文献类型：期刊文献

英文题名：Towards Rigorous Corpus-Based AI-Human Text Comparisons: A Methodological Synthesis With a Genre-Informed Illustration

作者：Zhang, Yan[1]

机构：[1]East China Univ Sci & Technol, Sch Foreign Languages, Shanghai, Peoples R China

年份：2025

外文期刊名：INTERNATIONAL JOURNAL OF APPLIED LINGUISTICS

收录：;WOS:【SSCI(收录号:WOS:001639018700001)，A&HCI(收录号:WOS:001639018700001)】；

语种：英文

外文关键词：AI-generated; corpus-based; comparative; EAP writing; methodology; (sic)(sic)(sic); (sic)(sic)(sic); AI(sic)(sic)(sic)(sic); (sic)(sic)(sic)(sic)(sic)(sic)

摘要：Corpus-based comparisons of AI- and human-authored academic texts extend established research traditions in English for Academic Purposes (EAP) but also present new methodological challenges arising from the real-world use of AI tools for communication. Addressing concerns about transparency and comparability, this paper pursues a dual focus: first, to synthesize current methodological practices in corpus-based AI-human text comparisons; and second, to illustratehow methodological decisions can be implemented transparently and with theoretical grounding, drawing on insights from the synthesis. Drawing on 14 recent studies, the synthesis examines rationales for linguistic feature selection, corpus construction, theoretical alignment, and pedagogical relevance. Building on these findings, a methodological checklist (Appendix A) is proposed and demonstrated through an an illustrative case comparing ChatGPT- and human-authored lay summaries, showing how linguistic features such as cohesion and syntactic complexity can be operationalized, justified, and aligned with both theoretical frameworks and genre conventions. The synthesis reveals emerging methodological trends and proposes directions for enhancing transparency and rigor in future corpus-based AI-human text comparisons. (sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(AI)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic), (sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(EAP)(sic)(sic)(sic)(sic)(sic)(sic)(sic), (sic)(sic)AI(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic).(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic), (sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic):(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)AI(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic);(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic),(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic).(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)14(sic)(sic)(sic)(sic)(sic)(sic)(sic), (sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic),(sic)(sic)(sic)(sic)(sic)(sic),(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic).(sic)(sic)(sic)(sic)(sic)(sic), (sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)((sic)(sic)(sic)A), (sic)(sic)ChatGPT(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic), (sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic), (sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic).(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic), (sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic), (sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic),(sic)(sic),(sic)(sic)(sic)(sic)(sic)(sic)(sic)AI(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic).