Elamur Research⌕
☰
home / papers / 2003 / A Categorial Variation Database for English

A Categorial Variation Database for English

Paper · 2003 · Nizar Habash and Bonnie Dorr

Summary

English words often have derivational relatives in other parts of speech, such as a verb, a noun and an adjective built on the same root. No large-scale English resource recorded these cross-category links, and WordNet linked words only within a single part of speech. Habash and Dorr built CatVar, a database that groups English lexemes into clusters of categorial variants. They started from seed links taken from curated lexicons (Englex, NOMLEX, LDOCE). They added lexeme coverage from the LCS databases, WordNet 1.6 and the Brown Corpus portion of the Penn Treebank. Porter-stem clustering then grew the seeds into full clusters. Every link records which source produced it. Human annotators judged a sample of clusters, which gave high precision and good recall. A linkability analysis showed that about 30% of CatVar's links cannot be found with the Porter stemmer. The resource works in both directions: it can collapse word forms for analysis and expand them for generation. It has been used in machine translation, headline generation and bilingual alignment.

Key points

  • CatVar 2.0 has 62,232 clusters covering 96,368 lexemes (nouns 62%, adjectives 24%, verbs 10%, adverbs 4%).
    • About half of the clusters contain a single word.
    • Multi-word clusters range from 2 to 27 words and follow a Zipf-like distribution.
  • A separate verb–preposition database (e.g., enter/into, avoid/away from) was built from shared LCS primitives. It has 242 clusters linking 230+ verbs and 29 prepositions.
  • Evaluation used 400 adjudicated clusters containing 808 words.
    • Precision was 91.82% and recall 72.46%, giving an F-score of 81.00%.
    • Recall rises to 87.16% (F 89.43%) when words that were present in CatVar but placed in other clusters are not counted as missing.
  • The linkability analysis covered 48,867 linked words, 14,731 clusters and 89,638 links.
    • About 12% of links involve identical forms across categories ("natural" links).
    • About 70% can be recovered by the Porter stemmer.
    • The remaining roughly 30% come only from CatVar.
  • CatVar was applied in three systems:
    • GHMT, where it constrains when a light verb can merge with its argument;
    • HeadGen, where it helps check that a headline contains an event;
    • DUSTer, for untangling translation divergences during bilingual alignment.
  • For IR use, CatVar needs an inflectional analyzer that first reduces words to their lexemes.

Interesting ideas

  • I1. Each link is tagged with the source that produced it, so an application can use only the subset it needs; HeadGen scored higher Bleu with the NOMLEX-related slice
  • I2. One shared resource serves both reduction (many forms to one cluster) and expansion (one root to its valid variants), avoiding the over- and under-coverage of a one-way stemmer or a free-running generator
  • I3. A light verb plus an event noun can be merged into one verb only if the noun has a verb variant in the database (e.g., "make … development" becomes "develop")
  • I4. "Linkability" splits a resource's links by which cheap method could also find them (identical form, Porter stem), isolating what the resource really adds over a baseline
  • I5. In evaluation, words "missing" from a cluster but present elsewhere in the database are separated from truly absent words, distinguishing linking errors from coverage gaps

For Mova

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

  • I1 (world): Compiled knowledge (verb classes, ontology links) could carry provenance tags, letting each reader choose trusted sources and keeping every inference traceable to a resource.
  • I2 (back): Generation could draw nominal, verbal or adjectival realizations of a concept from the same clusters the analyzer uses, so text → graph → text round-trips stay symmetric.
  • I3 (frames): Phrases like "made a decision" or "gave a promise" could be mapped to the frame of the related verb and its VerbNet roles when a categorial link exists.
  • I4 (learn): Any induced lexicon or repair-rule set could be scored by the share of its output that a trivial baseline could not produce, which acts as a built-in negative control.
  • I5 (learn): When evaluating silver data or lexical resources, wrongly attached items could be reported apart from absent ones, so repair work targets the right failure type.

