Answers are grounded in this paper and cite their sources.
Exploratory case study investigates generative AI's role in teaching metaphorical expressions to Georgian learners, suggesting enhanced instructional strategies.
This study explores the intersection of phraseodidactics and generative artificial intelligence (AI) within the context of teaching French as a Foreign Language (Français Langue Étrangère - FLE) to native Georgian speakers. Drawing on Conceptual Metaphor Theory (CMT) and cognitive approaches to second language (L2) vocabulary acquisition, this exploratory classroom case study investigates how intermediate university learners ($N = 15$) comprehend, interpret, and produce L2 metaphorical expressions when supported by large language models (LLMs). Over a two-week action research intervention, participants engaged with a structured pedagogical framework alternating between autonomous AI consultation and human teacher mediation. Written production artifacts, interaction logs, and learner reflection journals were systematically evaluated using a dual-coded analytical rubric across four primary dimensions: frequency of use, contextual adequacy, conceptual complexity, and originality. The empirical findings reveal a distinct performance gradient dictated by cross-linguistic transparency; learners demonstrated high operational mastery of transparent expressions sharing conceptual structures with Georgian or English (M = 74.3%), but experienced sharp drops in accuracy when handling culturally unique, opaque idioms (M = 29.6%). Crucially, qualitative analysis demonstrates that while LLMs provide effective low-friction validation for transparent forms, their automated definitions fail to override literal image interference in opaque structures without explicit intercultural scaffolding. Furthermore, the study identifies a robust correlation between idiomatic opacity and the triggering of high-level metacognitive monitoring behaviors. Based on these insights, a three-phase instructional integration model is proposed to optimize technological scaffolding while safeguarding the irreplaceable role of human pedagogical mediation.
No takes yet. Share an insight, caveat, or question.
Natia et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: