前期出版
前期出版
頁數：55﹣110
華裔認同協商的爭吵劇場： 基於YouTube評論語料庫的論述分析
研究論文
作者(中)
柯籙晏
關鍵詞(中)
大語言模型輔助內容分析、身分協商、原生數位資料、媒介事件、語料庫輔助論述分析、諸華裔社群
關鍵詞(英)
Large Language Model-Assisted Content Analysis, identity negotiation, natively digital data, media events, Corpus-Assisted Discourse Studies, ethnic Chinese communities
中文摘要
網路防火長城的隔離，導致中國境內用戶在全球主流網站上缺席；而這個結構因素，阻礙了作為整體的全球華裔社群的線上認同協商活動。然而，在2021年發表，諷刺激進中國國族主義網民「小粉紅」的YouTube音樂影片〈玻璃心〉，意外引發大量（自稱）中國用戶冒險翻牆參與評論，與全球其他用戶展開激烈的華裔身分認同協商。
本研究將〈玻璃心〉與其209,744則評論，視為一個原生線上媒介事件與資料邊界明確的全量語料庫，透過語料庫輔助的論述分析（Corpus-Assisted Discourse Studies, CADS），針對其中華裔身分協商的各種論述進行系統分析，並提供方便理解這些論述之間關係的地圖。
研究結果顯示，〈玻璃心〉評論是一個爭奪彼此華裔相關身分邊界的即興爭吵劇場，參與的草根主角們透過各種語藝實踐，宣告自我華裔相關身分，並且爭論這些集體身分的實際內涵。而這些爭吵的焦點，是一套強調漢人血緣、華語、中文與中共黨國政治意識形態一體化的中國人身分敘事。無論是認同或拒斥，都是透過回應這套敘事的主張，使彼此身分的實質邊界浮現。
本研究將〈玻璃心〉與其209,744則評論，視為一個原生線上媒介事件與資料邊界明確的全量語料庫，透過語料庫輔助的論述分析（Corpus-Assisted Discourse Studies, CADS），針對其中華裔身分協商的各種論述進行系統分析，並提供方便理解這些論述之間關係的地圖。
研究結果顯示，〈玻璃心〉評論是一個爭奪彼此華裔相關身分邊界的即興爭吵劇場，參與的草根主角們透過各種語藝實踐，宣告自我華裔相關身分，並且爭論這些集體身分的實際內涵。而這些爭吵的焦點，是一套強調漢人血緣、華語、中文與中共黨國政治意識形態一體化的中國人身分敘事。無論是認同或拒斥，都是透過回應這套敘事的主張，使彼此身分的實質邊界浮現。
英文摘要
Background and Objectives
Due to the Great Firewall—China’s nationwide internet censorship system—mainland Chinese internet users rarely participate at scale in identity negotiations among ethnic Chinese communities mediated by major global platforms. In October 2021, Malaysian Chinese singer Namewee and Australian Chinese singer Kimberley Chen released their Mandopop music video Fragile on YouTube, satirizing China’s radical nationalist netizens, termed the “Little Pinks.” The video drew large numbers of mainland Chinese users who circumvented their government’s firewall to comment on it, sparking disputes with overseas users over questions of ethnic Chinese identity.
This research treats the identity negotiation mediated by the Fragile comment section as a natively online media event with clear data boundaries. Its emergence, interaction, and meaning-making all unfolded within YouTube, constituting a relatively closed observable unit - namely, a corpus of over 200,000 comments posted under the Fragile music video. Accordingly, the research employs CADS to present these identity negotiation discourses and to chart a boundary map of their relations.
Literature Review
This paper reviews three bodies of literature. First, identity theory: drawing on social constructivism, identity is treated as a category shaped, reproduced, and mobilized in social interaction. Second, mediated collective identity: this body of literature traces research on national and ethnic identities mediated by print, radio, and television, highlighting the distinctive duality of online space relative to these mass media—namely, that this space can be both a site for negotiating identities across physical boundaries and a space of algorithmically segregated identities. Third, empirical literature on online ethnic Chinese identity negotiation: pointing out that the methods used in these studies, including surveys, in-depth interviews, and digital ethnography, as well as keyword-driven mining of large-scale online data, face limitations in addressing the simultaneously integrated and segregated duality of online space, and proposing a census corpus approach to circumvent the following limitations: (1) using natively digital data to capture negotiation processes directly, avoiding the limits of retrospective accounts; (2) treating the corpus as a population rather than a sample, sidestepping issues of sampling representativeness; and (3) collecting all comments from a complete event rather than filtering by keywords, avoiding omissions that keyword selection entails. The Fragile video thus offers a rare opportunity to address the dual methodological and field-related gaps in the literature on online ethnic Chinese identity.
