Elamur Research⌕
☰
home / papers / 2026 / A Dual-View Analysis of Multiple Languages in Colonial Newspapers

A Dual-View Analysis of Multiple Languages in Colonial Newspapers

Paper · 2026 · Zhan Su, Xiaoya Chen, Fengran Mo et al.

Summary

Colonial-era newspapers record how racial categories were built and maintained, but they are hard to analyse computationally. Annotated data is scarce, OCR output is noisy, and the texts are multilingual and archaic. The authors study racialization from two complementary angles.

  • Event extraction: framed as question answering. They built contextual (CQA) and visual (VQA) QA benchmarks from 18th–19th-century newspapers in Dutch, English-French and Spanish. These cover how enslavers described enslaved people and how the enslaved expressed their own situation. Gemini generated the QA data from page PDFs, and human annotators then checked every pair against its source.
  • Semantic change: they trained yearly Dutch embeddings with the TWEC (compass) framework over 1816–1882. This period spans the 1863 abolition of slavery in the Dutch colonies.

LLMs answer well when given clean text, but even strong multimodal models struggle when the input is images of the original pages. The embedding analysis suggests racialization did not disappear after abolition. Instead, the vocabulary of control moved onto new labour and identity terms, such as "workers." This matters for historians who want to use NLP on noisy, low-resource archives.

Key points

  • Corpora: De Curaçaosche Courant (Dutch, 3249 issues, ~13000 pages, 1816–1882), Quebec Gazette (French-English, 51 issues, ~200 pages, 1765–1807) and Revista Economica (Spanish, Havana, 50 issues, ~400 pages, 1878–1882).
  • Models evaluated:
    • CQA: Qwen-72B and GPT-4o.
    • VQA: Qwen3-VL-8B, Qwen2.5-VL-72B and GPT5.2.
    • QA pairs are built around seven attributes (e.g., name, age, time, reason) and scored with F1.
  • Embedding alignment: Procrustes alignment error rises with the time gap from 1817: 0.087 (to 1840), 0.112 (to 1847) and 0.134 (to 1870). Over the same intervals, the number of compass words drops from 487 to 456.
  • Diachronic Bias Shift (1817→1870):
    • "slaaf" moves from +0.35 to −0.02, a shift away from the property/commodification framing.
    • "arbeider" rises from +0.15 to +0.37.
    • "vreemdeling" stays criminalized, with no significant change.
    • Significance was tested by bootstrap with 1000 iterations; "slaaf" reaches p < 0.001.
  • Resources: the code is released at github.com/shuishen112/historical_nlp. The authors acknowledge limitations: coverage of only four languages, possible bias inherited from modern LLMs, and sparse data in some years.

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

🎓Université de MontréalMontreal · Canada68 works · 79,382 citations🎓University of CopenhagenCopenhagen · Denmark115 works · 15,292 citations🎓Tianjin UniversityTianjin · China41 works · 1,128 citations🎓Erasmus University RotterdamRotterdam · Netherlands3 works · 289 citations🎓Zhengzhou University of Light IndustryZhengzhou · China3 works · 78 citations🎓Halmstad UniversityHalmstad · Sweden3 works · 37 citations

 Fields, topics and concepts

📚 Shelves 7The 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 worksSocial Sciences · 1,772 worksGeneral Social Sciences · 227 worksLiterature and Literary Theory · 519 works

🧭 Topics 1Research subjects and languages the work is about.Multilingual & cross-lingual NLP · 828 works

Cite

Cite as Zhan Su, Xiaoya Chen, Fengran Mo, Ida L. Vos, Prayag Tiwari, Yazhou Zhang, Qian Zheng and Natália da Silva Perez. 2026. A Dual-View Analysis of Multiple Languages in Colonial Newspapers. Findings of the Association for Computational Linguistics: ACL 2026.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.findings-acl.1029

