Elamur Research⌕
☰
home / papers / 2026 / AmazoniaNLP: A Survey of Extreme Low-Resource Languages in the Peruvian-Brazilian Amazon

AmazoniaNLP: A Survey of Extreme Low-Resource Languages in the Peruvian-Brazilian Amazon

Summary

This survey examines how prepared five Indigenous languages of the Peru–Brazil Amazon borderlands are for NLP and speech technology. The languages are Matsés, Amahuaca, Kashinawa (Huni Kuĩ), Ticuna and Kukama-Kukamiria, and they cover the Panoan and Tupí-Guaraní families plus an isolate. The authors searched academic databases, language archives such as DoBeS/TLA, the California Language Archive, DoReCo and SIL, and code and model hubs such as GitHub and Hugging Face. Their queries combined language names, ISO codes and spelling variants. For each resource found, they recorded its type, task or modality, location, scale and any stated access or licensing terms. The picture is uneven and scattered. Grammars, dictionaries and archived recordings exist for most of the languages, but datasets ready for model training or evaluation, with clear reuse rights, are scarce. Inconsistent spelling across national traditions and patchy licensing make the gap worse. The authors argue that the most useful near-term work is "foundational resource engineering": turning documentation into normalized, aligned and well-described computational assets, guided by community priorities rather than by new modeling.

Key points

  • Amahuaca has the only UD treebank among the five languages. It contains 202 manually annotated sentences, with 1,028 word-level and 1,928 morph-level units (Angulo et al., 2025).
  • Kukama-Kukamiria has a gold-standard UMR release of 105 sentence-level annotations across 2 documents. Of these, 86 also carry document-level annotation (Bonn et al., 2024).
  • Ticuna illustrates the gap between documentation and NLP-ready data. The California Language Archive holds more than 1,396 hours of recordings, but only 33 hours are transcribed.
  • Kashinawa has 286 TLA/DoBeS archive bundles:
    • 146 are open, 138 restricted and 2 registered;
    • they include 180 WAV bundles and 88 ELAN bundles;
    • a DoReCo aligned speech corpus also exists.
  • Meta's MMS covers all five languages for ASR. Kashinawa and Ticuna additionally have TTS and language identification (LID) support. The authors recommend:
    • orthographic normalization across national spelling traditions;
    • extending UD and UMR annotation to the other languages;
    • visibility in venues such as SIGUL, AmericasNLP, IWSLT and the LRE Map.

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

Topics: Low-resource languages

Cite

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.63317/3sbz29bn779p

@inproceedings{zevallos2026amazonianlp,
  title = {{AmazoniaNLP: A Survey of Extreme Low-Resource Languages in the Peruvian-Brazilian Amazon}},
  author = {Zevallos, Rodolfo Joel and Carraro, Fabrício and Ortega, John E.},
  booktitle = {Proceedings of the SIGUL 2026 Joint Workshop with ELE, EURALI, and DCLRL: Towards Inclusivity and Equality: Language Resources and Technologies for Under-Resourced and Endangered Languages},
  year = {2026},
  doi = {10.63317/3sbz29bn779p},
  url = {https://doi.org/10.63317/3sbz29bn779p}
}

