Exploring student perceptions of language learning affordances of Large Language Models: A Q methodology study

Amid the ongoing discussion about the potential of LLMs (Large Language Models) to facilitate language learning, there has been a broad spectrum of views in academia. However, little is known about the different viewpoints of students and what contributes to these differences. In light of this, this study adopts Q-methodology, a mixed-methods approach to examine human subjectivity, to investigate the perceived language learning affordances of LLMs among Chinese students. 26 undergraduates from two universities were asked to rank statements related to the potential affordances and limitations of LLMs in terms of language learning. Factor analysis of these ranks revealed three distinct viewpoints: prudent skeptics, enthusiastic embracers and conservative adopters. The skeptics question the learning opportunities that LLMs can provide and voice concerns regarding the absence of human element; the embracers are generally pro-LLMs and they advocate for personalized learning opportunities that LLMs can offer; the conservatives cautiously acknowledge the supplementary role of LLMs, with a focus on users' ability to ask proper questions. The findings highlight the need to address key issues such as scaffolding, AI literacy and individual difference when utilizing LLMs in language learning contexts. While the nature of Q methodology may limit the generalizability of the results, the study contributes valuable insights into student perceptions regarding LLM affordances for language learning. This in-depth and nuanced understanding of student perceptions can not only help address student concerns and expectations, but also inform the curriculum development, pedagogical practice, and educational policy for integrating LLM into language education.