What is the project about?
Everybody talks about "climate change", but only certain people in certain contexts talk about "climate hysteria", "climate religion", "climate deniers" and so on. Linguists call these word pairs "compounds". We are, therefore, interested in "climate compounds" that carry a pragmatic "load" in the sense that they are used by discourse participants to label other participants in a positive or (more frequently) in a derogatory way. Following the idea of the German klimadiskurs.info site, we undertake steps toward an English glossary that characterizes the usage contexts of those compounds.
What is your research question?
- Find textual data that can be attributed to “activist” or “scepticist” discourse
- Build machinery for analyzing those texts: which climate compounds occur how often, by which community, in what contexts
- Find a way to (semi-)automatically build glossary entries from the analysis results of a compound
What methods, data sets, and tools are used?
Data sets: some we gather ourselves (e.g. reddit posts); some we find already on the web (e.g., pre-classified tweets)
Methods: collocation analysis (what are the context words of compounds), sentiment/emotion analysis, sarcasm analysis, text classification (ideology), …
What are your preliminary findings and/or expected outcomes?
Expected outcomes are an initial data collection for running analyses, scripts for preprocessing texts and extracting information, evaluations of existing classifiers (e.g. emotions, sarcasm), and - optionally - ideas on crafting glossary entries in a semi-automatic way from the various analyses of a compound.
Who is part of the team?
Thomas Haider (Uni Passau), Yi-Sheng Hsu (Uni Potsdam), Ruangrin Idallitsakool (Uni Potsdam), Tom Krümmel (Uni Potsdam), Tim Repke (MCC), Manfred Stede (Uni Potsdam), Clara Wicharz (Uni Potsdam), Robin Winkle (Uni Potsdam), René Wolf (Uni Potsdam)