Elamur Research⌕
☰
home / papers / 1998 / A surface-based approach to identifying discourse markers and elementary textual units in unrestricted texts

A surface-based approach to identifying discourse markers and elementary textual units in unrestricted texts

Summary

The paper addresses how to split unrestricted text into elementary, clause-like units and how to detect which cue phrases actually serve a discourse function. Such fine-grained segmentation matters for RST-style discourse parsing, anaphora resolution and summarization, where whole sentences are often too coarse. Marcu treats connectives as procedural instructions for building a text representation. He began with a manual study of more than 450 candidate cue phrases, using fragments sampled from the Brown corpus. For each cue he recorded its punctuation context, its usage type (sentential, discourse or pragmatic), its position, and an associated segmentation "action". From this study he hand-built regular expressions and a set of eleven actions. A shallow analyzer applies them left to right, using no tagging or parsing. On unseen texts labelled by human judges, it reached roughly 81% recall and 90% precision for both markers and unit boundaries. The method runs in linear time. Its main weakness is that surface cues cannot always separate sentential from discourse uses.

Key points

  • Corpus study: more than 7600 Brown corpus fragments were sampled, averaging about 17 per cue phrase and reaching 60 for highly ambiguous cues such as and. The algorithm itself drew on the analysis of 2200 of these fragments.
  • Actions: the eleven actions are NOTHING, NORMAL, COMMA, NORMAL_THEN_COMMA, END, MATCH_PAREN, COMMA_PAREN, MATCH_DASH, SET_AND, SET_OR and DUAL. Parenthetical spans are delimited and attached to their host unit rather than treated as separate elementary units.
  • Evaluation data: a C++ implementation was tested on three texts not used in development: a Scientific American text (5036 words), a Time article (1588 words) and a Brown Corpus narrative (583 words). Three judges segmented them, and a boundary counted as valid when at least two of them agreed.
  • Results: discourse markers were found with 80.8% recall and 89.5% precision, beating Hirschberg and Litman (1993) and its follow-ups. Clause-like boundaries reached 81.3% recall and 90.3% precision.
  • Limitations and future work: errors arise where only syntax can decide, for example a sentential but. Future work would add part-of-speech tagging and syntactic parsing.

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

5 papers of our library cite it (of the 140673 whose bibliographies we read); 2 of them have a page here, 3 not yet.

Cite

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

@inproceedings{marcu1998surface,
  title = {{A surface-based approach to identifying discourse markers and elementary textual units in unrestricted texts}},
  author = {Marcu, Daniel},
  booktitle = {Discourse Relations and Discourse Markers},
  year = {1998},
  url = {https://aclanthology.org/W98-0301/}
}

References

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

  1. Jean Caron. 1997. Toward a procedural approach of the meaning of connectives. In Jean Costermans and Michel Fayol, editors, Processing Interclausal Relationships. Studies in the Production and Comprehension of Text, pages 53-74. find
  2. Dan Cristea and Bonnie L. Webber. 1997. Expectations in incremental discourse processing. In Proceedings of the 35th Annual Meeting of the Association for Computational Linguistics (ACL/EACL-97), pages 88-95, Madrid, Spain, July 7-12. find
  3. Bruce Fraser. 1996. Pragmatic markers. Pragmatics, 6(2): 167-190. find
  4. Barbara Grosz and Julia Hirschberg. 1992. Some intonational characteristics of discourse structure. In Proceedings of the International Conference on Spoken Language Processing. find
  5. Barbara J. Grosz and Candace L. Sidner. 1986. Attention, intentions, and the structure of discourse. Computational Linguistics, 12(3): 175-204, July-September. find our card
  6. Marti A. Hearst. 1997. TextTiling: Segmenting text into multi-paragraph subtopic passages. Computational Linguistics, 23(1 ):33--64, March. find
  7. Julia Hirschberg and Diane Litman. 1993. Empirical studies on the disambiguation of cue phrases. Computational Linguistics, 19(3):501-530. find
  8. Diane J. Litman. 1996. Cue phrase classification using machine learning. Journal of Artificial Intelligence Research, 5:53-94. find
  9. 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
  10. Daniel Marcu. 1997a. From discourse structures to text summaries. In Proceedings of the ACL'97/EACL'97 Workshop on Intelligent Scalable Text Summarization, pages 82-88, Madrid, Spain, July 11. find
  11. Daniel Marcu. 1997b. The rhetorical parsing, summarization, and generation of natural language texts. Ph.D. thesis, Department of Computer Science, University of Toronto, December. find
  12. Leo G.M. Noordman and Wietske Vonk. 1997. Toward a procedural approach of the meaning of connectives. In Jean Costermans and Michel Fayol, editors, Processing Interclausal Relationships. Studies in the Production and Comprehension of Text, pages 75-93. Lawrence Erlbaum Associates. find
  13. Rebbeca J. Passonneau and Diane J. Litman. 1997. Discourse segmentation by human and automated means. Computational Linguistics, 23(1):103-140, March. find
  14. Eric V. Siegel and Kathleen R. McKeown. 1994. Emergent linguistic rules from inducing decision trees: Disambiguating discourse clue words. In Proceedings of the Twelfth National Conference on Artificial Intelligence (AAAI-94), volume 1, pages 820--826, Seattle, July 31 - August 4. find
  15. Simone Teufel and Marc Moens. 1998. Sentence extraction and rhetorical classification for flexible abstracts. In Working Notes of the AAA1-98 Spring Symposium on Intelligent Text Summarization, Stanford, March 23-25. find
  16. Teun A. van Dijk. 1979. Pragmatic connectives. Journal of Pragmatics, 3:447--456. find
  17. Yaakov Yaari. 1997. Segmentation of expository texts by hierarchical agglomerative clustering. In Proceedings of the Conference on Recent Advances in Natural Language Processing (RANLP-97), Bulgaria. find

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/marcu-1998-surface-based-approach-identifying-discourse/