(Pre-)Annotation of Topic-Focus Articulation in Prague Czech-English Dependency Treebank

What it is about

The paper reports the first step of annotating topic-focus articulation (TFA) in the Czech part of the Prague Czech-English Dependency Treebank. The annotation follows the guidelines of the Prague Dependency Treebank, based on Functional Generative Description. Each node is marked as contextually bound non-contrastive (t), contextually bound contrastive (c) or contextually non-bound (f). An automatic pre-annotation has 12 steps, using rules such as coreference, the "anaphor" relation and the verb's position. It marks over 40% of the nodes. Its distinction between bound and non-bound nodes is about 96% correct on manually checked texts. The annotator changed the pre-annotated value in 294 cases, 3.7% of the pre-annotated nodes. Only 0.8% of the "t" nodes and 5.4% of the "f" nodes were changed, and the "c" value was never set automatically.

Ideas

• I1. Write the TFA pre-annotation as a list of rules in prag, one rule per step, each with a count of nodes it marks and a count of later changes, as in the paper's Table 1 and Table 2. This makes each rule's error rate visible.
• I2. Use the coreference links from coref as the "given" test for a node: a node whose referent has an earlier mention is a candidate for the bound value "t". Measure the step's precision on a checked sample.
• I3. Use the information-structure labels in word order. In en/src/lin.rs, a focus-final rule could be tested against the linearisation the corpus shows.
• I4. Keep the "c" value out of automatic output, as the paper does, and report it as a gap, not a zero.

For Mova

The chapter is 29. Ellipsis, word order and information structure. The code is prag/src/roles.rs and en/src/lin.rs, with coref for the given test. The measure is the rule-by-rule agreement with a hand-checked sample, plus the number of changed values per rule, as in Table 2. The Czech corpus is not used. The ideas are for English only.

Topics: Coreference resolution · Treebanks

Cite

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

@inproceedings{mirovsku2013pre,
  title = {{(Pre-)Annotation of Topic-Focus Articulation in Prague Czech-English Dependency Treebank}},
  author = {Mírovský, Jiří and Rysová, Kateřina and Rysová, Magdaléna and Hajičová, Eva},
  booktitle = {Proceedings of the Sixth International Joint Conference on Natural Language Processing},
  year = {2013},
  url = {https://aclanthology.org/I13-1007/}
}

