Elamur Research⌕
☰
home / papers / 2004 / Discourse-level Annotation for Investigating Information Structure

Discourse-level Annotation for Investigating Information Structure

Paper · 2004 · Ivana Kruijff-Korbayova and Geert-Jan M. Kruijff

Summary

The paper tackles a practical obstacle to studying information structure (IS): the field has many competing, loosely defined dichotomies such as Topic-Focus, Theme-Rheme and Given-New, so existing annotations are hard to compare or reuse. Within the MULI (MUltilingual Information structure) pilot project, the authors avoid labelling any theory-specific IS category. Instead, they enrich German and English financial-news extracts from the Tiger treebank and the Penn treebank with descriptive, theory-neutral features on three levels:

  • Syntax: marked constructions and voice.
  • Prosody: ToBI/GToBI labels on read-aloud recordings.
  • Discourse: properties of discourse referents, plus coreference and bridging links.

The paper concentrates on the discourse scheme. It covers semantic type and sort, denotation characteristics, quantification and familiarity of entities evoked mainly by nominal-like expressions. The authors also describe how stand-off layers produced by different tools are aligned. The main contribution is the scheme and resource design rather than an experimental result. Because the annotation stays descriptive, researchers can test the predictions of competing IS theories against the same data and combine it with other neutral resources.

Key points

  • Corpus size: The MULI corpus has 250 German sentences (about 3,500 tokens) drawn from 21 consecutive Tiger treebank texts. It also has 320 English sentences (about 7,000 tokens) from 10 Penn treebank texts, all in the financial news genre.
  • Syntax and prosody layers:
    • The syntax layer marks clefts, pseudo-clefts, reversed pseudo-clefts, extraposition, fronting, expletives and voice (active, mediopassive, passive).
    • The prosody layer was recorded with one native speaker per language and annotated in EMU using ToBI and GToBI conventions.
    • All German texts were annotated prosodically, but only some English examples.
  • Discourse referents:
    • Referents are typed as object, property, eventuality or textual entity.
    • Objects get finer sorts, such as person, organization, location, quantity and date/time.
    • Denotational uses (unique, specific, or unspecific/generic) are separated from non-denotational, attributive uses such as predicative NPs.
  • Anaphoric links: Coreference and bridging links build on MUC-7, DRAMA and related classifications. The relation type is stored on the anaphor's markable.
  • Tools and formats:
    • Prosody is annotated in EMU, syntax in XML-Spy with a DTD, and discourse in MMAX.
    • Every layer is a separate stand-off XML file pointing to a shared base token file, with the word as the common unit.
    • Punctuation and pauses are stored as attributes of the preceding word.
    • Some cases remain unresolved: compounds carrying several accents, and spoken number forms that differ from the written ones.

Interesting ideas

  • I1. Annotate observable cues rather than theoretical labels, and compute any theory's categories later from those cues; this keeps the data reusable when theories change.
  • I2. Distinguish expressions that pick out instances from expressions that merely attribute a property, and treat generic readings as reference to types.
  • I3. Keep each annotation level in its own stand-off file keyed to a shared token base, even when the levels could be merged into one hierarchy.
  • I4. When aligning modalities whose units differ, attach elements that one side lacks to the neighbouring word as attributes. Examples are punctuation, which is absent in speech, and pauses, which are absent in writing.
  • I5. Record the type of an anaphoric relation, identity or bridging, on the anaphor itself, so each mention carries its own explanation of how it connects back.

For Mova

How these ideas could serve Mova, our small open language model — its files, stages and tests:

  • I1 (frames): Store raw cues on frame slots, such as definiteness, voice, marked word order and familiarity. Givenness or topic can then be derived by an explicit rule layer that can be inspected and swapped.
  • I2 (world): A predicative NP should add a property to an existing entity rather than create a new one. Generic NPs should link to ontology types instead of individual entities.
  • I3 (learn): Silver outputs from each witness parser and teacher tier can be kept as separate layers over the same token IDs. Disagreements can then be compared and traced without destructive merging.
  • I4 (back): In the text → graph → text compressor, punctuation and spacing could be stored as attributes of the preceding token. Regenerated text could then be compared word-aligned with the original.
  • I5 (frames): The coreference step can tag each mention with its link kind. The world builder can then attach bridged entities, such as parts or roles of an earlier entity, instead of merging them.

