis and framework consultation until the news articles under investigation were fully exposed for evaluative language and the linguistic evidence of evaluation accommodated to the Appraisal framework. Figure 2 shows the analytical approach used in my research. 16

30 2. FRAMEWORK CONSULTATION 1. DATA ANALYSIS Figure 2. Analytical Approach: Repeated Cycles of Data Analysis and Framework Consultation A similar approach was used by Economou (2009), whose research, the extension of the Appraisal framework to visual language, also involved applying general appraisal options to a corpus of news photos and proceeding to further delicacy in a repeated cycle of analysis and system-building (Economou, 2009, [Abstract]). 4. Data In this chapter, I introduce the two text sets that comprise my data: written news articles on Hillary Clinton s gate scandal published by Fox News and New York Magazine. First, I offer an overview of the two media outlets Fox News and New York Magazine in 4.1., including their websites where the data were pulled from and the situational characteristics of their news articles. In 4.2., I describe the data collection process. In 4.3., I give an overview of the data Fox News and New York Magazine: Overview Foxnews.com is an online news website that belongs to Fox News Channel, one of most widely available cable news channels in the United States. While Fox News self-declares to be Fair & Balanced (Fox News, 2016), it is commonly held to be the only media outlet on a national level that propagates conservative bias (Goldberg, 2007). The other two highest- 17

31 rated cable news channels are CNN and MSNBC, which are viewed as belonging more to the liberal media (n.a., 2014). Fox News is seen as more aligning with values that closely represent Republican voices. In this respect, Fox News is in a unique position on a national scale: not only is it believed to hold Republican views, it is also the sole major, national media outlet to do so. This requires Fox News to assume a number of roles and represent a group of ideological positions that are perhaps more evenly distributed among Democratic media outlets: it needs to express its distance from Democratic values, further a Republican political standpoint, but remain informative and preferably unbiased in news reporting at the same time. Media outlets held to advance Democratic values, on the other hand, have more options in positioning themselves along the clines of factual news propagation and opinionated commentary publishing, because there are more liberal media outlets available to assume these roles. As compared to Fox News, there is less data available on New York Magazine. Nymag.com self-declares to cover the new, the undiscovered, the next in politics, culture, food, fashion, and behavior nationally, through a New York lens (n.a., 2016). Nymag.com is the online version of the bi-weekly published print magazine. In 2013, the online media outlets that belong under the magazine were reported to have increased their number of online visitors considerably (Pompeo, 2013). While the magazine is usually not measured for bias in researches, media critic Eric Alterman, the author of What Liberal Media?, described the outlet as being in the heart of liberal country (Alterman, 2003, para. 11) foxnews.com and nymag.com: Overview gate news articles were collected on the websites and Alexa Internet, a major data analytics company situated in the United States, offers web traffic details and insights into the visitor metrics of websites (Alexa.com, 2016a, 2016b). Table 8 shows the web traffic data of foxnews.com and nymag.com, as downloaded in March

32 ALEXA INTERNET METRICS FOR AND Alexa Internet Metric foxnews.com nymag.com Global rank (volume of visitors and page views combined) Rank in the United States Percent of visitors in the United States 85.9% 77.4% Top search word to send visitors to this website fox news fox news Physical base New York, NY, US New York, NY, US Visitorship: Gender male female male female Visitorship: Education no college some college grad. school college no college some college grad. school college Table 8. Alexa Internet Metrics for foxnews.com and nymag.com The data show that foxnews.com has a considerably bigger reader base than nymag.com, as it is the 44 th most visited website in the United States. Its audience is more U.S.-based than that of nymag.com, and, interestingly, people frequently find their way to both websites by searching for the keywords fox + news. The diagrams provided by Alexa Internet show that the readership of foxnews.com is demographically more balanced along gender and education levels than that of nymag.com. This could reflect the more specific role of nymag.com and its catering for a narrower audience as compared to foxnews.com, which captures a wider audience Situational Characteristics of Fox News and New York Magazine News Articles In explaining the context of Fox News and New York Magazine articles, another necessary step is to introduce their written news articles on Hillary Clinton s gate scandal along the contextual properties, the production circumstances, and the purpose these articles serve - in other words, their situational characteristics. Table 9 summarises the situational characteristics of the Fox News and the New York Magazine text sets. These properties were 19

33 identified following Biber and Conrad s (2009) framework for analyzing situational characteristics of registers and genres (Biber & Conrad, 2009, p. 40). SITUATIONAL CHARACTERISTICS OF FOX NEWS/ NEW YORK MAGAZINE ARTICLES Situational characteristics Fox News articles New York Magazine articles Participants Relations among participants Channel Production circumstances Setting Communicative purposes Topic Addressor Addressee Interactiveness Social roles Personal relationship Shared knowledge Mode Medium Shared time Shared place General Specific Factuality General Specific One or more journalists, often unidentified A wide range of readers possibly scattered across different places and times; 85.9% of all visitors come from the United States. Limited Status/power differences possible Strangers Low degree One or more journalists A wide range of readers possibly scattered across different places and times; 77.4% of all visitors come from the United States. Writing Published online News articles are revised and edited before publication, sometimes corrected and updated after the first publication. Time is usually shared between the writer and the reader as news articles are expected to be read on the day of publication or soon afterwards; otherwise, date markers and timestamps indicate the release of the article. Online space shared; no physical place shared Propagating news; explaining news information Reporting on the developments of Hillary Clinton s controversy The factual representation of information is assumed. Politics Hillary Clinton, data security, privacy issues, controversy Table 9. Situational Characteristics of Fox News and New York Magazine Articles While researching the situational characteristics of Fox News and New York Magazine articles, an important difference arose between the two media outlets: the writers on Fox News often remain unnamed. This decreases the interactiveness between journalists and their audiences as the readers cannot address the writer of the given news article. From an evaluative standpoint, the lack of a named author is crucial because it leads the readers to either align or not with the collective philosophy represented by the media outlet. When the author is named, on the other hand, the audience has a choice to align or not with the 20

34 journalist s position, a stand that might not necessarily reflect the mentality of the whole media outlet. As for interactiveness, the readership has the ability to react to the published articles by way of posting comments under the articles and ing journalists (except for when the journalists remain unnamed, which often occurs in Fox News articles). Nevertheless, the written mode does not allow for direct interaction between the participants, resulting in limited interactiveness between writers and readers. Also of note is the low degree of shared knowledge between writers and readers. While some shared contextual knowledge is assumed to exist between the participants, the purpose of news propagation indicates more knowledge on the journalist s part, which is then passed onto the reader. As a result, the journalist is generally assumed to have information that has been unknown to the general readership, while the audience is assumed to have at least some knowledge and interest in the given topic Data Collection In this section, I present the method that I used to select data, written news articles on Hillary Clinton s gate scandal published by Fox News and New York Magazine, on foxnews.com and nymag.com Timeframe Primary data collection began by determining a timeframe within which news articles discussing Hillary Clinton would be collected on the two media outlets. The timeframe of my choice runs from 12 April 2015 to 1 February The start date marks Clinton s public and official announcement of running for president in 2016 on her YouTube channel (Clinton, 2015). 1 February 2016 is the date of the first Democratic caucus, held in Iowa (Putnam, 2015). Caucuses (and primaries) are part of the presidential nomination process, which lead the Democratic and the Republican political parties to the nomination of their presidential candidate based on a voting system. Although caucuses and primaries last until June 2016, The Washington Post predicted that the earlier contests will winnow the field of candidates enough that a candidate is very likely to claim enough delegates to clinch the nomination prior to the final contest. [ ] The presidential nomination 21

35 process has tended to produce essentially two leading candidates, a frontrunner and an alternative to the frontrunner. These candidates will emerge in the invisible primary or certainly after the first few contests (Putnam, 2015, para ). As of the middle of May 2016, around the time that this thesis is finalised, Donald Trump has emerged as the most likely Republican presidential nominee, while no winner has emerged from the Democratic party. As of 18 May 2016, Clinton has won 1,767 delegate votes as opposed to her Democratic opponent Bernie Sanders 1,488 delegate votes. In order for a candidate to be nominated by the Democratic Party, one must win 2,383 delegate votes (Andrews, Bennett, & Parlapiano, 2016). Bernie Sanders announced that he would stay in the race for the nomination until the Democratic Convention in July 2016, where party delegates decide whether to nominate Clinton or Sanders for President of the United States (Alcindor, 2016). Thus, whether Hillary Clinton becomes the Democratic presidential nominee remains undecided until July Defining Written News Articles Data collection proceeded by distinguishing news articles from other types of posts published online on Fox News and New York Magazine. Written news articles are understood in this thesis as written texts whose main purpose is to propagate the news. Additionally, these texts need to be the first textual references in the news archives of Fox News and New York Magazine to discuss a given gate news event. Furthermore, they must be published on a separate subpage on either website, i.e. one news article equals one subpage on foxnews.com or nymag.com. As a final criterion, it is important for any two news articles discussing the same news event to have been published relatively close to each other in time on the two media outlets. These delimiting criteria helped identify and exclude from the data a) video news published under news headlines on Fox News, and b) commentary articles published under news headlines on either Fox News or New York Magazine. Category a) video news were in many cases the first mentions of a given news event on Fox News. This is owing to the fact that Fox News is primarily an American television channel, and, consequently, multimodal news reporting is one of its inherent features. In cases where the news were first propagated exclusively in video format, I continued 22

36 searching for the first written article that discusses the given news event. Eventually, these articles were selected to be part of my data even if some audiovisual material was embedded in them. Category b) commentaries were distinguished from news articles proper under the following criteria. Commentaries were assumed to be dated later than when the given news information broke and to be not the first in a row of articles to discuss the news event. Both media outlets were expected to first propagate a news event and then provide more subjective commentary on them. This assumption is not to exclude the possibility that subjective commentary may be included in the selected news articles themselves. Furthermore, to ensure that the selected news articles indeed have the property of news propagation, each selected article was identified either as being the source that broke the news or as directly referencing a news source that broke the given news event. These predetermined criteria served their delimiting purpose effectively in most cases exemplifies this. However, there were instances where further research was necessary to identify a text string as a news article without a doubt details this Selecting News Articles Using Predefined Criteria The following is an example of the selection procedure where the predefined criteria successfully filtered non-relevant articles. On 4 September 2015, Clinton was interviewed by MSNBC s Andrea Mitchell, where Clinton s controversy arose as one of the main topics under discussion (MSNBC, 2015). Both Fox News and New York Magazine reported on the event. Appendix 1 shows New York Magazine s news archive sorted after the date of publication and filtered for news about Hillary Clinton (Hillary Clinton. Most recent articles, 2016). To find the news article reporting on the interview, first every post published before 4 September, in this case The Relatable Hillary, was determined as not carrying the potential to propagate this news event. Next, articles not pertaining to the topic in question, in this case The Sanders Surge whose main topic of discussion is not Clinton or her scandal, were disqualified from being the news article of my interest. 23

37 The initial exclusions reduced the list of articles to the ones that indeed attend to the Mitchell-Clinton interview. There was one article published on the day of the interview (Clinton on s), but that does not automatically grant it the news propagation properties as the news may be distributed later than the day of the event. On 8 and 17 September, two more articles related to the scandal were published, To Clarify, and Hillary Clinton is Hurt, respectively. The question remains as to which of the three articles is the one first distributing the news on Clinton s interview with Andrea Mitchell. The predetermined criteria verified that Clinton on s is the article propagating the interview: the text includes a direct link to MSNBC s source video, whereas To Clarify, posted four days later, references Clinton on s. Furthermore, while Hillary Clinton is Hurt, posted on 17 September, seems to discuss the scandal on the face of it, it in fact details a televised Clinton interview with Jimmy Kimmel. As a result, Clinton on s was selected and added to the list of news articles as primary data Selecting News Articles Using Extended Criteria In some cases, following the predefined criteria to select content that carries the news did not suffice. As a result, details beyond what had been preset by the criteria had to be scrutinized. This applies in particular to Fox News content for two major reasons. First, Fox News articles are assigned timestamps that reveal the day of publication but not the exact hour, as opposed to content on New York Magazine. This prevents delicate filtering for the first textual reference that holds relevant content. Second, posts are not tagged for topic, resulting in all articles containing the search words Hillary + Clinton to show up in the search results, a quantity that amounts to thousands of posts in the selected timeframe. Again, the tagging feature is implemented in the New York Magazine archive, rendering search for Clintonrelated content much faster and efficient than in the rather cumbersome Fox News archive (Results found for Hillary Clinton, 2016). For these reasons, it became standard procedure that I would first identify news content on New York Magazine, and then match it to a corresponding news article published on Fox News. If no match was found, the New York Magazine news article was discarded, as well. If a match was found, extended criteria were used to identify one Fox News article as the propagator of the news event in the written mode. 24

38 One instance where the predefined criteria proved insufficient to select news content was when on 22 May 2015 the State Department released Clinton s about the 2012 Benghazi terror attacks. Both Fox News and New York Magazine reported on the event. New York Magazine reported on it on 22 May under the headline State Department Releases Hillary s Benghazi s in Textbook Case of Holiday News Dump. To find the Fox News counterpart, I started by listing articles containing the search words Hillary + Clinton published on 22 May and later. Appendix 2 shows Fox News s news archive sorted after the date of publication and filtered for news about Hillary Clinton (Results found for Hillary Clinton, 2016). Headlines not containing Hillary or Clinton were ruled out first, followed by the exclusion of those articles whose image thumbnails featured a play button, which indicates exclusively audiovisual content. It is important to mention that posts without a play button in their image thumbnails may include video content, as well. However, the lack of the icon indicates content primarily in the written mode on Fox News, making the selection of written articles easier. Having discarded 8 of the 10 articles shown in Appendix 2 following this method nominated Another day of damaging disclosures, and Internal s show Clinton got detailed intel, as the potential bearers of the news. Because the news articles are not sorted alphabetically, one might assume that the exact time of publication is reflected in the order of the posts, with posts toward the bottom indicating earlier publication times. This assumption, together with a Fox News First tag on Another day of damaging disclosures point to this article as the first report on the release of the Benghazi s. In the end, Another day of damaging disclosures turned out to be a newsletter-like aggregation of headlines outlining interesting developments in politics in the recent past and the foreseeable future. While the article refers to the release of the Benghazi s, it does so in the future tense, saying, later today we will get the first wave of her s from a secret server (Stirewalt, 2015). As a result, Internal s show Clinton got detailed intel was selected as the news article propagating the news and corresponding to the one published by New York Magazine. 25

