---
title: "Languages Through the Looking Glass of BPE Compression (2023)"
description: "This study asks which kinds of structure BPE exploits when compressing text and whether these… 40 references, cited by 12 papers here."
authors: ["Ximena Gutierrez-Vasques", "Christian Bentz", "Tanja Samardžić"]
year: 2023
venue: "Computational Linguistics"
source: "https://aclanthology.org/2023.cl-4.5/"
doi: "10.1162/coli_a_00489"
licence: "CC BY-NC-ND 4.0 (Computational Linguistics, MIT Press)"
licence_url: "https://creativecommons.org/licenses/by-nc-nd/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/vasques-2023-languages-through-looking-glass-bpe/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/vasques-2023-languages-through-looking-glass-bpe/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2023](https://md.research.elamur.ai/free/en/papers/2023/page/9/) / Languages Through the Looking Glass of BPE Compression

# Languages Through the Looking Glass of BPE Compression

**Authors:** [Ximena Gutierrez-Vasques](https://md.research.elamur.ai/free/en/authors/ximena-gutierrez-vasques/), Christian Bentz and [Tanja Samardžić](https://md.research.elamur.ai/free/en/authors/tanja-samardzic/) · **Year:** 2023 · **Venue:** [Computational Linguistics](https://md.research.elamur.ai/free/en/venues/cl/) ([Computational Linguistics 2023](https://md.research.elamur.ai/free/en/venues/cl/2023/)) · **Publisher:** [The MIT Press](https://md.research.elamur.ai/free/en/publishers/the-mit-press/) · **Licence:** [CC BY-NC-ND 4.0 (Computational Linguistics, MIT Press)](https://creativecommons.org/licenses/by-nc-nd/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/2023.cl-4.5/) · [PDF](https://aclanthology.org/2023.cl-4.5.pdf) · [DOI 10.1162/coli\_a\_00489](https://doi.org/10.1162/coli_a_00489) · [OpenAlex](https://openalex.org/W4383533102)

**Citations:** 14 citations (OpenAlex, 2026-10-14) · [12](https://md.research.elamur.ai/free/en/papers/vasques-2023-languages-through-looking-glass-bpe/cited-by/) in our library, of the 149,439 papers whose bibliographies we read

## Summary
BPE subword tokenization is often dismissed as a statistical preprocessing step with no linguistic value. This study asks which kinds of structure BPE exploits when compressing text and whether these differ systematically between languages.

The authors apply BPE to 47 typologically diverse languages. The main dataset is the Parallel Bible Corpus (PBC), and JW300 and UDHR serve as robustness checks. For each subword created during the early merges, they score two properties inspired by morphological productivity:
- **Productivity:** how many distinct word types contain the subword.
- **Idiosyncrasy:** whether the subword appears in only a few word types that are very frequent.

These scores place each language as a vector in a BPE-based productivity space, which is then compared with WALS features. Strongly inflecting languages favour productive, affix-like subwords early on. Weakly inflecting languages rely more on idiosyncratic units such as frequent whole words. Both strategies compress text effectively.

The resulting vectors reflect known typological distinctions. This shows that typological knowledge can be induced from raw, unannotated text, which is useful for quantitative typology and for multilingual NLP, where database features are often incomplete.

## Key points
- **Data:** 47 languages from three parallel corpora (PBC, JW300, UDHR). For languages shared with PBC, vectors of the same language in different corpora lie closer together (Euclidean distance) than the overall average distance, despite differences in corpus size and register.
- **Idiosyncrasy measure:** it is the corpus frequency divided by the number of word types containing the subword. In English BPE, `the</w>` occurs in 3 types and scores 486.33, while `ed</w>` occurs in 270 types and scores 3.38.
- **Number of merges:** BPE spaces were built at 10, 50, 200 and 1000 merges. The earliest merges shape the language layout most, and the space barely changes after about 200 merges. WordPiece spaces change considerably more across merges.
- **Agreement with WALS:** fit was measured with silhouette coefficients for BPE200, which range from about 0.73 (Halh Mongolian, Russian) down to about −0.76 (Swahili).
- **Requirements and extras:** the method needs neither annotations nor external linguistic resources. The authors also tested a transliterated version of Korean. The work was funded by SNSF grant 176,305.

