Elamur Research⌕
☰
home / papers / 2008 / A Discourse Resource for Turkish: Annotating Discourse Connectives in the METU Corpus

A Discourse Resource for Turkish: Annotating Discourse Connectives in the METU Corpus

Summary

The paper lays the groundwork for turning the METU Turkish Corpus (MTC), until now a sentence-level resource, into one annotated at the discourse level. It follows the principles of the Penn Discourse TreeBank (PDTB). The authors plan to mark explicit discourse connectives and their two arguments, ARG1 and ARG2. These arguments denote abstract objects such as events, facts or propositions.

Turkish grammar lacks the clean coordinating/subordinating split used for English. The authors therefore grouped candidate connectives by shared morpho-syntactic properties and arrived at three grammatical types divided into five classes. Using corpus examples, they describe for each class:

  • where the connective sits,
  • how its arguments are ordered,
  • which syntactic units can act as arguments.

They also discuss free word order, punctuation, and how much text an argument should span.

The paper reports no annotated data or agreement figures yet. Its contribution is the annotation scheme and a preliminary inventory of connectives. This is an early step toward a PDTB-style discourse resource for a morphologically rich, free word-order language. It also responds to MTC users who had asked for discourse-level annotation.

Key points

  • The MTC: about 2 million words in 520 text samples from 291 sources written between 1990 and 2000, spanning several genres. Its METU-Sabancı TreeBank subset has 5,600 sentences with morphological and dependency annotation. The paper reports 81 requests to use the MTC and 31 for the TreeBank.
  • Five connective classes:
    • simple coordinating conjunctions (e.g., ama, çünkü, ve, the particle dA),
    • paired conjunctions (hem..hem, ne..ne),
    • simplex subordinators, i.e., converb suffixes such as –(y)ken and –(y)ArAk,
    • complex subordinators (a postposition plus a nominalizing suffix, e.g., için),
    • anaphoric connectives (e.g., ne var ki, üstelik).
  • Two arguments only: every connective takes exactly two arguments. Unlike in the PDTB, non-finite clauses are treated as arguments, because Turkish marks them with suffixes.
  • Argument order: coordinating conjunctions usually take ARG1 before ARG2. The particle dA and both kinds of subordinators typically take ARG2 before ARG1.
  • Planned approach and open questions:
    • annotation will target text spans rather than syntactic trees,
    • implicit connectives will be added later, with punctuation as a cue,
    • the authors will test whether nominalizing suffixes can disambiguate connective senses, as they do for için.

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: Discourse analysis · Treebanks

Cited by

33 papers of our library cite it (of the 140673 whose bibliographies we read); 13 of them have a page here, 20 not yet.

Cite

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

@inproceedings{zeyrek2008discourse,
  title = {{A Discourse Resource for Turkish: Annotating Discourse Connectives in the METU Corpus}},
  author = {Zeyrek, Deniz and Webber, Bonnie},
  booktitle = {Proceedings of the 6th Workshop on Asian Language Resources},
  year = {2008},
  url = {https://aclanthology.org/I08-7009/}
}

References

10 references, 2 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Nikhil Dinesh, Alan Lee, Eleni Miltsakaki, Rashmi Prasad, Aravind Joshi and Bonnie Webber (2005). Attribution and the (Non-)Alignment of Syntactic and Discourse Arguments of Connectives. Proceedings of the ACL Workshop on Frontiers in Corpus Annotation II: Pie in the Sky. find
  2. Ann Arbor, Michigan. June 2005. find
  3. Katherine Forbes-Riley, Bonnie Webber and Aravind Joshi (2006). Computing Discourse Semantics: The Predicate-Argument Semantics of Discourse Connectives in D-LTAG. Journal of Semantics 23, pp. 55—106. find
  4. Aslı Göksel and Celia Kerslake (2005). Turkish: A Comprehensive Grammar. London and New York: Routledge. find
  5. Kornfilt, Jacklin (1997). Turkish. London and New York: Routledge. find
  6. PDTB-Group (2006). The Penn Discourse TreeBank 1.0 Annotation Manual. Technical Report IRCS 06-01, University of Pennsylvania. find
  7. Rashmi Prasad, Samar Husain, Dipti Sharma and Aravind Joshi (2008). Towards an Annotated Corpus of Discourse Relations in Hindi. The Third International Joint Conference on Natural Language Processing, January 7-12, 2008. find our card
  8. Bilge Say, Deniz Zeyrek, Kemal Oflazer and Umut Özge (2002). Development of a Corpus and a TreeBank for Present-day Written Turkish. Proceedings of the Eleventh International Conference of Turkish Linguistics, Eastern Mediterranean University, Cyprus, August 2002. find
  9. Bonnie Webber, Aravind Joshi, Matthew Stone and Alistair Knott (2003). Anaphora and Discourse Structure. Computational Linguistics 29 (4) 547-588. find our card
  10. Appendix: A preliminary list of explicit discourse connectives found in the MTC belonging to five syntactic classes and their English equivalents Simple coordinating conjunctions ama fakat çünkü dA halbuki oysa önce sonra ve veya ya da veyahut English equivalent but but because and, but despite desp… find

↑ back to top