@inproceedings{su2026dual,
  title = {{A Dual-View Analysis of Multiple Languages in Colonial Newspapers}},
  author = {Zhan Su, Xiaoya Chen, Fengran Mo, Ida L. Vos, Prayag Tiwari, Yazhou Zhang, Qian Zheng and Natália {da Silva Perez}},
  booktitle = {Findings of the Association for Computational Linguistics: ACL 2026},
  year = {2026},
  doi = {10.18653/v1/2026.findings-acl.1029},
  url = {https://doi.org/10.18653/v1/2026.findings-acl.1029}
}

References

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

  1. H. Samy Alim. 2016. Introducing raciolinguistics: Racing language and languaging race in hyper- racial times. In Raciolinguistics, pages 1–30. find
  2. Nadav Borenstein, Natália da Silva Perez, and Isabelle Augenstein. 2023a. Multilingual event extraction from historical newspaper adverts. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 10304–10325. find
  3. Nadav Borenstein, Karolina Stańczak, Thea Rolskov, Natacha Klein Käfer, Natália da Silva Perez, and Isabelle Augenstein. 2023b. Measuring intersectional biases in historical documents. In Findings of the Association for Computational Linguistics: ACL 2023, pages 2711–2730. find
  4. Aylin Caliskan, Joanna J Bryson, and Arvind Narayanan. 2017. Semantics derived automatically from language corpora contain human-like biases. Science, 356(6334):183–186. find
  5. Ruirui Chen, Chengwei Qin, Weifeng Jiang, and Dongkyu Choi. 2024. Is a large language model a good annotator for event extraction? In Proceedings of the AAAI conference on artificial intelligence, volume 38, pages 17772–17780. find
  6. Komala Subramanyam Cherukuri, Pranav Abishai Moses, Aisa Sakata, Jiangping Chen, and Haihua Chen. 2025. Large language models for oral history understanding with text classification and sentiment analysis. arXiv preprint arXiv:2508.06729. — [arXiv:2508.06729](https://arxiv.org/abs/2508.06729) find
  7. James Costa and Zorana Sokolovska. 2025. Historical foundations: Some threads for integrating and interrogating historiography in critical sociolinguistics. Critical sociolinguistics: dialogues, dissonances, developments. find
  8. Mark J Hill and Simon Hengchen. 2019. Quantifying the impact of dirty ocr on historical text analysis: Eighteenth century collections online as a case study. Digital Scholarship in the Humanities, 34(4):825– 843. find
  9. Natália da Silva Perez and Nadav Borenstein. 2025. Annotating "privacy" to train semi-supervised event extraction models for historical newspapers. Current Research in Digital History, 8. find
  10. Frederik Hogenboom, Flavius Frasincar, Uzay Kaymak, and Franciska De Jong. 2011. An overview of event extraction from text. DeRiVE@ ISWC, pages 48–57. find
  11. John Dagdelen, Alexander Dunn, Sanghoon Lee, Nicholas Walker, Andrew S Rosen, Gerbrand Ceder, Kristin A Persson, and Anubhav Jain. 2024. Structured information extraction from scientific text with large language models. Nature communications, 15(1):1418.
  12. Valerio Di Carlo, Federico Bianchi, and Matteo Palmonari. 2019. Training temporal word embeddings with a compass. In Proceedings of the AAAI conference on artificial intelligence, volume 33, pages 6326–6334. find
  13. Xinya Du, Alexander M Rush, and Claire Cardie. 2021. Grit: Generative role-filler transformers for document-level event entity extraction. In Proceedings of the 16th conference of the european chapter of the association for computational linguistics: main Volume, pages 634–644. find
  14. Nikhil Garg, Londa Schiebinger, Dan Jurafsky, and James Zou. 2018. Word embeddings quantify 100 years of gender and ethnic stereotypes. Proceedings of the National Academy of Sciences, 115(16):E3635– E3644. find