39 4.3. Data Overview In this section, I give an overview of the data and provide calculations of word count. 43 news articles on New York Magazine were selected with a matching 43 news articles on Fox News that, when paired up from the opposite news website, propagate Clinton-related news in the period 12 April 2015 to 1 February Out of the 43 pairs, 20 pertain primarily to the Clinton campaign, including reports about her White House bid, campaign details, public speeches, debates, and poll results. 14 pairs attend to Clinton s controversy. 9 pairs are concerned with other Clinton-related matters that have been examined in the campaign context, but nevertheless mainly attend to topics such as Clinton s wealth, the Clinton Foundation, a book called Time Bomb that looks to reveal exclusive information on the Foundation, and a congressional hearing with Clinton testifying in relation to the 2012 Benghazi attacks. Table 10 shows the 2x14 news articles that comprise my primary data. Where the publication dates of the given pair differed, the first date shows that of Fox News, while the second one shows that of New York Magazine. FOX NEWS AND NEW YORK MAGAZINE ARTICLES ON GATE # Date Fox News article headlines New York Magazine headlines May Internal s show Clinton got detailed intel on 'planned' Benghazi hit July Blumenthal gave diplomatic advice to Hillary Clinton as early as 2009, s show July Clinton facing new calls to turn over server after IGs request probe August FBI investigating security of Hillary Clinton's private server August State Dept. accused of stiff-arming intel September watchdog over Hillary s Former Clinton aide who helped set up server to plead Fifth Amendment to avoid subpoena September Clinton sorry for confusion, stops short of apology for actions September FBI reportedly recovers deleted s September/ 25 September from Clinton server State Dept. receives Clinton chain apparently not included in pages turned over State Department Releases Hillary s Benghazi s in Textbook Case of Holiday News Dump 15 Fun Tidbits From Hillary Clinton s s Justice Department Asked to Investigate Hillary s s Clinton Gives In, Turns Over Server to Justice Department If Clinton Has a Backup of Her s, the FBI Would Like to See It Aide Who Set Up Clinton s Server Will Plead the Fifth Clinton on s: I Am Sorry That This Has Been Confusing to People The FBI Managed to Recover Hillary Clinton s Personal s Defense Department Finds More Clinton s 26

40 10 07 October FBI probe of Hillary Clinton s expands to second tech company November Intelligence officials reportedly walk back Clinton 'top secret' claims January/ 08 January 19 January/ 20 January State Department releases over 3,000 Clinton s on New Year's Eve Inspector General: Clinton s had intel from most secretive, classified programs January Official: Some Clinton s 'too damaging' to release Hillary Is Done Playing Defense, But gate Won t Go Away Report: Clinton Never Sent Classified Information From Her Private Account Hillary Clinton Reacts to Benghazi in Newest Batch of s New Report Says Some s Sent on Clinton s Server Were Super Top Secret The State Department Confirms Clinton s Server Hosted Top Secret Information Table 10. The Dates of Publication and Titles of the 2x14 gate News Articles The overview reveals two peculiarities that require closer inspection: one concerning the contents of article pair #02, and one pertaining to the production circumstances of article pair #12. While most headline pairs show a clear connection regarding their contents, the headlines of article pair #02 show considerable discrepancy regarding the story that they cover, the Fox News headline signalling news propagation, while the New York Magazine one apparently displaying commentary. I address the issue in section Further, Table 10 shows that most article pairs were published on the same day by the two media outlets, satisfying the definition of news articles and other criteria as presented in section 4.2. In two cases, one media outlet published the news the day after the other (articles #09 and #13). Of more interest is article #12, which shows a week-long gap in publication date. The news discussing the State Department s end-of-the-year release of Clinton s sees a discrepancy of 7 days between its release date on Fox News and New York Magazine. While the delayed propagation of the news is not the main focus of my thesis paper, the marked time difference emerged as a potential contributor to the biased representation of gate, and, as such, it necessitates further investigation. I address this in section Further Commentary on Data: Article #02 The content mismatch in article pair #02 could result from wrongly selecting 15 Fun Tidbits From Hillary Clinton s s as the news article on New York Magazine discussing the 27

41 State Department s release of a second batch of Clinton s s, and that an earlier article that carried those news had been overlooked. A re-check of the New York Magazine news archive confirmed, however, that the article, posted at 4.53 a.m. on 1 July, was the first one on New York Magazine to discuss the release that was published online on 30 June (Virtual Reading Room Documents Search Results, 2015). In addition, the article provides a direct link to the State Department s database: by referencing the news source, the article meets the data selection criteria detailed in section 4.2. The data analysis later revealed that the language in the article indeed takes up evaluative resources in a degree different to those represented in the rest of the articles. Despite this, I decided to keep the article pair in my data for two reasons. First, the article pair satisfies the preset criteria. Second, it reports on a significant event in the development of Clinton s controversy, namely the State Department s release of the first extensive, large-scale batch of s containing over 1,900 correspondences. (The release prior to that was specific to the Benghazi attacks and included 296 pages of .) (Virtual Reading Room Documents Search Results, 2015). To conclude, I considered the article 15 Fun Tidbits From Hillary Clinton s s a news article, albeit with commentary voice infused in it more than in the rest of the articles Further Commentary on Data: Article #12 On the face of it, the publishing irregularity that surrounds article #12 seems to be the result of either one of two scenarios. First, a failure to identify a New York Magazine article is possible - one that corresponds to the State Department s end-of-the-year release of Clinton s and was published around 1 January Alternatively, New York Magazine possibly waited a week (to review the s first) before construing a conclusion and propagating the news accordingly. A re-check of the news archive of Fox News and New York Magazine, however, suggests a third possibility. According to The Washington Post, the State Department was obliged under court order to release 82% of all Clinton s by the end of 2015 (Helderman, 2015). However, the approaching holidays and being understaffed prevented the Department 28

42 from achieving the court-ordered goal and instead released 3,079 correspondences on 31 December 2015 (Virtual Reading Room Documents Search Results, 2015). As well, they announced a second batch to be released the following week in order to make up for the shortfall (Helderman, 2015, para. 5). The release of the remaining 1,262 s eventually took place at around 2 a.m. on 8 January 2016 (Virtual Reading Room Documents Search Results, 2016). New York Magazine did not report on the larger batch of s released on 31 December Instead, article #12 Hillary Clinton Reacts to Benghazi in Newest Batch of s was published 9 hours after the State Department released the rest of the document collection on 8 January Claire Landsbaum of New York Magazine offered in article #12 a reflection over these correspondences, suggesting that they re mostly boring administrative s (Landsbaum, 2016, para. 2). The journalist further claimed that the newest batch of Hillary Clinton s arrived a little late (they were supposed to be out December 31), but the State Department worked into the night and finally released 3,000 pages of fit n sparkly new content around 2 a.m. (Landsbaum, 2016, para. 2). The quoted text contains an external link that points to the State Department s document collection of the 1,262 s published on 8 January Fox News, on the other hand, used a different strategy to cover the irregular, two-step release of the batch. The unnamed journalist on 1 January 2016 published article #12 State Department releases over 3,000 Clinton s on New Year's Eve, reporting on the release of the 3,079 correspondences. Following the State Department s release of the second batch of s on 8 January 2016, Fox News published two more articles detailing the contents of the document collection of the 1,262 remaining s. The release and the s were referenced in the first of these articles as the exchange was one of more than 1,200 messages released overnight, under a court order dating back to last year (Tomlinson, 2016). In the second article, the release is referenced as a recently released (FoxNews.com, 2016). The unusual practices in covering the end-of-the-year release of Clinton s points to the potential where New York Magazine underreported the release process of the Clinton s and misinformed the public about data volume. Conversely, Fox News can be assessed 29

43 as overreporting gate, producing articles with disproportionately great frequency and on incomplete batches of releases. The above detour was necessary in order for me to support my decision in selecting a pair of news articles with a week-long time difference in their publication date. Furthermore, the above examples show Hillary Clinton s controversy can be mitigated or aggravated beyond the level of language in order to spread or cover up the scandal. Last, it justifies once again and prompts a linguistic investigation into gate articles to see whether partisan bias is reflected in how the journalists report on the controversy Data Measurements I performed calculations on the 2x14 gate articles to see to what extent the two media outlets deal with the controversy. Word count measurements revealed that the 28 articles contain 16,608 words in total, 65.2% of which was produced by Fox News. It is important to note here that one necessary adjustment to the scope of my investigation was to sort the 2x14 articles on the sentence level and differentiate between those sentences that directly reference Hillary Clinton and those that do not. I eventually narrowed my data to include only those sentences where Hillary Clinton is explicitly mentioned. This step was crucial to reduce my data scope and to shift the focus of the research to how the controversy is described against the backdrop of Clinton s political persona and her run for presidency. The 28 news articles published on Fox News and New York Magazine consist of 653 sentences. After selecting the ones that directly reference Clinton, 410 sentences remained. These 410 sentences comprise my data. The 410 sentences, the core of my data analysis, consist of 11,278 words, 61% of which were published by Fox News. Table 11 below presents this information. A keyword density analysis performed on the 410 sentences revealed that the most frequently mentioned words in both text sets are Clinton, s, and Clinton s, with State Department, server, and information present in the top ten keywords list of both groups of sentences. A marked difference is that classified and secretary are the eighth and tenth most frequently used keywords in the Fox News sentences, respectively, while neither of the two are in the top ten 30

44 list of New York Magazine keywords. The opposite applies to the keyword Hillary, which ranks nine of the top ten keywords list of New York Magazine sentences but is not present in the same list for Fox News. WORD COUNT OF ARTICLES ON GATE Words in Fox News articles Words in New York Magazine articles Clinton Clinton Total Clinton Clinton Total Σ % 36.4% 63.6% 100% 24% 76% 100% Table 11. Word Count of Articles on gate As far as direct references to Hillary Clinton are concerned, 57% of all Fox News sentences reference Clinton directly; this number rises to 72% in the case of New York Magazine sentences. The sentences that directly reference Hillary Clinton and are thus added to my analysis amount to 239 and 171 for Fox News and New York Magazine, respectively. In search of direct references to Hillary Clinton, three distinct categories were identified as carriers of such references: name references, including Clinton or Hillary s, political positions, such as The 2016 Democratic frontrunner, secretary of state, or former Secretary Clinton s, and function words, such as she, her, herself, and in some direct quotes I and you. In order for a sentence to be included in my data, it had to contain at least one direct reference of these kinds. Table 12 is an overview of the top ten occurrences throughout the 410 sentences. MENTIONS OF HILLARY CLINTON IN ARTICLES ON GATE Fox News mentions New York Magazine mentions Top 10 references Occurrences Top 10 references Occurrences Clinton 116 Clinton 71 her (possessive determiner) 96 Clinton s 54 she (personal pronoun) 65 her (possessive determiner) 53 Clinton s 54 she (personal pronoun) 48 I (personal pronoun) 13 secretary of State 14 Hillary Clinton 12 Hillary Clinton s 12 Hillary Clinton s 12 Hillary Clinton 11 secretary of state 12 I (personal pronoun) 11 31

45 her (personal pronoun) 6 her (personal pronoun) 7 herself 5 Hillary 5 Number of all mentions 452 Number of all mentions 309 Table 12. Mentions of Hillary Clinton in gate Articles In total, 452 direct references were identified in Fox News articles and 309 in New York Magazine ones. Fox News used twice as many different expressions to reference Clinton (56) than New York Magazine (28). This is owing to the fact that Fox News news articles use a range of expressions that reference Clinton s past and prospective affiliations and positions. Examples include presidential candidate, top diplomat, First Lady, former Secretary Clinton, the secretary s, then-secretary of State Hillary Clinton, etc. During the analysis, the three categories of references were expanded to include expressions that do not fit in any of the three categories. Examples include the two (referencing Hillary Clinton and her daughter), their end user (a reference to Hillary Clinton to whom services were done by the tech company Datto), one of the three amigos (a nickname for Clinton during the 1980 s according to Fox News), and liar (used by some respondents to a Quinnipiac questionnaire that looked into what words come to people s mind when they hear the words Hillary Clinton). In addition, those sentences were also added to my data where there is a lack of direct references to Clinton, but she is the speaker of the given utterance, and thus her presence is certain. To sum up, my primary data consist of 239 Fox News and and 171 New York Magazine sentences that directly reference Clinton and were published as part of gate news articles. 5. Design of the Analytical Framework As emphasised in section 3.2. The Appraisal Framework: Evaluation in Verbal Language, certain linguistic patterns found across the gate news articles grouped along organizing principles different to the subcategories proposed by Martin and White s (2005) Appraisal framework. For instance, feelings (Attitude Affect) were found to be sourced to the writer, to Clinton, or a third party. In some cases, they were sourced through more than one actor, 32

46 as in Top aide Huma Abedin didn't think Clinton would be excited, where excitement is assigned to Clinton through a third party. While Martin and White make a distinction between authorial and non-authorial Affect, multiple sources are not examined. In addition, distinguishing third party-sourced Affect from Clinton-sourced Affect is specifically important with regard to the controversy news, where Clinton plays a central role. As for Engagement, certain linguistic elements such as scare quotes do not belong to one clearcut category but rather border on two or three categories. Scare quotes and other formatting conventions are not discussed in Martin and White s Appraisal framework. Thirdly, registering Graduation yielded the emergence of subcategories that are prevalent throughout my data but do not necessarily correspond to the Graduation system established by Martin and White. As mentioned in The Appraisal Framework: Graduation Overview, Graduation Focus had zero occurrences in my data. Further, Martin and White s Graduation Force attends to imprecise reckonings (a handful of s) among others, whereas my data showed that overly precise reckonings (31,105 s) also have evaluative potential. Thus, in order to accommodate standout linguistic phenomena across my data, adjustments were made to the Appraisal framework. This was necessary in order to provide a full account of evaluative language on Fox News and New York Magazine specific to Clinton and her controversy. Therefore, the expanded layout of the Appraisal framework is one specific to my data and does not completely reflect Martin and White s system (Martin & White, 2005). From sections 5.1. to 5.3., I move from domain to domain in the Appraisal framework and open up all categories and subcategories to provide a full account of the framework. In section 5.4., I show the complete map of the Appraisal framework modified for the text sets written news articles on Hillary Clinton s gate scandal published by Fox News and New York Magazine. Further, I present lexical items that I deemed non-evaluative and thus did not assign them to any of the Appraisal framework domains. 33

47 5.1. Design of the Appraisal Framework: Attitude Attitudinal meanings needed adjustment in the Affect category. Judgement values and assessments of Appreciation, on the other hand, were registered according their layout proposed by Martin and White (Martin & White, 2005). Figure 3 shows Attitude (A) and its categories. AFFECT AUTHORIAL NON-AUTHORIAL A JUDGEMENT SOCIAL ESTEEM SOCIAL SANCTION APPRECIATION REACTION COMPOSITION VALUATION Figure 3. Attitude Categories in Fox News and New York Magazine gate News Articles Attitude Affect Affect, the signalling of emotional responses is subcategorised by White according to whether the source of Affect is authorial or non-authorial (White, 2015). Authorial Affect means that the writer presents his or her own emotional responses to a given utterance. In Non-authorial Affect, the writer remains the voice behind the feeling, i.e. they take some responsibility for the affect value presented in the text, but it is not the author s emotions that are described. Rather, they are associated with other individuals or groups of people. With regard to my data, three additional subcategories were opened under Nonauthorial Affect: feelings sourced to Clinton, feelings sourced to a third party i.e. neither the author nor Clinton, and multiple non-authorial sources, i.e. when the described feeling is filtered through one actor and then at least another one. Figure 4 shows the fully expanded subcategories under Affect with examples taken from my data. 34

48 A AFFECT AUTHORIAL Clinton's FOMO is more understandable when you learn that (New York Magazine) Figure 4. Affect Subcategories in Fox News and New York Magazine gate News Articles Clinton has said she wants the department to release the s (Fox News); Then [Clinton] got stood up by the entire cabinet, confirming her worst fears (New York Magazine) [officials at Datto] did not want to run into a legal problem(fox News); Naturally, Republicans were giddy about Clinton's concession. (New York Magazine) In the interview, Mitchell also asked Clinton how she feels that a Quinnipiac poll found (Fox News); Top aide Huma Abedin didn't think Clinton would be excited (New York Magazine) 35 Source: Clinton Source: 3 rd Party Multiple Sources NON-AUTHORIAL

