---
title: "Are idioms surprising? — Philipp 2023"
description: "The authors ask whether information theory can help separate idiomatic from literal uses of English verb-noun constructions (VNC), such as blow whistle…"
authors: ["J. Nathanael Philipp", "Michael Richter", "Erik Daas", "Max Kölbl"]
year: 2023
venue: "Proceedings of the 19th Conference on Natural Language Processing (KONVENS 2023)"
source: "https://aclanthology.org/2023.konvens-main.15/"
licence: "CC BY 4.0 (ACL Anthology)"
licence_url: "https://creativecommons.org/licenses/by/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/papers/philipp-2023-are-idioms-surprising/"
---
[Elamur Research](https://md.research.elamur.ai/) · [this page in HTML](https://research.elamur.ai/free/papers/philipp-2023-are-idioms-surprising/)

[home](https://md.research.elamur.ai/) / [papers](https://md.research.elamur.ai/free/papers/) / [2023](https://md.research.elamur.ai/free/papers/2023/) / Are idioms surprising?

# Are idioms surprising?

**Authors:** J. Nathanael Philipp, Michael Richter, Erik Daas, Max Kölbl · **Year:** 2023 · **Venue:** Proceedings of the 19th Conference on Natural Language Processing (KONVENS 2023) · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/2023.konvens-main.15/) · [PDF](https://aclanthology.org/2023.konvens-main.15.pdf)

**Citations:** 3 in our library, of the 140673 papers whose bibliographies we read

## Summary
The authors ask whether information theory can help separate idiomatic from literal uses of English verb-noun constructions (VNC), such as *blow whistle* or *take heart*. They start from the idea that idioms break semantic expectations and should therefore produce more surprisal and more uneven information flow. They used 1,997 idiomatic and 535 literal sentences from Peng et al. (2018). To score each word, they applied the Topic Context Model (TCM), which estimates semantic surprisal from LDA topic distributions. The LDA was trained on a reference set of Wortschatz Leipzig news and Wikipedia sentences. Information flow was measured with local Uniform Information Density (UID_LOCAL), and groups were compared using Welch tests and Cohen's d.

Across whole sentences, idioms and literals differed significantly, but the effect was tiny. Unexpectedly, idioms looked like the reference corpus, while literals were the group that departed from it. When only the VNC tokens were examined, the effect grew and idioms showed smaller surprisal jumps. The authors conclude that semantic surprisal can discriminate idioms from literals, and that everyday language may be more idiomatic than assumed.

## Key points
- **Data:** VNC sentences come from BNC, COCA, COHA and GloWbE, originally compiled by Peng et al. (2018). Idiomaticity is treated as a binary label rather than a gradual property.
- **Model setup:** TCM uses an LDA with 100 topics. The reference corpus combines English news (2020) and Wikipedia (2016) from Wortschatz Leipzig, with 1M sentences each.
- **Sentence-level results:**
  - Idioms vs. literals: t = 2.19, p = .029, Cohen's d = .022.
  - Reference vs. literals: significant (p = .025).
  - Reference vs. idioms: not significant (p = .69).
- **VNC-only analysis:** Restricting the comparison to the 12 VNC types (793 idiom vs. 637 literal surprisal values) gave t = −1.955, p = .05, d = .104. Idioms showed UID_LOCAL values nearer to zero, meaning more uniform information flow.
- **Limitations:** The reference corpora contain only isolated sentences, so TCM cannot draw on wider document context. The study covers English only, and the authors suggest longer texts, larger corpora and other languages as future work.