*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 Zurich](https://md.research.elamur.ai/free/en/universities/university-of-zurich/) — University, Zurich, Switzerland · 78 works · 12,260 citations
- [University of Tübingen](https://md.research.elamur.ai/free/en/universities/university-of-tubingen/) — University, Tübingen, Germany · 75 works · 41,081 citations

📚 Shelves (4): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 16,667 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 16,194 works · [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 16,039 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 15,891 works

🧭 Topics (1): [✂️ Tokenization & segmentation](https://md.research.elamur.ai/free/en/topics/tokenization/) · 119 works

## Cite

Cite as: Ximena Gutierrez-Vasques, Christian Bentz and Tanja Samardžić. 2023. Languages Through the Looking Glass of BPE Compression. Computational Linguistics.

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

```bibtex
@article{gutierrezvasques2023languages,
  title = {{Languages Through the Looking Glass of BPE Compression}},
  author = {Ximena Gutierrez-Vasques, Christian Bentz and Tanja Samardžić},
  journal = {Computational Linguistics},
  year = {2023},
  doi = {10.1162/coli\_a\_00489},
  url = {https://doi.org/10.1162/coli_a_00489}
}
```

- [← previous IfQA: A Dataset for Open-domain Question Answering under Counterfactual Presuppositions](https://md.research.elamur.ai/free/en/papers/yu-2023-ifqa-dataset-open-domain-question/)
- [next → Large Language Models are Strong Zero-Shot Retriever](https://md.research.elamur.ai/free/en/papers/shen-2023-large-language-models-strong-zero/)
- [↑ back to papers of 2023](https://md.research.elamur.ai/free/en/papers/2023/page/9/)