49 Attitude Judgement Judgement, the evaluation of behaviour, falls into two distinct subcategories: Social Esteem and Social Sanction. Social Esteem covers those positive or negative evaluations that could boost or diminish the esteem that one is given in society. These judgements mostly answer to behaviours and deeds that do not have a lawful/unlawful quality to them, but can easily result in the actor being raised or lowered in the esteem of their community (White, 2015). Social Sanction, on the other hand, attends to morality or legality and comes into play when rules have been broken and misconduct is under discussion. People evaluated along Social Sanction are potentially caught up in some degrees of (im)moral/(un)lawful conduct. Social Esteem can be further broken down into three subcategories: evaluations targeted at how unusual someone is (normality), how able they are (capacity), and how resolute they are (tenacity). Also on the level of subcategories, Social Sanction differentiates between evaluations of someone s truthfulness (veracity) and their ethical behaviour (propriety). Figure 5 shows the fully expanded subcategories under Judgement with examples taken from my data (examples of tenacity are taken from Martin & White for a lack of tenacity evaluations in my data) (Martin & White, 2005, p. 53). 36

50 A Figure 5. Judgement Subcategories in Fox News and New York Magazine gate News Articles Tenacity [brave], [reckless] Capacity well-trained, struggle, jump on (Fox News); tech-savvy, quickly challenged, quickly responded (New York Magazine) Propriety inappropriately, should have known, avoid public scrutiny (Fox News); holiday news dump, should, sorry (New York Magazine) 37 Normality controversial, unusual (Fox News); controversial, famously (New York Magazine) Veracity Clinton hedged, liar, secretive (Fox News); shady, liar, dishonest (New York Magazine) SOCIAL ESTEEM SOCIAL SANCTION JUDGEMENT

51 Attitude Appreciation Appreciation, the assessment of objects, things, natural phenomena, and performances, is established by Martin and White along our reactions to things (do they catch our attention; do they please us?), their composition (balance and complexity), and their value (how innovative, authentic, timely, etc.) (Martin & White, 2005, p. 56). Figure 6 shows the fully expanded subcategories under Appreciation with examples taken from my data. 38

52 A APPRECIATION REACTION COMPOSITION VALUATION a flattering feature (New York Magazine) classified programs (Fox News); secure network (New York Magazine) insufficient evidence (Fox News); the most significant findings (New York Magazine) Figure 6. Appreciation Subcategories in Fox News and New York Magazine gate News Articles 39

53 5.2. Design of the Appraisal Framework: Engagement The Engagement domain, which attends to whether assertions are sourced internally or externally, sees a distinction between Monoglossia and Heteroglossia as introduced in The Appraisal Framework: Engagement Overview. Categories under Heteroglossia, utterances where alternative voices or positions are invoked, are detailed in this section. Figure 7 lists Engagement (E) and its categories. MONOGLOSSIA E HETEROGLOSSIA CONTRACT DISCLAIM PROCLAIM HETEROGLOSSIA EXPAND ATTRIBUTE ENTERTAIN Figure 7. Engagement Categories in Fox News and New York Magazine gate News Articles Heteroglossic utterances influence the dialogic space in either of two ways. The first possibility is Heteroglossia Expand, where heteroglossic assertions are used to open up dialogic space and signal that the position they represent is only one out of a range of possible positions. In this way, the dialogic space becomes expanded (Martin & White, 2005, p. 103). Section details expansion. The second possibility for heteroglossic assertions to impact the dialogic space is to restrict the scope of alternative voices or positions (Heteroglossia Contract). In this way, the dialogic space is closed down, i.e. contracted. Section details contraction. During data analysis, I found language that can be interpreted as both expanding and contracting the dialogic space. These ambiguous linguistic features are discussed in section

54 Engagement Heteroglossia Expand There are two ways to show that an assertion expands dialogism, i.e. that it is only one of many possibilities. First, the authorial voice can disassociate itself from the utterance by way of sourcing, i.e. attributing it to an external voice (Expand Attribute). Attribution is often realised grammatically through reported speech. Attributions can go two different ways: the authorial voice can acknowledge or distance itself from the attributed statement. When the authorial voice does not explicitly state where it stands with respect to the displayed proposition, the voice acknowledges the utterance (Attribute Acknowledge). When the author attributes a proposition to an external voice and in doing so refuses to take responsibility for the attributed utterance, he or she distances him- or herself from the proposition (Attribute Distance). Attribute Acknowledge is commonly realised by instances such as he said, he believes, he noted, according to her, in her view, etc. Attribute Distance is carried through wordings that stress that the assertion has not been generally accepted. While some may agree with it, other voices do not accept it. Instances such as he claimed, it s rumored that, she maintained, and sometimes he insisted, carry the idea that at least some people do not align with or believe the statement. The second way to construe an utterance as but one of a range of alternatives is by way of entertaining them (Expand Entertain). By using wordings of probability (might, probably, possibly, I believe that, it s possible that - expressing that there are alternatives, as well), evidentiality (it seems, the evidence suggests, apparently), and permission/obligation (must, have to), the authorial voice displays the proposition as one possibility only but explicitly boosts or decreases the validity of it via wordings of entertainment. As such, while the authorial voice opens up the possibility that the reader might think differently about the matter, they entertain, i.e. explicitly and subjectively align with the presented proposition. In sum, reported speech and wordings of different levels of probability announce a proposition to be but one of a number of propositions to align with. When the authorial voice selects this assertion, they do not overtly indicate their stance with respect to it (Acknowledge), step back from it (Distance), or advance their own assessment of it 41

55 (Entertain). When attributing, the authorial voice grounds the utterance in the subjectivity of an external voice. When entertaining, the authorial voice grounds the proposition in his or her own subjectivity. Figure 8 shows the fully expanded subcategories under Heteroglossia Expand with examples taken from my data. 42

56 E HETEROGLOSSIA EXPAND ATTRIBUTE ENTERTAIN Acknowledge according to, said, told (Fox News); report, said (New York Magazine) Probability perhaps (Fox News); presumably (New York Magazine) Distance maintain, claim (Fox News); claim (New York Magazine) Evidentiality suggest (Fox News); ostensibly (New York Magazine) Obligation must, has to (New York Magazine) Figure 8. Heteroglossia Expand Subcategories in Fox News and New York Magazine gate News Articles 43

57 Engagement Heteroglossia Contract Similarly to Expand, Heteroglossia Contract offers two main routes to affect the range of possible dialogic alternatives. First, the authorial voice can announce one claim to hold fully and thus excludes other positions or voices from offering an alternative (Contract Proclaim). Second, the authorial voice can reject a certain claim or show it as replacing one that was expected in its place (Contract Disclaim). In Proclaim Concur, the authorial voice explicitly agrees with the single external voice or position that has been presented. This can take place by using wordings such as of course, naturally, not surprisingly, certainly, etc. These wordings position the proposition as the generally agreed and accepted viewpoint. When the authorial voice introduces an external position and announces them as valid, correct, or undeniable, they use wordings such as show, demonstrate, prove, reveal, confirm, corroborate, disclose, confide, etc. These fall under the subcategory Proclaim Endorse. An important distinction to make here is the one between endorsing and acknowledging an utterance. In Contract Proclaim Endorse, the authorial voice explicitly aligns with a statement and takes responsibility for them, thereby rejecting alternative voices as holding true. In contrast, in Expand Attribute Acknowledge, the authorial voice disassociates from the proposition and takes no responsibility for what has been said. To say Internal s show Clinton got detailed intel on 'planned' Benghazi hit is to nominate the statement as undeniably valid and taken for granted. To say Clinton said she loves swimming, chocolate, House of Cards, sleep, and... , the author reports on what has been said and expresses that the proposition is valid in the subjectivity of the external voice from Clinton s viewpoint in this case. A third way to proclaim that a proposition is warrantable is for the authorial voice to intervene in the utterance and place authorial emphasis on the proposition (Proclaim Pronounce). In speech, primary stress is one way to introduce these interpolations. In writing, formatting including emboldening or italicising text parts can contribute to this effect. Also, wordings such as indeed, really, as well as the helping verbs does, did, was, were, etc. in positive sentences to express emphatic assertion (e.g. The judge has yet to rule on that matter, but he did set hearings for February ) can be used to pronounce a proposition. Proclaim 44

58 Pronounce puts solidarity with the audience at risk by confronting and defeating alternative voices or positions. Unless the reader subscribes to the one pronounced statement and agrees with alternatives being shut down, the writer risks solidarity with his or her audience. The second scenario in dialogic contraction concerns the authorial voice dismissing alternative voices or positions (Contract Disclaim). When the authorial voice disclaims alternatives, they show either denial or counterexpectation. In the first case (Disclaim Denial), the author uses negation to introduce the alternative positive position into the dialogue and then reject it. This effect is unavoidable when negating a statement because the negative invariably carries with it the positive (Martin & White, 2005, p. 118). Consequently, when the authorial voice says Clinton never sent classified information, an alternative voice is invoked (stating that Clinton sent classified information) and shown to not hold, i.e. excluded from the scope of alternative voices. This does not work reciprocally, i.e. the positive does not necessarily carry with it the negative. The second way to disclaim alternatives is to introduce a proposition and present it as replacing another proposition that was expected in its place (Disclaim Counter). This is grammatically delivered via conjunctions and connectives (even though, although, however, yet, but), as well as comment adjuncts and adverbials (surprisingly, even, only, still). The contrary position invoked by the authorial voice is shown to not hold and shut out of the alternative propositions. Disclaim Counter puts solidarity between the writer and the reader at risk because the author takes the replacing proposition for granted and not up for discussion. As a result, if the reader does not subscribe to the notion advanced by the author, the solidarity between them is subject to damage. In this respect, Counter is similar to Proclaim Pronounce. To sum up, dialogic contraction singles out one assertion as legitimate and warrantable (Contract Proclaim) or announces a range of propositions as invalid (Contract Disclaim), thereby closing down the dialogic space for alternative voices or positions. Figure 9 shows the fully expanded subcategories under Heteroglossia Contract with examples taken from my data. 45

59 E HETEROGLOSSIA CONTRACT DISCLAIM PROCLAIM Denial not criminal, never sent, didn t (Fox News); never, not targeting, no s (New York Magazine) Concur undoubtedly, clearly, surely, obviously (New York Magazine) Counter all but, however, yet (Fox News); yet, but, still, any (New York Magazine) Endorse show, reveal, confirm (Fox News); realize, reveal, show (New York Magazine) Pronounce really (Fox News); especially, also of note, (New York Magazine) Figure 9. Heteroglossia Contract Subcategories in Fox News and New York Magazine gate News Articles 46

60 Engagement Heteroglossia Borderline There arose one standout linguistic feature across the data under my investigation that seems to override the construction of the Appraisal framework system along the duality of expansion and contraction: formatting conventions, used by the journalists in the data to infuse voices into their propositions, often represent features of both Expand and Contract. The salient formatting choices that potentially carry evaluative meanings in the two text sets under my scrutiny include a) the punctuation marks parentheses, the em dash, the colon, and b) quote marks. I traced category a) punctuation marks for heteroglossic features and eventually categorised them under Proclaim Pronounce for the following reasons. Parentheses and the em dash are used throughout Fox News and New York Magazine articles to introduce a shift in voice toward commentary and separate that content from the rest of the proposition. Regardless of how subjective the isolated content is, these markers introduce a pause to the text, a separation that can be assumed to be authorial interventions. The author in these instances foregrounds the information separated by punctuation marks and nominates it as more important than the rest of the assertion or at least requiring special attention. In this respect, the same applies to capitalized wordings followed by a colon: they might not infuse commentary voice into the text, but they act as significance nominators of text parts. These punctuation marks achieve an effect very similar to what has been discussed under Proclaim Pronounce: they mark overt authorial intervention in which the writer makes their subjective role stand out and takes responsibility for marking up assertions as more significant than other propositions. Category b) quote marks proved more difficult to categorise along Martin and White s description of heteroglossic subcategories (Martin & White, 2005, p. 92). As both single and double quotes occur frequently in the texts under investigation, a reasonable starting point to explore the functions fulfilled by quotes is to look into the difference between these two types of quote marks. Throughout my data, double quotes are used in body text, while single quotes are used either inside double quotes to mark a quote inside a quote, or in article headlines. As for headline punctuation practices, Malcolm Gibson of the University of Kansas points to the 47

61 following rule: Headline punctuation is normal with two significant exceptions: Use periods for abbreviations only, and use single quotes where you would use double quotes in a story (Gibson, n.d., para. 16). Further, according to Beth Hill of The Editor s Blog, single quotation marks are used in newspaper headlines when quotation marks are required. (Space is limited for newspapers; they cut punctuation wherever they can.) (Hill, 2014, para. 17). This explains the presence of both types of quote marks in my data and removes this distinction from the scope of my discussion with regard to evaluative language. However, a visually less pronounced difference within double quote marks requires further investigation. Throughout my data, I encountered a small set of quote marks that are put to use to mark something different from the verbatim account of some external voice. These quote marks isolate very few words only and seem to encompass hints toward criticism, irony, vague accusations, disbelief, or pronunciation with regard to the quoted material. These effects are all the more identifiable against the backdrop of proper quotes in these texts, which usually cover rather long strings of words or even multiple sentences. Double quotes of this special usage have been labeled as scare quotes. As The News Manual points out on the issue, Scare quotes are words or short phrases which are placed between quotation marks when they really do not belong. Usually, the writer is trying to add stress to the words or to suggest something other than their obvious meaning. The simplest reason for scare quotes is to add emphasis; A more common use of the scare quote is to suggest that the word or phrase should not be taken at face value. It is often used to suggest disbelief or actual disagreement with the words as they are being used (Henshall & Ingram, 2008, para. 92). A further comment comes from Mark Davidson, the author of Right, Wrong, and Risky: A Dictionary of Today's American English Usage, who advises to not use quotation marks for a word or phrase that the speaker happened to use unless the word or phrase is particularly dramatic or otherwise unusual (Davidson, 2005, p. 446). To sum up the above, when the author relies on scare quotes, they might do so to add stress, suggest disbelief/disagreement, or add a dramatic tone to the dialogue. Based on my 48

62 data, a number of effects can come into play ranging from actually neutral quotes to the advancing of factuality, the pronunciation of alignment, disalignment, or the distancing from what has been externally sourced. An important consideration here is that scare quotes infuse the text with ambiguity as to where the author stands with regard to the proposition. From this standpoint, scare quotes have strong evaluative potential because they pass the responsibility of registering alignment or disalignment entirely to the audience. The proposition remains ambiguous and subjective in a way that the exact nature of authorial stance remains unknown, thereby risking solidarity between writer and reader in case the reader interprets the writing as a proposition that he or she does not subscribe to. As such, scare quotes are problematic to assign to heteroglossic subcategories and require the contextual background to pinpoint their assumed evaluative effect. On the face of it, scare quotes can be categorised under Contract Proclaim Endorse or Pronounce as they mark explicit authorial interventions in support of the externally sourced material, making them similar in function to other punctuation marks. However, in New York Magazine s scare quoted reference to Clinton s controversy in the sentence Four months after we first learned about gate, on Tuesday night the State Department finally released the first batch of Hillary Clinton's s, it is safe to assume that the authorial voice does not make an attempt at endorsing the label gate or pronouncing it as highly warrantable. Rather, there seems to be a sense of disalignment with (Expand Attribute Distance), or even disapproval of the use of the term gate. It can also be seen as a means of questioning the existence of the scandal (Contract Disclaim Denial). Scare quotes are up for discussion in journalistic discourse in particular, where register conventions foreshadow factual information sharing, and while scare quotes seem to foreground facts on the surface, they in fact bring evaluative meanings into the assertion. Fox News s scare quoted reference in the below sentence exemplifies this. 49