References

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

  1. João Capistrano de Abreu. 1914. Rã-txa hu-ni-ku-ı̃: a lingua dos caxinauás do Rio Ibuaçu, affluente do Muru (Prefeitura de Tarauacá). Typographia Leuzinger, Rio de Janeiro. find
  2. Gilles Adda, Sebastian Stüker, Martine Adda-Decker, Odette Ambouroue, Laurent Besacier, David Blachon, Hélène Bonneau-Maynard, Pierre Godard, Fatima Hamlaoui, Dmitry Idiatov, Guy-Noël Kouarata, Lori Lamel, Emmanuel-Moselly Makasso, Annie Rialland, Mark Van de Velde, François Yvon, and Sabine Zerbian… find
  3. Candy Angulo, Pilar Valenzuela, and Roberto Zariquiey. 2025. Universal Dependencies for Amahuaca. In Proceedings of the Eight Workshop on the Use of Computational Methods in the Study of Endangered Languages, pages Denis Bertet. 2021. Tikuna: a ten-toneme language in amazonia. Amerindia, 43:55–101. find our card
  4. Damian Blasi, Antonios Anastasopoulos, and Graham Neubig. 2022. Systematic inequalities in language technology performance across the world’s languages. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 5486–5505, Dublin, Irelan… find our card
  5. Rishi Bommasani, Drew A Hudson, Ehsan Adeli, Russ Altman, Simran Arora, Sydney von Arx, Michael S Bernstein, Jeannette Bohg, Antoine Bosselut, Emma Brunskill, et al. 2021. On the opportunities and risks of foundation models. arXiv preprint arXiv:2108.07258. find our card
  6. Julia Bonn, Matthew J. Buchholz, Jayeol Chun, Andrew Cowell, William Croft, Lukas Denk, Sijia Ge, Jan Hajič, Kenneth Lai, James H. Martin, Skatje Myers, Alexis Palmer, Martha Palmer, Claire Benet Post, James Pustejovsky, Kristine Stenzel, Haibo Sun, Zdeňka Urešová, Rosa Vallejos, Jens E. L. Van Gyse… find
  7. Eliane Camargo. 2002. Ergatividade cindida em caxinawá. In Ana Suelly Arruda Câmara Cabral and Aryon Dall’Igna Rodrigues, editors, Línguas indígenas brasileiras: fonologia, gramática e história: Atas do I Encontro Internacional do Grupo de Trabalho sobre Línguas Indígenas da ANPOLL. EDUFPA, Belém. find
  8. Chapman University (The Voice of Wilkinson). 2023. A historic event for the amahuaca people of peru. States that 330 people speak Amahuaca. Accessed 2026-02-14. find
  9. Emily Clem. 2019. Amahuaca ergative as agreement with multiple heads. Natural Language & Linguistic Theory, 37(3):785–823. find
  10. Emily Clem. 2023. The expression of time in amahuaca switch-reference clauses. Languages, 8(2):134. find
  11. DoBeS (Documentation of Endangered Languages). Cashinahua: Language description page. Mentions a Cashinahua ethnic community of around six thousand and discusses sociolinguistic tendencies. Accessed 2026-02-14. find
  12. David M. Eberhard, Gary F. Simons, and Charles D. Fennig, editors. 2024. Ethnologue: Languages of the World, 27 edition. SIL International, Dallas, Texas. find
  13. David W. Fleck. 2024. Why is matses an ono-matopoetic language? Anthropological Linguistics, 64(1–2):1–35. Issue dated Spring/Summer 2022; published on Project MUSE in 2024. Accessed 2026-02-15. find
  14. David William Fleck. 2003. A grammar of Matses. Ph.D. thesis. find
  15. David William Fleck. 2013. Panoan languages and linguistics.(Anthropological papers of the American Museum of Natural History, no. 99). American Museum of Natural History. find
  16. IBGE. 2022. Censo demográfico 2022: resultados sobre línguas indígenas (tikuna). Accessed 2026-02-14. find
  17. Instituto Socioambiental. 2024. Povos indígenas no brasil: Huni kuin (kaxinawá). Accessed: 2026-02-19. find
  18. Instituto Socioambiental (ISA). Matsés (indigenous peoples in brazil). Accessed 2026-02-14. find
  19. Pratik Joshi, Sebastin Santy, Amar Budhiraja, Kalika Bali, and Monojit Choudhury. 2020. The state and fate of linguistic diversity and inclusion in the NLP world. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 6282–6293, Online. Association for Comp… find our card
  20. Manuel Mager, Özlem Çetinoğlu, and Katharina Kann. 2020. Tackling the low-resource challenge for canonical segmentation. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 5237–5250, Online. Association for Computational Linguistics. find
  21. Manuel Mager, Ximena Gutierrez-Vasques, Gerardo Sierra, and Ivan Meza-Ruiz. 2018. Challenges of language technologies for the indigenous languages of the Americas. In Proceedings of the 27th International Conference on Computational Linguistics, pages 55–69, Santa Fe, New Mexico, USA. Association fo… find
  22. Manuel Mager, Elisabeth Mager, Katharina Kann, and Ngoc Thang Vu. 2023. Ethical considerations for machine translation of indigenous languages: Giving a voice to the speakers. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 48… find
  23. Alexandre Magueresse, Vincent Carles, and Evan Heetderks. 2020. Low-resource languages: A review of past work and future challenges. arXiv preprint arXiv:2006.07264. — [arXiv:2006.07264](https://arxiv.org/abs/2006.07264) find
  24. Ministerio de Cultura de Perú. 2024a. Base de datos de pueblos indígenas u originarios: Amahuaca. Accessed: 2026-02-19. find
  25. Ministerio de Cultura de Perú. 2024b. Base de datos de pueblos indígenas u originarios: Kashinawa. Accessed: 2026-02-19. find
