Elamur Research
☰
home / papers / 2008 / The Penn Discourse TreeBank 2.0.

The Penn Discourse TreeBank 2.0.

What it is about

PDTB-2.0 is a lexically grounded annotation of discourse relations over the 1-million-word Wall Street Journal corpus. Each relation has two abstract-object arguments (Arg1, Arg2), and an explicit connective (18,459 tokens) or an inserted implicit connective (16,224) marks it; the rest are AltLex, EntRel and NoRel. Each relation has a three-level sense (class, type, subtype), an attribution record (source, type of attribution, scopal polarity, determinacy), and its arguments can lie far from the connective. Agreement is 90.2% exact-match for explicit arguments and 85.1% for implicit ones, and 94%, 84% and 80% at the class, type and subtype levels. The paper is a resource description with statistics, not a method.

Ideas

For Mova

The chapter is languages/en/36-discourse-relations.md (sections 4, 6, 7 and 11). The code is the frame and link code in prag and the connective handling in en (grep for the connective list before adding anything). The PDTB text and its data are LDC and NC: tests only, never in a release, and no PDTB examples in the repo.

Library summary

The Research library's own summary of the paper, in our words — not the paper's text.

Written for the Research library (claude-opus-5-5) from the paper text, in our own words: the license "CC BY-NC-SA 3.0 (ACL Anthology, before 2016)" does not allow redistributing the text itself. Original — aclanthology.org/L08-1093.pdf

Summary

PDTB-2.0 is the second release of the Penn Discourse TreeBank. It adds discourse-level annotation to the roughly 1-million-word Wall Street Journal corpus, the same texts already covered by the Penn Treebank syntax and Propbank. The annotation is tied to words in the text. When a connective such as "because" or "as a result" signals a relation, the annotators mark it and its two text spans, Arg1 and Arg2. When two adjacent sentences are related but no connective is present, they insert a fitting "Implicit" connective. Every relation also gets a sense from a hierarchical scheme, and the source of each relation and each argument (attribution) is recorded. The scheme does not commit to any single theory of discourse structure, so researchers from different frameworks can use it and test their theories on it. The paper reports corpus statistics, especially on where Arg1 is located and how far it extends, and lists what changed since PDTB-1.0. The resource matters because it puts syntax, semantics and discourse on one corpus. That makes it useful for training and evaluation in summarization, question answering, text generation and connective sense disambiguation.

Key points

  • PDTB-2.0 was distributed through the LDC (LDC2008T05) in February 2008, after a first preliminary release in April 2006. It includes alignments with the Penn Treebank, but not with Propbank.
  • Explicit connectives come from three classes: subordinating conjunctions, coordinating conjunctions and discourse adverbials. Modified forms ("only because"), conjoined forms and paired forms ("either..or") are also annotated.
  • There are 18,459 Explicit connective tokens. For about 61% of them, Arg1 is in the connective's own sentence, about 30% in the sentence just before, and 9% (1,666) in an earlier, non-adjacent sentence. Those distant cases mostly involve discourse adverbials, sentence-initial coordinators and "also."
  • Arguments follow a minimality principle: each contains only the text needed to interpret the relation. Material that helps but isn't required is marked separately as Sup1/Sup2.
  • Attribution is annotated for each relation and each argument. It supports subjectivity analysis and multi-perspective QA, and it shows how sentence structure and discourse structure interact.

Topics: Discourse analysis · Treebanks

Cite

The paper itself, not this card. BibTeX from the library's record of the work. DOI 10.63317/55sr8v9kexr7

@inproceedings{prasad2008penn,
  title = {{The Penn Discourse TreeBank 2.0.}},
  author = {Prasad, Rashmi and Dinesh, Nikhil and Lee, Alan and Miltsakaki, Eleni and Robaldo, Livio and Joshi, Aravind and Webber, Bonnie},
  booktitle = {Proceedings of the Sixth International Conference on Language Resources and Evaluation (LREC'08)},
  year = {2008},
  doi = {10.63317/55sr8v9kexr7},
  url = {https://doi.org/10.63317/55sr8v9kexr7}
}