63 In the [MSNBC] interview, Mitchell also asked Clinton how she feels that a Quinnipiac poll found that the first words that came to mind among voters when asked about her were liar, untrustworthy and crooked Here, the authorial voice can be surmised at the minimum to have used quote marks to introduce external voices and reference them properly without explicitly stating his or her own stance on the quoted material (Expand Attribute Acknowledge). However, a second check of the Quinnipiac poll results otherwise unreferenced by the Fox News article reveals the Fox News statement to be false. In fact, untrustworthy was the third word and crooked the nineteenth to have come to voters mind when asked about Clinton (Malloy et al., 2016). The writer of the quoted sentence used quote marks to infuse the assertion with factuality on the face of it, whereas he or she in fact subjectively selected an arbitrary group of words and claimed them to be the first in the list. Consequently, the quoted material is in fact more a statement that the authorial voice aligns with and thus promotes it (Contract Proclaim Pronounce) but without explicitly admitting it. To sum up, most formatting choices made by the authors were assigned to the Proclaim Pronounce category, but a set of scare quotes were retained as having both expansive and contractive potential and thus relocating the responsibility of interpretation to the audience. These cases were assigned under Heteroglossia during a calculation of frequencies of the Appraisal domains in the two text sets, but were left out of calculations targeting Contract and Expand and their subcategories. Figure 10 shows borderline examples taken from my data. 50

64 E Figure 10. Heteroglossia Borderline Subcategories in Fox News and New York Magazine gate News Articles Proclaim Pronounce "several" of Clinton's s contained classified intelligence information -- and at least one of them was made public, EXCLUSIVE: (Fox News); But her scandal worsened again on Tuesday (with some help from Republicans), Official: (New York Magazine) Contract Proclaim Endorse/Pronounce OR Expand Attribute Acknowledge/Distance so-called homebrew server, the extremely serious investigation (Fox News); after we first learned about gate, fit n sparkly content (New York Magazine) 51 FORMATTING: PUNCTUATION FORMATTING: SCARE QUOTES HETEROGLOSSIA BORDERLINE

65 5.3. Design of the Appraisal Framework: Graduation Graduation, which is suggested by Martin and White to overarch Attitude and Engagement rather than being a separate third domain (Martin & White, 2005, p. 136), required several adjustments with regard to my data. As stated in section The Appraisal Framework: Graduation Overview, instances of Focus, one of the two major fields of Graduation, did not occur in my data. While Force, the second Graduation category, occurred frequently, I decided to recategorise the domain according to what gradable linguistic features were salient in my data. The reason behind this recategorisation is the nature of Graduation. Gradability is intrinsic in language, and, as such, categories within it blend, and fuzzy boundaries occur around what can be regarded as a graded utterance and what not. Taking the main distinctive feature of Graduation into consideration (down-scaling/mitigating and upscaling/intensifying the force of a proposition), graded language can be traced in any proposition and removed until we arrive at a decomposed assertion built from irreducible meanings. This is the natural semantic metalanguage, a very small set of words such as good/bad, yes/no, big/small, etc. (Goddard, 2008, p. 33). However, the goal of journalistic discourse is not to present stories by way of these semantic primitives. As a result, while acknowledging that strictly speaking words such as never and stated are up-scaled meanings when compared to their semantic primitive counterparts not and said, I decided to assess as graded those assertions only which show striking differences in graded meanings when Fox News wordings and New York Magazine wordings are compared. For example, word choice in the below two sentences clearly point at graded meanings that contribute to the seriousness versus the triviality of the controversy. Clinton has been under fire through much of 2015 about her use of a private, unsecured server as secretary of state, specifically over the security of her server, and her incomplete retention of her s. (Fox News) The questions surrounding Clinton s private server haven t taken up much oxygen in the Democratic primary race save for inspiring a popular meme. (New York Magazine) 52

66 Fox News presents an up-scaled substitute for criticism (under fire), imprecise reckoning of time to elongate the presence of gate in time (through much of 2015), upscales security issues (private, unsecured server) and goes on to isolate some content as being significant (specifically). The adjective order in private, unsecured server requires further explanation here. As Enrica Rosato (2013) points out in her thesis Adjective order in English: A semantic account with cross-linguistic applications, relative descriptions of a noun concerning quality are syntactically placed farther from the base noun than intrinsic features such as color, origin, or material (Rosato, 2013, p. 26). The default word order in the Fox News utterance would thus be unsecured private server, a pattern that is followed in general by both media outlets in their description of Clinton s server. Via the irregular adjective order in private, unsecured server, the concept of the insecurity of the server is up-scaled by assessing the server as objectively, irrefutably unsecured. As for the New York Magazine sentence, it down-scales the scandalous nature of the controversy (haven t taken up much oxygen) and trivialises it by down-scaling the scandal to have inspired a popular (Internet) meme. Applying this comparative approach of graded wordings on Fox News and New York Magazine gave rise to three distinct Graduation tendencies in my data that contribute to intensifying meanings and, as a result, overplay the gravity of the propositions, and to softening meanings and thus underplaying the gravity of the news content. The three subcategories are Reckoning, Scaling, and Isolation. While registering these instances in my data, only those utterances were marked as graduated where the authors selected configurations from the language that overtly alter meanings along grades. Figure 11 lists these Graduation (G) categories. 53

67 RECKONING MASS PROXIMITY G SCALING SCANDAL REPETITION ISOLATION Figure 11. Graduation Categories in Fox News and New York Magazine gate News Articles From to , I present the Graduation categories Reckoning, Scaling, and Isolation Graduation Reckoning Reckoning deals with how accurately or precisely factual data are described across the text sets under investigation. Along the cline of preciseness, two main subcategories arose. Reckoning Mass comes into play when the journalists describe volumes of s, documents, and the level of preciseness in doing so. Rather than marking up all reckonings as graduated meanings, I focussed on those instances where amounts are made unnecessarily precise (31,105 s) or are clearly understated (a handful of s) in a way that they imply evaluative meanings that contribute to the aggravation or mitigation of a proposition. The second subcategory within Reckoning was named Proximity. Proximity attends to imprecise and precise descriptions of time and space. Again, I selected instances where reckonings of this type, especially times and dates, are blurred or sharpened to the level that they carry meanings beyond face value. Instances of proximity were registered when dates were replaced by imprecise counterparts or were extended with further description to link events in time that could boost an incriminating tone (Blumenthal gave diplomatic advice to Hillary Clinton as early as 2009). The same applies to overly precise descriptions of time where the hour and the minute are disclosed for an exchange or an release date. These are apparently no constructive additions to the text but can be used to evaluate, for example, the administrative burden imposed by the server investigation ( the State 54

68 Department worked into the night and finally released 3,000 pages of fit n sparkly new content around 2 a.m.). Figure 12 shows the fully expanded subcategories under Graduation Reckoning with examples taken from my data. 55

69 G RECKONING MASS some, several, thousands of (Fox News); many, a handful of s (New York Magazine) precision 3,105 s (Fox News); 31,830 s (New York Magazine) PROXIMITY recently (Fox News); a little late (New York Magazine) precision September 11, 2012, at 11 p.m. (New York Magazine) Figure 12. Graduation Reckoning Subcategories in Fox News and New York Magazine gate News Articles 56

70 Graduation Scaling Scaling was registered when the authorial voice scaled up or down their general descriptors usually via adjectives or adverbs (e.g. highly unusual). When referring to the gate controversy, the authors sometimes replaced these references with words that explicitly aggravate (e.g. even more sensitive intelligence) or mitigate (e.g. saga, stories) what is at stake in the context of the scandal. Therefore, I opened up the subcategory Scaling Context, which attends to graded descriptions surrounding the controversy proper. Another type of scaling was identified when individual words expressed that an action was done over and over or when two synonymous words stood next to each other to intensify the gravity of the proposition (Scaling Repetition). Figure 13 shows the fully expanded subcategories under Graduation Scaling with examples taken from my data. 57

71 G SCALING highly unusual, growing questions (Fox News); a clearer picture (New York Magazine) Scandal criminal investigation, major issues (Fox News); saga, stories, a distraction (New York Magazine) Repetition repeated questions, repeatedly denied (Fox News) Figure 13. Graduation Scaling Subcategories in Fox News and New York Magazine gate News Articles 58

72 Graduation Isolation Finally, a major theme that runs through my data is related to when the authorial voice singles out some information and provides extra description about it (Isolation). This was recognised as a frequently used technique to separate content (usually contents or gate developments) and nominate them as more significant than other information. When the author makes prominent content related to the Benghazi attacks, the wealth of Clinton aides as against Clinton s yoga schedule and her daughter s wedding plans, the authorial voice fronts content in a way that contributes to the seriousness or triviality of the controversy. Two sentences to describe this are taken from Fox News and New York Magazine, respectively. In another exchange, Billionaire George Soros, a major donor to liberal causes, confided to a former Clinton aide that he made the wrong choice in supporting Barack Obama in the 2008 primaries over Clinton. (Fox News) When Clinton fractured her elbow, Senator Barbara Mikulski expressed sympathy for her inability to do lady things. (New York Magazine) From this standpoint, Isolation is related in functionality to formatting choices discussed under Engagement. As opposed to text formatting choices, however, the authorial voice here promotes or relegates content in a way that is interwoven in the narrative of the news article. Because of this, the presence of the subjective author is not as overtly stated in Graduation Isolation as it is in Engagement Contract Proclaim Pronounce. Yet, singling out content can contribute to an overall impression of all the Clinton s as containing texts that can potentially implicate Clinton morally or legally versus comprising private content irrelevant to the public. Figure 14 shows the fully expanded subcategories under Graduation Isolation with examples taken from my data. 59

73 G ISOLATION another reason, at least one Clinton (Fox News); yet another gate update (New York Magazine) Figure 14. Graduation Isolation Subcategories in Fox News and New York Magazine gate News Articles 60

74 5.4. Design of the Appraisal Framework: Summary And Out-of-Scope Lexical Items Figure 15 on the next page shows the full layout of the Appraisal framework adjusted to the text sets written news articles on Hillary Clinton s gate scandal published by Fox News and New York Magazine. Having opened up all subcategories under the Appraisal framework, one consideration remains to be discussed before I turn to the research results. A small set of lexical items were taken out of the scope of my research as not qualifying for evaluative language. Table 13 presents this set. OUT-OF-SCOPE LEXICAL ITEMS IN ARTICLES ON GATE Nouns Verbs Adjectives batch (of s) release, receive, send ( s) private, personal, work-related, top-secret ( s) Table 13. Out-of-Scope Lexical Items in Articles on gate These lexical items were deemed neutral and not carrying evaluation. These items occur with high frequency not only across my data, but also in other media outlets discussing gate that I relied on as secondary literature. As pervasive items in the narrative of Clinton s controversy, I considered these words commonly used descriptors and activity verbs used to report on gate news. 61

75 Systemic Functional Linguistics (SFL) Interpersonal mode APPRAISAL FRAMEWORK A G E Figure 15. The Full Layout of the Appraisal Framework SCANDAL REPETITION MASS PROXIMITY ATTRIBUTE ENTERTAIN PROBABILITY EVIDENTIALITY OBLIGATION ACKNOWLEDGE DISTANCE CONCUR ENDORSE PRONOUNCE FORMATTING DISCLAIM PROCLAIM 62 DENY COUNTER REACTION COMPOSITION VALUATION SOCIAL ESTEEM NORMALITY CAPACITY TENACITY SOCIAL SANCTION VERACITY PROPRIETY AUTHORIAL NON-AUTHORIAL CLINTON 3 RD PARTY MULTIPLE SOURCES

76 6. Results In this chapter, I first present the frequencies of occurrences of the three Appraisal framework domains Attitude, Engagement, and Graduation, and their categories across the Fox News and New York Magazine data in sections 6.1. to 6.4. Sections 6.5 to 6.7. present and discuss selected examples of all categories and their subcategories in detail. In 6.8., I present the Appraisal framework analysis on exemplary sentences from Fox News and New York Magazine, paying special attention to sentence parts that carry implicit evaluations Frequencies of Attitude, Engagement, and Graduation This section presents frequencies of the three Appraisal framework domains across the two text sets written news articles on Hillary Clinton s gate scandal published by Fox News and New York Magazine Normed Rates of Occurrence To arrive at comparable frequencies across the datasets, I calculated normalised occurrences. This was necessary because the Fox News data consist of 6898 words, while the New York Magazine one of Without normalised computation, the number of occurrences would not reflect the discrepancy in data volume. For instance, meanings of Attitude were registered 66 times in the Fox News data and 68 times in the New York Magazine one. The two numbers imply nearly equal use of attitudinal meanings in the two sets of texts. However, when the difference in data volume is considered, i.e. normalised occurrences are calculated, the New York Magazine data display a 62% more use of attitudinal meanings than the Fox News one. Normed rates of occurrence are the rate at which a feature occurs in a fixed amount of text (Biber & Conrad, 2009, p. 62). In my calculations, I chose 1000 words as the fixed amount of text. Table 14 exemplifies this method. NORMED RATES OF OCCURRENCE occurrences of Attitude (Fox News) occurrences of Attitude (New York Magazine) ( ) 1000 ( ) Table 14. The Calculation of Normed Rates of Occurrence 63

77 Table 14 shows that there are 9.57 words expressing attitudinal meanings per 1000 words in the Fox News data. In contrast, there are words expressing attitudinal meanings per 1000 words in the New York Magazine data Occurrences in Total When the two datasets are viewed together, attitudinal meanings were the least employed evaluative resources to present Hillary Clinton s scandal, and Engagement the most frequently featured domain. When the two text sets are looked at separately, The New York Magazine data are shown to use each of the three domains more than the Fox News texts. Chart 1 presents these results. 60 Instances of Attitude, Engagement, and Graduation per dataset Fox News New York Magazine Attitude Engagement Graduation Chart 1. Instances of Attitude, Engagement, and Graduation per dataset In sum, Attitude categories are the least frequently used across my data (13%). Engagement was registered the most (47%), with Graduation categories constituting 40% of all evaluative meanings in Fox News and New York Magazine news articles. When looking at the two text sets separately, New York Magazine makes more frequent use of all of the three Appraisal framework domains than Fox News. 64

78 6.2. Frequencies of Attitude Categories The Attitude categories were found to distribute differently in the two text sets. Chart 2 below presents these results. New York Magazine features meanings of Affect over 7 times more often than Fox News. The Fox News data rely most on Judgement values, while New York Magazine texts use Appreciation the most often, twice as much as Fox News. This difference in category distribution shows that the moral and legal implications of gate are much more accentuated on Fox News than on New York Magazine. In turn, the latter lays emphasis on the emotional aspect of the scandal and attends more to the things, objects, and performances around Clinton rather than the moral issues surrounding the controversy. Frequencies of Attitude categories 50% - 50% + Fox News 90% - 10% + 70% - 30% + New York Magazine 32% - 68% + 70% - 30% + 37% - 63% + Affect Judgement Appreciation Chart 2. Frequencies of Attitude Categories Chart 2 also shows the distribution of negative versus positive attitudinal meanings in the two text sets. 90% of Judgement values, the most common Attitude category used by Fox News authors, are negative on Fox News. 63% of all Appreciation values, the most frequently used Attitude category by New York Magazine journalists, are positive in New York Magazine articles. In sum, the Fox News data contain more negative evaluations in each Attitude category than the New York Magazine data. 65

