Discourse Analysis of Pairwise Twitter Hashtags
Event Type
DescriptionViral hashtags frequently emerge on Twitter to discuss political topics and bring attention to acts of injustice. Response hashtags are often created subsequently to express divergent stances (e.g. #AllLivesMatter in response to #BlackLivesMatter). Understanding the nature of online disagreement can provide insight into how individuals debate and navigate sensitive topics. We evaluate the potential for classification between hashtag pairs by characterizing discourse surrounding opposing hashtags using Linguistic Inquiry and Word Count (LIWC), Bidirectional Encoder Representations (BERT) word embeddings, and measures of entropy and divergence. Analyses found uniformity in topic, diction, and syntax, but observed distinctions in sentiment and psychological processes.