Elamur Research⌕
☰
home / papers / 2013 / A Corpus-Based Tool for Exploring Domain-Specific Collocations in English

A Corpus-Based Tool for Exploring Domain-Specific Collocations in English

Paper · 2013 · Ping-Yu Huang, Chien-Ming Chen, Nai-Lung Tsao et al.

Summary

General academic vocabulary lists such as Coxhead's Academic Word List (AWL) assume that one shared set of words serves students in every field. Following Hyland and Tse's critique, the authors argue that these words behave differently by discipline: they occur at different rates, combine with different partners, and sometimes carry different senses. To give EAP learners discipline-specific phrasing, they built TechCollo, an online tool for looking up collocations within one domain and comparing them across domains or against general English. It draws on six medium-sized specialized corpora covering medicine, engineering and law. For each domain there is one corpus derived from Wikipedia and one built from journal articles, and two general reference corpora are also available. Candidate word pairs are ranked by raw frequency, traditional MI, and a sense-normalized variant called normMI. A pilot study compared AWL nouns in the medical and engineering Wikipedia corpora. Many of these nouns were unevenly distributed and took field-specific verb collocates, which supports the case for collocation lists organized by discipline.

Key points

  • The domain corpora are the Medical, Engineering and Legal Wiki Corpora (MWC, EWC, LWC), built from WaCkypedia_EN, which came already tagged and parsed with TreeTagger and MaltParser. Alongside them are the Medical, Engineering and Legal Academic Corpora (MAC, EAC, LAC), built from 280, 408 and 106 articles taken from sixty journals and processed with Stanford CoreNLP.
  • Each domain corpus holds roughly 1.8–5.5M tokens. The general comparison corpora are Wikipedia (about 834M tokens) and the BNC (about 95M tokens).
  • normMI (Wible et al., 2004) adjusts traditional MI by dividing a lexeme's frequency by its number of WordNet senses. It rests on Yarowsky's one-sense-per-collocation assumption, and its purpose is to stop collocations of very frequent words, such as make a decision, from being undervalued.
  • In the pilot, area combined with rub and scratch only in medical texts, where it referred to a patch of body surface. The two fields also chose different verbs for similar meanings, for example confer vs offer benefit and undergo vs undertake research.
  • Planned work includes per-discipline word lists with frequency indexing and tests of paradigmatic modifiability (Wermter and Hahn) and of combined association measures (Pecina) to raise extraction precision.

Interesting ideas

  • I1. Normalizing an association score by a word's sense count keeps frequent, polysemous words from losing out in collocation ranking, since frequency spread over many senses is not penalized as raw MI would penalize it.
  • I2. A distinctive collocate can reveal a domain-specific sense: a noun that pairs with verbs found only in one field, such as area with rub, is probably being used in that field's meaning.
  • I3. The comparison view separates collocates into "shared" and "only in domain A or B" columns, which makes cross-corpus differences easy to inspect by hand.
  • I4. Each domain has two corpora of different kinds, encyclopedic and journal-based, plus general reference corpora, so domain effects can be separated from register effects.
  • I5. Different fields prefer different verbs for the same idea, so a single default verb choice can sound unidiomatic.

For Mova

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

  • I1 (world): Sense-normalized association scores could weight the verb–noun preferences stored in compiled knowledge so that light verbs like make or take keep meaningful links.
  • I2 (frames): When choosing a frame or WordNet sense for an argument noun, the governing verb's domain-restricted collocates could serve as explainable evidence for one sense over another.
  • I3 (learn): The same three-way split could be used to compare parser or frame outputs on two registers, showing what generalizes and what is specific to one corpus.
  • I4 (learn): Silver data and evaluations could likewise be grouped by register within a topic, so that improvements are checked on both encyclopedic and expert text before they are trusted.
  • I5 (back): During generation from the meaning graph, the system could pick the verb realization preferred in the text's domain, for example undertake research in engineering-style output.

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

🔬Academia SinicaTaipei · Taiwan10 works · 760 citations🔬Institute of Information Science, Academia SinicaTaipei · Taiwan7 works · 51 citations🎓National Central UniversityTaoyuan City · Taiwan2 works · 53 citations

Also on the byline: Ming Chi University of Technology.

 Fields, topics and concepts

📚 Shelves 4The field of science this work belongs to (OpenAlex fields).Arts and Humanities · 16,039 worksLanguage and Linguistics · 15,891 worksPsychology · 1,552 worksDevelopmental and Educational Psychology · 499 works

🧭 Topics 2Research subjects and languages the work is about.English · 2,263 worksIdioms & multiword expressions · 217 works

Cite

Cite as Ping-Yu Huang, Chien-Ming Chen, Nai-Lung Tsao and David Wible. 2013. A Corpus-Based Tool for Exploring Domain-Specific Collocations in English. PACLIC 27 Workshop on Computer-Assisted Language Learning.

