The analysis of argumentation topoi

Abstract This paper examines the role of argumentation topoi in discourse analysis, focusing on their identification, annotation, and automated classification. While topos analysis is a well-established method for uncovering collective patterns of reasoning in public discourse, its application in large-scale digital discourse analysis poses methodological challenges. This study investigates the so-called utility topos in bioethical debates, which encompasses both medical and economic arguments about benefit. We collaboratively developed annotation guidelines to ensure consistency and measured inter-annotator agreement. We then trained a BERT-based model on the gold-standard annotations and used it to classify utility topoi in German texts. The results indicate that while explicit argument structures are detectable, implicit and context-dependent reasoning remains difficult to capture. These findings highlight the need for refined annotation guidelines and contextual modelling in automated discourse analysis. The study contributes to the integration of qualitative and computational methods in argument analysis.