Elamur Research⌕
☰
home / papers / 2025 / Non Idiomatic Conventionalised Expressions: A New Pain in the Neck?

Non Idiomatic Conventionalised Expressions: A New Pain in the Neck?

Summary

The paper argues that NLP has overlooked a class of fixed phrases such as "as far as I know" or "a moment or two." These phrases can be read literally, yet their literal reading misses part of what they convey. The authors call them Non-Idiomatic Conventionalised Expressions (NICEs) and place them between fully idiomatic and fully compositional multiword expressions.

To find candidates, they combine two signals from the British National Corpus (BNC):

  • Association: greedy PMI-based merging of word pairs, followed by MDL pruning.
  • Non-compositionality: a word2vec-based score that ranks how far each candidate departs from its parts.

After excluding named entities, compounds and idioms, three linguists screened the top-ranked candidates by hand.

Two substitution-based measures, naturalness and semantic shift, show that NICEs behave differently from ordinary phrase types. A small study with Google Translate finds that these expressions are often rendered literally or only partially, especially into Hindi and Telugu. The authors conclude that NICEs deserve treatment as a separate phenomenon in NLP.

Key points

  • Candidate extraction: PMI merging with threshold θ = 10, then MDL-based pruning. The ranking score is the cosine distance between a candidate's word2vec vector and the normalised sum of its component vectors. Contextualised embeddings (BERT) failed to reflect compositionality.
  • Manual filtering: of the top 300 ranked candidates, all three annotators agreed on 90 as NICEs. The data and MT evaluations are released at github.com/neshkatrapati/nice-data.
  • Substitution tests: NICEs were compared against random Adjective-Noun pairs, Verb-Determiner-Noun combinations and random trigrams. The naturalness measure is based on corpus counts of nearest-neighbour substitutions. The semantic shift measure uses SBERT sentence embeddings. NICEs had lower means across all quartiles on both measures.
  • Machine translation: 220 BNC sentences were translated with Google Translate and rated on a 0–2 scale, with one annotator per language. NICE translations were judged acceptable in 89.7% of cases for French, 59% for Hindi and 51.3% for Telugu. Even in French, 24 of the 90 NICEs had less-than-perfect scores.
  • Limitations: the authors acknowledge a small, non-exhaustive expression set, few annotators and limited language coverage.

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

Cite

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

@inproceedings{katrapati2025non,
  title = {{Non Idiomatic Conventionalised Expressions: A New Pain in the Neck?}},
  author = {Katrapati, Ganesh and Shrivastava, Manish},
  booktitle = {Proceedings of the 39th Pacific Asia Conference on Language, Information and Computation},
  year = {2025},
  url = {https://aclanthology.org/2025.paclic-1.42/}
}