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

 Institutions

🎓Saarland UniversitySaarbrücken · Germany148 works · 18,287 citations

 Fields, topics and concepts

📚 Shelves 4The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 worksArts and Humanities · 16,039 worksLanguage and Linguistics · 15,891 works

🧭 Topics 3Research subjects and languages the work is about.English · 2,263 worksCoreference resolution · 574 worksDiscourse analysis · 466 works

Cite

Cite as Ivana Kruijff-Korbayova and Geert-Jan M. Kruijff. 2004. Discourse-level Annotation for Investigating Information Structure. Proceedings of the Workshop on Discourse Annotation.

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

@inproceedings{kruijffkorbayova2004discourse,
  title = {{Discourse-level Annotation for Investigating Information Structure}},
  author = {Ivana Kruijff-Korbayova and Geert-Jan M. Kruijff},
  booktitle = {Proceedings of the Workshop on Discourse Annotation},
  year = {2004},
  doi = {10.3115/1608938.1608944},
  url = {https://doi.org/10.3115/1608938.1608944}
}

References

36 references, 4 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. S. Bauman, C. Brinckmann, S. Hansen-Schirra, G.-J. Kruijff, I. Kruijff-Korbayová, S. Neumann, and E. Teich. 2004a. Multi-dimensional annotation of linguistic corpora for investigating information structure. In Proc. of the Workshop on Frontiers in Corpus Annotation, held at the NAACL-HLT 2004 Conference. find
  2. S. Bauman, C. Brinckmann, S. Hansen-Schirra, G.-J. Kruijff, I. Kruijff-Korbayová, S. Neumann, E. Teich, E. Steiner, and H. Uszkoreit. 2004b. The muli project: Annotation and analysis of information structure in German and English. In Proc. of the LREC 2004 Conference.
  3. M. E. Beckmann and J. Hirschberg. 1994. The ToBI annotation conventions. Ms. and accompanying speech materials, Ohio State University. find
  4. D. Biber, S. Johansson, G. Leech, S. Conrad, and E. Finegan. 1999. The Longman Grammar of Spoken and Written English. Longman, Harlow. find
  5. S. Bird and M. Liberman. 2001. A formal framework for linguistic annotation. Speech Communication, 33(1-2):23–60. S. Brants, S. Dipper, P. Eisenberg, S. Hansen, E. König, W. Lezius, C. Rohrer, G. Smith, and H. Uszkoreit. to appear. TIGER: Linguistic interpretation of a German corpus. Journal of Language and Computation (JLAC), Special Issue. ↗
  6. E. Buráňová, E. Hajičová, and P. Sgall. 2000. Tagging of very large corpora: Topic-focus articulation. In Proc. of the 18th Conference on Computational Linguistics (COLING’2000), July 31 - August 4 2000, pages 139– 144. Universität des Saarlandes, Saarbrücken, Germany. find
  7. J. Carletta, N. Dahlbäck, N. Reithinger, and M. A. Walker. 1997. Standards for dialogue coding in natural language processing. Report on the dagstuhl seminar, Discourse Resource Initiative, February 3–7. find
  8. S. Cassidy and J. Harrington. 2001a. Multi-level annotation in the emu speech database management system. Speech Communication, 33(1-2):61–78. ↗
  9. S. Cassidy and J. Harrington. 2001b. Multi-level annotation in the EMU speech database management system. Speech Communication, 33(1-2):61–78. ↗
  10. P. Eisenberg. 1994. Grundriss der deutschen Grammatik, 3. Aufl. Metzler, Stuttgart, Weimar. find
  11. K. Erk, A. Kowalski, S. Padó, and M. Pinkal. 2003. Towards a resource for lexical semantics: A large german corpus with extensive semantic annotation. In Proc. of ACL 2003, Sapporo, Japan. M. Grice, S. Baumann, and R. Benzmüller. in press. German intonation in autosegmental-metrical phonology. In Sun-Ah Jun, editor, Prosodic Typology: Through Intonational Phonology and Transcription. OUP. find
  12. J. Gundel, N. Hedberg, and R. Zacharski. 1993. Cognitive status and the form of referring expressions in discourse. Language, (69):274–307. ↗
  13. M. A.K. Halliday. 1985. Introduction to Functional Grammar. Edward Arnold, London, U.K. find