79 6.3. Frequencies of Engagement Categories As for Engagement, the dialogism of the data, the majority of sentences in both text sets are heteroglossic and a smaller amount monoglossic. The Fox News data have twice as many monoglossic sentences than New York Magazine ones, indicating that Fox News journalists advanced their own interpretation of the controversy as opposed to New York Magazine journalists, who referenced more external sources. Monoglossia was found to override heteroglossia once: one New York Magazine sentence expressed explicit evaluation when presenting a Hillary Clinton quote as an assertion which the presidential front-runner famously said. The full sentence goes, The [Clinton-Mitchell] interview took place a day before the 20th anniversary of Hillary Clinton's U.N. speech in Beijing, where she famously said, It is time for us to say here in Beijing, and the world to hear, that it is no longer acceptable to discuss women's rights as separate from human rights. The New York Magazine journalist inserts this information in article #07 Clinton on s: I Am Sorry That This Has Been Confusing to People in relation to a Clinton interview with Andrea Mitchell targeting Clinton s usage. While there is a four-minute-long discussion of Clinton s U.N. speech in the interview (MSNBC, 2015) lasting from 23:00 to 26:50, the New York Magazine article details the speech on women s right in four paragraphs and two embedded videos, making up for 40% of the article in terms of word count. Against this backdrop, the unusual emphasis placed by the journalist on Clinton s fight for women s rights can be viewed as removing the focus from the main topic of the interview, gate. Furthermore, introducing a detailed description of Clinton s humanitarian deeds as famous is a way to praise her work and express sympathy or even admiration. As a result, while the quoted material points at the sentence being heteroglossic, I categorised the assertion as monoglossic, in which the writer positively evaluates a quoted material and thus positions the audience to take up similar standpoints. Chart 3 presents the ratio of monoglossic versus heteroglossic assertions. Heteroglossic utterances were further investigated to identify heteroglossic subcategories across the text sets. 66

80 Monoglossic vs. Heteroglossic utterances Fox News New York Magazine Monoglossia Heteroglossia Chart 3. Monoglossic vs. Heteroglossic Utterances To further break down Heteroglossia, an overwhelming majority of heteroglossic utterances feature Attribute where the authorial voice remains neutral or expresses distance from the evoked external voice. Most attributions were carried by said, reported, and claimed. Entertain was registered the least, with New York Magazine relying on it over four times more than Fox News, lending certain utterances a hypothetical, uncertain value via the words perhaps, probably, raises the possibility, and likely. Proclaim was used twice as much by New York Magazine than by Fox News. Examples include obviously, really, surely, naturally, and also of note. The calculations include the formatting choices previously discussed in section Engagement Heteroglossia Borderline. Scare quotes were not added to the calculations as they potentially belong to more than one subcategory. In total, scare quotes were registered 13 times across the Fox News data and 5 times in the New York Magazine data. 67

81 Disclaim shows the most balanced usage of engagement resources between the two text sets. Disclaim is also the second most frequently used category, mainly because most details around gate were still up for discussion at the time when the articles were published, resulting in the authorial voices frequently refuting the claims coming from external voices. Chart 4 shows the frequencies of Engagement categories. Frequencies of Engagement categories Fox News New York Magazine Disclaim Proclaim Entertain Attribute Chart 4. Frequencies of Engagement Categories 68

82 6.4. Frequencies of Graduation Categories All Graduation categories are more prevalent in New York Magazine news articles than in Fox News ones. The Reckoning category appears the most often in both sets of texts, while Scaling and Isolation are both less frequently used. The two media outlets make near equal use of Scaling and Isolation. Chart 5 presents these results. Frequencies of Graduation categories Fox News New York Magazine Reckoning Scaling Isolation Chart 5. Frequencies of Graduation Categories 6.5. Attitude: A Closer Look Sections 6.5. to 6.7. present the way the categories of the Appraisal domains are used across my data. To illustrate my points, I use examples from each text set. An extended list of examples is in Appendix Attitude Affect Attitude Affect was not only found to be used in the two text sets to a different extent, but it is also used to create different feelings and dispositions toward Clinton and the controversy. The Fox News data made scarce use of Affect, with none of the affect values being authorial-sourced. In most cases, they express desire as in what Clinton, a third party, or a multiple sourced actor wants to do about the rising controversy, e.g. Clinton has 69

83 said she wants the department to release the s. In one instance, the authorial voice describes how employees at Datto, Inc., a data security company, were surprised that Clinton s server was backing up to two backup servers, and how they did not want to run into a legal problem. Affect values were found most frequently in multiple sourced assertions. The New York Magazine dataset shows a different distribution of affect values across their subcategories. The authorial voice steps up as the source of a given feeling more often (understandable, surprised), and third parties as well as multiple sourced actors show a greater occurrence than in the Fox News data. Apart from the authorial voice introducing their own feelings into the assertions more often, a marked difference is in the amount of Clinton-sourced feelings that New York Magazine presents the reader with. One part of assigning emotions to Clinton appears where the author singles out, dissects, and presents some of the released s (Graduation Isolation). The journalists interpret the content of the s in a way that Clinton s feelings are foregrounded. Examples include Clinton [ ] got worried that everyone was hanging out with her, Clinton liked Syrian dictator Bashar al-assad, Clinton said she loves swimming, and Clinton receives a cheerful note from her BFF Senator Mikulski. In other cases, the author ventures to guess and describe how Clinton might be feeling in the course of the scandal developments. Examples include Clinton [ ] was being a tad too optimistic, Clinton s remorse, Hillary Clinton will surely be relieved to learn, and the use of the informal internet slang FOMO (fear of missing out) in Clinton's FOMO is more understandable when you learn that (FOMO, 2011). In sum, the affect values presented on New York Magazine describe Clinton in a human context as surrounded by friends and having feelings in everyday life, and they thus make Clinton relatable as a human being. Some of the negative affect values also attain to this goal, as in [Clinton] got stood up and confirming [Clinton s] worst fears. As opposed to this presentation, Fox News removes the emotional aspect of the controversy and does not aim to describe Clinton from this perspective. Appendix 3 shows more examples of Affect. 70

84 Attitude Judgement Attitude Judgement is the only category in any domain that is more frequently used by Fox News than by New York Magazine. It is notable that only 10% of all judgement values are positive evaluations in the Fox News data. These are usually assigned to third parties such as the Republicans (e.g. Congressional Republicans seized on Clinton s reversal and the Republican National Committee quickly jumped on Clinton s remarks, where seized on and quickly jumped on are evaluations of Social Esteem Capacity). The majority of negative evaluations, making up for 90% of all instances of judgement, are descriptive of Clinton and her use of a private server. For instance, Social Esteem Normality describes the outof-the-ordinary nature of Clinton s usage (e.g. Clinton controversially conducted official State Department business and Clinton s unusual usage of a private account), while some are targeted at people with strong ties to Clinton (e.g. the controversial confidant Blumenthal). As for Judgement subcategories in the Fox News texts, instances of Social Sanction far outnumber those of Social Esteem, which indicates that Fox News journalists often form opinions on moral and legal implications. In Social Sanction Veracity, Clinton is evaluated as hedging questions and stating that she couldn t be taken at her word, assigning to her an untruthful quality. Social Sanction Propriety is the most frequently used subcategory across the Fox News data. Evaluations here often target the release process (e.g. document dumps, which stands for releasing bad news or documents on a Friday afternoon in an attempt to avoid media scrutiny (Friday News Dump, n.d.). In other cases, Clinton is the subject of Social Propriety evaluations (e.g. [Clinton s use of a private account] to avoid public scrutiny, Clinton repeatedly denied any wrongdoing, and Clinton had handled classified materials inappropriately). Thirdly, Social Propriety describes Clinton s usage in the passive mode (e.g. whether classified information was improperly shared, mishandled). An interesting pattern that runs through the Fox News data is the morphological negation of positive judgement values to construe negative judgements. Words such as improperly, inappropriately, and unusual, are constructions where a lexical item is used to deny the truth of an expression (Hamawand, 2009, p. 1). In this way, these words create 71

85 effects similar to Engagement Disclaim in that they introduce an alternative, expected course of events (proper, appropriate, usual), which are then proven to not hold. Judgement values were less frequently used by New York Magazine. In that set of texts, positive judgement values were registered 30% of the time, evaluating Clinton s campaign team which quickly challenged the story that a criminal investigation against Clinton was under preparation (Social Esteem Capacity), and Republicans who quickly responded to the news that at least four Clinton s contained classified information (Social Esteem Capacity). The 70% negative judgement values evaluate people around Clinton (e.g. controversial aide Sidney Blumenthal, Social Esteem Normality), the representation of the scandal in the media (e.g. In yet another gate update that sounds shady but proves nothing, Social Sanction Veracity), and the release process of the s (e.g. Textbook Case of Holiday News Dump, Social Sanction Propriety). Judgement evaluations of Clinton include her Beijing speech (Clinton famously said, positive evaluation, Social Esteem Normality), and the Quinnipiac poll results that are words coming from external voices ( liar and dishonest, negative evaluation, Social Sanction Veracity). In sum, Fox News texts rely on more Judgement than New York Magazine ones. Fox News assigns negative evaluations of Veracity and Propriety to Clinton the most, while New York Magazine evaluates negatively the representation of the scandal in the media more often. Appendix 4 shows more examples of Judgement Attitude Appreciation Attitude Appreciation is also used in a different manner by Fox News and New York Magazine. In the Fox News data, 70% of all instances of Appreciation are negative. There are zero occurrences of Reaction, the emotional response to things or performances. Composition, the balance and complexity of things, is the most frequently occurring Appreciation subcategory in the Fox News data. It mostly evaluates the contents of Clinton s server via words such as detailed, sensitive, classified, and upgraded. Few adjectives 72

86 show Valuation, where the authorial voice assesses the value of a thing (insufficient evidence, an awkward time for Clinton, and Clinton s ill-fated 2008 campaign). Appreciation is not only used more frequently by New York Magazine, but the distribution of its subcategories is also highly different from that of Fox News. In the New York Magazine data, 37% of all Appreciation values are negative. The data show the most reliance on Reaction. While some of these are descriptive of Clinton (e.g. a flattering feature on Clinton and her fun side), the majority evaluate the content (e.g. [stories that say] nice things about [Clinton], fun tidbits, bureaucratic boringness, boring administrative s, a cheerful note, fit n sparkly new content, thrilling update, and less than remarkable update). Some of these seem positive evaluations at face value (fun, thrilling), but when placed in the context of the details that they are used to describe (e.g. Clinton s daily errands, her yoga schedule, and her daughter s wedding plans), they create the impression of sarcastic descriptions of an overblown investigation that sheds light on Clinton s personal matters - uninteresting to the public. Composition values (e.g. secure network, sensitive s, technical, and opaque) occur the least in the New York Magazine texts. Assessments of Valuation occur the second most often, and they mostly describe Clinton s performance (e.g. a rare interview, [Clinton s] last big television appearance, Clinton s pointed criticism of China s policy toward women, and Clinton rounded off her summer in the best way possible: by giving back). To sum up, Appreciation in Fox News is represented by Composition the most, evaluating Clinton s server by questioning the confidentiality level of the contents. New York Magazine, on the other hand, relies the most on Reaction, with the authors evaluating Clinton s media presence positively and describing the contents as carrying minute, personal, uninteresting details on Clinton s everyday life. Appendix 5 shows more examples of Appreciation Engagement: A Closer Look In section 6.3., I provided an overview of how mono- and heteroglossic utterances are distributed across the Fox News and the New York Magazine data. In this section, I detail occurrences of heteroglossic assertions in its four categories Attribute, Entertain, Proclaim, and Disclaim. 73

87 Engagement Attribute The majority of heteroglossic utterances in both text sets take place when an utterance is attributed to an external voice. While the majority of attributions occur by way of acknowledging other positions (Attribute Acknowledge), of more interest is distancing, where the journalist introduces an external voice to the text and disaligns with its proposition. In the Fox News data, 91% of Attribute Distance was registered as Clintonsourced, i.e. the journalist reports on what she has said and expresses disalignment with the proposition. Examples include [Clinton] went on to reiterate her claim that her use of was fully above board and allowed by the State Department, She also claims that she never sent or received s marked classified, Clinton insists the information was not classified at the time, and Clinton has maintained that she turned over all relevant federal records before deleting her s off her sever (Maintain, n.d.). Attribute Distance is infrequent in the New York Magazine data. The authorial voice distances itself from a Clinton statement on one occasion (Clinton has insisted that she never sent or received classified information over her private account). Distance in the direction of third parties is more common (e.g. General I. Charles McCullough III had claimed that two s sent from Clinton s private account contained top-secret information ). In summation, the writers of both datasets rely heavily on attributing propositions through Acknowledge, which is a neutral, or at least not explicitly side-taking means of introducing alternative positions to the texts. When the writers introduce other voices and express disalignment with them, Fox News journalists more often distance themselves from what Clinton has said, while New York Magazine journalists do so from what a third party has stated. Appendix 6 shows more examples of Attribute Engagement Entertain Instances of Entertain, when the authorial voice selects one voice or position and attaches high probability to it, are four times less prevalent in the Fox News data than in the New York Magazine one. Fox News often uses probability markers (perhaps, assumption) to describe data security concerns around gate (e.g. Only [Clinton] and perhaps a small 74

88 circle of advisers know the content of the discarded communications, and the intelligence agencies are operating on the assumption there are more copies of the Clinton s out there). Evidentiality markers (apparently, suggests), which stress that Clinton failed to turn over all of her work-related s to the authorities, are also frequent (e.g. State Dept. receives Clinton chain apparently not included in pages turned over, and at least one suggests she directed a subordinate to work around the rules). New York Magazine uses over three times as many probability markers as Fox News. In these instances, the writer surmises that some groupings of people are interested in Clinton s everyday matters and wish criminal evidence against Clinton to emerge from the chains. Examples include probably, presumably, and possibility in the following sentences. Clinton is scheduled to testify in front of the House Select Committee on Benghazi on October 22 a date conservatives are probably looking forward to. Those hoping to find a smoking gun will probably have to wait for next month's dispatch. presumably the hunt for Clinton's yoga schedules continues. This raises the possibility that someday the public will be able to learn all about her yoga schedules and Chelsea's wedding-planning process. A lower number of words show evidentiality (e.g. ostensibly personal s), and even less express obligation (e.g. Independent experts have suggested that actual charges against Clinton are highly unlikely, as publicly available evidence has to indicate that she broke any laws). When using Entertainment, New York Magazine frequently addresses its presumed audience directly and in a colloquial way (e.g. despite what you re likely to hear in the coming year, and Cliffhanger: You ll have to read the next batch of s to find out whether Hillary and Huma ever figured out how to use the fax machine.). This, along with previously exemplified acronyms and Internet slang words (BFF, FOMO), add a casual, non-formal tone to the New York Magazine articles. In sum, the use of Entertain in the two text sets is directed at different themes that comprise Clinton s scandal. Fox News journalists take guesses at how Clinton s unsecured communications exposed confidential information and thus put the United States in jeopardy. They also emphasise with markers of evidentiality that Clinton deleted half of her communications without turning them over to the Sate Department. New York 75