## References

40 references, 8 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Adelaar, Willem. 2004. Quechua. In Geert Booij, Christian Lehmann, Joachim Mugdan, and Stavros Skopeteas, editors, Morphology: An International Handbook of Inflection and Word-formation, volume 2. Walter de Gruyter, Berlin, pages 1454–1463. Agić, Željko and Ivan Vulić. 2019. JW300: A wide-coverage… — [doi:10.18653/v1/p19](https://doi.org/10.18653/v1/p19) [find](https://scholar.google.com/scholar?q=Adelaar%2C+Willem%2E+2004%2E+Quechua%2E+In+Geert+Booij%2C+Christian+Lehmann%2C+Joachim+Mugdan%2C+and+Stavros+Skopeteas%2C+editors%2C+Morphology%3A+An+International+Handbook+of+Inflection+and+Word%2Dformation%2C+volume+2%2E+Wal)
2. Al Roumi, Fosca, Sébastien Marti, Liping Wang, Marie Amalric, and Stanislas Dehaene. 2021. Mental compression of spatial sequences in human working memory using numerical and geometrical primitives. Neuron, 109(16):2627–2639. https://doi.org/10.1016/j.neuron .2021.06.009, PubMed: 34228961 Amrhein, C… [↗](https://doi.org/10.1016/j.neuron)
3. Berg, Thomas. 2014. On the relationship between type and token frequency. Journal of Quantitative Linguistics, 21(3):199–222. https://doi.org/10.1080/09296174 .2014.911505 Bickel, Balthasar and Johanna Nichols. 2007. Inflectional morphology. In Timothy Shopen, editor, Language Typology and Syntactic… — [doi:10.1080/09296174](https://doi.org/10.1080/09296174) [↗](https://doi.org/10.1080/09296174)
4. Bickel, Balthasar, Johanna Nichols, Taras Zakharko, Alena Witzlack-Makarevich, Kristine Hildebrandt, Michael Rießler, Lennart Bierkandt, Fernando Zúñiga, and John B. Lowe. 2022. The autotyp database (v1.0.1). [find](https://scholar.google.com/scholar?q=Bickel%2C+Balthasar%2C+Johanna+Nichols%2C+Taras+Zakharko%2C+Alena+Witzlack%2DMakarevich%2C+Kristine+Hildebrandt%2C+Michael+Rie%C3%9Fler%2C+Lennart+Bierkandt%2C+Fernando+Z%C3%BA%C3%B1iga%2C+and+John+B%2E+Lowe%2E+2022%2E+The+autotyp+database+%28)
5. Bickel, Balthasar and Fernando Zúñiga. 2017. The word in polysynthetic languages: Phonological and syntactic challenges. In Michael Fortescue, Marianne Mithun, and Nicholas Evans, editors, The Oxford Handbook of Polysynthesis. Oxford University Press, Oxford, pages 158–185. https://doi.org/10.1093/o… [find](https://scholar.google.com/scholar?q=Bickel%2C+Balthasar+and+Fernando+Z%C3%BA%C3%B1iga%2E+2017%2E+The+word+in+polysynthetic+languages%3A+Phonological+and+syntactic+challenges%2E+In+Michael+Fortescue%2C+Marianne+Mithun%2C+and+Nicholas+Evans%2C+editors%2C+The+Oxford+)
6. Bjerva, Johannes, Yova Kementchedjhieva, Ryan Cotterell, and Isabelle Augenstein. 2019. A probabilistic generative model of linguistic typology. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1… [find](https://scholar.google.com/scholar?q=Bjerva%2C+Johannes%2C+Yova+Kementchedjhieva%2C+Ryan+Cotterell%2C+and+Isabelle+Augenstein%2E+2019%2E+A+probabilistic+generative+model+of+linguistic+typology%2E+In+Proceedings+of+the+2019+Conference+of+the+North+Amer)
7. SIGTYP 2020 shared task: Prediction of typological features. In Proceedings of the Second Workshop on Computational Research in Linguistic Typology, pages 1–11. https://doi.org/10.18653/v1/2020.sigtyp-1.1 Blevins, James P., Petar Milin, and Michael Ramscar. 2017. The Zipfian paradigm cell filling p… [↗](https://doi.org/10.18653/v1/2020.sigtyp-1.1)
