Elamur Research⌕
☰
home / papers / 2011 / A Prague Markup Language profile for the SemTi-Kamols grammar model

A Prague Markup Language profile for the SemTi-Kamols grammar model

Summary

The authors look for a convenient way to annotate a Latvian syntactic treebank using the hybrid SemTi-Kamols model. The model was designed for languages with relatively free word order. It combines dependencies with elements of constituency structure: x-word units, which include analytical verb forms, prepositional phrases, coordination and similar constructions. Instead of writing a tool from scratch, the researchers developed a profile of the Prague Markup Language (PML) format, borrowed from the Prague Dependency Treebank (PDT). They adapted the TrEd tree editor to this profile. The w and m layers were taken from PDT almost unchanged. On the analytical layer a, they added artificial container nodes: the child nodes of a container are treated as constituents of the parent node, and the remaining children are treated as its dependents. The TrEd extension includes schemas, validation macros and two tree display modes. To demonstrate that the approach is viable, the authors annotated 100 sentences from "Sophie's World". The work shows that the PDT infrastructure can be applied to a more complex hybrid model. This is also useful for other languages with free word order.

Key points

  • Three container types are distinguished: coordination, x-word proper, and PMC (punctuation mark constructs). PMC is a new addition to the model; it represents in the tree the punctuation marks around direct speech, subordinate clauses, parenthetical insertions and similar constructions. Omitted sentence members, such as an elliptical predicate, are marked with "empty" nodes that have no corresponding token.
  • On the m layer, the authors reused the PDT categories for correcting spelling errors almost entirely and added one new category: it marks cases where two tokens form a single morphological unit. The tectogrammatical layer t is not covered in the paper.
  • The TrEd module is distributed under the GPL license (eksperimenti.ailab.lv/tred). It has two modes. Full view is intended for annotators and highlights likely incorrect roles in red. Compact view shows containers as colored edges, but does not allow editing.
  • The prototype treebank covers the first 100 sentences of J. Gaarder's novel "Sophie's World". Morphology was annotated semi-automatically, and the trees were built manually by a linguist. They were then reviewed jointly by 2–3 linguists and the architect of the PML schema.
  • Future plans: combine the PDT tools with the rule-based partial parser SemTi-Kamols and build a larger treebank for training a statistical parser.

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: Treebanks

Cite

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

@inproceedings{pretkalnia2011prague,
  title = {{A Prague Markup Language profile for the SemTi-Kamols grammar model}},
  author = {Pretkalniņa, Lauma and Nešpore, Gunta and Levāne-Petrova, Kristīne and Saulīte, Baiba},
  booktitle = {Proceedings of the 18th Nordic Conference of Computational Linguistics (NODALIDA 2011)},
  year = {2011},
  url = {https://aclanthology.org/W11-4645/}
}

References

10 references, 0 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Guntis Bārzdiņš, Normunds Grūzītis, Gunta Nešpore, Baiba Saulīte. 2007. Dependency-Based Hybrid Model of Syntactic Analysis for the Languages with a Rather Free Word Order. Proceedings of the 16th Nordic Conference of Computational Linguistics (NODALIDA), pp. 13–20 find
  2. Adriane Boyd, Markus Dickinson, Detmar Meurers. 2008. On Detecting Errors in Dependency Treebanks. Research on Language and Computation 6(2), pp. 113–137. find
  3. Noam Chomsky. 1957. Syntactic Structures. find
  4. The Hague: Mouton Jan Hajič, Alena Böhmová, Eva Hajičová, Barbora Vidová Hladká. 2000. The Prague Dependency Treebank: A Three-Level Annotation Scenario. A. Abeillé (ed.): Treebanks: Building and Using Parsed Corpora, Amsterdam: Kluwer, pp. 103–127. find
  5. Jan Hajič, Barbora Vidová Hladká, Petr Pajas. 2001. The Prague Dependency Treebank: Annotation Structure and Support. Proceedings of the IRCS Workshop on Linguistic Databases, Proceedings of the IRCS Workshop on Linguistic Databases, Philadelphia, USA, pp. 105–114. find
  6. Timo Järvinen, Pasi Tapanainen. 1998. Towards an implementable dependency grammar. Proceedings of the Workshop on Processing of Dependency-Based Grammars, pp. 1–10. find
  7. Gunta Nešpore, Baiba Saulīte, Guntis Bārzdiņš, Normunds Grūzītis. 2010. Comparison of the SemTi-Kamols and Tesnière’s Dependency Grammars. Proceedings of the 4th International Conference on Human Language Technologies — the Baltic Perspective, Frontiers in Artificial Intelligence and Applications, V… find
  8. Petr Pajas, Jan Štěpánek. 2006. XML-Based Representation of Multi-Layered Annotation in the PDT 2.0. Proceedings of the LREC Workshop on Merging and Layering Linguistic Information (LREC 2006), pp. 40–47. find
  9. Lucien Tesnière. 1959. Éléments de syntaxe structurale. Klincksieck, Paris. (Translation to Russian: Теньер Л. 1988. Основы структурного синтак-сиса. Ред. В.Г. Гак. Москва, Прогресс.) find
  10. ISSN 1736-6305 Vol. 11 http://hdl.handle.net/10062/16955 ↗

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/a-2011-prague-markup-language-profile-semti/