References

49 references, 12 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Timothy Baldwin, Colin Bannard, Takaaki Tanaka, and Dominic Widdows. 2003. An empirical model of multiword expression decomposability. pages 89– 96. Association for Computational Linguistics. find our card
  2. BNC. 2007. British national corpus, XML edition. Oxford Text Archive. find
  3. Gerlof Bouma. 2009. Normalized (pointwise) mutual information in collocation extraction. find
  4. Afsaneh Fazly, Paul Cook, and Suzanne Stevenson. 2009b. Unsupervised type and token identification of idiomatic expressions. Computational Linguistics, 35(1):61–103. find our card
  5. Afsaneh Fazly and Suzanne Stevenson. 2006. Automatically constructing a lexicon of verb phrase idiomatic combinations. find our card
  6. EACL 2006 - 11th Conference of the European Chapter of the Association for Computational Linguistics, Proceedings of the Conference, pages 337–344. find
  7. Nicoletta Calzolari, Charles J. Fillmore, Ralph Grishman, Nancy Ide, Alessandro Lenci, Catherine MacLeod, and Antonio Zampolli. 2002. Towards best practice for multiword expressions in computational lexicons. In Proceedings of the Third International Conference on Language Resources and Evaluation (… find
  8. Hessel Haagsma, Johan Bos, and Malvina Nissim. 2020. MAGPIE: A large corpus of potentially idiomatic expressions. find our card
  9. LREC 2020 - 12th International Conference on Language Resources and Evaluation, Conference Proceedings, (May):279–287. find
  10. Kenneth Ward Church and Patrick Hanks. 1989. Word association norms, mutual information, and lexicography. 16(1):76–83. find our card
  11. Ken Hyland. 2008. As can be seen: Lexical bundles and disciplinary variation. English for Specific Purposes, 27(1):4–21. find
  12. Jean Pierre Colson. 2017. The idiomsearch experiment: extracting phraseology from a probabilistic network of constructions. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 10596 LNAI:16–28. find
  13. Ray Jackendoff. 1996. guage faculty. find
  14. Silvio Cordeiro, Carlos Ramisch, Marco A. Idiart, and Aline Villavicencio. 2016. Predicting the compositionality of nominal compounds: Giving word embeddings a hard time. In Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 1986– 1… find
  15. Florian Coulmas. 1979. On the sociolinguistic relevance of routine formulae. Journal of Pragmatics, 3(3-4):239–266. find
  16. A. P. Cowie. 1993. Oxford Advanced Learners Dictionary of Current English, 5th edition. Oxford University Press, Oxford. find
  17. Mathias Creutz and Krista Lagus. 2002. Unsupervised discovery of morphemes. pages 21–30. find
  18. Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. Bert: Pre-training of deep bidirectional transformers for language understanding. In NAACL. find
  19. Amanda Edmonds. 2014. Conventional expressions. Studies in Second Language Acquisition, 36(1):69– 99. find
  20. Afsaneh Fazly, Paul Cook, and Suzanne Stevenson. 2009a. Unsupervised type and token identification of idiomatic expressions. Computational Linguistics, 35(1):61–103. find our card
  21. Hessel Haagsma, Malvina Nissim, and Johan Bos. 2019. Casting a Wide Net: Robust Extraction of Potentially Idiomatic Expressions. pages 1–34. find
  22. The architecture of the lan-Graham Katz and Eugenie Giesbrecht. 2006. Automatic Identification of Non-compositional Multiword Expressions Using Latent Semantic Analysis. MWE ’06, pages 12–19, Stroudsburg, PA, USA. Association for Computational Linguistics. find
  23. Douwe Kiela and Stephen Clark. 2013. Detecting compositionality of multi-word expressions using nearest neighbours in vector space models. In Proceedings of the 2013 Conference on Empirical Methods in Natural Language Processing, pages 1427–1432, Seattle, Washington, USA. Association for Computation… find
  24. Anoop Kunchukuttan, Pratik Mehta, and Pushpak Bhattacharyya. 2018. The iit bombay english–hindi parallel corpus. In Proceedings of the Eleventh International Conference on Language Resources and Evaluation (LREC 2018), Miyazaki, Japan. European Language Resources Association (ELRA). find our card
  25. Ronald Langacker. 2008. Cognitive Grammar: A Basic Introduction, volume 12. Oxford University Press. find
  26. Hang Li and Naoki Abe. 1992. Clustering Words with MDL Priciple. Journal of Natural Language Processing, 4(2):71–88. find