References

24 references, 9 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Lynn Carlson, Daniel Marcu, and Mary Ellen Okurowski. 2001. Building a discourse-tagged corpus in the framework of rhetorical structure theory. In Proceedings of the 2nd SIGDIAL Workshop on Discourse and Dialogue, Eurospeech 2001. find our card
  2. 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. In Proceedings of the ACL Workshop on Frontiers in Corpus Annotation II: Pie in the Sky. 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, 23:55–106. find
  4. Jaakko Hintikka. 1971. Semantics for propositional attitudes. In L. Linsky, editor, Reference and Modality, pages 145–167. Oxford. find
  5. Laurence Horn. 1978. Remarks on neg-raising. In Peter Cole, editor, Syntax and Semantics 9: Pragmatics. Academic Press, New York. find
  6. Lauri Karttunen. 1971. Some observations on factivity. Papers in Linguistics, 4:55–69. find
  7. Andrew Kehler. 2002. Coherence, Reference, and the Theory of Grammar. CSLI Publications, Stanford. find
  8. Carol Kiparsky and Paul Kiparsky. 1971. Fact. In D. D. Steinberg and L. A. Jakobovits, editors, Semantics: An Interdisciplinary Reader in Philosophy, Linguistics and Psychology, pages 345–369. Cambridge University Press, Cambridge. find
  9. Alan Lee, Rashmi Prasad, Aravind Joshi, Nikhil Dinesh, and Bonnie Webber. 2006. Complexity of dependencies in discourse: Are dependencies in discourse more complex than in syntax? In Proceedings of the 5th International Workshop on Treebanks and Linguistic Theories. find
  10. Beth Levin. 1993. English verb classes and alternations: a preliminary investigation. University of Chicago Press. find
  11. William C. Mann and Sandra A. Thompson. 1988. Rhetorical structure theory. Toward a functional theory of text organization. Text, 8(3):243–281. ↗ our card
  12. Mitchell P. Marcus, Beatrice Santorini, and Mary Ann Marcinkiewicz. 1993. Building a large annotated corpus of english: The Penn Treebank. Computational Linguistics, 19(2):313–330. ↗ our card
  13. Eleni Miltsakaki, Rashmi Prasad, Aravind Joshi, and Bonnie Webber. 2004. Annotating discourse connectives and their arguments. In Proceedings of the HLT/NAACL Workshop on Frontiers in Corpus Annotation, pages 9– 16. find
  14. Eleni Miltsakaki, Nikhil Dinesh, Rashmi Prasad, Aravind Joshi, and Bonnie Webber. 2005. Experiments on sense annotation and sense disambiguation of discourse connectives. In Proceedings of the Fourth Workshop on Treebanks and Linguistic Theories. find
  15. Martha Palmer, Daniel Guildea, and Paul Kingsbury. 2005. The Proposition Bank: an annotated corpus of semantic roles. Computational Linguistics, 31(1):71–106. find our card
  16. Rashmi Prasad, Aravind Joshi, Nikhil Dinesh, Alan Lee, Eleni Miltsakaki, and Bonnie Webber. 2005. The Penn Discourse TreeBank as a resource for natural language generation. In Proceedings of the Corpus Linguistics Workshop on Using Corpora for NLG. find our card
  17. Rashmi Prasad, Nikhil Dinesh, Alan Lee, Aravind Joshi, and Bonnie Webber. 2006. Annotating attribution in the Penn Discourse TreeBank. In Proceedings of the COLING/ACL Workshop on Sentiment and Subjectivity in Text, pages 31–38. find our card
  18. Ivan A. Sag and Carl Pollard. 1991. An integrated theory of complement control. Language, 67(1):63–113. find
  19. PDTB-Group. 2006. The Penn Discourse TreeBank 1.0 Annotation Manual. Technical Report IRCS-06-01, Institute for Research in Cognitive Science, University of Pennsylvania. find our card
  20. PDTB-Group. 2008. The Penn Discourse TreeBank 2.0 Annotation Manual. Technical Report IRCS-08-01, Institute for Research in Cognitive Science, University of Pennsylvania. find our card
  21. Bonnie Webber and Aravind Joshi. 1998. Anchoring a lexicalized tree-adjoining grammar for discourse. In Manfred Stede, Leo Wanner, and Eduard Hovy, editors, Discourse Relations and Discourse Markers: Proceedings of the Conference, pages 86–92. Association for Computational Linguistics. Bonnie Webber and Rashmi Prasad. To appear. Discourse structure: Swings and roundabouts. In Cathrine Fabricius-Hansen and Bergljot Behrens, editors, Linguistics, Special issue on Structuring Information in Discourse: the Explicit/Implicit Dimension. find
  22. Bonnie Webber, Aravind Joshi, Matthew Stone, and Alistair Knott. 2003. Anaphora and discourse structure. Computational Linguistics, 29(4):545–587. find our card
  23. Janyce Wiebe, Theresa Wilson, and Claire Cardie. 2005. Annotating expressions of opinions and emotions in language. Language Resources and Evaluation, 1(2). find
  24. Florian Wolf and Edward Gibson. 2005. Representing discourse coherence: A corpus-based study. Computational Linguistics, 31(2). find