Elamur Research⌕
☰
home / papers / 2015 / Recovering discourse relations: Varying influence of discourse adverbials

Recovering discourse relations: Varying influence of discourse adverbials

Summary

Discourse relations may be marked by a connective or left for the reader to infer. This paper asks whether a reader can perceive an additional, unmarked relation when only one explicit discourse adverbial is present. It also asks whether that relation follows from the adverbial or from the passage content. The authors report Phase I of a crowd-sourcing study that builds on an earlier MSc project by Jiang (2013). The study examines four adverbials, namely instead, in general, in fact and after all. Participants saw passages with a gap before the adverbial and chose one of six conjunctions to fill it, or selected None at all or Other. Some passages originally contained a conjunction (explicit), and others had none (implicit). Participants seldom declined to insert a conjunction, and they agreed fairly consistently on their choices. After all strongly attracted because. The other adverbials yielded passage-dependent choices, so content rather than the adverbial drives the inference. This matters for discourse parsing in Language Technology, which may need to look for implicit relations even when an explicit connective is present.

Key points

  • The response options were and, because, before, but, or, so, plus None at all and Other. A separate rating let participants flag stylistic awkwardness, which removed the ambiguity of "None" in Jiang's design. Each participant judged 50 passages.
  • The six conjunctions are largely unambiguous at a coarse level: 86.7–99.6% of their tokens in the Penn Discourse TreeBank carry their dominant sense label.
  • In implicit passages, None was rare: 26/348 for after all, 20/231 for in fact, 13/290 for in general and 10/290 for instead. After all drew because in 245 of 348 responses.
  • Instead did not mainly attract the contrastive but (103 responses); so was chosen more often (119).
  • Mean agreement on the top response was 0.706 for after all, 0.581 for in fact, 0.503 for in general and 0.717 for instead. A Google NGRAMS survey guided adverbial selection, and the study ran from August 2014 to June 2015.

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

Cited by

10 papers of our library cite it (of the 140673 whose bibliographies we read); 4 of them have a page here, 6 not yet.

Cite

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/w15-2703

@inproceedings{rohde2015recovering,
  title = {{Recovering discourse relations: Varying influence of discourse adverbials}},
  author = {Rohde, Hannah and Dickinson, Anna and Clark, Chris and Louis, Annie and Webber, Bonnie},
  booktitle = {Proceedings of the First Workshop on Linking Computational Models of Lexical, Sentential and Discourse-level Semantics},
  year = {2015},
  doi = {10.18653/v1/w15-2703},
  url = {https://doi.org/10.18653/v1/w15-2703}
}

References

15 references, 3 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. M.J. Cuenca and M.J. Marin. 2009. Co-occurrence of discourse markers in Catalan and Spanish oral narrative. Journal of Pragmatics, 41(5):899–914. find
  2. Bruce Fraser. 2013. Combinations of contrastive discourse markers in english. International Review of Pragmatics, 5:318–340. find
  3. G. Gazdar. 1979. Pragmatics: Implicature, Presupposition, and Logical Form. Academic Press, London. find
  4. Xi Jiang. 2013. Predicting the use and interpretation of implicit and explicit discourse connectives. Master’s thesis, Linguistics and English Language, University of Edinburgh. MSc in English Language. find
  5. Andrew Kehler. 2002. Coherence, Reference, and the Theory of Grammar. CSLI Publications, Stanford. find
  6. Alistair Knott. 1996. A Data-driven Methodology for Motivating a Set of Coherence Relations. Ph.D. thesis, Department of Artificial Intelligence, University of Edinburgh. find
  7. William C. Mann and Sandra A. Thompson. 1988. Rhetorical Structure Theory: Toward a functional theory of text organization. Text, 8(3):243–281. find our card
  8. Sarah Oates. 2000. Multiple discourse marker occurrence: Creating hierarchies for natural language generation. In Proceedings, ANLP-NAACL 2000 Student Research Workshop, pages 41–45, Seattle WA. find
  9. Rebecca Passonneau and Bob Carpenter. 2014. The benefits of a model of annotation. Transactions of the Association of Computational Linguistics, 2(1):311–326. find
  10. Rashmi Prasad, Nikhil Dinesh, Alan Lee, Eleni Miltsakaki, Livio Robaldo, Aravind Joshi, and Bonnie Webber. 2008. The Penn Discourse TreeBank 2.0. In Proceedings, 6th International Conference on Language Resources and Evaluation, pages 2961– 2968, Marrakech, Morocco. find our card
  11. Rashmi Prasad, Bonnie Webber, and Aravind Joshi. 2014. Reflections on the penn discourse treebank, comparable corpora and complementary annotation. Computational Linguistics, 40(4):921–950. find our card
  12. Evan Sandhaus. 2008. New York Times corpus: Corpus overview. LDC catalogue entry LDC2008T19. find
  13. The PDTB Research Group. 2008. The Penn Discourse TreeBank 2.0 Annotation Manual. Available at http://www.seas.upenn.edu/˜pdtb/, or as part of the download of LDC2008T05. ↗
  14. Bonnie Webber, Alistair Knott, and Aravind Joshi. 1999. Multiple discourse connectives in a lexicalized grammar for discourse. In Third International Workshop on Computational Semantics, pages 309– 325, Tilburg, The Netherlands. find
  15. Deniz Zeyrek. 2014. On the distribution of contrastive-concessive discourse connectives ama (but/yet) and fakat (but) in written Turkish. In P. Suihkonen and L.J. Whaley, editors, On Diversity and Complexity of Languages Spoken in Europe and North and Central Asia. find

↑ back to top

This page in Markdown →