*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](https://md.research.elamur.ai/free/shelves/linguistics/)

## Cite

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

```bibtex
@inproceedings{philipp2023are,
  title = {{Are idioms surprising?}},
  author = {Philipp, J. Nathanael and Richter, Michael and Daas, Erik and Kölbl, Max},
  booktitle = {Proceedings of the 19th Conference on Natural Language Processing (KONVENS 2023)},
  year = {2023},
  url = {https://aclanthology.org/2023.konvens-main.15/}
}
```

- [← previous Are Frequent Phrases Directly Retrieved like Idioms? An Investigation with Self-Paced Reading and Language Models](https://md.research.elamur.ai/free/papers/rambelli-2023-are-frequent-phrases-directly-retrieved/)
- [next → Are Large Language Models Robust Coreference Resolvers?](https://md.research.elamur.ai/free/papers/le-2023-large-language-models-robust-coreference/)
- [↑ back to papers of 2023](https://md.research.elamur.ai/free/papers/2023/)

## References

28 references, 5 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. T Florian Jaeger and Roger P Levy. 2007. Speakers optimize information density through syntactic reduction. In Advances in neural information processing systems, pages 849–856. [find](https://scholar.google.com/scholar?q=T+Florian+Jaeger+and+Roger+P+Levy%2E+2007%2E+Speakers+optimize+information+density+through+syntactic+reduction%2E+In+Advances+in+neural+information+processing+systems%2C+pages+849%E2%80%93856%2E)
2. Martijn Bentum. 2021. Listening with great expectations: A study of predictive natural speech processing. Ph.D. thesis, [Sl]:[Sn]. [find](https://scholar.google.com/scholar?q=Martijn+Bentum%2E+2021%2E+Listening+with+great+expectations%3A+A+study+of+predictive+natural+speech+processing%2E+Ph%2ED%2E+thesis%2C+%5BSl%5D%3A%5BSn%5D%2E)
3. David M. Blei, Andrew Y. Ng, and Michael I. Jordan. 2003. Latent dirichlet allocation. Journal of machine Learning research, 3(Jan):993–1022. [find](https://scholar.google.com/scholar?q=David+M%2E+Blei%2C+Andrew+Y%2E+Ng%2C+and+Michael+I%2E+Jordan%2E+2003%2E+Latent+dirichlet+allocation%2E+Journal+of+machine+Learning+research%2C+3%28Jan%29%3A993%E2%80%931022%2E) [our card](https://md.research.elamur.ai/free/papers/blei-2003-latent-dirichlet-allocation/)
4. Jacob Cohen and Jacob Willem Cohen. 1988. Statistical power analysis for the behavioral sciences, 2. ed. edition. Erlbaum, Hillsdale, NJ [u.a.]. Lit-eraturverz. S. 553 - 558. [find](https://scholar.google.com/scholar?q=Jacob+Cohen+and+Jacob+Willem+Cohen%2E+1988%2E+Statistical+power+analysis+for+the+behavioral+sciences%2C+2%2E+ed%2E+edition%2E+Erlbaum%2C+Hillsdale%2C+NJ+%5Bu%2Ea%2E%5D%2E+Lit%2Deraturverz%2E+S%2E+553+%2D+558%2E)
5. Michael Xavier Collins. 2014. Information density and dependency length as complementary cognitive models. Journal of psycholinguistic research, 43(5):651–681. [find](https://scholar.google.com/scholar?q=Michael+Xavier+Collins%2E+2014%2E+Information+density+and+dependency+length+as+complementary+cognitive+models%2E+Journal+of+psycholinguistic+research%2C+43%285%29%3A651%E2%80%93681%2E)
6. Paul Cook, Afsaneh Fazly, and Suzanne Stevenson. 2008. The vnc-tokens dataset. In Proceedings of the LREC Workshop Towards a Shared Task for Multiword Expressions (MWE 2008), pages 19–22. Citeseer. [find](https://scholar.google.com/scholar?q=Paul+Cook%2C+Afsaneh+Fazly%2C+and+Suzanne+Stevenson%2E+2008%2E+The+vnc%2Dtokens+dataset%2E+In+Proceedings+of+the+LREC+Workshop+Towards+a+Shared+Task+for+Multiword+Expressions+%28MWE+2008%29%2C+pages+19%E2%80%9322%2E+Citeseer%2E)