89 Magazine writers, on the other hand, entertain other voices in the controversy. They bring in voices of people who aim to find incriminating evidence in the Clinton s. Entertaining these with probability markers is a way to mitigate the accusatory tone that the writers use when describing these groups of people. Appendix 7 shows more examples of Entertain Engagement Proclaim Proclaim, when the author presents one external voice as highly warrantable and dismisses the rest, occurred 50% less in the Fox News data than in the New York Magazine texts. As for the Proclaim subcategories, Proclaim Concur, which expresses explicit standing with that one external position, was registered zero times in the Fox News texts. Proclaim Endorse occurs frequently and describes what the Clinton s showed or revealed about the presidential candidate, her use of s, and her personal matters. Proclaim Pronounce appeared once lexically in the Fox News texts, carried by really in the sentence Republican critics have demanded to know if any of those s were really work-related s that should have been turned over to the State Department Formatting conventions were also assigned to the Pronounce subcategory. In the Fox News data, formatting choices are used to boost the credibility of a given utterance, as exemplified by the dash in the following sentence. However, the inspectors general wrote in a subsequent memo last week that several of Clinton s s contained classified intelligence information -- and at least one of them was made public. Other formatting choices that attain the same goal include words in all caps followed by a colon, as in EXCLUSIVE: Hillary Clinton s s on her unsecured, homebrew server contained intelligence from the U.S. government s most secretive and highly classified programs, and Official: Some Clinton s too damaging to release. These Fox News pronouncements invariably reinforce the incriminating load of the journalists utterances. In the New York Magazine texts, all of the three Proclaim subcategories occur with a higher frequency than in the Fox News data. There, instances of Proclaim Concur are carried by undoubtedly, clearly, surely, and obviously. Some of these are directed at groups of people who are presumably interested in finding incriminating evidence against Clinton (e.g. While there are undoubtedly teams of people scouring the documents for information 76

90 that will torpedo Clinton's 2016 bid, and many especially those critical of Hillary Clinton's use of a personal account during her tenure at the State Department are sure to interpret the discovery of these s as a sign that there may be more work-related s Clinton failed to make available to the public). Often, words of Concur reinforce the casual, non-serious tone in the description of the releases and contents (e.g. Clearly, being secretary of State has its perks, such as directing underlings to track down that rad rug you saw during your trip to China, and Hillary Clinton will surely be relieved to learn that the FBI has recovered the 60,000 s she carelessly deleted.). On one occasion, the authorial voice shows giddiness to be a natural part of Republican conduct by saying Naturally, Republicans were giddy about Clinton's concession. Similarly to its use on Fox News, Proclaim Endorse occurs most often around s that show or reveal details about Clinton, her work as secretary of state, and personal matters. While Proclaim Pronounce was registered on one word in the Fox News data, it occurred lexically nine times in New York Magazine, making the authorial presence more emphasised throughout the New York Magazine articles. Examples include especially, the helping verb did, and also of note. For example, the authorial voice of New York Magazine made a statement about Clinton s October Benghazi hearing, adding that it is a date conservatives are probably looking forward to given the slow and steady stream of Clinton news from the past few months, especially since none of the news has done anything to sink her campaign yet. Also, when news broke that Clinton s server was backing up to two backup servers, and, as a result, retrieving her previously deleted chains became a possibility, a New York Magazine journalist wrote It s also worth noting that this information comes to us via a GOP [=Republican] senator a day after Clinton released an ad attacking congressional Republicans. In this instance, the authorial voice openly refutes a claim which came from the Republicans and stresses its lack of credibility via It s also worth noting that. The use of formatting conventions to advance the author s own stand with an issue, also assigned to Proclaim Pronounce, was registered almost five times more across New York Magazine texts than in Fox News ones. Standout examples include the use of 77

91 parentheses in But [Clinton s] scandal worsened again on Tuesday (with some help from Republicans). Another example is the em dash in the last batch which offered yet another look at the less than exciting conversations inspired by bureaucracy was published Monday as well as in an FBI investigation into the server [was set off] an investigation that is not targeting Clinton. Nominating information as highly credible by using the colon and adding Attitude Appreciation also occurred frequently. Examples include Report: Clinton Never Sent Classified Information From Her Private Account, and Bonus content: She expresses sincere surprise that a State Department employee is using his personal account, where bonus expresses positive appreciation. These inserts of commentary nature stress the insignificance of the Clinton contents, the existence of people hoping to see Clinton prosecuted, and that Clinton did not break the law by using a personal server. Appendix 8 shows more examples of Proclaim Engagement Disclaim Fox News and New York Magazine data show balance in the amount of Engagement Disclaim values that they present. Markers of disclaim, however, are targeted at different voices in Fox News than in New York Magazine. Throughout the Fox News data, Disclaim Denial and Counter are frequently used to illustrate what Clinton has claimed, thereby signalling that an opposing statement was expected in place of what she has said. Examples include Clinton has repeatedly denied sending or receiving any classified information on her personal account and Clinton, the Democratic presidential frontrunner, has repeatedly denied any wrongdoing related to her private server. These utterances express that some might rightfully think that Clinton did send or receive classified information, and that Clinton has committed wrongdoing. On other occasions, utterances where Clinton is said to not have had top-secret information on her private server are further infused with Attribute Distance in order for the journalist to advance his or her doubtful tone about the Clinton statement, e.g. She also claims (Distance) that she never sent or received (Denial) s marked classified. Disclaim Denial and Counter are used differently in New York Magazine texts. There is a marked acquitting tone in the following negated sentences. Some of the information in Clinton's s were classified after the fact, but it wouldn't have been a breach of 78

92 protocol to them at the time. The U.S. intelligence community no longer believes that Clinton sent messages containing top-secret information. In these cases, the authorial voice itself argues that Clinton did not break any laws or introduces external positions via Attribute Acknowledge that state that Clinton did not have classified information on her private server. The authorial voice also relies on Attribute Acknowledge to introduce a proposition that Clinton has denied, e.g. Clinton and her campaign team have argued that the s didn't contain classified information when sent In sum, Disclaim instances occur frequently across both text sets, but they are used in different configurations to advance different perspectives on Clinton s scandal. Fox News journalists mostly infuse Clinton statements with denial and often add to these presentations their own disalignment via Attribute Distance. New York Magazine authors negate propositions that would otherwise incriminate Clinton if they were written in the positive, and introduce statements which Clinton has denied via Attribute Acknowledge. Appendix 9 shows more examples of Disclaim Graduation: A Closer Look As stated in section 6.4., Graduation categories occurred more frequently in the New York Magazine data than in the Fox News texts. In this section, I detail how Reckoning, Scaling, and Isolation are used by the two media outlets in their presentation of Clinton s controversy. Appendix 10 shows examples of Graduation on Fox News. Appendix 11 shows examples of Graduation on New York Magazine Graduation Reckoning Precise and imprecise reckonings of the data volume, dates, space, and other quantifiable data are used differently by the two media outlets. Fox News journalists often up-scale the volume of chains by imprecise reckonings, e.g. potentially hundreds of classified s, and public release of thousands of Clinton s. Other wordings that signal impreciseness include some, more, partial, enough, and just a fraction. For instance, in the clauses there are more copies of the Clinton s out there, and even releasing a partial would provide enough clues to trace back to the original, the journalist implies that a big number of Clinton s contain classified information, and the release of even small parts of them could allow the identification of special access programs intelligence, 79

93 i.e. top-secret information. In the case of overly precise reckonings, e.g. The latest batch of 3,105 s includes 275 documents upgraded to classified, the precise presentation of large numbers adds a shock value to the utterance to underscore claims that the Clinton s indeed have confidential information that could jeopardize the country s security. When reckonings are connected to time, the imprecise presentation of dates act to elongate the scandal, aggrandising its coverage in the media, as in Clinton has been under fire through much of 2015 about her use of a private, unsecured server As for Reckoning in the other set of texts, New York Magazine uses overly imprecise reckonings when the journalist selects some exemplary content and, via quantifying these, he or she renders the examples representative of the entirety of the chains. For example, a release is described as containing mostly boring administrative s lots of Clinton asking aides to print things for her In another example, the author says Many of Hillary Clinton's contributions to the threads are limited to Pls print or thx. In these utterances, mostly, lots of, and many highlight the representativeness of the examples on the scale of all the s. Imprecise reckonings are also used to shift focus from an act to the quantities involved in the act in, for instance, two consecutive sentences that discuss the fact that the State Department found s that Clinton had not turned over to the authorities....the Defense Department just found a handful of s that she didn t turn over with the rest of her workrelated s. The s in question, fewer than ten of them, were sent between Clinton both before and shortly after she became secretary of State and now-retired general David Petraeus... Here, the imprecise reckonings a handful of s and fewer than ten of them play down the amount of s not turned over, thereby acquitting the act of not turning s over by presenting them as only very few. A similar effect is advanced by the imprecise reckoning of time shortly after she became secretary of State, a vaguely acquitting assertion proposing that the act of not turning s over is justifiable by how soon they were sent or received after Clinton took position. Time reckonings also stand out when they are unnecessarily precise, as in the following sentence. The newest batch of Hillary Clinton s arrived a little late (they were supposed to be out December 31), but the State Department worked into the night and finally 80

94 released 3,000 pages of "fit n sparkly" new content around 2 a.m. Here, it is implied that the controversy apparently aimless as it is sparkled by documents of bureaucratic boringness, according to other New York Magazine articles has placed undue burden on the State Department. The journalist describes that officials had to make every endeavour on New Year s Eve and the week after to release 3,000 threads, a goal that was met around 2 a.m. Overall, reckonings are used throughout the text sets to blow up or reduce the amount of potentially incriminating s, to move the focus of an utterance from the described act to quantities, and to present gate as long-lasting and demanding. In general, Fox News uses reckonings to advance the seriousness of gate and its ramifications, whereas New York Magazine describes the disproportionate public attention paid to gate as compared to the insignificance of the contents of the s Graduation Scaling Scaling marks a more explicitly evaluative description of Clinton s usage and the public s response to gate. Carried usually by adjectives and adverbs, Fox News journalists tend to scale up the controversy by way of discussing the highly unusual private server, deep concerns about the contents, the revelation adds to the growing questions, and an exchange in September 2010 [ ] showed considerable confusion over [Clinton s] practices. In these instances, the authorial voice aggravates the described objects and things by up-scaling them via highly, deep, growing, and considerable. To refer to the scandal proper, Fox News uses up-scaled utterances including major issues and major document dumps. The above examples stand out especially when compared to Scaling used in the New York Magazine data. Examples here include actual charges against Clinton are highly unlikely, [the s] which many hope will provide a clearer picture of Clinton's tenure, and conservatives are probably looking forward to [the Benghazi hearings] given the slow and steady stream of Clinton news from the past few months. In these examples, actual, highly, clearer, and slow and steady carry scaled meanings. In the first sentence, scaled meanings imply a reduced likelihood of legal consequences for Clinton. The second sentence is sourced out to conservatives, a frequently dismissed group of people throughout the New 81

95 York Magazine articles, as exemplified in Engagement Entertain and Engagement Proclaim. Slow and steady scales down the effectiveness of gate, which has done [nothing] to sink [Clinton s] campaign yet. In addition, to refer to the scandal proper, New York Magazine often relies on wordings such as Clinton s saga, story, cliffhanger, inspiring a popular meme, and the whole gate scandal was a distraction. The first four wordings are in play to reinforce the scandal as belonging to the yellow press, taking away from the seriousness of the controversy. The tabloid quality assigned to the scandal by New York Magazine writers is reinforced by the sentence If you're hungry for more Clinton news, get ready for next Wednesday, when the new batch will be released to the public, where devouring gate like a scandal story implicitly evaluates the controversy as overblown, exaggerated, and sensationalised. The last example, the whole gate scandal was a distraction, discusses the controversy in the past tense, implying that the talk about Clinton s usage diminished and is no longer up for discussion. Because this sentence comes from 15 Fun Tidbits From Hillary Clinton s s, published 1 July 2015, this tense choice is unmotivated in retrospect. Repetition is a subcategory of Scaling which looks to up- or down-scale information through the recycling of the same information. Repetition was found prevalent in Fox News articles, especially with regard to statements about the job position and ranks of an external actor. Presenting people s official titles seems accusatory when used differently from the norm (e.g. Clinton knowingly sent classified information during her tenure as America's top diplomat on Fox News, where the usual reference to Clinton is secretary of state), or have an undertone that justifies Clinton s usage (e.g. top White House staffers at the time, including David Axelrod and Rahm Emanuel, were aware that she was using a personal address on New York Magazine). In one salient example, official titles are recycled in a Fox News article to up-scale the seriousness of the assertion. The sentence is as follows. Clinton, the former first lady, senator from New York and top diplomat now running for the Democratic presidential nomination, announced Tuesday that she had told aides to turn over the actual server to the Justice Department, giving in to months of demands that she relinquish the device she used 82

96 to store her correspondence while secretary of state. Here, the unnecessarily long listing of Clinton s past and present positions is used in a vaguely accusatory way to imply how much is at stake in the controversy. In sum, Scaling is used by Fox News journalists to aggravate Clinton s controversy and overplay the severity of her use of a private server. New York Magazine takes a different approach in that it expresses the ineffectiveness of the questions around Clinton s usage for a lack of incriminating evidence Graduation Isolation Isolation is the least frequently used Graduation subcategory in both text sets. In the case of Isolation, the authorial voice, while detailing Clinton s s, nominates one as exemplary and describes that one at length. Based on the offered description, the seriousness of the scandal is aggravated or mitigated. Examples in the Fox News dataset include at least one and another in the following sentences. Fox News reported Friday that at least one Clinton contained information identified as "HCS-O," which is the code for intelligence from human spying. Republican National Committee (RNC) Chairman Reince Priebus seized upon the news of the upgraded s as another reason the 2016 presidential candidate couldn't be taken at her word. Here, the Fox News author selects one attribute as especially important and expresses that these features may also apply to other s. In a similar vein, New York Magazine in one example describes an gate content as yet another update, implying that other updates carry the same quality. In In yet another gate update that sounds shady but proves nothing, the writer up-scales the dubious and ineffective character of the scandal. In another example, content is isolated via only: They re mostly boring administrative s lots of Clinton asking aides to print things for her but 66 were retroactively classified, and the only marked secret features Clinton saying, Wow not good. The approach here is the opposite of how the Fox News sentences use Isolation. The New York Magazine journalist selects the sole that has been marked classified during the review of the Clinton s, and by way of a scare quote and Clinton s quoted words, the writer makes the statement that the has been marked up as secret unnecessarily and without good reason. 83

97 To sum up, Fox News propagates select details to imply that top-secret information potentially passed through Clinton s private server. On the contrary, New York Magazine isolates and presents content that discusses Clinton s everyday office work (e.g. printing, quick letters of response), and shows one classified as apparently containing top-secret information Implicit Meaning Making In this section, I show analysis of evaluative meanings on the sentence level. For this demonstration, I selected 10 Fox News and 10 New York Magazine sentences from the data that make outstanding use of evaluative resources. I highlight in blue wordings that explicitly express Attitude, in red those that render the utterance heteroglossic, and in green Graduation categories. Those sentence parts are underlined where the interplay of evaluative resources construe implicitly opinionated assertions, but a single word cannot be pinpointed as explicitly carrying the evaluative meaning. These are implicitly evaluative propositions. In my explanations, I focus on how the underlined sentence parts contribute to the opinionated discussion of Clinton and her controversy Implicit Meaning Making on Fox News gate articles on Fox News tend to use a vaguely accusatory tone where nation-wide security risks are implied resulting from Clinton s usage. Fox News journalists tend to stress the significance and seriousness of the investigations into Clinton s threads and frequently point out that they possibly contain classified information, such as the government s closely held secrets. The journalists often describe this sensitive information as potentially exposed to the prying eyes of foreign intelligence, as they hypothesise that Clinton failed to encrypt her homebrew server. The writers also often emphasise that the controversy has left its damaging mark on Clinton s past year, and, as more and more revelations come to light, Clinton s presidential bid may be at risk. Another salient aspect in the Fox News articles concerns the journalists treatment of authority. The writers often present Clinton as having the power to direct subordinates to work around the rules. When external voices are introduced to a proposition, the journalists 84