  27. Tomas Mikolov, Kai Chen, Greg Corrado, and Jeffrey Dean. 2013. Efficient estimation of word representations in vector space. arXiv preprint. find
  28. Rosamund Moon. 1998. Fixed Expressions and Idioms in English: A Corpus-Based Approach. Oxford University Press, Oxford. find
  29. Rosamund Moon. 2002. Fixed Expressions and Idioms in English: A Corpus-Based Approach (review), volume 78. find
  30. Anoop Nandakumar, Bahar Salehi, and Timothy Baldwin. 2018. How well can we predict multiword expression compositionality using embedding-based In Proceedings of the Australasian methods? Language Technology Association Workshop 2018, pages 73–81. ALTA. find
  31. Geoffrey Nunberg, Ivan A. Sag, and Thomas Wasow. 1994. Idioms. Language, 70(3):491–538. find
  32. Giancarlo Salton, Robert Ross, and John Kelleher. 2017. Idiom type identification with smoothed lexical features and a maximum margin classifier. In Proceedings of the International Conference Recent Advances in Natural Language Processing, RANLP 2017, pages 642–651, Varna, Bulgaria. INCOMA Ltd. find
  33. Caroline Sporleder and Linlin Li. 2009a. Unsupervised recognition of literal and non-literal use of idiomatic expressions. In Proceedings of the 12th Conference of the European Chapter of the ACL (EACL 2009), pages 754–762, Athens, Greece. Association for Computational Linguistics. find our card
  34. Pavel Pecina and Pavel Schlesinger. 2006. Combining association measures for collocation extraction. In Proceedings of the COLING/ACL 2006 Main Conference Poster Sessions, pages 651–658, Sydney, Australia. Association for Computational Linguistics. find
  35. Caroline Sporleder and Linlin Li. 2009b. Unsupervised recognition of literal and non-literal use of idiomatic expressions. In Proceedings of the 12th Conference of the European Chapter of the ACL (EACL 2009), pages 754–762, Athens, Greece. Association for Computational Linguistics. find our card
  36. Thomas Pickard. 2020. Comparing word2vec and glove for automatic measurement of mwe compositionality. In Proceedings of the Joint Workshop on Multiword Expressions and Electronic Lexicons (MWE+ELex 2020), pages 95–100, Barcelona, Spain (Online). find
  37. Caroline Sporleder, Linlin Li, Philip John Gorinski, and Xaver Koch. 2010. Idioms in context: The IDIX corpus. Proceedings of the 7th International Conference on Language Resources and Evaluation, LREC 2010, pages 639–646. find
  38. Gowtham Ramesh, Sumanth Doddapaneni, Aravinth Bheemaraj, Mayank Jobanputra, Raghavan AK, Ajitesh Sharma, Sujit Sahoo, Harshita Diddee, Mahalakshmi J, Divyanshu Kakwani, Navneet Kumar, Aswin Pradeep, Srihari Nagaraj, Kumar Deepak, Vivek Raghavan, Anoop Kunchukuttan, Pratyush Kumar, and Mitesh Shantad… find our card
  39. Andrea Tyler. 2005. Cognitive grammar. Studies in Second Language Acquisition, 27:650 – 651. find
  40. A Wray. 2000. Formulaic sequences in second language teaching: principle and practice. Applied Linguistics, 21(4):463–489. find
  41. Ke Wu, Jiangsheng Yu, Hanpin Wang, and Fei Cheng. 2010. Unsupervised text pattern learning using minimum description length. 2010 4th International Universal Communication Symposium, IUCS 2010 - Proceedings, pages 161–166. find
  42. Stefanie Wulff. 2013. Words and idioms. find
  43. Siva Reddy, Diana McCarthy, and Suresh Manandhar. 2011. An Empirical Study on Compositionality in Compound Nouns. find our card
  44. IJCNLP 2011 - Proceedings of the 5th International Joint Conference on Natural Language Processing, pages 210–218. find
  45. Nils Reimers and Iryna Gurevych. 2019. Sentence-bert: Sentence embeddings using siamese bert-networks. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing. Association for Computational Linguistics. find
  46. Ivan A. Sag, Timothy Baldwin, Francis Bond, Ann A. Copestake, and Dan Flickinger. 2002. Multiword expressions: A pain in the neck for nlp. In CICLing. find our card
  47. Bahar Salehi, Paul Cook, and Timothy Baldwin. 2015. A word embedding approach to predicting the compositionality of multiword expressions. find
  48. NAACL HLT 2015 - 2015 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Proceedings of the Conference, pages 977–983. find
  49. Carlos A. Yorio. 1989. Idiomaticity as an indicator of second language proficiency, page 5572. Cambridge University Press. find

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/katrapati-2025-non-idiomatic-conventionalised-expressions-new/