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

🎓University of Maryland, College ParkCollege Park · United States132 works · 25,380 citations

 Fields, topics and concepts

📚 Shelves 4The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 worksArts and Humanities · 16,039 worksLanguage and Linguistics · 15,891 works

🧭 Topics 1Research subjects and languages the work is about.English · 2,263 works

Cite

Cite as Nizar Habash and Bonnie Dorr. 2003. A Categorial Variation Database for English. Proceedings of the 2003 Human Language Technology Conference of the North American Chapter of the Association for Computational Linguistics.

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

@inproceedings{habash2003categorial,
  title = {{A Categorial Variation Database for English}},
  author = {Nizar Habash and Bonnie Dorr},
  booktitle = {Proceedings of the 2003 Human Language Technology Conference of the North American Chapter of the Association for Computational Linguistics},
  year = {2003},
  doi = {10.3115/1073445.1073458},
  url = {https://doi.org/10.3115/1073445.1073458}
}

References

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

  1. E.L. Antworth. 1990. PC-KIMMO: A Two-Level Processor for Morphological Analysis. Dallas Summer Institute of Linguistics. find
  2. S. Bangalore and O. Rambow. 2000. Exploiting a probabilistic hierarchical model for generation.
  3. Marine Carpuat, Grace Ngai, Pascale Fung, and Kenneth Church. 2002. Creating a Bilingual Ontology: A Corpus-Based Approach for Aligning WordNet and HowNet. In Proceedings of the 1st Global WordNet Conference, Mysore, India. find
  4. D. Cruse. 1986. Lexical Semantics. Cambridge University Press. find
  5. Zhendong Dong. 2000. HowNet Chinese-English Conceptual Database. Technical Report Online Software Database, Released at ACL. http://www.keenage.com. ↗
  6. Bonnie J. Dorr, Gina-Anne Levow, and Dekang Lin. 2000. Building a Chinese-English Mapping between Verb Concepts for Multilingual Applications. In Proceedings of the Fourth Conference of the Association for Machine Translation in the Americas (AMTA), Cuernavaca, Mexico, pages 1–12. ↗
  7. Bonnie J. Dorr, Lisa Pearl, Rebecca Hwa, and Nizar Habash. 2002. DUSTer: A Method for Unraveling Cross-Language Divergences for Statistical Word-Level Alignment. In Proceedings of the Fifth Conference of the Association for Machine Translation in the Americas, AMTA-2002, Tiburon, California. find
  8. Bonnie J. Dorr. 2001. LCS Verb Database. Technical Report Online Software Database, University of Maryland, College Park, MD. http://www.umiacs.umd.edu/˜bonnie/LCS Database Docmentation.html. ↗
  9. Christiane Fellbaum. 1998. WordNet: An Electronic Lexical Database. MIT Press. http://www.cogsci.princeton.edu/˜wn [2000, September 7]. ↗
  10. Rebecca Green, Lisa Pearl, Bonnie J. Dorr, and Philip Resnik. 2001. Mapping WordNet Senses to a Lexical Database of Verbs. In Proceedings of the 39th Annual Meeting of the Association for Computational Linguistics, pages 244–251, Toulouse, France. find
  11. Nizar Habash and Bonnie J. Dorr. 2002. Handling Translation Divergences: Combining Statistical and Symbolic Techniques in Generation-Heavy Machine Translation. In Fifth Conference of the Association for Machine Translation in the Americas, AMTA-2002, Tiburon, California. find
  12. Nizar Habash, Bonnie J. Dorr, and David Traum. 2002. Efficient Language Independent Generation from Lexical Conceptual Structures. Machine Translation. find
  13. Nizar Habash. 2002. Generation-Heavy Machine Translation. In Proceedings of the International Natural Language Generation Conference (INLG’02) Student Session, New York. find
  14. M.A. Hearst. 1998. Automated Discovery of WordNet Relations. In Christiane Fellbaum, editor, WordNet: an Electronic Lexical Database. MIT Press. ↗
