EFL learners’ perceptual perezhivaniya and actual writing revision behaviors mediated by GenAI: A sociocultural theory perspective

Abstract

In recent years, Generative Artificial Intelligence (GenAI) has gained prominence as a valuable tool in enhancing EFL learners’ experiences in Second Language Acquisition (SLA). However, empirical studies specifically investigating on GenAI’s role in mediating the revision stage of EFL learners’ L2 writing processes remain scarce. Notably, research exploring the connection between learners’ perceived perezhivaniya—derived from Vygotskyan theory, denoting an individual’_’s subjective lived experiences—and their actual revision behaviors is particularly limited. To address this lacuna, we employed interviews, Q-methodology, and the Translog screen recording tool to investigate the revision processes of nineteen EFL learners at a university in Southern China as they revised L2 continuation writing tasks using GenAI. A key contribution of this study is the development of a novel seven-dimensional L2 writing revision taxonomy (word choice, content, discourse, syntax, errors, alignment, and typographic elements), which guided both the Q-questionnaire design and the Translog data coding. This framework uniquely facilitated a comparison between perceived perezhivaniya and actual revision behaviors by integrating online revision with offline reflective perceptions, thereby capturing an essential dynamic aspect of the concept of perezhivanie. Results indicated a strong consistency between most of the learners’ perceived experiences and their actual revision behaviors, in content, alignment, and typographic elements. However, for some learners, alignment—vital for continuation writing—was frequently overlooked despite its perceptual recognition, highlighting significant pedagogical implications.