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

@inproceedings{huang2013corpus,
  title = {{A Corpus-Based Tool for Exploring Domain-Specific Collocations in English}},
  author = {Ping-Yu Huang, Chien-Ming Chen, Nai-Lung Tsao and David Wible},
  booktitle = {PACLIC 27 Workshop on Computer-Assisted Language Learning},
  year = {2013},
  url = {https://aclanthology.org/Y13-2005/}
}

References

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

  1. Marco Baroni, Silvia Bernardini, Adriano Ferraresi, and Eros Zanchetta. 2009. The WaCky Wide Web: A Collection of Very Large Linguistically Processed Web-crawled Corpora. Language Resources and Evaluation 43(3): 209-226. ↗
  2. Caroline Barrière. 2009. Finding Domain Specific Collocations and Concordances on the Web. Proceedings of the Workshop on Natural Language Processing Methods and Corpora in Translation, Lexicography, and Language Learning. find
  3. Laurie Bauer. 1993. Manual of Information to Accompany the Wellington Corpus of Written New Zealand English. Victoria University of Wellington. find
  4. British National Corpus, Version 2 (BNC World). 2001. Distributed by Oxford University Computing Services on behalf of the BNC Consortium. URL: http://www.natcorp.ox.ac.uk/ ↗
  5. Qi Chen and Guang-Chun Ge. 2007. A Corpus-Based Lexical Study on Frequency and Distribution of Coxhead’s AWL Word Families in Medical Research Articles (RAs). English for Specific Purposes 26(4): 502-514. ↗
  6. Kenneth Ward Church and Patrick Hanks. 1990. Word Association Norms, Mutual Information, and Lexicography. Computational Linguistics 16(1): 22-29. ↗
  7. Andrew Cohen, Hilary Glasman, Phyllis R. Rosenbaum-Cohen, Jonathan Ferrara, and Jonathan Fine. 1979 Reading English for Specialized Purposes: Discourse Analysis and the Use of Student Informants. TESOL Quarterly, 34: 551-564. ↗
  8. Nigel Collier, Chikashi Nobata, and Junichi Tsujii. 2002. Automatic Acquisition and Classification of Terminology Using a Tagged Corpus in the Molecular Biology Domain. Terminology 7(2): 239-257. ↗
  9. Averil Coxhead. 2000. A New Academic Word List. TESOL Quarterly, 34(2): 213-238. ↗
  10. Mark Davies. 2008. The Corpus of Contemporary American English (COCA): 400+ Million Words, 1990-present. http://www.americancorpus.org ↗
  11. Paul Farrell. 1990. Vocabulary in ESP: A Lexical Analysis of the English of Electronics and a Study of Semi-Technical Vocabulary. CLCS Occasional Paper No. 25. find
  12. John Flowerdew. 1993. Concordancing as a Tool in Course Design. System 21(2): 231-244. ↗
  13. Nilgün Hancioğlu, Steven Neufeld, and John Eldridge. 2008. Through the Looking Glass and into the Land of Lexico-Grammar. English for Specific Purposes 27(4): 459-479. ↗
  14. Dan Klein and Christopher D. Manning. 2003. Accurate Unlexicalized Parsing. Proceedings of the 41st Annual Meeting on Association for Computational Linguistics-Volume 1.
  15. Ken Hyland and Polly Tse. 2007. Is there an “academic vocabulary”?." TESOL Quarterly 41(2): 235-253. find
  16. I. S. P. Nation. 2001. Learning Vocabulary in Another Language. Cambridge University Press, Cambridge, UK. ↗
  17. Stig Johansson. 1978. Manual of Information to Accompany the Lancaster-Oslo/Bergen Corpus of British English, for Use with Digital Computers. University of Oslo. Oslo, Norway. find
  18. Pavel Pecina. 2008. A Machine Learning Approach to Multiword Expression Extraction. Proceedings of the LREC MWE 2008 Workshop. find
  19. Thomas C. Rindflesch, Lawrence Hunter, and Alan R. Aronson. 1999, Mining Molecular Binding Terminology from Biomedical Text. Proceedings of the AMIA Symposium. American Medical Informatics Association. find
  20. Frank Smadja. 1993. Retrieving Collocations from Text: Xtract. Computational Linguistics 19(1): 143-177.
  21. Cucu Sutarsyah, Paul Nation, and Graeme Kennedy. 1994. How Useful Is EAP Vocabulary for ESP? A Corpus Based Case Study. RELC Journal 25(2): 34-50. find
  22. Viphavee Vongpumivitch, Ju-yu Huang, and Yu-Chia Chang. 2009. Frequency Analysis of the Words in the Academic Word List (AWL) and Non-AWL Content Words in Applied Linguistics Research Papers. English for Specific Purposes 28(1): 33-41. ↗
  23. Jeremy Ward. 2009. A Basic Engineering English Word List for Less Proficient Foundation Engineering Undergraduates. English for Specific Purposes 28(3): 170-182. ↗
  24. Joachim Wermter and Udo Hahn. 2004. Collocation Extraction Based on Modifiability Statistics. Proceedings of the 20th International Conference on Computational Linguistics. Association for Computational Linguistics. find
  25. Joachim Wermter and Udo Hahn. 2005. Paradigmatic Modifiability Statistics for the Extraction of Complex Multi-word Terms. Proceedings of the Conference on Human Language Technology and Empirical Methods in Natural Language Processing. Association for Computational Linguistics. find
  26. David Wible, Chin-Hwa Kuo, and Nai-Lung Tsao. 2004. Improving the Extraction of Collocations with High Frequency Words. Proceedings of International Conference on LREC. find
  27. Huizhong Yang. 1986. A New Technique for Identifying Scientific/Technical Terms and Describing Science Texts. Literary and Linguistic Computing 1: 93-103. find
  28. David Yarowsky. 1995. Unsupervised Word Sense Disambiguation Rivaling Supervised Methods. Proceedings of the 33rd Annual Meeting on Association for Computational Linguistics. Association for Computational Linguistics.

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/huang-2013-corpus-based-tool-exploring-domain/