Elamur Research⌕
☰
home / papers / 2018 / The EcoLexicon English Corpus as an open corpus in Sketch Engine

The EcoLexicon English Corpus as an open corpus in Sketch Engine

Paper · 2018 · Pilar Leon-Arauz, Antonio San Martin and Arianne Reimerink

Summary

Specialists who need real evidence about domain language rarely have time to build large corpora themselves. General reference collections such as the BNC also cover technical genres poorly. The authors present the EcoLexicon English Corpus (EEC), 23.1 million words of contemporary environmental writing. The LexiCon group first assembled it as an internal knowledge-extraction source for EcoLexicon, a terminological knowledge base. Each text carries XML metadata, partly drawn from Dublin Core, describing:

  • subject domain
  • expertise of the intended reader
  • English variety
  • genre
  • editor type
  • year
  • country

The corpus was POS-tagged with TreeTagger (Penn Treebank tagset) and published in Sketch Engine as an open corpus, so no subscription is needed. It ships with the EcoLexicon Semantic Sketch Grammar (ESSG), a custom CQL rule set that captures five terminological relations. Most of the paper is a guided tour of possible queries: text-type filtering, CQL patterns, variant counts, semantic word sketches, sketch diffs and word lists. There is no quantitative evaluation. The contribution is a free resource where pragmatic, syntactic and semantic constraints can be combined, which serves terminologists, translators and domain experts.

Key points

  • The EEC has 23.1M words from 1973–2016, covering American, British and Euro English. Readers are split into expert, semi-expert and general public. Combined with the Spanish corpus also used by EcoLexicon, the EEC brings the total to more than 50M words.
  • The Sketch Engine version uses TreeTagger 3.3 with the Penn Treebank tagset. It includes ready-made subcorpora for American English, British English, 1973–1999, 2000–2009 and 2010–2016.
  • The ESSG targets generic-specific, part-whole, location, cause and function relations. It is separate from the default sketch grammar and can be downloaded for reuse on other corpora.
  • Illustrative findings:
    • nitrogen is sketched as a pollutant in Air Quality Management but as a nutrient in Biology.
    • liquefy is more frequent in expert texts.
    • photovoltaic concentrates in Renewable Energy.
  • Through the Sketch Engine API the corpus feeds EcoLexiCAT, a CAT tool that also draws on EcoLexicon, BabelNet and IATE. Planned work adds semantic tags for semantic categories and argument structure.

Interesting ideas

  • I1. A small hand-written pattern grammar over tagged text can harvest typed concept relations (is-a, part-of, location, cause, function) along with the sentences that justify them.
  • I2. A term's relational profile depends on subdomain, as with nitrogen being a pollutant in one field and a nutrient in another.
  • I3. Tagging texts by audience expertise makes it possible to compare how usage shifts from lay to expert writing.
  • I4. Sketch diff compares the collocational profiles of two lemmas, of one lemma across two subcorpora, or of singular versus plural forms.
  • I5. A single query with alternations can gather spelling and compounding variants of a multiword term, such as geologic/geological and timescale/time scale, and then count which form dominates.

For Mova

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

  • I1 (world): Similar rules written over UD trees rather than flat POS strings could populate the knowledge graph. Each edge would carry its rule and source sentence, so every fact stays explainable.
  • I2 (world): Compiled ontology facts could be stored with a domain or context label. The world builder could then prefer the reading that fits the genre of the current text.
  • I3 (answer): Reader evaluations could be split by register, for example school-level versus technical passages, to see where comprehension breaks down.
  • I4 (learn): The same contrast applied to dependency-relation profiles of a lemma could compare witness parsers or treebanks. This would surface systematic disagreements worth turning into repair rules.
  • I5 (tokens): Variant families of multiword terms could be mapped to one canonical lemma, with frequency deciding which form is the default.

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

🎓Universidad de GranadaGranada · Spain7 works · 307 citations

Also on the byline: Université du Québec à Trois-Rivières.