98 tend to make mentions of the voices ranks and titles to lend authenticity to the utterance. Entities such as the FBI, lawmakers, and other officials are often described as A-team, senior, and top. This is to underscore the authenticity and trustworthiness of their announcements about the chains. Table 15 below shows 10 Fox News sentences marked up for explicitly and implicitly evaluative language. The sentences are in chronological order of publication. EVALUATIVE LANGUAGE ON FOX NEWS Sentence There is no evidence Clinton used encryption to shield the s or her personal server from foreign intelligence services or other potentially prying eyes. The FBI is investigating whether classified information that passed through Clinton's so-called homebrew server during her time as secretary of state was mishandled. The source also told Fox News an FBI A-team is leading the extremely serious investigation into Clinton's server and the focus includes a provision of the law pertaining to gathering, transmitting or losing defense information. The revelation adds to the growing questions related to the Democratic presidential front-runner's unusual usage of a private account and server while in government. The source familiar with the investigation said that like all major tech companies on the front lines, Datto has faced cyberattacks, another subject of great interest to the FBI in its probe of Clinton s server. The U.S. intelligence community has reportedly stepped away from claims that two s in Hillary Clinton s private server contained top secret information, according to POLITICO. Clinton has been under fire through much of 2015 about her use of a private, unsecured server as secretary of state, specifically over the security of her server, and her incomplete retention of her s. Evaluative meanings no, Disclaim Denial other, Isolation potentially, Entertain Probability prying, Judgement Propriety whether, Attribute Acknowledge scare quotes homebrew, Judgement Veracity mishandled, Judgement Propriety also, Reckoning told, Attribute Acknowledge scare quotes scare quotes extremely, Scaling quote marks, Attribute Acknowledge growing, Scaling unusual, Judgement Normality said, Attribute Acknowledge all, Reckoning major, Appreciation Composition another, Isolation great, Scaling reportedly, Attribute Acknowledge two, Reckoning according to, Attribute Acknowledge much, Reckoning unsecured, Appreciation Valuation specifically, Isolation incomplete, Appreciation Composition EXCLUSIVE: Hillary Clinton s s on her unsecured, homebrew server contained intelligence from the U.S. government's most secretive and highly classified programs, formatting, Proclaim Pronounce unsecured, Appreciation Valuation homebrew, Judgement Veracity most, Scaling 85

99 according to an unclassified letter from a top inspector general to senior lawmakers. That indicates a level of classification beyond even top secret, the label previously given to two s found on her server, and brings even more scrutiny to the presidential candidate s handling of the government s closely held secrets. Despite Clinton s recent public statements about not knowing how the technology works, at least one suggests she directed a subordinate to work around the rules. secretive, Judgement Veracity highly, Scaling classified, Valuation Composition according to, Attribute Acknowledge unclassified, Valuation Composition even, Disclaim Counter quote marks two, Reckoning even, Disclaim Counter more, Scaling closely, Reckoning despite, Disclaim Counter recent, Reckoning not, Disclaim Denial at least one, Isolation suggests, Entertain Evidentiality Table 15. Ten Fox News Sentences with Implicitly Evaluative Language Implicit Meaning Making on New York Magazine New York Magazine articles tend to highlight the media s unnecessary and undue scrutiny on gate. In doing so, the authors often discuss personal information found in the chains in detail, attending to how Clinton received sympathy s after she fractured her elbow and became unable to do lady things. This and the repeated mentions of Clinton s yoga schedules and her private thoughts on home furnishings are made prominent when the journalists bring in external voices who say that gate was a distraction from the real issues facing our country. To further this perspective, the journalists sometimes introduce popular culture references, e.g. describing the contents as moments that could have been stolen from an episode of Veep, an HBO comedy series (HBO, n.d.). New York Magazine lays focus on Hillary Clinton s character, often presenting her feelings, private thoughts, and everyday matters. Throughout the articles, she is also construed as a fighter who attacks Republicans, gears up for fights, and goes on the offensive. Against this backdrop, gate is presented as a sensationalised burden that has no real substance yet annoyingly prevents Clinton from focussing on her campaign efforts. Table 16 shows the 10 selected New York Magazine sentences marked up for explicitly and implicitly evaluative language. The sentences are in chronological order of their publication date. 86

100 EVALUATIVE LANGUAGE ON NEW YORK MAGAZINE Sentence Some people said the whole gate scandal was a distraction from the real issues facing our country, but we think they'll change their tune once they read Clinton's nickname for Senator Dianne Feinstein, and more of her private thoughts on home furnishings. When Clinton fractured her elbow, Senator Barbara Mikulski expressed sympathy for her inability to do lady things. In yet another gate update that sounds shady but proves nothing, a lawyer for Hillary Clinton's former aide Bryan Pagliano said Wednesday that his client intends to invoke his Fifth Amendment right in response to questions from the House committee investigating the 2012 Benghazi attack. The FBI is still investigating, and Clinton's personal s may be protected by public-information laws, but this raises the possibility that someday the public will be able to learn all about her yoga schedules and Chelsea's wedding-planning process. If Hillary Clinton hoped that saying sorry would make all of her problems go away, she was being a tad too optimistic especially since the Defense Department just found a handful of s that she didn't turn over with the rest of her work-related s. However, many especially those critical of Hillary Clinton's use of a personal account during her tenure at the State Department are sure to interpret the discovery of these s as a sign that there may be more work-related s Clinton failed to make available to the public. Evaluative meanings some, Reckoning said, Attribute Acknowledge whole, Scaling but, Disclaim Counter think, Attribute Acknowledge will, Entertain Probability expressed, Attribute Acknowledge sympathy, Attitude Affect inability, Judgement Capacity yet, Disclaim Counter another, Isolation shady, Judgement Veracity but, Disclaim Counter nothing, Disclaim Denial, Reckoning said, Attribute Acknowledge still, Disclaim Counter may, Entertain Probability but, Disclaim Counter possibility, Entertain Probability someday, Reckoning will, Entertain Probability able, Judgement Capacity all, Reckoning if, Attribute Acknowledge hoped, Attitude Affect saying, Attribute Acknowledge sorry, Judgement Propriety would, Entertain Probability all, Reckoning problems, Scaling tad too, Scaling optimistic, Attitude Affect dash, Proclaim Pronounce especially, Proclaim Pronounce just, Reckoning a handful, Reckoning didn t, Disclaim Denial the rest, Isolation however, Disclaim Counter many, Reckoning dash, Proclaim Pronounce especially, Proclaim Pronounce those, Reckoning critical, Judgement Normality sure, Proclaim Concur interpret, Attribute Acknowledge these, Isolation may, Entertain Probability failed, Judgement Capacity 87

101 Based on the last few releases, they will mostly entail bureaucratic boringness, sharing links to stories that say nice things about the secretary of State, and other moments that could have been stolen from an episode of Veep. From showing off her fun side on Saturday Night Live, to attacking Republicans over their Benghazi obsession, to gearing up for a potential primary fight against Joe Biden, everything Hillary Clinton has done in recent days suggests that she's ready to go on the offensive and move past the issues that plagued the first six months of her campaign. The questions surrounding Clinton s private server haven t taken up much oxygen in the Democratic primary race, save for inspiring a popular meme. Possibly because the details are technical and opaque, Democratic voters have, to this point, been entirely uninterested in another alleged Clinton scandal. last few, Isolation will, Entertain Probability mostly, Reckoning bureaucratic boringness, Appreciation Reaction say, Attribute Acknowledge nice things, Appreciation Reaction other, Isolation could, Entertain Probability fun side, Appreciation Reaction everything, Reckoning suggests, Entertain Evidentiality ready, Judgement Capacity first six, Reckoning much, Reckoning save for, Disclaim Counter popular, Appreciation Reaction possibly, Entertain Probability technical, Appreciation Composition opaque, Appreciation Composition entirely, Scaling uninterested, Attitude Affect another, Isolation alleged, Appreciation Valuation Table New York Magazine Sentences with Implicitly Evaluative Language Sections and detailed the approaches used by Fox News and New York Magazine in their presentation of Clinton s public image and the scandal. Explicit and implicit meaning making shows that Fox News approaches Clinton s usage in a dismissive voice, while New York Magazine in a defensive voice. Between 22 May 2015 and 29 January 2016, these voices remained largely unchanged. This lack of change is notable because this period saw Clinton s usage grow from a political scandal into a morally and legally questioned controversy, resulting in probes and investigations. I address this lack of change of voices in more detail in chapter 7. Discussion. 88

102 7. Discussion This thesis paper aimed to answer the following research question: Is there evaluative language in Fox News and New York Magazine news articles discussing Hillary Clinton s controversy? If so, how is evaluative language expressed and how does it build media bias? I found that Fox News uses negative evaluative language in its coverage of Clinton s scandal. The authorial voice on Fox News overplays the importance of gate, presenting it as a security risk to the United States, and expresses that this is the result of Hillary Clinton s unusual and improper use of a homebrew private server. This evaluation is often expressed by wordings of negative Judgement, by way of bringing sources into the narrative that make dismissive statements about Clinton s usage, showing disbelief of Clinton s own claims, up-scaling the amount of s involved in the security breach, and detailing threads that contained potentially top-secret information. I found that New York Magazine, on the other hand, uses negative evaluative language to report on the dismissive opinions surrounding Clinton s scandal. The authorial voice on New York Magazine underplays the relevance of the controversy, showing that it is a sensationalised and overblown matter in the media, and that it takes away attention from the important issues that the United States face. New York Magazine uses values of positive Affect and Appreciation to present Clinton as a relatable person dealing with everyday matters and having everyday feelings, and details correspondences that underscore how her s contain personal information irrelevant to the public. Moreover, New York Magazine journalists often name the Republicans as a group of people who wish to find incriminating evidence against Clinton in the threads, and stress that gate was a futile attempt to undermine Clinton s presidential bid. These types of evaluations lead to a biased representation of Clinton and her controversy both on Fox News and on New York Magazine. Fox News is biased against Clinton as it overplays the importance and gravity of her practices, often expressing concerns over how Clinton put the nation at risk. The negative disposition of Fox News journalists toward Clinton results also from the largely dismissive tone of their language. New York Magazine is biased toward Clinton as it underplays gate, stressing that it is 89

103 largely overblown in other media only to hinder her run for presidency. New York Magazine journalists discuss Clinton s usage in a defending and permissive way and dismiss the media s exaggerated focus on it. These two divergent ways of presenting gate are notable especially because the controversy did not only create a public divide, but it has also grown to operate on more levels than a political scandal. First and foremost, Clinton as a public figure has been from the beginning deeply embedded in the controversy. Thus, voices of approval and disapproval may be grounded in whether the given journalist has an endorsing or opposing attitude toward Clinton, including the sum of her public image and her political persona (ideology). Second, opinions on gate have also been positioned along morality: questions have been raised about how ethical it is from a United States official to store sensitive information on a private server instead of a governmental one. A third layer to the scandal pertains to legal concerns: once the FBI s involvement in gate was revealed in July 2015, the filing of criminal charges against Clinton became a possibility. Despite the perspectives of morality and legality gaining relevance later in the scandal, the narratives on Fox News and New York Magazine have remained largely unaffected between 22 May 2015 (the publication date of the first article under investigation) and 29 January 2016 (the publication date of the last article under investigation). Fox News advances a dismissing voice and New York Magazine a defending or permissive voice when discussing Clinton s practices. The two media outlets show little resilience in reassessing their narratives of gate throughout the development of the scandal. As Figure 16 shows, the voices around gate in the two media outlets are grounded in Clinton s political stand and public image. The arising legal concerns from May 2015 to January 2016 did little to affect how the journalists described the scandal later. 90

104 DISMISSIVE VOICE legality morality personality (political ideology) DEFENSIVE VOICE Figure 16. Voices and Issues at Stake in Clinton's Controversy The persistent use of the same narrative over time applies in particular to New York Magazine. Following Clinton s controversial remarks about wiping her server, the intelligence community s finding of several dozen s containing intelligence beyond topsecret, and the FBI s launching of its own investigation into Clinton s usage, the last article in the New York Magazine data still described gate as another alleged Clinton scandal that Democratic voters have been entirely uninterested in. To compare the two media outlets, I found New York Magazine to be more biased than Fox News in its presentation of the controversy. First, New York Magazine news articles contain more evaluative language from a metric standpoint. New York Magazine journalists tend to infuse their subjective viewpoints into their texts in a more salient way. These viewpoints defend Clinton s usage, and, because of the growing moral and legal implications around the issue, a gradually more defensive tone was necessary for the New York Magazine narrative to remain unchanged. Despite lesser amounts of evaluative language in its news articles, Fox News also persistently uses one approach to interpret gate developments. For example, when the intelligence community announced that two Clinton s were found to not have contained classified information after a flawed first review, Fox News reported Intelligence officials 91

105 reportedly walk back Clinton top secret claims, implicitly suggesting that the s could in fact contain top-secret communication, and it might only be the latest assessment that claims otherwise. Another marked difference between the two media outlets is in their expression of partisan bias. In the investigated data, Fox News seldom referred to Democrats or Republicans. Instances where this happened include positive evaluations of the Republican Party s agile response to gate developments (Judgement Capacity). While New York Magazine sometimes evaluates the Republican Party similarly, it makes its political stand very explicit throughout the news articles. New York Magazine refers to giddy Republicans and blames them for aggrandising gate. The journalists also express their alignment with the Democrats when they stress the Democrats lack of interest in the overblown scandal made up of uninteresting details about Clinton s yoga schedule that have surfaced. The media outlet also tends to address its audience directly with the frequent use of you. The combination of explicit partisan references and the directly addressed audience positions New York Magazine readers and prompts them to take stands similar to those of the journalists. When, for example, a New York Magazine journalist describes gate as a shallow creation of the yellow press and goes on to say that If you're hungry for more Clinton news, get ready for next Wednesday, the journalist positions his or her audience to subscribe to the writer s ideas and, ideally, dismiss the sensationalised scandal. In sum, I found that Fox News exhibits negative bias toward Clinton. Also, New York Magazine shows positive bias toward Clinton and makes its Democratic partisan bias explicit, thereby positioning its readership to also stand with the Democrats in the controversy. For future research, it is important to note that further refinements can be made to the Appraisal framework analysis of gate news articles. Each Appraisal framework category could be further broken down into finer subcategories in order to account quantitatively for who or what exactly is targeted in each subcategory. Measuring who of what political affiliation is evoked in heteroglossic utterances, and how those assertions are presented by the authorial voice is one consideration. This could shed light on how frequently Republican and Democratic voices are introduced to the texts and whether the journalists 92

106 align or disalign with these voices. As Tawnya Adkins-Covert (2007) points out, one of the most common methods for measuring bias emphasizes who is quoted/given voice in news coverage of social and political issues (Covert & Wasburn, 2007, p. 691). The question arises as to what the ratio is among heteroglossic propositions uttered by Clinton, a Democratic third party, or a Republican third party. To what extent do writers introduce these voices with dismissal (Distance, Disclaim), permission, or endorsement (Acknowledge, Endorse, Entertain)? To answer these questions, the dataset must be extended to include all sentences of the 2x14 gate news articles rather than investigating the ones with references to Clinton only. In addition, background checks of the political affiliations of these people are also necessary to find out what their political stand is. In many cases, both Fox News and New York Magazine report on what the Republican chairman has said in relation to gate. However, the political affiliations of most of the other actors are not explicitly stated in the articles. Assessing the importance of background information also requires further investigation. The Fox News data at one point remind the readers of Clinton s partnership at a law firm in the 1980s, and the journalist refers to her as one of three amigos. New York Magazine, on the other hand, recalls how the Mitchell-Clinton interview took place a day before the 20th anniversary of Hillary Clinton s U.N. speech in Beijing and goes on to elaborate on how she famously stepped up for women s rights. These two instances introduce background information directly irrelevant to gate. What other background information of such nature is included in the two media outlets news articles? What purpose do they serve, and how do they help the authorial voice shape Clinton s image through evaluative resources? This again requires world knowledge and the scrutiny of background information. Lastly, scare quotes also require further analysis. A large-scale investigation of journalistic corpora including scare quotes could help categorise this ambiguous formatting choice in the Appraisal framework to see what effect(s) the authorial voice attempts to make, and to what extent these equivocal scare quotes endanger solidarity between the writer and the reader. 93