7. Katherine A DeLong, Thomas P Urbach, and Marta Kutas. 2005. Probabilistic word pre-activation during language comprehension inferred from electrical brain activity. Nature neuroscience, 8(8):1117–1121. [find](https://scholar.google.com/scholar?q=Katherine+A+DeLong%2C+Thomas+P+Urbach%2C+and+Marta+Kutas%2E+2005%2E+Probabilistic+word+pre%2Dactivation+during+language+comprehension+inferred+from+electrical+brain+activity%2E+Nature+neuroscience%2C+8%288%29%3A1117%E2%80%931121)
8. Ayush Jain, Vishal Singh, Sidharth Ranjan, Rajakrishnan Rajkumar, and Sumeet Agarwal. 2018. Uniform information density effects on syntactic choice in hindi. In Proceedings of the Workshop on Linguistic Complexity and Natural Language Processing, pages 38–48. [find](https://scholar.google.com/scholar?q=Ayush+Jain%2C+Vishal+Singh%2C+Sidharth+Ranjan%2C+Rajakrishnan+Rajkumar%2C+and+Sumeet+Agarwal%2E+2018%2E+Uniform+information+density+effects+on+syntactic+choice+in+hindi%2E+In+Proceedings+of+the+Workshop+on+Linguist)
9. Max Kölbl, Yuki Kyogoku, J. Nathanael Philipp, Michael Richter, Clemens Rietdorf, and Tariq Yousef. 2020. Keyword Extraction in German: Information-theory vs. Deep Learning. In Proceedings of the 12th International Conference on Agents and Artificial Intelligence - Volume 1: NLPinAI,, pages 459–464.… [find](https://scholar.google.com/scholar?q=Max+K%C3%B6lbl%2C+Yuki+Kyogoku%2C+J%2E+Nathanael+Philipp%2C+Michael+Richter%2C+Clemens+Rietdorf%2C+and+Tariq+Yousef%2E+2020%2E+Keyword+Extraction+in+German%3A+Information%2Dtheory+vs%2E+Deep+Learning%2E+In+Proceedings+of+the+12th)
10. Max Kölbl, Yuki Kyogoku, J. Nathanael Philipp, Michael Richter, Clemens Rietdorf, and Tariq Yousef. 2021. The semantic level of shannon information: Are highly informative words good keywords? a study on german. In Roussanka Loukanova, editor, Natural Language Processing in Artificial Intelligence -… [find](https://scholar.google.com/scholar?q=Max+K%C3%B6lbl%2C+Yuki+Kyogoku%2C+J%2E+Nathanael+Philipp%2C+Michael+Richter%2C+Clemens+Rietdorf%2C+and+Tariq+Yousef%2E+2021%2E+The+semantic+level+of+shannon+information%3A+Are+highly+informative+words+good+keywords%3F+a+study)
11. M Teresa Espinal and Jaume Mateu. 2019. Idioms and phraseology. In Oxford Research Encyclopedia of Linguistics. [find](https://scholar.google.com/scholar?q=M+Teresa+Espinal+and+Jaume+Mateu%2E+2019%2E+Idioms+and+phraseology%2E+In+Oxford+Research+Encyclopedia+of+Linguistics%2E)
12. Roger Levy. 2008. Expectation-based syntactic comprehension. Cognition, 106(3):1126–1177. [find](https://scholar.google.com/scholar?q=Roger+Levy%2E+2008%2E+Expectation%2Dbased+syntactic+comprehension%2E+Cognition%2C+106%283%29%3A1126%E2%80%931177%2E)
13. Afsaneh Fazly, Paul Cook, and Suzanne Stevenson. 2009. Unsupervised type and token identification of idiomatic expressions. Computational Linguistics, 35(1):61–103. [find](https://scholar.google.com/scholar?q=Afsaneh+Fazly%2C+Paul+Cook%2C+and+Suzanne+Stevenson%2E+2009%2E+Unsupervised+type+and+token+identification+of+idiomatic+expressions%2E+Computational+Linguistics%2C+35%281%29%3A61%E2%80%93103%2E) [our card](https://md.research.elamur.ai/free/papers/fazly-2009-unsupervised-type-token-identification-idiomatic/)