  15. Jack Grieve, Sara Bartl, Matteo Fuoli, Jason Grafmiller, Weihang Huang, Alejandro Jawerbaum, Akira Murakami, Marcus Perlman, Dana Roemling, and Bodo Winter. 2025. The sociolinguistic foundations of language modeling. Frontiers in Artificial Intelligence, 7:1472411. find
  16. Igor Grossmann, Matthew Feinberg, Dawn C Parker, Nicholas A Christakis, Philip E Tetlock, and William A Cunningham. 2023. Ai and the transformation of social science research. Science, 380(6650):1108–1109. find
  17. Kunal Handa, Alex Tamkin, Miles McCain, Saffron Huang, Esin Durmus, Sarah Heck, Jared Mueller, Jerry Hong, Stuart Ritchie, Tim Belonax, and 1 others. 2025. Which economic tasks are performed with ai? evidence from millions of claude conversations. arXiv preprint arXiv:2503.04761. find
  18. Jakob Hauser, Daniel Kondor, Jenny Reddish, Majid Benam, Enrico Cioni, Federica Villa, James Bennett, Daniel Hoyer, Pieter Francois, Peter Turchin, and 1 others. 2024. Large language models’ expert-level global history knowledge benchmark (hist-llm). Advances in Neural Information Processing Systems… find
  19. Kaiyu Huang, Fengran Mo, Xinyu Zhang, Hongliang Li, You Li, Yuanchi Zhang, Weijian Yi, Yulong Mao, Jinchen Liu, Yuzhuang Xu, and 1 others. 2026. A survey on large language models with multilingualism: Recent advances and new frontiers. Artificial Intelligence Review. find
  20. Mark Humphries, Lianne C Leddy, Quinn Downton, Meredith Legace, John McConnell, Isabella Murray, and Elizabeth Spence. 2024. Unlocking the archives: large language models achieve state-of-the-art performance on the transcription of handwritten historical documents. Available at SSRN. find
  21. Aaron Hurst, Adam Lerer, Adam P Goucher, Adam Perelman, Aditya Ramesh, Aidan Clark, AJ Ostrow, Akila Welihinda, Alan Hayes, Alec Radford, and 1 others. 2024. Gpt-4o system card. arXiv preprint arXiv:2410.21276.
  22. Tristan Karch, Jakhongir Saydaliev, Isabella Di Lenardo, and Frederic Kaplan. 2025. Llm agents for interactive exploration of historical cadastre data: framework and application to venice. Computational Humanities Research, 1:e11. find
  23. Andrey Kutuzov, Lilja Øvrelid, Terrence Szymanski, and Erik Velldal. 2018. Diachronic word embeddings and semantic shifts: a survey. arXiv preprint arXiv:1806.03537.
  24. Viet Dac Lai, Minh Van Nguyen, Heidi Kaufman, and Thien Huu Nguyen. 2021. Event extraction from historical texts: A new dataset for black rebellions. In Findings of the Association for Computational Linguistics: ACL-IJCNLP 2021, pages 2390–2400. find
  25. Maria Levchenko. 2025. Evaluating llms for historical document ocr: A methodological framework for digital humanities. arXiv preprint arXiv:2510.06743. find
  26. Simon Levis Sullam, Giorgia Minello, Rocco Tripodi, and Massimo Warglien. 2022. Representation of jews and anti-jewish bias in 19th century french public discourse: Distant and close reading. Frontiers in big Data, 4:723043. find
  27. Fayuan Li, Weihua Peng, Yuguang Chen, Quan Wang, Lu Pan, Yajuan Lyu, and Yong Zhu. 2020. Event extraction as multi-turn question answering. In Findings of the Association for Computational Linguistics: EMNLP 2020, pages 829–838.
  28. Ying Lin, Heng Ji, Fei Huang, and Lingfei Wu. 2020. A joint neural model for information extraction with global features. In Proceedings of the 58th annual meeting of the association for computational linguistics, pages 7999–8009.