107 8. Conclusion In sum, this thesis looked at how political journalists use language to advance their own opinions in news stories, thereby creating bias. I first introduced Hillary Clinton s controversy, then Martin and White s (2005) Appraisal framework, an analytical framework used to identify evaluative language. Second, I introduced two media outlets, Fox News and New York Magazine, whose news articles on gate I investigated. Next, I detailed the Appraisal framework adjusted for the data under investigation. By analyzing news article texts in the framework, I found that Fox News uses dismissive language in its description of Clinton s usage, showing its negative bias toward Clinton. I also found that New York Magazine defends Clinton s usage in its language and shows positive bias toward Clinton and negative bias toward Republicans. These findings shed light on the responsibility of political journalists in presenting news stories through evaluative language and, by those means, creating bias and shaping how stories across the political landscape reach the public. As Leon V. Sigal (1973) remarks, news is not what happens but what someone says has happened or will happen (Sigal, 1973, p. 15). Through evaluative language, political journalists load into their news articles their own evaluations of factual news information. This results in a slanted retelling of events. This, in return, can help journalists win over large audiences to support their cases. Winning or losing support of one s audience is especially significant during political election periods, when politicians are constantly tested in public. The fact that Hillary Clinton s chances of winning in the 2016 presidential elections of delegate counts have lessened in the spring of 2016 is a result of a number of factors. The presentation of her controversy in the press is one of these factors, especially because gate has continued to be highly publicised and divisive throughout 2016 so far. Another important finding of the research is the identification of linguistic resources that are used to defend public figures and their deeds in front of a wide audience. This can be relevant in the field of public reputation work. Using these linguistic resources can help protect, for instance, a client who is caught up in controversies and negative publicity. Because Clinton s controversy has political, moral and legal repercussions which, throughout the Fox News and the New York Magazine articles, have been infused with 94

108 dismissing and defending tones, my research uncovers evaluative language that can be used to protect and attack public figures from political, moral, and legal perspectives. In this way, this thesis showed what types of evaluative language can be useful when a public figure faces challenges in the press. Furthermore, this thesis is interesting with respect to crisis communication work by showing how language can be adjusted to aggravate or mitigate issues. Linguistic resources of aggravation and mitigation can make or break scandals. The evaluative language identified in this thesis also helps over- or underplay key actors and entities involvement and responsibilities in these controversies. Lastly, my research revealed that formatting conventions previously outside the scope of the Appraisal framework carry evaluative potential beyond the word level. This thesis showed instances where the use of quote marks, parentheses, and dashes can be used to infuse written language with the writer s opinions in a subtle and often ambiguous way. 95

109 References Primary Sources Ed Henry. (2015). FBI probe of Hillary Clinton s expands to second tech company. Retrieved January 21, 2016, from FoxNews.com. (2015). Blumenthal gave diplomatic advice to Hillary Clinton as early as 2009, s show. Retrieved January 21, 2016, from FoxNews.com. (2015). Clinton sorry for confusion, stops short of apology for actions. Retrieved January 21, 2016, from FoxNews.com. (2015). Clinton facing new calls to turn over server after IGs request probe. Retrieved January 21, 2016, from FoxNews.com. (2015). FBI investigating security of Hillary Clinton s private server. Retrieved January 21, 2016, from FoxNews.com. (2015). FBI reportedly recovers deleted s from Clinton server. Retrieved January 21, 2016, from FoxNews.com. (2015). Former Clinton aide who helped set up server to plead Fifth Amendment to avoid subpoena. Retrieved January 21, 2016, from xiii

110 FoxNews.com. (2015). Intelligence officials reportedly walk back Clinton top secret claims. Retrieved January 21, 2016, from FoxNews.com. (2015). Internal s show Clinton got detailed intel on planned Benghazi hit. Retrieved January 21, 2016, from FoxNews.com. (2016). State Department releases over 3,000 Clinton s on New Year s Eve. Retrieved January 21, 2016, from FoxNews.com. (2015). State Dept. accused of stiff-arming intel watchdog over Hillary s. Retrieved January 21, 2016, from FoxNews.com. (2015). State Dept. receives Clinton chain apparently not included in pages turned over. Retrieved January 21, 2016, from Fuller, J. (2015). Clinton on s: I Am Sorry That This Has Been Confusing to People. Retrieved January 21, 2016, from Fuller, J. (2015). Defense Department Finds More Clinton s. Retrieved January 21, 2016, from xiv

111 Fuller, J. (2016). New Report Says Some s Sent on Clinton s Server Were Super Top Secret. Retrieved January 21, 2016, from Fuller, J. (2015). State Department Releases Hillary s Benghazi s in Textbook Case of Holiday News Dump. Retrieved January 21, 2016, from Hartmann, M. (2015). Aide Who Set Up Clinton s Server Will Plead the Fifth. Retrieved January 21, 2016, from Hartmann, M. (2015). Clinton Gives In, Turns Over Server to Justice Department. Retrieved January 21, 2016, from Hartmann, M. (2015). If Clinton Has a Backup of Her s, the FBI Would Like to See It. Retrieved January 21, 2016, from Hartmann, M. (2015). 15 Fun Tidbits From Hillary Clinton s s. Retrieved January 21, 2016, from Hartmann, M. (2015). Hillary Is Done Playing Defense, But gate Won t Go Away. Retrieved January 21, 2016, from xv

112 Hartmann, M. (2015). The FBI Managed to Recover Hillary Clinton s Personal s. Retrieved January 21, 2016, from Hartmann, M., & Fuller, J. (2015). Justice Department Asked to Investigate Hillary s s. Retrieved January 21, 2016, from Herridge, C., & Browne, P. K. (2016). Inspector General: Clinton s had intel from most secretive, classified programs. Retrieved January 21, 2016, from Herridge, C., & Browne, P. K. (2016). Official: Some Clinton s too damaging to release. Retrieved January 30, 2016, from Landsbaum, C. (2016). Hillary Clinton Reacts to Benghazi in Newest Batch of s. Retrieved January 21, 2016, from Levitz, E. (2015). Report: Clinton Never Sent Classified Information From Her Private Account. Retrieved January 21, 2016, from Levitz, E. (2016). The State Department Confirms Clinton s Server Hosted Top Secret Information. Retrieved January 30, 2016, from xvi

113 Secondary Sources About the Select Committee Select Committee on Benghazi. (n.d.). Retrieved January 10, 2016, from Alcindor, Y. (2016). Bernie Sanders to Reassess Candidacy After Tuesday s Vote, but He'll Stay in Race. Retrieved May 17, 2016, from Alexa.com. (2016a). How popular is foxnews.com? Alexa Traffic Ranks. Retrieved March 20, 2016, from Alexa.com. (2016b). How popular is nymag.com? Alexa Traffic Ranks. Retrieved March 20, 2016, from Alterman, E. (2003). What Liberal Media? Retrieved January 10, 2016, from Andrews, W., Bennett, K., & Parlapiano, A. (2016) Delegate Count and Primary Results. Retrieved May 18, 2016, from Bednarek, M. (2006). Evaluation in media discourse: Analysis of a newspaper corpus. London: Continuum. Biber, D., & Conrad, S. (2009). Register, genre, and style. Cambridge: Cambridge University Press. CBS News. (2015). Face the Nation Transcripts September 20: Clinton and Paul. Retrieved January 10, 2016, from Chait, J. (2015). Why Hillary Clinton Is Probably Going to Win the 2016 Election. Retrieved November 6, 2015, from xvii

114 Clinton, H. [Hillary Clinton]. (2015, April 12). Getting Started Hillary Clinton [Video file]. Retrieved April 1, 2016 from Covert, T. A., & Wasburn, P. C. (2007). Measuring Media Bias: A Content Analysis of Time and Newsweek Coverage of Domestic Social Issues, Social Science Q Social Science Quarterly, 88(3), doi: /j x Covert, T. A., & Wasburn, P. C. (2009). Media bias?: A comparative study of Time, Newsweek, the National review, and the progressive coverage of domestic social issues, Lanham, MD: Lexington Books Davidson, M. (2005). Right, wrong, and risky: A dictionary of today's American English usage. New York: W.W. Norton & Company. Dilanian, K. (2016). Hillary Clinton s Held Info Beyond Top Secret: IG. Retrieved January 29, 2016, from Dowd, M. (2016). Hillary Is Not Sorry. Retrieved April 21, 2016, from Economou, D. (2009). Photos in the news: Appraisal analysis of visual semiosis and verbal-visual intersemiosis. Sydney: University of Sydney. FAQs on AP's Vote-Counting Network. (2016). Retrieved April 1, 2016, from Ferrechio, S. (2015). Court orders Clinton release every 30 days. Retrieved January 10, 2016, from Fishel, J. (2016). State Department Halts Its Clinton Investigation, Defers to FBI. Retrieved April 21, 2016, from xviii

115 FOMO. (2011). Macmillan Dictionary. In Macmillan Dictionary. Retrieved February 1, 2016, from Fox News. (2016). Retrieved January 10, 2016, from FoxNews.com. (2016). Clinton about non-secure fax re-ignites concerns about her sidestepping security measures. Retrieved January 21, 2016, from Friday News Dump. (n.d.). Taegan Goddard's Political Dictionary. In Taegan Goddard's Political Dictionary. Retrieved February 1, 2016, from Geis, M. L. (1987). The language of politics. New York: Springer-Verlag. Gerstein, J. (2015). Source: Key Clinton s did not contain highly classified secrets. Retrieved January 10, 2016, from Gibson, M. (n.d.). Writing Effective Headlines. Retrieved March 20, 2016, from Goddard, C. (2008). Natural Semantic Metalanguage: The state of the art. In Cross- Linguistic Semantics Studies. doi: /slcs god Hamawand, Z. (2009). The semantics of English negative prefixes. London: Equinox Publishing. HBO. (n.d.). HBO: Veep: About. Retrieved January 21, 2016, from Helderman, S. R. (2015). State Department misses court-ordered goal on Clinton release. Retrieved January 21, 2016, from xix

116 Henshall, P., & Ingram, D. (2008). Chapter 8: Quotes. Retrieved March 20, 2016, from Volume 1/volume1_08.htm Hill, B. (2014). Single Quotation Marks And Their Uses The Editor s Blog. Retrieved March 20, 2016, from Hillary Clinton. Most recent articles. (2016). Retrieved January 7, 2016, from Hommerberg, C., & Don, A. C. (2015). Appraisal and the language of wine appreciation : a critical discussion of the potential of the Appraisal framework as a tool to analyse specialised genres. Functions of Language, 2(22), Kiely, E. (2015). Clinton s Brag. Retrieved January 10, 2016, from Landsbaum, C. (2016). Hillary Clinton Reacts to Benghazi in Newest Batch of s. Retrieved January 21, 2016, from Leopold, J. (2016). There Are 1,800 Reasons Why the Controversy Over Hillary Clinton s s Is Far From Over. Retrieved March 20, 2016, from Maintain. (n.d.). Oxford Advanced Learner's Dictionary. In Oxford Advanced Learner's Dictionary. Retrieved February 1, 2016, from Malloy, T., Associates, R., & Relations, P. (2016). BIDEN RUNS BETTER THAN CLINTON AGAINST TOP REPUBLICANS, QUINNIPIAC UNIVERSITY NATIONAL POLL FINDS; TRUMP GOP LEAD GROWS AS CLINTON DEM LEAD SHRINKS. Martin, J., & White, P.R.R. (2005). The language of evaluation: Appraisal in English. Basingstoke: Palgrave Macmillan. xx

117 MSNBC. (2015, September 4). Hillary Clinton Interview: One-On-One Andrea Mitchell MSNBC [Video file]. Retrieved January 10, 2016, from N.A. (2014). More News Media Sources Trusted by Those on the Left. Retrieved January 10, 2016, from N.A. (2016). NYMag. Retrieved January 10, 2016, from New York Daily News. (2015). TRANSCRIPT: Hillary Clinton takes responsibility for Benghazi attack, defends actions in emotional testimony before Congress. Retrieved January 10, 2016, from Peek, L. (2015). Hillary Clinton and 2016: It s about the money, stupid. Retrieved November 6, 2015, from it-about-money-stupid.html# Pompeo, J. (2013). New York mag hires new online deputy, as traffic grows. Retrieved January 10, 2016, from Putnam, J. (2015, April 28). South Carolina Republicans eyeing February 20 presidential primary date. Retrieved February 1, 2016, from Remorse. (n.d.). Collins Dictionary. In Collins Dictionary. Retrieved February 1, 2016, from Results found for Hillary Clinton. (2016). Retrieved January 7, 2016, from results/search?q=hillary%20clinton&ss=fn&sort=latest&min_date= &max_date= &start=5020 xxi

118 Rosato, E. (2013). Adjective order in English: A semantic account with crosslinguistic applications. Schmidt, M. S. (2015). Hillary Clinton Used Personal Account at State Dept., Possibly Breaking Rules. Retrieved January 10, 2016, from Schmidt, M. S., & Apuzzo, M. (2015). Inquiry Sought in Hillary Clinton s Use of . Retrieved January 10, 2016, from Schmidt, M. S., & Chozick, A. (2015). Using Private , Hillary Clinton Thwarted Record Requests. Retrieved January 10, 2016, from Sigal, L. V. (1973). Reporters and officials: The organization and politics of newsmaking. Lexington, MA: D. C. Heath and Company. Stirewalt, C. (2015). Another day of damaging disclosures for Hillary. Retrieved January 21, 2016, from The New York Times. (2015, March 10). Hillary Clinton Answers Questions on Controversy [FULL] The New York Times [Video file]. Retrieved January 10, 2016, from The Wall Street Journal. (2015). Updated Timeline: Hillary Clinton s Troubles. Retrieved January 10, 2016, from Trochim, W. M. K. (2006). Deduction & Induction. Retrieved January 10, 2016, from xxii

119 Tomlinson, L. (2016). Clinton fumed at staff after addressing Tunisian official by wrong name, s show. Retrieved January 21, 2016, from Virtual Reading Room Documents Search Results. (2015, May 22). Retrieved February 1, 2016, from Virtual Reading Room Documents Search Results. (2015, June 30). Retrieved February 1, 2016, from Virtual Reading Room Documents Search Results. (2015, December 31). Retrieved February 1, 2016, from Virtual Reading Room Documents Search Results. (2016, January 7). Retrieved February 1, 2016, from White, P. R. R. (2015). 1. Attitude/Affect 5. Retrieved January 26, 2016, from 04.htm#P57_12861 White, P. R. R. (2015). 2. Attitude/Judgement 1. Retrieved January 26, 2016, from Word Counter: The Word Count Calculator. (n.d.). Retrieved January 25, 2016, from xxiii

120 Appendices Appendix 1. Part of the New York Magazine News Archive xxiv

121 Appendix 2. Part of the Fox News News Archive xxv

122 xxvi