Elamur Research⌕
☰
home / papers / 2020 / Universal Dependency Treebank for Xibe

Universal Dependency Treebank for Xibe

Summary

The authors build the first syntactically annotated corpus for Xibe, a Tungusic language spoken in Xinjiang. The language has fewer than 40 thousand speakers, and UNESCO classifies it as severely endangered. Until now, Xibe had neither a corpus nor any computational tools. The team collected 810 sentences of written Xibe: examples from the grammar reference by Setuken (2009) and news texts from the newspaper Cabcal News. All sentences were annotated manually from scratch following the UD scheme: lemmas, parts of speech, morphological features and dependencies. In the CoNLL-U format, each sentence contains the original in Xibe script, a Latin transliteration and an English translation. A large part of the paper is devoted to language-specific decisions. These cover Chinese loanwords, which are written syllable by syllable as separate tokens, the attributive use of the case marker i, the topic marker oci, and relative and adverbial clauses. The paper contains no parsing experiments. The main contribution is the treebank itself and the documented annotation guidelines, which lay the groundwork for NLP and for documenting an endangered language.

Key points

  • Size: 810 sentences, 15,489 tokens. 544 sentences come from the grammar (5,773 tokens, 10.6 per sentence on average). 266 sentences come from 9 issues of Cabcal News (9,716 tokens, 36.5 on average, the longest sentence has 92 tokens).
  • Annotation was done in UD Annotatrix. The grammar part was annotated by two people: the first annotator handled 464 sentences, the second handled 80, which were then checked by the first. The news texts were annotated by a single annotator. The grammar examples were transliterated manually, the newspaper texts with a Python script.
  • Chinese loanwords written syllable by syllable are combined into an MWE using the flat relation. Bound syllables receive the tag X, while a free syllable receives the part of speech that matches the syntactic function of the whole word. Example: gung he cgo "republic".
  • Converbs (8 types of suffixes) are labeled advcl. Participial constructions with the nouns ba, erin, fon, turgun in the dative case receive obl. A postposition after a participle (for example, manggi "after") is tagged SCONJ with the mark relation.
  • Future plans include expanding the corpus with texts from Cabcal News and school textbooks, and exploring parsing approaches for this extremely low-resource language.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

Fields: Linguistics & languages

Topics: Treebanks

Cited by

5 papers of our library cite it (of the 140673 whose bibliographies we read); 1 of them have a page here, 4 not yet.

Cite

The paper itself, not this page. BibTeX from the library's record of the work.

@inproceedings{zhou2020universal,
  title = {{Universal Dependency Treebank for Xibe}},
  author = {Zhou, He and Chung, Juyeon and Kübler, Sandra and Tyers, Francis},
  booktitle = {Proceedings of the Fourth Workshop on Universal Dependencies (UDW 2020)},
  year = {2020},
  doi = {10.48448/nmnk-jw85},
  url = {https://doi.org/10.48448/nmnk-jw85}
}

References

19 references, 1 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Liliya M Gorelova. 2002. Manchu Grammar. Brill. find
  2. Songjie Gu. 2016. A literature review on Sibe language. Manchu Studies, 2(2):83–87. find
  3. Zengyi Hu. 1994. A Comprehensive Manchu-Chinese Dictionary. Xinjiang People’s Publishing House, Urumqi, Xinjiang, China. find
  4. Taeho Jang. 2008. Sibe Grammar. The Nationalities Publishing House of Yunnan, Kunming, Yunnan, China. find
  5. Herman Leung, Rafaël Poiret, Tak-sum Wong, Xinying Chen, Kim Gerdes, and John Lee. 2016. Developing universal dependencies for Mandarin Chinese. In Proceedings of the 12th Workshop on Asian Language Resources, pages 20–29, Osaka, Japan. find
  6. Shulan Li. 1979. A survey on the Sibe language. Minority Languages of China, 6(3):221–232. find
  7. Shulan Li. 1982. Possession category in Sibe. Minority Languages of China, 6(5):50–57. find
  8. Shulan Li. 1985. Adverbials in Sibe. Minority Languages of China, 6(5):12–25. find
  9. Shulan Li. 1988. Auxilaries in sibe. Minority Languages of China, 6(6):27–32. find
  10. Ryan McDonald, Joakim Nivre, Yvonne Quirmbach-Brundage, Yoav Goldberg, Dipanjan Das, Kuzman Ganchev, Keith Hall, Slav Petrov, Hao Zhang, Oscar Täckström, Claudia Bedini, Núria Bertomeu Castelló, and Jungmee Lee. 2013. Universal dependency annotation for multilingual parsing. In Proceedings of the 51… find
  11. Joakim Nivre, Marie-Catherine De Marneffe, Filip Ginter, Yoav Goldberg, Jan Hajic, Christopher D Manning, Ryan McDonald, Slav Petrov, Sampo Pyysalo, Natalia Silveira, et al. 2016. Universal dependencies v1: A multilingual treebank collection. In Proceedings of the Tenth International Conference on L… find our card
  12. Jerry Norman. 1974. A sketch of Sibe morphology. Central Asiatic Journal, 18(3):159–174. find
  13. Setuken. 2009. General Introduction to Xibe Grammar. Xinjiang People’s Publishing House, Urumqi, Xinjiang, China. find
  14. Francis Tyers, Mariya Sheyanova, and Jonathan Washington. 2017. UD Annotatrix: An annotation tool for universal dependencies. In Proceedings of the 16th International Workshop on Treebanks and Linguistic Theories, pages 10–17. find
  15. Tak-sum Wong, Kim Gerdes, Herman Leung, and John Lee. 2017. Quantitative comparative syntax on the Cantonese-Mandarin Parallel Dependency Treebank. In Proceedings of the Fourth International Conference on Dependency Linguistics, pages 266−–275, Pisa, Italy. find
  16. Xinjiang Ethnic Language Work Committee. 1992. nei fon sibe šu tacin gisun i arara kooli (Modern Literary Xibe Orthography). Xinjiang People’s Publishing House, Urumqi, Xinjiang, China. find
  17. Koichi Yasuoka. 2019. Universal dependencies treebank of the four books in Classical Chinese. In DADH2019: 10th International Conference of Digital Archives and Digital Humanities, pages 20–28, Taipei, Taiwan. find
  18. Veronika Zikmundová. 2013. Spoken Sibe: Morphology of the Inflected Parts of Speech. Karolinum Press. find
  19. A Tagset, Relations and Features This treebank uses 17 Universal POS Tags, 30 Universal Dependency relations, relation subtypes, and 20 features. A.1 Universal POS Tags ADJ PART A.2 A.2.1 AUX PUNCT CCONJ SCONJ DET SYM INTJ VERB NOUN X NUM Universal Dependency Relations advcl compound mark vocative a… find

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/papers/zhou-2020-universal-dependency-treebank-xibe/