Exploring sentiment across disciplines and argumentative moves: A sentiment analysis of open-access comments

Abstract

Sentiment analysis, a computational method originating from natural language processing, has recently gained interest in applied linguistics as a tool for examining evaluative language in academic discourse. This study applies sentiment analysis to analyzing open-access comments (OA comments), a novel academic genre designed to engage a broad readership across disciplines. Studying sentiment in these comments is crucial, as it reveals how scholars express not only factual information but also their emotion and attitudes towards the topics under discussion. The corpus includes 361 open-access comments published in Nature. The results reveal significant differences in sentiment scores across hard and soft science disciplines and in different argumentative moves. These findings highlight the potential of sentiment analysis as a promising method to explore open-assess comments as a unique academic genre, deepening our understanding of academic writing and informing academic writing pedagogy, particularly in emerging hybrid genres such as OA comments.