Argument Corpora

All of our corpora are now available through the AIFdb Corpus Management tools, available online at corpora.aifdb.org

We have compiled various different corpora of argument:

• AIFdb2013 – A 12,000 node corpus (where a node is roughly a proposition or an inference) of analysed argumentation, incorporating all the analyses of the AraucariaDB corpus plus many more arguments from various sources including users argublogging, and working with Arvina, plus other datasets such as Archelogos.
• MM2012 – A 65,000 word corpus of unanalysed dialogue transcripts from the BBC Radio 4 Moral Maze programme.
• AraucariaDBpl – A corpus of analysed arguments in Polish
• AraucariaDB2005 – A corpus of over 650 analysed arguments described in Reed, C. (2006) “Preliminary Results from an Argument Corpus” in Eloína Miyares Bermúdez and Leonel Ruiz Miyares (eds) Linguistics in the Twenty First Century, Cambridge Scholars Press, (ISBN 9781904303862), pp185-196.
• Mediation discourse – A corpus of dispute mediation transcripts

These corpora are free for academic use (for non-academic use, please contact Prof. Chris Reed). If you find them useful, please cite either (Reed, 2006) for the AraucariaDB corpus, or for the others (until we publish something formally), the following:

Lawrence, J. & Reed, C. (2014) “AIFdb Corpora” in Parsons, S., Oren, N., Reed, C. & Cerutti, F. (eds) Proceedings of the Fifth International Conference on Computational Models of Argument (COMMA 2014), IOS Press, Pitlochry, pp465-466. Available online.