An Appraisal Analysis Of New York Times' Reports On Riots In China And India

In order to explore the writing styles of different news reports and to learn the means of conveying information, opinions and ideologies by modern influential media, a small linguistic corpus is built up under the framework of Appraisal Theory in the light of the theories and methods practiced in the field of Critical Discourse Analysis, with the help of UAM Corpus Tool developed by Mick O'Donnell. The research focuses on the news reports on social riots happening in Urumqi of China and Punjab of India last year. A collection of related news writings published by New York Times within 13 months from the outbreak of those social events has been made, analyzed, and annotated carefully according to the Appraisal Framework. The work of this stage is qualitative. The quantitative statistics is provided by the professional statistic software. Reliable results may be obtained by combining the qualitative and quantitative researches.The Appraisal Theory has originated and developed from the dissatisfaction about the deficiency of traditional language studies in the field of semantics and rhetoric. Appraisal Theory, exploring language from the functional and cognitive perspectives, aims at describing the means by which the speaker/writer expresses their sentiments, evaluates things and interacts with others by the language. It contains three basic categories: Attitude, Engagement and Graduation. And in turn, each category has a number of sub-categories and different dimensions.Two kinds of annotation are employed in the research, namely document-coding and segment-coding. First, each text in the corpus is coded as a whole document, and so, according to the contents of the news, they are classified into two sets of data: news in China (the reports on the Urumqi riots in the early July, 2009) and news in India (the reports on the riots in Punjab of India in the late May, 2009). Then all the materials are analyzed word by word and sentence by sentence to assign the related Appraisal values to each linguistic segment. The corpus tool software is used throughout the whole annotating process, and it can provide both general descriptive calculations and statistics for a feature-based comparison between multiple sets of data and multiple layers of features. Levels of significance of the differences between the sets are displayed both in terms of T-test and Chi-squared results.Investigation shows that the writings all embody complex appraisal values, possibly due to the fact that they are the reports on kind of complicated, violent social phenomenon. The attitudinal analysis reveals that, there are significant differences in terms of Appreciation, Explicitness and Attitude-Polarity between the two sets of data. Considering the attitudinal positioning, there is clear bias or negative evaluation on the Chinese authorities and Xinhua News Agency. Other Attitudinal values bear much resemblance between the two sets. The small proportion of Affect values may be determined by the so-called objectivity required by journalism, while the larger proportion of Judgment features may be related to the news contents. Both the riots have the nature of ethnic and religious conflicts.The similarities in Engagement resources may also be determined by the doctrines advocated in journalism such as objectivity and truthfulness. The reports on the riots emphasize the sharing of language power by quoting from different parties and providing sophisticated references. Riots are intense and violent social actions. The big part of explicit lexical tokens for Graduation is Force. However, its sub-category Extent demonstrates clear difference between the two sets.In spite of the small scale of the corpus, the linguistic materials selected here do demonstrate the values of the appraisal resources, and herein, detailed and comprehensive analyses are conducted. Generally, the research can be trusted and the reliability of statistics can be verified to a certain degree by the fact that all the conclusions regarding the differences are drawn at the 98% level of significance.