8. Bostrom, Kaj and Greg Durrett. 2020. [Byte pair encoding is suboptimal for language model pretraining](https://md.research.elamur.ai/free/en/papers/bostrom-2020-byte-pair-encoding-is-suboptimal/). In Findings of the Association for Computational Linguistics: EMNLP 2020, pages 4617–4624. https://doi.org/10.18653/v1/2020 .findings-emnlp.414 Bruce, Les, et al. 1984. The Alamblak Language of Papu… [↗](https://doi.org/10.18653/v1/2020)
9. Bybee, Joan. 2003. Phonology and Language Use, volume 94. Cambridge University Press. [find](https://scholar.google.com/scholar?q=Bybee%2C+Joan%2E+2003%2E+Phonology+and+Language+Use%2C+volume+94%2E+Cambridge+University+Press%2E)
10. Bybee, Joan. 2010. [Language, Usage and Cognition](https://md.research.elamur.ai/free/en/papers/clark-2022-canine-pre-training-efficient-tokenization/). Cambridge University Press. https://doi.org/10.1017/CBO9780511750526 Choenni, Rochelle and Ekaterina Shutova. 2022. Investigating Language Relationships in Multilingual Sentence Encoders Through the Lens of Linguistic Typology. Computational Linguis… [↗](https://doi.org/10.1017/CBO9780511750526)
11. Creutz, Mathias and Krista Lagus. 2002. [Unsupervised discovery of morphemes](https://md.research.elamur.ai/free/en/papers/creutz-2002-unsupervised-discovery-morphemes/). In Proceedings of the ACL-02 Workshop on Morphological and Phonological Learning, pages 21–30. https://doi.org/10.3115/1118647.1118650 Cuskley, Christine, Francesca Colaiori, Claudio Castellano, Vittorio Loreto, Martina Pu… [↗](https://doi.org/10.3115/1118647.1118650)
12. The World’s Writing Systems. Oxford University Press, Oxford. Devlin, Jacob, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. [BERT: Pre-training of deep bidirectional transformers for language understanding](https://md.research.elamur.ai/free/en/papers/devlin-2019-bert-pre-training-deep-bidirectional/). In Proceedings of the 2019 Conference of the North American Chapter of the Associat…
13. Dixon, Robert M. W. and Alexandra Y. Aikhenvald. 2003. Word: A cross-linguistic Typology. Cambridge University Press, Cambridge. https://doi.org/10.1017/CBO9780511486241 Dixon, Robert M. W. 1988. A Grammar of Boumaa Fijian. University of Chicago Press, Chicago. [↗](https://doi.org/10.1017/CBO9780511486241)
14. Domingo, Miguel, Mercedes García-Martínez, Alexandre Helle, Francisco Casacuberta, and Manuel Herranz. 2023. How much does tokenization affect neural machine translation? In Computational Linguistics and Intelligent Text Processing, pages 545–554, Springer Nature Switzerland, Cham. [https://doi.org](https://doi.org) /… — [doi:10.1007/978-3-031-24337-0_38](https://doi.org/10.1007/978-3-031-24337-0_38) [↗](https://doi.org)
15. Dryer, Matthew S. and Martin Haspelmath. 2013. The World Atlas of Language Structures Online. Zenodo. https://doi.org/10 .5281/zenodo.7385533 Ehret, Katharina. 2016. An Information-Theoretic Approach to Language Complexity: Variation in Naturalistic Corpora. Ph.D. thesis, Albert-Ludwigs-Universität… [find](https://scholar.google.com/scholar?q=Dryer%2C+Matthew+S%2E+and+Martin+Haspelmath%2E+2013%2E+The+World+Atlas+of+Language+Structures+Online%2E+Zenodo%2E+https%3A%2F%2Fdoi%2Eorg%2F10+%2E5281%2Fzenodo%2E7385533+Ehret%2C+Katharina%2E+2016%2E+An+Information%2DTheoretic+Approach+)