  14. Z. Hlavsa. 1975. Denotace objektu a jejı́ prostředky v současné češtině [Denotating of objects and its means in contemporary Czech], volume 10 of Studie a práce lingvistické [Linguistic studies and works]. Academia. find
  15. N. Ide, P. Bonhomme, and L. Romary. 2000. Xces: An xml-based standard for linguistic corpora. pages 825– 830, Athens, Greece. find
  16. H. Kamp and U. Reyle. 1993. From discourse to logic. Kluwer Academic Publishers, Dordrecht, the Netherlands. ↗
  17. Geert-Jan M. Kruijff 2001. A Categorial-Modal Logical Architecture of Informativity: Dependency Grammar Logic & Information Structure, Faculty of Mathematics and Physics, Charles University. Prague, Czech Republic. find
  18. I. Kruijff-Korbayová and M. Steedman. 2003. Discourse and information structure. Journal of Logic, Language and Information: Special Issue on Discourse and Information Structure, 12(3):249–259. ↗
  19. M. Marcus, G. Kim, M. Ann Marcinkiewicz, R. MacIntyre, A. Bies, M. Ferguson, K. Katz, and B. Schasberger˙ 1994. The Penn treebank: Annotating predicate argument structure. In Proc. of the Human Language Technology Workshop, San Francisco, Morgan Kaufmann.
  20. D. McKelvie, A. Isard, A. Mengel, M.B. Moller, M. Grosse, and M. Klein. 2001. The MATE workbench — an annotation tool for XML coded speech corpora. Speech Communication, 33(1-2):97–112. ↗
  21. C. Müller and M. Strube. 2001. Annotating anaphoric and bridging relations with MMAX. In Proc. of the 2nd SIGdial Workshop on Discourse and Dialogue, pages 90–95, Aalborg, Denmark, 1–2 September. http://www.eml.villa-bosch.de/english/Research/NLP/sigdial ↗
  22. C. Müller and M. Strube. 2003. Multi-level annotation in mmax. In Proc. of the 4th SIGdial Workshop on Discourse and Dialogue, Sapporo, Japan, 4-5 July. http://www.eml.villa-bosch.de/english/Research/NLP/Publications. ↗
  23. R. Passoneau. 1996. Instructions for applying discourse reference annotation for multiple applications (DRAMA). draft, December 20. find
  24. J. Pierrehumbert and J. Hirschberg. 1990. The meaning of intonational contours in the interpretation of discourse. In P.R. Cohen, J. Morgan, and M.E. Pollack, editors, Intentions in Communication, pages 271–311. MIT press. ↗
  25. Massimo Poesio, Renate Henschel, Janet Hitzeman, Rodger Kibble, Shane Montague, and Kees van Deemter 1999. Towards An Annotation Scheme For Noun Phrase Generation In Proc. of the EACL Workshop on Linguistically Interpreted Corpora. Bergen, Norway. find
  26. Massimo Poesio 2000. The GNOME Annotation Scheme Manual Available online http://www.hcrc.ed.ac.uk/~gnome/anno manual.html E. Prince. 1981. Toward a taxonomy of given-new information. In P. Cole, editor, Radical Pragmatics, pages 223–256. Academic Press. ↗
  27. R. Quirk, S. Greenbaum, G. Leech, and J. Svartik. 1985. A comprehensive grammar of the English language. Longman, London. find
  28. P. Sgall, E. Hajičová, and J. Panevová. 1986. The meaning of the sentence in its semantic and pragmatic aspects. Reidel, Dordrecht, The Netherlands. find
  29. W. Skut, B. Krenn, T. Brants, and H. Uszkoreit. 1997. An annotation scheme for free word order languages. In Applied Natural Language Processing 1997, pages 88–95.
  30. M. Steedman. 2000. Information structure and the syntax-phonology interface. Linguistic Inquiry, 31(4):649–689. ↗
  31. E. Teich, S. Hansen, and P. Fankhauser. 2001. Representing and querying multi-layer annotated corpora. pages 228–237, Philadelphia. find
  32. E. Vallduvı́ and E. Engdahl. 1996. The linguistic realisation of information packaging. Linguistics, 34:459– 519. find
  33. B. L. Webber. 1983. So what can we talk about now? M.I.T. Press. find
  34. B. L. Webber. 1991. Structure and ostension in the interpretation of discourse deixis. Language and Cognitive Processes, 6(2):107–135. ↗
  35. H. Weinrich. 1993. Textgrammatik der deutschen Sprache. Dudenverlag, Mannheim u.a. find
  36. L. Wind. 2002. Manual zur Annotation von anaphorischen und Bridging-relationen. European Media Laboratory GmbH, August 9. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/korbayova-2004-discourse-level-annotation-investigating-information/