14. Roger Levy and T Florian Jaeger. 2007. Speakers optimize information density through syntactic reduction. Advances in neural information processing systems, 19:849. [find](https://scholar.google.com/scholar?q=Roger+Levy+and+T+Florian+Jaeger%2E+2007%2E+Speakers+optimize+information+density+through+syntactic+reduction%2E+Advances+in+neural+information+processing+systems%2C+19%3A849%2E)
15. Geoffrey Nunberg, Ivan A Sag, and Thomas Wasow. 1994. Idioms. Language, 70(3):491–538. [find](https://scholar.google.com/scholar?q=Geoffrey+Nunberg%2C+Ivan+A+Sag%2C+and+Thomas+Wasow%2E+1994%2E+Idioms%2E+Language%2C+70%283%29%3A491%E2%80%93538%2E)
16. Anna Feldman and Jing Peng. 2013. Automatic detection of idiomatic clauses. In Computational Linguistics and Intelligent Text Processing: 14th International Conference, CICLing 2013, Samos, Greece, March 24-30, 2013, Proceedings, Part I 14, pages 435–446. Springer. [find](https://scholar.google.com/scholar?q=Anna+Feldman+and+Jing+Peng%2E+2013%2E+Automatic+detection+of+idiomatic+clauses%2E+In+Computational+Linguistics+and+Intelligent+Text+Processing%3A+14th+International+Conference%2C+CICLing+2013%2C+Samos%2C+Greece%2C+Ma)
17. Jing Peng, Anna Feldman, and Ekaterina Vylomova. 2018. Classifying idiomatic and literal expressions using topic models and intensity of emotions. arXiv preprint arXiv:1802.09961. — [arXiv:1802.09961](https://arxiv.org/abs/1802.09961) [find](https://scholar.google.com/scholar?q=Jing+Peng%2C+Anna+Feldman%2C+and+Ekaterina+Vylomova%2E+2018%2E+Classifying+idiomatic+and+literal+expressions+using+topic+models+and+intensity+of+emotions%2E+arXiv+preprint+arXiv%3A1802%2E09961%2E+%E2%80%94+%5BarXiv%3A1802%2E09961%5D)
18. Dirk Goldhahn, Thomas Eckart, Uwe Quasthoff, et al. 2012. Building large monolingual dictionaries at the leipzig corpora collection: From 100 to 200 languages. In LREC, volume 29, pages 31–43. [find](https://scholar.google.com/scholar?q=Dirk+Goldhahn%2C+Thomas+Eckart%2C+Uwe+Quasthoff%2C+et+al%2E+2012%2E+Building+large+monolingual+dictionaries+at+the+leipzig+corpora+collection%3A+From+100+to+200+languages%2E+In+LREC%2C+volume+29%2C+pages+31%E2%80%9343%2E) [our card](https://md.research.elamur.ai/free/papers/goldhahn-2012-building-large-monolingual-dictionaries-leipzig/)
19. John Hale. 2001. A probabilistic earley parser as a psycholinguistic model. In Proceedings of the second meeting of the North American Chapter of the Association for Computational Linguistics on Language technologies, pages 1–8. Association for Computational Linguistics. [find](https://scholar.google.com/scholar?q=John+Hale%2E+2001%2E+A+probabilistic+earley+parser+as+a+psycholinguistic+model%2E+In+Proceedings+of+the+second+meeting+of+the+North+American+Chapter+of+the+Association+for+Computational+Linguistics+on+Langu) [our card](https://md.research.elamur.ai/free/papers/hale-2001-probabilistic-earley-parser-as-psycholinguistic/)
20. John Hale, David Lutz, Wen-Ming Luh, and Jonathan Brennan. 2015. Modeling fmri time courses with linguistic structure at various grain sizes. In Proceedings of the 6th workshop on cognitive modeling and computational linguistics, pages 89–97. [find](https://scholar.google.com/scholar?q=John+Hale%2C+David+Lutz%2C+Wen%2DMing+Luh%2C+and+Jonathan+Brennan%2E+2015%2E+Modeling+fmri+time+courses+with+linguistic+structure+at+various+grain+sizes%2E+In+Proceedings+of+the+6th+workshop+on+cognitive+modeling+a)
