1 Introduction

Emphasis was placed on the importance of Corporate Social Responsibility (CSR) reporting in corporate information disclosure, highlighting its impact on investor decision-making. Huawei and British Telecom (BT), as representatives of CSR reports in the telecommunications industry, have their authenticity and effectiveness verified by third-party auditing agencies. The article uses the sentiment lexicon method to conduct sentiment analysis on the English CSR reports of these two companies, and combines critical discourse analysis to explore the ideological differences in emotional expression in the reports. The aim is to provide guidance for the writing of English CSR reports by multinational companies in China and help promote the international dissemination of corporate image.

2 Research Status

In recent years, sentiment analysis has received attention in business discourse research, involving natural language processing, intelligence science, and management. Researchers use algorithm design to conduct sentiment analysis on texts such as online reviews, annual reports, and e-commerce evaluations of products. Fuoli (2012) explored the internal and external relationships and image shaping of companies in CSR reports, while Wang Lifei and Bu Han (2017) compared the emotional tendencies of annual reports of Chinese and American companies. Shao Shanshan and Wang Lifei (2019) analyzed the emotional differences in English Chinese evaluation discourse in e-commerce. However, there is currently a lack of research on sentiment mining in CSR reports. This article will use the LIWC-22 Sentiment Dictionary to compare the emotional differences in CSR reports between Huawei and BT companies, and explore the underlying motivations from a critical discourse analysis perspective. The research aims to answer two questions: 1) the differences in emotional relationship construction between Huawei and BT CSR reports; 2) The ideological motivations behind these differences.

3 Research Design

Introduced the research design, including the self built corpus, research tools, and research steps. The corpus contains 20 English CSR reports from Huawei and BT companies from 2012 to 2021. Analysis shows that Huawei's reports have high vocabulary complexity, low use of pronouns, and high use of conjunctions, which may be related to non English native speakers. LIWC-22 Emotion Dictionary is used to identify emotions and psychological processes in text, selecting 16 indicators for analysis. The research steps include downloading reports, text processing, obtaining emotional indicators, and data analysis to reveal the differences in emotional expression between Chinese and foreign CSR reports.

4 Results and Discussion

In the "Results and Discussion" section of the paper, the researcher conducted an sentiment analysis comparison of English social responsibility reports from multinational corporations in China and abroad. Through statistical analysis, it was found that there are significant differences in the use of personal pronouns between Huawei and BT companies. BT company reported that the first person plural pronoun "we" is used more frequently than Huawei company, while Huawei company tends to use the first person plural pronoun "we" to express the community relationship between the company and readers, as well as the third person plural pronoun "they" to refer to third parties outside the company and readers. The statistical analysis of emotional process words shows that both companies have more positive emotional words than negative emotional words in their reports, indicating that the companies tend to establish connections with potential investors through positive emotional resources. Huawei's report uses positive sentiment words more frequently, while BT company provides a more comprehensive introduction to the challenges and risks faced by the company in the report. The analysis of social history words reveals the different focuses of the two companies on social themes such as family, friendship, politeness, and communication. BT company's performance on family and friendship themes is more prominent, while Huawei company emphasized its care for employee health during the epidemic.

5 Conclusion

By analyzing the English CSR reports of Huawei and BT companies through an emotional dictionary, significant differences were found in the use of personal pronouns, emotional process words, and social element descriptions between the two companies. BT reports more actively demonstrate social responsibility, while Huawei focuses on long-term emotional maintenance. Positive sentiment words decrease over time in Huawei reports, while BT reports have a more balanced use of sentiment words. The two companies also have different attitudes towards describing difficulties, which may be influenced by Chinese and English cultural values. The BT report constructs multidimensional subject emotional relationships, while Huawei highlights humanistic care during the pandemic era. The research provides reference for the writing of English CSR reports for multinational corporations in China, and suggests considering the target language readers' habits and cultural backgrounds, effectively conveying information, and enhancing the international competitiveness of enterprises. The research relies on the LIWC-22 sentiment dictionary, and in the future, open source programming languages can be used to further investigate the sentiment semantic network.