 Fields, topics and concepts

📚 Shelves 4The field of science this work belongs to (OpenAlex fields).Computer Science · 9,085 worksArtificial Intelligence · 8,888 worksArts and Humanities · 9,296 worksLanguage and Linguistics · 9,206 works

🧭 Topics 1Research subjects and languages the work is about.English · 1,373 works

Cite

Cite as Pilar Leon-Arauz, Antonio San Martin and Arianne Reimerink. 2018. The EcoLexicon English Corpus as an open corpus in Sketch Engine. arXiv:1807.05797.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.48550/arxiv.1807.05797 · arXiv 1807.05797

@misc{leonarauz2018ecolexicon,
  title = {{The EcoLexicon English Corpus as an open corpus in Sketch Engine}},
  author = {Pilar Leon-Arauz, Antonio San Martin and Arianne Reimerink},
  year = {2018},
  doi = {10.48550/arxiv.1807.05797},
  eprint = {1807.05797},
  archivePrefix = {arXiv},
  url = {https://doi.org/10.48550/arxiv.1807.05797}
}

References

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

  1. Faber, P., León-Araúz, P. & Reimerink, A. (2016). EcoLexicon : New Features and Challenges. In GLOBALEX 2016: Lexicographic Resources for Human Language Technology in conjunction with LREC 2016, pp. 73–80. find
  2. Jakubíček, M., Kilgarriff, A., McCarthy, D. & Rychlý, P. (2010). Fast syntactic searching in very large corpora for many languages. Proceedings of the PACLIC 24, pp. 741–747. find
  3. Kilgarriff, A., Baisa, V., Bušta, J., Jakubíček, M., Kovář, V., Michelfeit, J., Rychlý, P. & Suchomel, V. (2014). The Sketch Engine: ten years on. Lexicography, 1(1), pp. 7–36. http://doi.org/10.1007/s40607-014-0009-9 ↗
  4. Kilgarriff, A., Kovár, V., Krek, S., Srdanovic, I. & Tiberius, C. (2010). A Quantitative Evaluation of Word Sketches. In A. Dykstra & T. Schoonheim (Eds.), Proceedings of the 14th EURALEX International Congress. pp. 372–379. Leeuwarden/Ljouwert, The Netherlands: Fryske Akademy. find
  5. Kosem, I., Gantar, P., Logar, N. & Krek, S. (2014). Automation of Lexicographic Work Using General and Specialized Corpora: Two Case Studies. In A. Abel, C. Vettori & N. Ralli (Eds.), Proceedings of the XVI EURALEX International Congress: The User in Focus, pp. 355–364. Bolzano/Bozen: Institute for Specialised Communication and Multilingualism. find
  6. León-Araúz, P. & Reimerink, A. (2018). Evaluating EcoLexiCAT: a Terminology-Enhanced CAT Tool. In Proceedings of the 11th International Language Resources and Evaluation Conference (LREC2018). Miyazaki: ELRA. find
  7. León-Araúz, P., Reimerink, A. & Faber, P. (2017). EcoLexiCAT: a Terminology-enhanced Translation Tool for Texts on the Environment. In I. Kosem, C. Tiberius, M. Jakubíček, J. Kallas, S. Krek & V. Baisa (Eds.), Electronic lexicography in the 21st century. Proceedings of eLex 2017 conference, pp. 321–341. Leiden: Lexical Computing. find
  8. León-Araúz, P. & San Martín, A. (2018). The EcoLexicon Semantic Sketch Grammar: from Knowledge Patterns to Word Sketches. In I. Kerneman & S. Krek (Eds.), Proceedings of the LREC 2018 Workshop “Globalex 2018 – Lexicography & WordNets”, pp. 94–99. Miyazaki: Globalex. find
  9. León-Araúz, P., San Martín, A. & Faber, P. (2016). Pattern-based Word Sketches for the Extraction of Semantic Relations. In Proceedings of the 5th International Workshop on Computational Terminology, pp. 73–82. Osaka. find
  10. Meyer, I. (2001). Extracting knowledge-rich contexts for terminography. In D. Bourigault, C. Jacquemin & M.-C. L’Homme (Eds.), Recent advances in computational terminology, pp. 279–302. Amsterdam/Philadelphia: John Benjamins. ↗
  11. San Martín, A., Cabezas-García, M., Buendía Castro, M., Sánchez Cárdenas, B., León-Araúz, P. & Faber, P. (2017). Recent Advances in EcoLexicon. Dictionaries: Journal of the Dictionary Society of North America, 38(1), pp. 96–115. ↗
  12. Sinclair, J. (1991). Corpus, Concordance, Collocation. Oxford: Oxford University Press. find
  13. Teubert, W. (2005). My version of corpus linguistics. International Journal of Corpus Linguistics, 10(1), 1–13. http://doi.org/10.1075/ijcl.10.1.01teu Acknowledgements This research was carried out as part of projects FF2014-52740-P, Cognitive and Neurological Bases for Terminology-enhanced Translation (CONTENT), and FFI2017-89127-P, Translation-oriented Terminology Tools for Environmental Texts (TOTEM), funded by the Spanish Ministry of Economy and Competitiveness. ↗

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/leon-arauz-2018-ecolexicon-english-corpus-open-corpus/