Research Question
Theoretically, integrating media event theory and mediated ritual theory, this study treats the video as a grassroots conflictual media event viewed as an improvised agonistic theater that lacks a rigid script yet retains the theatrical structure of role, scene, and performance. Methodologically, as natively digital data, the comments under this single YouTube video constitute a bounded, observable census corpus, allowing researchers to observe the articulation, contestation, and reorganization of discourses without recourse to external data.
Preliminary analysis further shows that the Fragile corpus contains a high frequency of personal pronouns and ethnic Chinese identity-related terms. Drawing on the positioning theory, which treats personal pronouns as key linguistic indicators of identity claims and boundary work, this distribution suggests that interactions in the corpus are organized around self-identity claims and the ascription of identities to others. Accordingly, this study asks the following questions.
RQ: What statistically representative patterns of ethnic Chinese-related identity reference appear in the Fragile video comments? How do participants use these patterns to contest identity boundaries within the agonistic theater?
Method
This research employs CADS, introducing Large Language Model-Assisted Content Analysis (LACA) as a bridge between corpus statistics and discourse interpretation, addressing the statistical storytelling risk endemic to conventional CADS, wherein the statistical patterns identified in the corpus are representative, but the samples researchers extract to interpret these patterns are not. The procedural steps are: (1) corpus preprocessing; (2) word frequency and collocation analysis; (3) extracting keywords-in-context (KWIC) entries using first- and second-person pronouns as search terms; (4) drawing random samples independently for each search term based on a 95% confidence level and a 3% margin of error, yielding a total of 3,567 entries; (5) binary coding of the sampled entries by a large language model (LLM) to identify A-samples that claim self-identity or ascribe identity to others in ethnic Chinese-related terms, yielding 937 A-samples with inter-coder reliability Cohen’s κ ranging from .869 to .979; and (6) focused coding and relational analysis of A-samples based on grounded theory to identify narrative trajectories within the agonistic theater, then returning to the full corpus for cases reflecting who is arguing with whom, about what, and how.
Findings
The findings show that the performances within the improvised agonistic theater all take a standard identity narrative - a Chinese identity that opposes Western imperialism under the leadership of the Chinese Communist Party (CCP) party-state - as their reference point. In this discursive space, other identity narratives are not merely juxtaposed with the standard narrative, but are recognized, responded to, and evaluated in relation to it. Alternative identity narratives are deconstructions and recombinations of the elements the standard narrative asserts - ancestry (Han Chinese), language and script (Mandarin and Sinographs), state sovereignty (the People’s Republic of China), and political ideology (party-state ideology) - forcing the enacted boundaries of identity to surface in the tension between the standard narrative and alternative ones.
Discussion and Conclusion
From the perspective of the integrative identity negotiation theory, when incompatible values collide in a shared field, the resulting friction marks the enacted boundaries that separate participants. The Fragile agonistic theater can thus be read as a breaching experiment: through satire of radical Chinese nationalists, it forces the boundaries among ethnic Chinese communities into view. The boundary map produced here allows comments that appeared scattered or meaningless, including non-Chinese-language comments, comments seemingly unrelated to the disputes, and superficially flippant verbal attacks, to be located within the map.
By introducing LACA as a bridge between corpus analysis and discourse analysis within CADS, this paper methodologically supplies the alignment procedure that conventional CADS empirical work often lacks, reducing the risk of statistical storytelling and offering future researchers facing similar problems with massive online corpora a path. This study does not predict trends in ethnic Chinese identity. Rather, it maps the discursive trajectories of one specific natively online media event, providing a comparative basis for subsequent research.
Due to the Great Firewall—China’s nationwide internet censorship system—mainland Chinese internet users rarely participate at scale in identity negotiations among ethnic Chinese communities mediated by major global platforms. In October 2021, Malaysian Chinese singer Namewee and Australian Chinese singer Kimberley Chen released their Mandopop music video Fragile on YouTube, satirizing China’s radical nationalist netizens, termed the “Little Pinks.” The video drew large numbers of mainland Chinese users who circumvented their government’s firewall to comment on it, sparking disputes with overseas users over questions of ethnic Chinese identity.
This research treats the identity negotiation mediated by the Fragile comment section as a natively online media event with clear data boundaries. Its emergence, interaction, and meaning-making all unfolded within YouTube, constituting a relatively closed observable unit - namely, a corpus of over 200,000 comments posted under the Fragile music video. Accordingly, the research employs CADS to present these identity negotiation discourses and to chart a boundary map of their relations.
