Elamur Research⌕
☰
home / papers / 2026 / Coreference annotation for characterising characters in literary works in Portuguese

Coreference annotation for characterising characters in literary works in Portuguese

Original (Portuguese): Marcação de correferência para a caracterização de personagens em obras literárias em português

Paper · 2026 · Diana Santos, Luisa Mara Silva Lima and Emanoel Pires

Summary

Computational literary studies and distant reading need to follow characters through a whole text, but systems that rely only on proper names miss most mentions. In Portuguese, characters are often referred to through pronouns, forms of address or null (unexpressed) subjects. The authors build a first public Portuguese literary dataset that annotates character coreference, including the attribution of null subjects to characters, which earlier English and German resources did not handle. They started from three canonical Brazilian novels in the OBras corpus that already had character networks. The tabular AC/DC format was converted into BRAT so that literary scholars could annotate and review the texts easily. The results were then converted back into a new AC/DC column, which keeps corpus queries and concordances available. Annotating six characters showed three to eight times more mentions than proper names alone yield. The resource lays groundwork for character profiling by actions and attributes. The authors also plan two coreference modules, one open-source rule-based and one based on machine learning.

Key points

  • Novels annotated:
    • Úrsula (1859, Romantic; 53,568 units)
    • O Cortiço (1890, Naturalist; 95,472 units)
    • Canaã (1902, Pre-Modernist; 90,167 units)
  • Characters covered: João Romão, Rita Baiana, Úrsula, Tancredo, Milkau and Maria. The annotations are released in both BRAT and tabular (AC/DC) formats.
  • What was marked: possessive pronouns, personal pronouns, forms of address and null subjects.
  • What was excluded: nominal anaphora, cataphora (mentions before the first proper name), unnamed characters and characters referred to only as part of a group.
  • Motivation for null subjects: prior counts with UDpipe on OBras found that about 41% of verbs in Brazilian literary texts lack an explicit subject. The share is likely higher in European Portuguese.
  • Pipeline context: the work builds on the DIP challenge and the PALAVRAS parser, which supplies explicit subjects. Verb lists linked to each character are intended for later profiling and comparison of characters.

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 OsloOslo · Norway51 works · 4,155 citations

Also on the byline: Universidade Estadual do Maranhão.

 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.Coreference resolution · 574 works

Cite

Cite as Diana Santos, Luisa Mara Silva Lima and Emanoel Pires. 2026. Marcação de correferência para a caracterização de personagens em obras literárias em português. Proceedings of the 17th International Conference on Computational Processing of Portuguese (PROPOR 2026) - Vol. 1.

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

@inproceedings{santos2026marcacao,
  title = {{Marcação de correferência para a caracterização de personagens em obras literárias em português}},
  author = {Diana Santos, Luisa Mara Silva Lima and Emanoel Pires},
  booktitle = {Proceedings of the 17th International Conference on Computational Processing of Portuguese (PROPOR 2026) - Vol. 1},
  year = {2026},
  url = {https://aclanthology.org/2026.propor-1.56/}
}

