Wine writing meets MIP and the process of vetting Generative AI-based research tools: Expert vs. non-expert figurative language in standard English and English as a lingua franca
Keywords
Loading...
Authors
Issue Date
2025-10-28
Language
en
Document type
Journal Title
Journal ISSN
Volume Title
Publisher
Title
ISSN
Volume
Issue
Startpage
Endpage
DOI
Abstract
Tropes in wine discourse have increasingly attracted scholarly attention, yet systematic analyses of figurative strategies across reviewer groups remain limited. This thesis examines how wine reviewers use literal and figurative resources; it compares reviews produced by L1 experts and L2 non-experts on two major online platforms to investigate cross-linguistic and cross-platform variation. Overlap was observed but distinct conceptual orientations emerged between groups: lexical metaphors produced by experts frequently drew on domains such as social interaction, whereas non-experts drew on e.g. quantity. Besides other analyses, the study assessed the potential of adopting LLMs to conduct large-scale future metaphor research with sufficient linguistic rigor. More specifically, five LLMs, including ChatGPT-05, were tested through prompt engineering. Although they may offer potential scalability in the future, only moderate reliability was achieved for ChatGPT-05, indicating that they currently fall short of capturing the culturally embedded and cognitively grounded nature of metaphor, locally hosted LLMs especially.
Description
Citation
Supervisor
Faculty
Faculteit der Letteren