Turkish Proposition Bank is the first semantically annotated corpus built for Turkish language. It is constructed upon IMST and later alinged with IMST Universal Dependencies (UD). It contains sense and semantic argument information of each predicate similar to
giy.01 means the “1st” sense of predicate lemma giy (wear) and labels A-A, A0 and A1 represents semantic arguments. Numbered arguments (A0-A4) of each predicate sense are registered in frame files and definitions of adjunct like arguments can be found in original English PropBank project.
Frame files can be download from this link.
Corpora constructed upon IMST and IMST-UD can be investigated from project’s corpus repo. Please check the LICENSE and README files before usage.
Annotated corpora are distributed from İTÜ NLP Group - Datasets page. The original version described in  is distributed under the tag Original. We also distribute the version compatible with the latest treebank release under the Latest tag. The same applies to Universal Dependency (UD) compatible annotations.
A part of this corpus has been annotated via crowd workers. To guide researchers we share our design files, screenshots, instructions and task results (aggregation files)[Verb Sense Annotation], [Semantic Role Annotation].
Turkish Semantic Role Labeling
We distribute resources built in scope of this project under Creative Commons BY-NC-SA 4.0 International license.
 Gözde Gül Şahin and Eşref Adalı. 2017. Annotation of semantic roles for the Turkish Proposition Bank. Language Resources and Evaluation (in press).
 Gözde Gül Şahin. 2016. Verb Sense Annotation for Turkish PropBank via Crowdsourcing. In Proceedings of 17th International Conference on Computational Linguistics and Intelligent Text Processing (CICLing 2016), Konya, Turkey, April 3-9, 2016, Revised Selected Papers, Part I. pages 496–506.
 Gözde Gül Şahin. 2016. Framing of Verbs for Turkish PropBank. In Proceedings of 1st International Conference on Turkic Computational Linguistics, TurCLing, Konya, Turkey, April 9, 2016
 Gözde Gül Şahin and Eşref Adalı. 2014. Using morphosemantic information in construction of a pilot lexical semantic resource for Turkish. In Proceedings of Workshop on Lexical and Grammatical resources for Language Processing, Dublin, Ireland.