  29. Mingyu Derek Ma, Xiaoxuan Wang, Po-Nien Kung, P Jeffrey Brantingham, Nanyun Peng, and Wei Wang. 2024. Star: boosting low-resource information extraction by structure-to-text data generation with large language models. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, pages… find
  30. Susanna Mäkinen. 2022. Stability and variation in the genre of runaway slave notices in american newpapers 1704–1865. Ennen ja nyt: Historian ti-etosanomat, 22(3):83–87. find
  31. Chuan Meng, Litu Ou, Sean MacAvaney, and Jeff Dalton. 2026. Revisiting text ranking in deep research. In Proceedings of the 49th International ACM SI-GIR Conference on Research and Development in Information Retrieval. find
  32. Fengran Mo, Yifan Gao, Chuan Meng, Xin Liu, Zhuofeng Wu, Kelong Mao, Zhengyang Wang, Pei Chen, Zheng Li, Xian Li, and 1 others. 2025. Uniconv: Unifying retrieval and response generation for large language models in conversations. In Proceedings of the 63rd Annual Meeting of the Association for Compu… find
  33. Fengran Mo, Zhan Su, Yuchen Hui, Jinghan Zhang, Jia Ao Sun, Zheyuan Liu, Chao Zhang, Tetsuya Sakai, and Jian-Yun Nie. 2026a. Opendecoder: Open large language model decoding to incorporate document quality in rag. arXiv preprint arXiv:2601.09028. — [arXiv:2601.09028](https://arxiv.org/abs/2601.09028) find
  34. Fengran Mo, Jinghan Zhang, Yuchen Hui, Jia Ao Sun, Zhichao Xu, Zhan Su, and Jian-Yun Nie. 2026b. Convmix: A mixed-criteria data augmentation framework for conversational dense retrieval. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 40, pages 15555–15563. find
  35. Simon P Newman. 2022. Freedom seekers: Escaping from slavery in restoration London. University of London Press. find
  36. Thien Nguyen and Ralph Grishman. 2018. Graph convolutional networks with argument-aware pooling for event detection. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 32. find
  37. Andrew Piper and Sophie Wu. 2025. Evaluating large language models for narrative topic labeling. In Proceedings of the 5th International Conference on Natural Language Processing for Digital Humanities, pages 281–291. find
  38. Rui Qi, Fengran Mo, Yufeng Chen, Xue Zhang, Shuo Wang, Hongliang Li, Jinan Xu, Meng Jiang, Jian-Yun Nie, and Kaiyu Huang. 2026. Language-coupled reinforcement learning for multilingual retrieval-augmented generation. Preprint, arXiv:2601.14896.
  39. John R Rickford. 2016. Raciolinguistics: How language shapes our ideas about race. Oxford University Press. find
  40. Christophe Rigaud, Antoine Doucet, Mickaël Coustaty, and Jean-Philippe Moreux. 2019. find
  41. Icdar 2019 competition on post-ocr text correction. In 2019 international conference on document analysis and recognition (ICDAR), pages 1588–1593. IEEE. find
  42. Natália da Silva Perez. 2025. Racialised language in colonial newspaper advertisements during the eighteenth and nineteenth centuries. In The Routledge Handbook of Information History, pages 95–109. find
  43. Arno Simons, Michael Zichert, and Adrian Wüthrich. 2025. Large language models for history, philosophy, and sociology of science: Interpretive uses, methodological challenges, and critical perspectives. arXiv preprint arXiv:2506.12242. — [arXiv:2506.12242](https://arxiv.org/abs/2506.12242) find
  44. Matthew Sims, Jong Ho Park, and David Bamman. 2019. Literary event detection. In Proceedings of the 57th annual meeting of the association for computational linguistics, pages 3623–3634. find
  45. Rachele Sprugnoli and Sara Tonelli. 2019. Novel event detection and classification for historical texts. Computational Linguistics, 45(2):229–265. find
