MFD-2-TM-de: German Moral Foundation Dictionary 2 by Translation Matrix

This dataset was created for the research paper: "Reiter-Haas, M., Kopeinik, S., & Lex, E. (2021, May). Studying moral-based differences in the framing of political tweets. In Proceedings of the International AAAI Conference on Web and Social Media (Vol. 15, pp. 1085-1089).". Available at https://ojs.aaai.org/index.php/ICWSM/article/view/18135/17938

The translation were created using a translation matrix of positive and negative valence words. The dictionary was used to measure effects in the embedding space for which it is suitable. However, it is to note that the dictionary does not work in a classical sense, as it contains duplicates (e.g., words of the opposite poles). While this was also the case in the original dictionary (e.g., for "wound"), the effect has become much more pronounced in the translated version.

Format:

Lines starting with % denote a new section that the following words belong to (e.g., care.vice for harm).

Otherwise, one word per line.

For implementation details, refer to the GitHub Repo