References

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

1. E. Bejček, J. Panevová, J. Popelka, P. Straňák, M. Ševčíková, J. Štěpánek, Z. Žabokrtský. 2012. Prague Dependency Treebank 2.5 – a revisited version of PDT 2.0. In: Proceedings of the 24th Inter Of course, it is a matter of discussion (and testing), how much effort of the human annotator such a pre-annotation saves and how to set the reliability limit for the rule selection. national Conference on Computational Linguistics (Coling 2012), Mumbai, India, pp. 231–246. find
2. S. Calhoun, M. Nissim, M. Steedman, J. Brenier. 2005. A Framework for Annotating Information Structure in Discourse. In: Proceedings of the Workshop on Frontiers in Corpus Annotations II: Pie in the Sky. Ann Arbor, Michigan: Association for Computational Linguistics, pp. 45–52. URL http://aclweb.org/anthology/W/W05/W05-0307. ↗ our card
3. P. Cook, F. Bildhauer. 2011. Annotating information structure. The case of ”topic”. In: S. Dipper & H. Zinsmeister (eds.), Beyond Semantics.Corpus based Investigations of Pragmatic and Discourse Phenomena, Ruhr Universität Bochum, Bochumer Linguistische Arbeitsberichte, pp. 45–56. URL http://www.linguistics.ruhr-uni-bochum.de/bla/be-yondsem2011/cook_final.pdf. ↗
4. S. Dipper, M. Götze, S. Skopeteas (eds.). 2007. Infor mation Structure in CrossLinguistic Corpora: An notation Guidelines for Phonology, Morphology, Syntax, Semantics and Information Structure, vol. 7 of Interdisciplinary Studies on Information Structure. Potsdam, Germany: Universitätsverlag Potsdam. URL http://www.sfb632.uni-potsdam.de/pub-lications/isis07.pdf. ↗
5. S. Dipper, M. Götze, M. Stede, T. Wegst. 2004. AN-NIS: A Linguistic Database for Exploring Information Structure. In Ishihara, S., Schmitz, M., Schwarz, A. (Eds.), Working Papers of the SFB632, Interdisciplinary Studies on Information Structure (ISIS) 1, pp. 245–279. Potsdam: University publishing house Potsdam. find
6. K. Donhauser. 2007. Zur informationsstrukturellen Annotation sprachhistorischer Texten. Sprache und Informationsverarbeitung 31, pp. 39–45. URL http://www.sfb632.uni-potsdam.de/publications/B4/donhauser_2007.pdf. ↗
7. J. Hajič, J. Panevová, E. Hajičová, P. Sgall, P. Pajas, J. Štěpánek, J. Havelka, M. Mikulová, Z. Žabokrtský, M. Ševčíková-Razímová. 2006. Prague Dependency Treebank 2.0. Software prototype, Linguistic Data Consortium, Philadelphia, PA, USA, ISBN 1-58563-370-4, http://www.ldc.u-penn.edu, Jul 2006. ↗
8. J. Hajič, E. Hajičová, J. Panevová, P. Sgall, O. Bojar, S. Cinková, E. Fučíková, M. Mikulová, P. Pajas, J. Popelka, J. Semecký, J. Šindlerová, J. Štěpánek, J. Toman, Z. Urešová, Z. Žabokrtský. 2012. Announcing Prague Czech-English Dependency Treebank 2.0. In: Proceedings of the 8th International Conference on Language Resources and Evalua tion (LREC 2012), European Language Resources Association, İstanbul, Turkey, ISBN 978-2-9517408-7-7, pp. 3153–3160. find
9. E. Hajičová, J. Havelka, K. Veselá. 2005. Corpus Evidence of Contextual Boundness and Focus. In: Proceedings of the Corpus Linguistics Conference Series, University of Birmingham, Birmingham, UK, ISSN 1747-9398. find
10. E. Hajičová, J. Panevová, P. Sgall. 2000. A Manual for Tectogrammatical Tagging of the Prague Dependency Treebank. Technical report tr200009, ÚFAL/CKL. URL http://ufal.mff.cuni.cz/pdt/Cor-pora/PDT_1.0/Doc/tmanual/tmanen.pdf. In cooperation with A. Böhmová, M. Ceplová and V. Řezníčková. Translated by Z. Kirschner, E. Hajičová and P. Sgall. ↗
11. E. Hajičová, B. H. Partee, P. Sgall. 1998. Topicfocus articulation, tripartite structures, and semantic content. Dordrecht, Boston: Kluwer Academic Publishers. find
12. M. P. Marcus, B. Santorini, M. A. Marcinkiewicz. 1993. Building a large annotated corpus of English: The Penn Treebank. Computational Linguistics, 19(2), pp. 313–330. find our card
13. M. Mikulová et al. 2005. Annotation on the tec togrammatical layer in the Prague Dependency Treebank. The Annotation Guidelines. Prague: UFAL MFF. Available at: http://ufal.mff.cuni.cz/pdt2.0/doc/manuals/en/t-layer/html/index.html. ↗
14. M. Nissim, S. Dingare, J. Carletta, M. Steedman. 2004. An annotation scheme for information status in dialogue. In: Proceedings of the 4thConference on Language Resources and Evaluation. Lisbon, Portugal. URL http://www.lrec-conf.org/proceed-ings/lrec2004/pdf/638.pdf. ↗
15. P. Paggio. Annotating Information Structure in a Corpus of Spoken Danish. 2006. In: Proceedings of the Fifth International Conference on Language Resources and Evaluation (LREC), pp. 1606–1609, Genoa, Italy. find
16. O. Postolache. 2005. Learning Information Structure in The Prague Treebank. In: Proceedings of the ACL Student Research Workshop, pp. 115–120, Ann Arbor, Michigan, June 2005. find
17. P. Sgall. 1967. Generative description of language and the Czech Declension (in Czech). Prague: Academia. find
18. P. Sgall, E. Hajičová, E. Benešová. 1973. Topic, focus and generative semantics (Vol. 1). Kronberg Taunus: Scriptor Verlag. find
19. P. Sgall, E. Hajičová, J. Panevová. 1986. The meaning of the sentence in its semantic and pragmatic aspects. Springer. find

This page in Markdown → https://md.mova.elamur.ai/free/en/docs/papers/mirovsky-2013-pre-annotation-topic-focus-articulation/