  15. David A. Hull and Gregory Grefenstette. 1996. Experiments in Multilingual Information Retrieval. In Proceedings of the 19th Annual International ACM SIGIR Conference on Research and Development in Information Retrieval. http://www.xerox.fr/people/grenoble/hull/papers/sigir96.ps. ↗
  16. Daniel Karp, Yves Schabes, Martin Zaidel, and Dania Egedi. 1992. A Freely Available Wide Coverage Morphological Analyzer for English. In Proceedings of Fourteenth International Conference on Computational Linguistics (COLING-92), Nantes, France. find
  17. K. Knight and S. Luk. 1994. Building a Large Knowledge Base for Machine Translation. In Proceedings of AAAI-94. find
  18. R. Krovetz. 1993. Viewing Morphology as an Inference Process,. In Proceedings of the Sixteenth Annual International ACM SIGIR Conference on Research and Development in Information Retrieval, pages 191–203. find
  19. Irene Langkilde and Kevin Knight. 1998. Generation that Exploits Corpus-Based Statistical Knowledge. In ACL/COLING 98, Proceedings of the 36th Annual Meeting of the Association for Computational Linguistics (joint with the 17th International Conference on Computational Linguistics), pages 704–710, Montreal, Canada.
  20. Marc Light. 1996. Morphological Cues for Lexical Semantics. In Proceedings of the 34th Annual Meeting of the Association for Computational Linguistics. find
  21. Catherine Macleod, Ralph Grishman, Adam Meyers, Leslie Barrett, and Ruth Reeves. 1998. NOMLEX: A Lexicon of Nominalizations. In Proceedings of EU-RALEX’98, Liege, Belgium. find
  22. Mitchell Marcus, Beatrice Santorini, and Mary Ann Marcinkiewicz. 1993. Building a Large Annotated Corpus of English: the Penn Treebank. Computational Linguistics, 19(2):313–330. ↗
  23. Igor Mel’čuk. 1988. Dependency Syntax: Theory and Practice. State University of New York Press, New York. find
  24. K. Papineni, S. Roukos, T. Ward, and W. Zhu. 2002. Bleu: a Method for Automatic Evaluation of Machine Translation. In Proceedings of Association of Computational Linguistics, Philadelphia, PA.
  25. M.F. Porter. 1980. An algorithm for suffix stripping. Program, 14(3):130–137. ↗
  26. P. Procter. 1983. Longman Dictionary of Contemporary English: Computer Codes for the Definition Space Other than the Subject Field. Longman Group LTD. find
  27. Margarita Alonso Ramos, Agnes Tutin, and Guy Lapalme. 1994. Lexical Functions of the Explanatory Combinatorial Dictionary for Lexicalization in Text Generation. In Patrick Saint-Dizier and Evelyne Viegas, editors, Computational Lexical Semantics. Cambridge University Press. ↗
  28. Philip Resnik. 1999. Disambiguating Noun Groupings with Respect to WordNet Senses. In S. Armstrong, K. Church, P. Isabelle, S. Manzi, E. Tzoukermann, and D. Yarowsky, editors, Natural Language Processing Using Very Large Corpora, pages 77–98. Kluwer Academic, Dordrecht. ↗
  29. J. Veronis. 1998. A study of polysemy judgements and inter-annotator agreement. In Programme and advanced papers of the Senseval workshop, Herstmonceux Castle, England. find
  30. Ellen M. Voorhees. 1993. Using WordNet to Disambiguate Word Senses for Text Retrieval. In Robert Korfhage, Edie Rasmussen, and Peter Willett, editors, Proceedings of the Sixteenth Annual International ACM SIGIR Conference on Research and Development in Information Retrieval, pages 171–180. ACM, June. find
  31. Jinxi Xu and W. Bruce Croft. 1998. Corpus-based stemming using cooccurrence of word variants. ACM Transactions on Information Systems, 16(1):61–81. ↗
  32. David M. Zajic, Bonnie J. Dorr, and Rich Schwartz. 2002. Automatic headline generation for newspaper stories. In Proceedings of the ACL-2002 Workshop on Text Summarization, Philadelphia, PA. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/habash-2003-categorial-variation-database-english/