References

28 references, 2 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Harald Baayen. 2008. Analyzing Linguistic Data: A practical introduction to Statistics using R. Cambridge University Press. find
  2. David Bamman, Olivia Lewke, e Anya Mansoor. 2020. A Dataset of Literary Coreference. Em Proceedings of the 12th Conference on Language Resources and Evaluation (LREC 2020), Marseille, 11–16 May 2020, página 44–54. find
  3. David Bamman, Brendan O’Connor, e Noah A. Smith. 2013. Learning latent personas of film characters. Em Proceedings of the 51st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), páginas 352–361, Sofia, Bulgaria. Association for Computational Linguistics.
  4. Cláudia Freitas, Diana Santos, Cristina Mota, Hugo Gonçalo Oliveira, e Paula Carvalho. 2009. Detection of relations between named entities: report of a shared task. Em Proceedings of the NAACL HLT Workshop on Semantic Evaluations: Recent Achievements and Future Directions, SEW-2009, páginas 129–137. find
  5. Cláudia Freitas, Elivis de Souza, e Luísa Rocha. 2019. Quantificando e qualificando o sujeito oculto em português. find
  6. Em STIL 2029, XII Symposium in Information and Human Language Technology and Collocates Events, páginas 289–293. find
  7. Cláudia Freitas e Elvis de Souza. 2021. Sujeito oculto às claras: uma abordagem descritivo-computacional. Revista de Estudos linguísticos, Belo Horizonte, 29(2):1033–1058. find
  8. David Bamman, Ted Underwood, e Noah A. Smith. 2014. A bayesian mixed effects model of literary character. Em Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), páginas 370–379.
  9. Suemi Higuchi, Diana Santos, Cláudia Freitas, e Alexandre Rademaker. 2019. "Distant reading Brazilian history. Em Proceedings of 4th Conference of The Association Digital Humanities in the Nordic Countries (Copenhagen, March 6-8 2019), páginas 190–200. find
  10. Eckhard Bick. 2000. The Parsing System "Palavras": Automatic Grammatical Analysis of Portuguese in a Constraint Grammar Framework. Tese de Doutora-mento, Aarhus University, Aarhus, Denmark. find
  11. Mary Kato, Ana Maria Martins, e Jairo Nunes. 2023. The Syntax of Portuguese. Cambridge University Press. find
  12. Eckhard Bick. 2010. A dependency based approach to anaphora resolution. Em Extended Activities Proceedings, 9th International Conference on Computational Processing of the Portuguese Language, Apr. 27-30. Porto Alegre, Brazil. find
  13. Eckhard Bick. 2023a. Attribution of Quoted Speech in Portuguese Text. Em Proceedings of the NoDaLiDa 2023 Workshop on Constraint Grammar - Methods, Tools and Applications (Tórshavn, Faroe Islands), páginas 1–9. find
  14. Eckhard Bick. 2023b. Extraction of Literary Character Information in Portuguese. Linguamática, 15(1):31– 40. find
  15. Daniela Barreiro Claro, Joaquim Santos, Marlo Souza, Renata Vieira, e Vládia Pinheiro. 2024. Extração de informação. Em H. M. Caseli e M. G. V. Nunes, editores, Processamento de Linguagem Natural: Conceitos, Técnicas e Aplicações em Português, 3 edição, book chapter 22. BPLN. find
  16. Regina Dalcastagnè. 2021. Ausências e estereótipos no romance brasileiro das últimas décadas: alterações e continuidades. Letras de Hoje, Porto Alegre, 56(1):109–143. find
  17. Regina Dalcastagnè. 2025. A personagem do romance brasileiro contemporâneo: 1990-2004. Estudos de Literatura Brasileira Contemporânea, 26:13–71. find
  18. Evandro Fonseca, Aline Aver Vanin, e Renata Vieira. 2023. Resolução de correferência. Em Processamento de Linguagem Natural: Conceitos, Técnicas e Aplicações em Português. 1.a edição. find
  19. Markus Krug, Lukas Weimer, Isabella Reger, Luisa Macharowsky, Stephan Feldhaus, Frank Puppe, e Fotis Jannidis. 2018. Description of a Corpus of Character References in German Novels - DROC [Deutsches ROman Corpus]. find
  20. Luisa Mara Silva Lima. 2026. A construção da personagem literária maranhense sob uma perspectiva computacional: uma análise baseada em correferên-cias e sujeitos nulos. Tese de Mestrado, Universidade Estadual do Maranhão. find
  21. F. Martins e Cláudia Freitas. 2019. Sujeitos ocultos em verbetes biográficos: Contornando dificuldades da extração automática de informações. Em XI Congresso Internacional da ABRALIN. find
  22. Franco Moretti. 2013. Distant Reading. Verso. find
  23. Cristina Mota e Diana Santos, editores. 2008. Desafios na avaliação conjunta do reconhecimento de entidades mencionadas: O Segundo HAREM. Linguateca. find
  24. Diana Santos. 2018. Galeria de personagens. find
  25. Diana Santos e Cláudia Freitas. 2019. Estudando personagens na literatura lusófona. Em STIL - Symposium in Information and Human Language Technology. find
  26. Diana Santos, Cláudia Freitas, e Eckhard Bick. 2018. OBras: a fully annotated and partially human-revised corpus of Brazilian literary works in public domain. Em CorLex. find
  27. Diana Santos, Cristina Mota, Emanoel Pires, Marcia Langfeldt, Rebeca Schumacher Fuão, e Roberto Willrich. 2023. DIP - Desafio de Identificação de Personagens: objectivo, organização, recursos e resultados. Linguamática, 15(1):3–30. find
  28. Jefferson Fontinelle da Silva. 2011. Resolução de correferência em múltiplos documentos utilizando aprendizado não supervisionado. Tese de Mestrado, USP. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/santos-2026-marcacao-de-correferencia-para-caracterizacao/