21. J. Nathanael Philipp, Max Kölbl, Erik Daas, Yuki Kyogoku, and Michael Richter. Perplexed by idioms. (in press). [find](https://scholar.google.com/scholar?q=J%2E+Nathanael+Philipp%2C+Max+K%C3%B6lbl%2C+Erik+Daas%2C+Yuki+Kyogoku%2C+and+Michael+Richter%2E+Perplexed+by+idioms%2E+%28in+press%29%2E)
22. J. Nathanael Philipp, Max Kölbl, Yuki Kyogoku, Tariq Yousef, and Michael Richter. 2022. One step beyond: Keyword extraction in german utilising surprisal from topic contexts. In Intelligent Computing, pages 774–786, Cham. Springer International Publishing. [find](https://scholar.google.com/scholar?q=J%2E+Nathanael+Philipp%2C+Max+K%C3%B6lbl%2C+Yuki+Kyogoku%2C+Tariq+Yousef%2C+and+Michael+Richter%2E+2022%2E+One+step+beyond%3A+Keyword+extraction+in+german+utilising+surprisal+from+topic+contexts%2E+In+Intelligent+Computing%2C)
23. Manali Pradhan, Jing Peng, Anna Feldman, and Bianca Wright. 2018. Idioms: Humans or machines, it’s all about context. In Computational Linguistics and Intelligent Text Processing: 18th International Conference, CICLing 2017, Budapest, Hungary, April 17–23, 2017, Revised Selected Papers, Part I 18, p… [find](https://scholar.google.com/scholar?q=Manali+Pradhan%2C+Jing+Peng%2C+Anna+Feldman%2C+and+Bianca+Wright%2E+2018%2E+Idioms%3A+Humans+or+machines%2C+it%E2%80%99s+all+about+context%2E+In+Computational+Linguistics+and+Intelligent+Text+Processing%3A+18th+International+C)
24. Tatjana Scheffler, Michael Richter, and Roeland van Hout. 2023. Tracing and classifying german intensifiers via information theory. Language Sciences, 96:101535. [find](https://scholar.google.com/scholar?q=Tatjana+Scheffler%2C+Michael+Richter%2C+and+Roeland+van+Hout%2E+2023%2E+Tracing+and+classifying+german+intensifiers+via+information+theory%2E+Language+Sciences%2C+96%3A101535%2E)
25. Claude Elwood Shannon. 1948. A mathematical theory of communication. The Bell system technical journal, 27(3):379–423. [find](https://scholar.google.com/scholar?q=Claude+Elwood+Shannon%2E+1948%2E+A+mathematical+theory+of+communication%2E+The+Bell+system+technical+journal%2C+27%283%29%3A379%E2%80%93423%2E)
26. Myron Tribus. 1961. Information theory as the basis for thermostatics and thermodynamics. [find](https://scholar.google.com/scholar?q=Myron+Tribus%2E+1961%2E+Information+theory+as+the+basis+for+thermostatics+and+thermodynamics%2E)
27. Bernard L Welch. 1938. The significance of the difference between two means when the population variances are unequal. Biometrika, 29(3/4):350– 362. [find](https://scholar.google.com/scholar?q=Bernard+L+Welch%2E+1938%2E+The+significance+of+the+difference+between+two+means+when+the+population+variances+are+unequal%2E+Biometrika%2C+29%283%2F4%29%3A350%E2%80%93+362%2E)
28. Ziheng Zeng and Suma Bhat. 2021. Idiomatic expression identification using semantic compatibility. Transactions of the Association for Computational Linguistics, 9:1546–1562. [find](https://scholar.google.com/scholar?q=Ziheng+Zeng+and+Suma+Bhat%2E+2021%2E+Idiomatic+expression+identification+using+semantic+compatibility%2E+Transactions+of+the+Association+for+Computational+Linguistics%2C+9%3A1546%E2%80%931562%2E) [our card](https://md.research.elamur.ai/free/papers/zeng-2021-idiomatic-expression-identification-semantic-compatibility/)