Literature Review
This paper reviews three bodies of literature. First, identity theory: drawing on social constructivism, identity is treated as a category shaped, reproduced, and mobilized in social interaction. Second, mediated collective identity: this body of literature traces research on national and ethnic identities mediated by print, radio, and television, highlighting the distinctive duality of online space relative to these mass media—namely, that this space can be both a site for negotiating identities across physical boundaries and a space of algorithmically segregated identities. Third, empirical literature on online ethnic Chinese identity negotiation: pointing out that the methods used in these studies, including surveys, in-depth interviews, and digital ethnography, as well as keyword-driven mining of large-scale online data, face limitations in addressing the simultaneously integrated and segregated duality of online space, and proposing a census corpus approach to circumvent the following limitations: (1) using natively digital data to capture negotiation processes directly, avoiding the limits of retrospective accounts; (2) treating the corpus as a population rather than a sample, sidestepping issues of sampling representativeness; and (3) collecting all comments from a complete event rather than filtering by keywords, avoiding omissions that keyword selection entails. The Fragile video thus offers a rare opportunity to address the dual methodological and field-related gaps in the literature on online ethnic Chinese identity.
Research Question
Theoretically, integrating media event theory and mediated ritual theory, this study treats the video as a grassroots conflictual media event viewed as an improvised agonistic theater that lacks a rigid script yet retains the theatrical structure of role, scene, and performance. Methodologically, as natively digital data, the comments under this single YouTube video constitute a bounded, observable census corpus, allowing researchers to observe the articulation, contestation, and reorganization of discourses without recourse to external data.
Preliminary analysis further shows that the Fragile corpus contains a high frequency of personal pronouns and ethnic Chinese identity-related terms. Drawing on the positioning theory, which treats personal pronouns as key linguistic indicators of identity claims and boundary work, this distribution suggests that interactions in the corpus are organized around self-identity claims and the ascription of identities to others. Accordingly, this study asks the following questions.
RQ: What statistically representative patterns of ethnic Chinese-related identity reference appear in the Fragile video comments? How do participants use these patterns to contest identity boundaries within the agonistic theater?
Method
This research employs CADS, introducing Large Language Model-Assisted Content Analysis (LACA) as a bridge between corpus statistics and discourse interpretation, addressing the statistical storytelling risk endemic to conventional CADS, wherein the statistical patterns identified in the corpus are representative, but the samples researchers extract to interpret these patterns are not. The procedural steps are: (1) corpus preprocessing; (2) word frequency and collocation analysis; (3) extracting keywords-in-context (KWIC) entries using first- and second-person pronouns as search terms; (4) drawing random samples independently for each search term based on a 95% confidence level and a 3% margin of error, yielding a total of 3,567 entries; (5) binary coding of the sampled entries by a large language model (LLM) to identify A-samples that claim self-identity or ascribe identity to others in ethnic Chinese-related terms, yielding 937 A-samples with inter-coder reliability Cohen’s κ ranging from .869 to .979; and (6) focused coding and relational analysis of A-samples based on grounded theory to identify narrative trajectories within the agonistic theater, then returning to the full corpus for cases reflecting who is arguing with whom, about what, and how.
Findings
The findings show that the performances within the improvised agonistic theater all take a standard identity narrative - a Chinese identity that opposes Western imperialism under the leadership of the Chinese Communist Party (CCP) party-state - as their reference point. In this discursive space, other identity narratives are not merely juxtaposed with the standard narrative, but are recognized, responded to, and evaluated in relation to it. Alternative identity narratives are deconstructions and recombinations of the elements the standard narrative asserts - ancestry (Han Chinese), language and script (Mandarin and Sinographs), state sovereignty (the People’s Republic of China), and political ideology (party-state ideology) - forcing the enacted boundaries of identity to surface in the tension between the standard narrative and alternative ones.
Discussion and Conclusion
From the perspective of the integrative identity negotiation theory, when incompatible values collide in a shared field, the resulting friction marks the enacted boundaries that separate participants. The Fragile agonistic theater can thus be read as a breaching experiment: through satire of radical Chinese nationalists, it forces the boundaries among ethnic Chinese communities into view. The boundary map produced here allows comments that appeared scattered or meaningless, including non-Chinese-language comments, comments seemingly unrelated to the disputes, and superficially flippant verbal attacks, to be located within the map.
By introducing LACA as a bridge between corpus analysis and discourse analysis within CADS, this paper methodologically supplies the alignment procedure that conventional CADS empirical work often lacks, reducing the risk of statistical storytelling and offering future researchers facing similar problems with massive online corpora a path. This study does not predict trends in ethnic Chinese identity. Rather, it maps the discursive trajectories of one specific natively online media event, providing a comparative basis for subsequent research.
36次下載
2026/ 秋
No.169
