1 Introduction

This section introduces the development background of the online live streaming industry, noting that internet hosts were once looked down upon but were officially recognized as a new national profession in 2024. Despite the industry's rapid growth, public perception remains negative, which is related to media discourse practices. As a mainstream media outlet, the discourse strategies of the People's Daily have a significant impact on public perception. This study aims to decode how it constructs the image of internet hosts through discourse.

2 Literature Review

This section reviews the current state of research on internet hosts in China, focusing on the two dimensions of "self-presentation" and "external evaluation." In terms of identity, it explores how hosts shape their media identity through role-playing and the resulting identity crises. Regarding stigmatization, it analyzes the influence of channels such as bullet comments and tipping on the spread of stigma, as well as the media's reinforcement of hosts' negative images in reporting. It also points out shortcomings in previous research, leading to this study’s objective and research questions: using reports from the People's Daily from 2006 to 2024 as a corpus and combining corpus linguistics with diachronic discourse analysis to explore the diachronic construction trajectory and social implications of the image of internet hosts.

3 Research Methods

This section introduces the methods for studying the discourse changes in the People's Daily reports on hosts. It adopts a corpus research approach, utilizing the Extended Units of Meaning (EUM) model, which includes five components: the core meaning, collocation, colligation, semantic preference, and semantic prosody, combined with marker analysis to examine linguistic patterns. The research steps involve corpus construction, text preprocessing, topic modeling, manual coding, collocation analysis, and linguistic pattern description. The corpus was collected from the "People's Daily Graphic Database," using "host" as a keyword for retrieval, resulting in 832 reports after screening, totaling approximately 9.94 million tokens.

4 Findings

This section provides an in-depth analysis of the discourse changes in the People's Daily reports on hosts from 2006 to 2024. The study first divides the development of China's internet host industry into three stages: the budding stage (2006–2015), the explosive growth stage (2016–2019), and the standardization stage (2020–2024). It details key events in each stage, such as the operation of the 9158 platform in 2005 marking the emergence of the show-based live streaming model, the explosive growth of live streaming in 2016, and the diversification of scenarios and assistance to traditional industries during the pandemic in 2020. The number of reports on internet hosts in the People's Daily corresponds to the three-stage industry development, showing an overall increasing trend, with surges in 2016 and 2020, followed by a decline and subsequent rise after 2020.

In terms of topic analysis, the study uses the LDA topic model to extract themes from host-related texts, identifying eight core themes: "Host + Industrial Upgrading," "Information Technology," "Industrial Economy," "Cultural Ecology," "New Forms of Entertainment," "Industry Regulation," "New Forms of Employment," and "Vulgar and Unhealthy Content." Among these, the "Host + Industrial Upgrading" theme saw a sharp increase in 2020 before declining but remains prominent, the "Industry Regulation" theme has increased year by year since 2016, the "Vulgar and Unhealthy Content" theme has decreased annually after 2020, and reports on the "New Forms of Employment" theme have grown yearly. The flow of these themes reflects changes in the People's Daily's focus in reporting on internet hosts.

Sentiment analysis shows that positive reports increased gradually from 2010 to 2015, with the highest proportion of negative reports in 2016. Afterward, negative reports increased annually but decreased in 2019. After 2020, the number of positive reports far exceeded neutral and negative reports. In terms of collocation and linguistic pattern analysis, texts from the budding stage were few and showed no significant collocations or linguistic patterns. Collocation analysis during the explosive growth stage revealed diverse collocates in the left-1 position and more limited collocates in the right-1 position. Linguistic pattern description and semantic prosody analysis indicated that industry regulation and vulgar/unhealthy content were focal points, while reports on AI technology applications for hosts were positive. Collocation analysis and linguistic pattern description for the standardization stage were not detailed in this text.

5 Conclusion

This section summarizes the study on the discourse changes in the People's Daily reports on internet hosts from 2006 to 2024. Based on a self-constructed corpus, the study analyzes the construction of the host image across three stages from a diachronic perspective. From 2006 to 2015, hosts emerged as a new form of entertainment, with reports predominantly positive. From 2016 to 2019, reports focused on criticism, shaping a negative image. After 2020, the discourse shifted to praise, incorporating narratives of positive energy and reshaping hosts as contributors to society. The study provides insights for discourse analysis, suggesting a shift from traditional analytical approaches, focusing on topics like employment and people's livelihoods, and helping practitioners understand social development and cultural trends.