Constructing a Gold Corpus of Annotated Youtube Comments for Discursive Strategies Span Classification

Abstract

Our research project is focused on the creation of a corpus of German YouTube comments relating to the topic of gender diversity. It aims at the identification of dominant arguments for and against gender diversity and the multiple discursive strategies, such as victim-blaming, denial, proposal, etc., used to convey them. This is crucial for understanding the pro and contra communication as well as developing linguistic strategies to refute misconceptions and fake information around the topic. This paper discusses the theoretical background, data collection, corpus creation, and the annotation process as a way of reconstructing discursive strategies identified by works in critical discourse analysis and bringing to light new ones.