  46. Zhan Su, Fengran Mo, Prayag Tiwari, Benyou Wang, Jian-Yun Nie, and Jakob Grue Simonsen. 2024. Mixture of latent experts using tensor products. arXiv preprint arXiv:2405.16671. — [arXiv:2405.16671](https://arxiv.org/abs/2405.16671) find
  47. Qwen Team and 1 others. 2024. Qwen2 technical report. arXiv preprint arXiv:2407.10671, 2(3). find
  48. Konstantin Todorov and Giovanni Colavizza. 2022. An assessment of the impact of ocr noise on language models. arXiv preprint arXiv:2202.00470. — [arXiv:2202.00470](https://arxiv.org/abs/2202.00470) find
  49. Hugo Touvron, Thibaut Lavril, Gautier Izacard, Xavier Martinet, Marie-Anne Lachaux, Timothée Lacroix, Baptiste Rozière, Naman Goyal, Eric Hambro, Faisal Azhar, and 1 others. 2023. Llama: Open and efficient foundation language models. arXiv preprint arXiv:2302.13971.
  50. Crina Tudor, Beáta Megyesi, and Robert Östling. 2025. Prompting the past: Exploring zero-shot learning for named entity recognition in historical texts using prompt-answering llms. In Proceedings of the 9th Joint SIGHUM Workshop on Computational Linguistics for Cultural Heritage, Social Sciences, Hu… find
  51. Jianshu Weng and Bu-Sung Lee. 2011. Event detection in twitter. In Proceedings of the international aaai conference on web and social media, volume 5, pages 401–408. find
  52. Wei Xiang and Bang Wang. 2019. A survey of event extraction from text. IEEE Access, 7:173111–173137. find
  53. Yongan Yu, Xianda Du, Qingchen Hu, Jiahao Liang, Jingwei Ni, Dan Qiang, Kaiyu Huang, Grant McKenzie, Renee Sieber, and Fengran Mo. 2025a. Weatherarchive-bench: Benchmarking retrieval-augmented reasoning for historical weather archives. arXiv preprint arXiv:2510.05336. — [arXiv:2510.05336](https://arxiv.org/abs/2510.05336) find
  54. A.2 Table 6 presents the hyperparameters used for TWEC training. For Word2vec, we use the gensim 2 to train the word2vec seperately for each year. All the training is conducted on an A100 80G GPU. find
  55. Yongan Yu, Qingchen Hu, Xianda Du, Jiayin Wang, Fengran Mo, and Renée Sieber. 2025b. Wximpact-bench: A disruptive weather impact understanding benchmark for evaluating large language models. In Findings of the Association for Computational Linguistics: ACL 2025, pages 4016–4035. *Table 6: TWEC Hyper… find
  56. Jinghan Zhang, Fengran Mo, Tharindu Cyril Weerasooriya, Ruimin Dai, Xiaoyan Han, Yanjie Fu, Dakuo Wang, and Kunpeng Liu. 2026. Starpo: Stability-augmented reinforcement policy optimization. Preprint, arXiv:2604.08905. — [arXiv:2604.08905](https://arxiv.org/abs/2604.08905) find
  57. Jinghan Zhang, Xiting Wang, Weijieying Ren, Lu Jiang, Dongjie Wang, and Kunpeng Liu. 2025. Ratt: A thought structure for coherent and correct llm reasoning. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 39, pages 26733–26741. find
  58. Shibingfeng Zhang and Giovanni Colavizza. 2025. Named entity recognition of historical text via large language model. arXiv preprint arXiv:2508.18090. — [arXiv:2508.18090](https://arxiv.org/abs/2508.18090) find
  59. Wenxuan Zhang, Yue Deng, Bing Liu, Sinno Pan, and Lidong Bing. 2024b. Sentiment analysis in the era of large language models: A reality check. In Findings of the Association for Computational Linguistics: NAACL 2024, pages 3881–3906.
  60. TWEC Model Configuration A.3 Parameter Value Embedding dimension (d) Context window (w) Minimum word frequency (f min ) Skip-gram negative samples Training epochs (T ) Learning rate Compass vocabulary size (\|C\|) CV stability threshold (τ stability ) 100 5 3 5 15 0.025 500 0.5 Computational Complexit… find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/su-2026-dual-view-analysis-multiple-languages/