16. Ehret, Katharina and Benedikt Szmrecsanyi. 2016. An information-theoretic approach to assess linguistic complexity. In Raffaela Baechler and Guido Seiler, editors, Complexity, Isolation and Variation. De Gruyter, Berlin. https://doi.org/10 ss.1515/9783110348965-004 Facundes, Sidi. 2014. Negation in… — [doi:10.1163/9789004257023_007](https://doi.org/10.1163/9789004257023_007) [find](https://scholar.google.com/scholar?q=Ehret%2C+Katharina+and+Benedikt+Szmrecsanyi%2E+2016%2E+An+information%2Dtheoretic+approach+to+assess+linguistic+complexity%2E+In+Raffaela+Baechler+and+Guido+Seiler%2C+editors%2C+Complexity%2C+Isolation+and+Variation%2E)
17. Hsu, and Stuart Semple. 2013. Compression as a universal principle of animal behavior. Cognitive Science, 37(8):1565–1578. https://doi.org/10.1111/cogs.12061, PubMed: 23895349 Fortescue, Michael. 1992. The development of morphophonemic complexity in eskimo languages. Acta Linguistica Hafniensia, 25… [↗](https://doi.org/10.1111/cogs.12061)
18. West Greenlandic. Croom Helm London. Gage, Philip. 1994. A new algorithm for data compression. The C Users Journal, 12(2):23–38. [find](https://scholar.google.com/scholar?q=West+Greenlandic%2E+Croom+Helm+London%2E+Gage%2C+Philip%2E+1994%2E+A+new+algorithm+for+data+compression%2E+The+C+Users+Journal%2C+12%282%29%3A23%E2%80%9338%2E)
19. Gallé, Matthias. 2019. Investigating the effectiveness of BPE: The power of shorter sequences. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP), pages 1375–1381. https://… [find](https://scholar.google.com/scholar?q=Gall%C3%A9%2C+Matthias%2E+2019%2E+Investigating+the+effectiveness+of+BPE%3A+The+power+of+shorter+sequences%2E+In+Proceedings+of+the+2019+Conference+on+Empirical+Methods+in+Natural+Language+Processing+and+the+9th+Int)
20. Göksel, Aslı and Celia Kerslake. 2004. Turkish: A Comprehensive Grammar. Routledge, London/New York. https://doi.org/10.4324/9780203340769 Goldsmith, John. 2001. Unsupervised learning of the morphology of a natural language. Computational Linguistics, 27(2):153–198. https://doi.org/10 .1162/0891201… — [doi:10.4324/9780203340769](https://doi.org/10.4324/9780203340769) [↗](https://doi.org/10.4324/9780203340769)
21. Greenberg, Joseph H. 1966. Language Universals: With special reference to feature hierarchies. [find](https://scholar.google.com/scholar?q=Greenberg%2C+Joseph+H%2E+1966%2E+Language+Universals%3A+With+special+reference+to+feature+hierarchies%2E)
22. Mouton, The Hague. Greenberg, Joseph H. 1960. A quantitative approach to the morphological typology of language. International Journal of American Linguistics, 26(3):178–194. https://doi.org/10.1086/464575 Grönroos, Stig Arne, Sami Virpioja, Peter Smit, and Mikko Kurimo. 2014. Morfessor FlatCat: An… — [doi:10.1086/464575](https://doi.org/10.1086/464575) [↗](https://doi.org/10.1086/464575)
23. Guerrero, Lilián. 2019. Supleción en yaqui y wixárika. Lingüística Mexicana, 1(2):119–140. [find](https://scholar.google.com/scholar?q=Guerrero%2C+Lili%C3%A1n%2E+2019%2E+Supleci%C3%B3n+en+yaqui+y+wix%C3%A1rika%2E+Ling%C3%BC%C3%ADstica+Mexicana%2C+1%282%29%3A119%E2%80%93140%2E)
24. Gutierrez-Vasques, Ximena, Christian Bentz, Olga Sozinova, and Tanja Samardzic. 2021. From characters to words: The turning point of BPE merges. In Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume, pages 3454–3468, Association f… [find](https://scholar.google.com/scholar?q=Gutierrez%2DVasques%2C+Ximena%2C+Christian+Bentz%2C+Olga+Sozinova%2C+and+Tanja+Samardzic%2E+2021%2E+From+characters+to+words%3A+The+turning+point+of+BPE+merges%2E+In+Proceedings+of+the+16th+Conference+of+the+European+C)
