Is Japanese CCGBank empirically correct? A case study of passive and causative constructions

Daisuke Bekki, Hitomi Yanaka


Abstract
The Japanese CCGBank serves as training and evaluation data for developing Japanese CCG parsers. However, since it is automatically generated from the Kyoto Corpus, a dependency treebank, its linguistic validity still needs to be sufficiently verified. In this paper, we focus on the analysis of passive/causative constructions in the Japanese CCGBank and show that, together with the compositional semantics of ccg2lambda, a semantic parsing system, it yields empirically wrong predictions for the nested construction of passives and causatives.
Anthology ID:
2023.tlt-1.4
Volume:
Proceedings of the 21st International Workshop on Treebanks and Linguistic Theories (TLT, GURT/SyntaxFest 2023)
Month:
March
Year:
2023
Address:
Washington, D.C.
Editors:
Daniel Dakota, Kilian Evang, Sandra Kübler, Lori Levin
Venues:
TLT | SyntaxFest
SIG:
SIGPARSE
Publisher:
Association for Computational Linguistics
Note:
Pages:
32–36
Language:
URL:
https://aclanthology.org/2023.tlt-1.4
DOI:
Bibkey:
Cite (ACL):
Daisuke Bekki and Hitomi Yanaka. 2023. Is Japanese CCGBank empirically correct? A case study of passive and causative constructions. In Proceedings of the 21st International Workshop on Treebanks and Linguistic Theories (TLT, GURT/SyntaxFest 2023), pages 32–36, Washington, D.C.. Association for Computational Linguistics.
Cite (Informal):
Is Japanese CCGBank empirically correct? A case study of passive and causative constructions (Bekki & Yanaka, TLT-SyntaxFest 2023)
Copy Citation:
PDF:
https://aclanthology.org/2023.tlt-1.4.pdf