  26. Ministerio de Cultura de Perú. 2024c. Base de datos de pueblos indígenas u originarios: Matsés. Accessed: 2026-02-19. find
  27. María Emilia Montes Rodríguez. 2004. Morfos-intaxis de la lengua Tikuna (Amazonía Colombiana). Number 15 in Lenguas aborígenes de Colombia: Descripciones. Universidad de los Andes, Bogotá. find
  28. María Emilia Montes Rodríguez. 2018. Género, clasificación y nombres ligados en tikuna (amazonia colombiana). Revista Brasileira de Linguística Antropológica, 6(1). Portal metadata labels the issue as v.6 n.1 (2014); published online 2018-12-21. find
  29. Christopher Moseley and Alexandre Nicolas, editors. 2010. Atlas of the World’s Languages in Danger, 3rd edition. Memory of Peoples. UN-ESCO, Paris, France. find
  30. Sanderson Castro Soares de Oliveira. 2024. Contribuições para a sociolinguística do kaxinawá, uma língua pluricêntrica. Raído, 18(46):378– 397. find
  31. Omnilingual ASR Team. 2025. Omnilingual ASR: Open-source multilingual speech recognition for 1600+ languages. find
  32. John E Ortega, Richard Castro Mamani, and Kyunghyun Cho. 2020. Neural machine translation with a polysynthetic low resource language. Machine Translation, 34(4):325–346. find
  33. John E Ortega and Krishnan Pillaipakkamnatt. 2018. Using morphemes from agglutinative languages like quechua and finnish to aid in low-resource translation. Technologies for MT of Low Resource Languages (LoResMT 2018), page 1. find
  34. Vineel Pratap, Andros Tjandra, Bowen Shi, Paden Tomasello, Arun Babu, Sayani Kundu, Ali Elkahky, Zhaoheng Ni, Apoorv Vyas, Maryam Fazel-Zarandi, Alexei Baevski, Yossi Adi, Xiaohui Zhang, Wei-Ning Hsu, Alexis Conneau, and Michael Auli. 2023. Scaling speech technology to 1,000+ languages. Accessed 202… find our card
  35. Endangered Languages Project. n.d.a. Cashinahua (elp language page). Accessed 2026-02-14. find
  36. Endangered Languages Project. n.d.b. Cocama-cocamilla (elp language page). Accessed 2026-02-14. find
  37. Endangered Languages Project. n.d.c. Ticuna (elp language page). Accessed 2026-02-14. find
  38. Surangika Ranathunga, En-Shiun Annie Lee, Marjana Prifti Skenduli, Ravi Shekhar, Mehreen Alam, and Rishemjit Kaur. 2023. Neural machine translation for low-resource languages: A survey. ACM Computing Surveys, 55(11):1–37. find
  39. Farhana Shahid, Mona Elswah, and Aditya Vashistha. 2025. Think outside the data: Colonial biases and systemic issues in automated moderation pipelines for low-resource languages. Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society, 8(3):2331–2344. find
  40. Amalia Skilton. 2021. Ticuna (tca) language documentation: A guide to materials in the california language archive. Language Documentation and Conservation, 15:153–189. find
  41. Amalia Horan Skilton. 2019. Spatial and Non-spatial Deixis in Cushillococha Ticuna. Phd dissertation, University of California, Berkeley. find
  42. Margarethe Sparing-Chávez. 2012. Aspects of amahuaca grammar: An endangered language of the amazon basin. Dallas: SIL International. find
  43. Rosa Vallejos. 2016. A grammar of Kukama-Kukamiria: A language from the Amazon, volume 13. Brill. find
  44. Rosa Vallejos. 2018. Kukama–kukamiria. International Journal of American Linguistics, 84(S1):S129–S147. find
  45. Jens E. L. Van Gysel, Meagan Vigus, Lukas Denk, Andrew Cowell, Rosa Vallejos, Tim O’Gorman, and William Croft. 2021. Theoretical and practical issues in the semantic annotation of four indigenous languages. In Proceedings of the Joint 15th Linguistic Annotation Workshop (LAW) and 3rd Designing Meani… find
  46. Shira Wein. 2025. Can uniform meaning representation help GPT-4 translate from indigenous languages? In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers), pages 278– 285, Vienna, Austria. Association for Computational Linguistics. find
  47. Rodolfo Zevallos and Nuria Bel. 2023. Hints on the data for language modeling of synthetic languages with transformers. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 12508–12522, Toronto, Canada. Association for Computationa… find
  48. Language Resource References DataScientia Foundation. 2023. Cashinahua UKC Lexicon. LMF-XML; small seed lexicon. License shown as CC BY-NC-SA. Accessed 2026-02-14. find
  49. DoBeS (Documentation of Endangered Languages). 2017. Cashinahua: Documentation project page. Accessed 2026-02-14. find
  50. Fleck, David William and Bëso, Fernando Shoque Uaquí and Huanán, Daniel Manquid Jiménez. 2012. Diccionario matsés-castellano: con índice alfabético castellano-matsés e índice semántico castellano-matsés. Tierra Nueva. find
  51. Meta AI (facebook). 2023. MMS language coverage overview (ASR/TTS/LID). Accessed 2026-02-20. find
  52. Reiter, Sabine. 2024. Cashinahua DoReCo dataset. In Seifart, Frank and Paschen, Ludger and Stave, Matthew (eds.), Language Documentation Reference Corpus (DoReCo) 2.0. Lyon: Laboratoire Dynamique Du Langage (UMR5596, CNRS & Université Lyon 2). Accessed 2026-02-14. find
  53. Rosa Vallejos-Yopán and Rosa Amías. 2015. Diccionario Kukama-Kukamiria / Castellano. FORMABIAP, Iquitos, Perú. 9.7MB. find

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/zevallos-2026-amazonianlp-survey-extreme-low-resource/