25. Understanding Morphology. Hodder Education. Hewitt, Brian George. 1995. Georgian: A Structural Reference Grammar, volume 2. John Benjamins Publishing. https://doi.org/10.1075/loall.2 Jenny, Mathias and San San Hnin Tun. 2016. Burmese: A Comprehensive Grammar. Routledge, New York/London. https://do… — [doi:10.1075/loall.2](https://doi.org/10.1075/loall.2) [↗](https://doi.org/10.1075/loall.2)
26. Kirby, Simon, Monica Tamariz, Hannah Cornish, and Kenny Smith. 2015. [Compression and communication in the cultural evolution of linguistic structure](https://md.research.elamur.ai/free/en/papers/lauscher-2020-zero-hero-limitations-zero-shot/). Cognition, 141:87–102. https://doi.org/10.1016/j.cognition.2015.03.016, PubMed: 25966840 Kriegeskorte, Nikolaus, Marieke Mur, and Peter A. Bandettini… [↗](https://doi.org/10.1016/j.cognition.2015.03.016)
27. Mager, Manuel, Arturo Oncevay, Elisabeth Mager, Katharina Kann, and Thang Vu. 2022. BPE vs. morphological segmentation: A case study on machine translation of four polysynthetic languages. In Findings of the Association for Computational Linguistics: ACL 2022, pages 961–971. https://doi.org/10.18653… [find](https://scholar.google.com/scholar?q=Mager%2C+Manuel%2C+Arturo+Oncevay%2C+Elisabeth+Mager%2C+Katharina+Kann%2C+and+Thang+Vu%2E+2022%2E+BPE+vs%2E+morphological+segmentation%3A+A+case+study+on+machine+translation+of+four+polysynthetic+languages%2E+In+Findings)
28. Mayer, Thomas, Bernhard Wälchli, Christian Rohrdantz, and Michael Hund. 2014. [From the extraction of continuous features in parallel texts to visual analytics of heterogeneous areal-typological datasets](https://md.research.elamur.ai/free/en/papers/mielke-2021-between-words-characters-brief-history/). In The Role of Functionally Oriented Computational Models, pages 13–38. https://doi.org/10.1075/…
29. Mielke, Sebastian J. and Jason Eisner. 2019. Spell once, summon anywhere: A two-level open-vocabulary language model. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 33, pages 6843–6850. https://doi.org/10.1609/aaai.v33i01 .33016843 Moran, Steven and Michael Cysouw. 2018. T… [↗](https://doi.org/10.1609/aaai.v33i01)
30. Moravcsik, Edith A. 2012. Introducing Language Typology. Cambridge University Press, Cambridge. https://doi.org/10 .1017/CBO9780511978876 Myung, I. J. 2001. Computational approaches to model evaluation. In International Encyclopedia of the Social & Behavioral Sciences. Pergamon, Oxford, pages 2453–2… [find](https://scholar.google.com/scholar?q=Moravcsik%2C+Edith+A%2E+2012%2E+Introducing+Language+Typology%2E+Cambridge+University+Press%2C+Cambridge%2E+https%3A%2F%2Fdoi%2Eorg%2F10+%2E1017%2FCBO9780511978876+Myung%2C+I%2E+J%2E+2001%2E+Computational+approaches+to+model+evaluatio)
31. Pelloni, Olga, Anastassia Shaitarova, Olga Sozinova, and Tanja Samardzic. 2022. Subword evenness (sue) as a predictor of cross-lingual transfer to low-resource languages. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing: Main Volume, pages 7428–7445. https://… [find](https://scholar.google.com/scholar?q=Pelloni%2C+Olga%2C+Anastassia+Shaitarova%2C+Olga+Sozinova%2C+and+Tanja+Samardzic%2E+2022%2E+Subword+evenness+%28sue%29+as+a+predictor+of+cross%2Dlingual+transfer+to+low%2Dresource+languages%2E+In+Proceedings+of+the+2022+Co)
32. Pires, Telmo, Eva Schlinger, and Dan Garrette. 2019. [How multilingual is multilingual BERT?](https://md.research.elamur.ai/free/en/papers/pires-2019-how-multilingual-is-multilingual-bert/) In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 4996–5001. https://doi.org/10.18653/v1/P19-1493 Ponti, Edoardo Maria, Helen O’Horan, Yevgeni Berzak, Ivan Vuli… [↗](https://doi.org/10.18653/v1/P19-1493)
