This corpus repo contains annotations of argumentative discourse units for 300 German tweets from 2019 on the climate change debate. Tweets are grouped in pairs of source and reply tweet, i.e. the latter is a direct reply to the former. All annotations refer to the reply tweet; the source tweet has only been used as context. The tweet pairs have been randomly collected from a larger set of tweets.
For copyright reasons we cannot distribute the actual tweet content. Instead with share the tweet IDs and the annotations. Currently only the global i.e. full tweet annotations are shared.
If not declared otherwise expert annotations are mediated as described in Iskender Et Al 2021.
Please cite the following papers when you use the corpus for your own work:
Schaefer & Stede (2020). Annotation and Detection of Arguments in Tweets. In Proceedings of the 7th Workshop on Argument Mining (at COLING 2020).
Iskender, Schaefer, Polzehl and Möller. Argument Mining in Tweets: Comparing Crowd and Expert Annotations for Automated Claim and Evidence Detection. In Proceedings of the 26th International Conference on Natural Language & Information Systems (NLDB 2021).