33. Rathi, Neil, Michael Hahn, and Richard Futrell. 2021. An information-theoretic characterization of morphological fusion. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 10115–10120. https://doi.org/10.18653/v1/2021 .emnlp-main.793 Reynar, Jeffrey C.,… [↗](https://doi.org/10.18653/v1/2021)
34. Rissanen, J. 1978. [Modeling by shortest data description](https://md.research.elamur.ai/free/en/papers/rust-2021-how-good-is-your-tokenizer/). Automatica, 14(5):465–471. https://doi.org/10.1016/0005 -1098(78)90005-5 Rousseeuw, Peter J. 1987. Silhouettes: A graphical aid to the interpretation and validation of cluster analysis. Journal of Computational and Applied Mathematics, 20:53–… [↗](https://doi.org/10.1016/0005)
35. Harcourt, Brace. Schuster, Mike and Kaisuke Nakajima. 2012. Japanese and Korean voice search. In 2012 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pages 5149–5152. https://doi.org/10.1109/ICASSP.2012 .6289079 Sennrich, Rico, Barry Haddow, and Alexandra Birch. 2… — [doi:10.1109/icassp.2012](https://doi.org/10.1109/icassp.2012) [↗](https://doi.org/10.1109/ICASSP.2012)
36. Shapiro, Pamela and Kevin Duh. 2018. BPE and charCNNs for translation of morphology: A cross-lingual comparison and analysis. arXiv preprint arXiv:1809.01301. — [arXiv:1809.01301](https://arxiv.org/abs/1809.01301) [find](https://scholar.google.com/scholar?q=Shapiro%2C+Pamela+and+Kevin+Duh%2E+2018%2E+BPE+and+charCNNs+for+translation+of+morphology%3A+A+cross%2Dlingual+comparison+and+analysis%2E+arXiv+preprint+arXiv%3A1809%2E01301%2E+%E2%80%94+%5BarXiv%3A1809%2E01301%5D%28https%3A%2F%2Farxiv%2Eorg%2Fab)
37. Shimelman, Aviva. 2017. A Grammar of Yauyos Quechua. [find](https://scholar.google.com/scholar?q=Shimelman%2C+Aviva%2E+2017%2E+A+Grammar+of+Yauyos+Quechua%2E)
38. Language Science Press, Berlin. Stolz, Thomas. 2015. Chamorro inflection. In The Oxford Handbook of Inflection. https://doi.org/10.1093/oxfordhb/9780199591428.013.20 Storer, James A. and Thomas G. Szymanski. 1978. The macro model for data compression. In Proceedings of the Tenth Annual ACM Symposiu… — [doi:10.1093/oxfordhb](https://doi.org/10.1093/oxfordhb) [↗](https://doi.org/10.1093/oxfordhb/9780199591428.013.20)
39. John Wiley & Sons, Ltd. Tamariz, Mónica and Simon Kirby. 2015. Culture: Copying, compression, and conventionality. Cognitive Science, 39(1):171–183. https://doi.org/10 .1111/cogs.12144, PubMed: 25039798 Ullman, Michael T. 1999. Acceptability ratings of regular and irregular past-tense forms: Evidenc… [find](https://scholar.google.com/scholar?q=John+Wiley+%26+Sons%2C+Ltd%2E+Tamariz%2C+M%C3%B3nica+and+Simon+Kirby%2E+2015%2E+Culture%3A+Copying%2C+compression%2C+and+conventionality%2E+Cognitive+Science%2C+39%281%29%3A171%E2%80%93183%2E+https%3A%2F%2Fdoi%2Eorg%2F10+%2E1111%2Fcogs%2E12144%2C+PubMed%3A+250397)
40. Korean: A Comprehensive Grammar. Routledge, New York. Ziv, Jacob and Abraham Lempel. 1977. A universal algorithm for sequential data compression. IEEE Transactions on Information Theory, 23(3):337–343. https://doi.org/10.1109/TIT.1977.1055714 Zúñiga, Fernando. 2017. Mapudungun. In Michael Fortescue… — [doi:10.1109/tit.1977.1055714](https://doi.org/10.1109/tit.1977.1055714) [↗](https://doi.org/10.1109/TIT.1977.1055714)

*Updated 30 October 2026*