Elamur Research⌕
☰
home / papers / 2023 / Designing the LECOR Learner Corpus for Romanian

Designing the LECOR Learner Corpus for Romanian

Summary

The paper describes the design and first stages of building LECOR. It is the first electronic corpus of Romanian as a foreign language. The texts are written or spoken by foreign students of various nationalities who spend a year studying Romanian intensively at the University of Bucharest. Previously, only two printed collections of learner texts existed for Romanian. Each sample in the corpus has two versions: the student's original and a version corrected by the teacher. Both versions are automatically assigned lemmas and morphological tags (UPOS, MSD). An algorithm then aligns them token by token. It checks whether the form or the lemma matches, compares words without diacritics, and takes the Levenshtein distance into account. Discrepancies are checked manually, labeled with basic error types, and stored in CoNLL-U Plus. NoSketch Engine is being adapted for searching the corpus together with metadata about the students and tasks. This is an interim report without final experimental results. The authors present a working processing pipeline, a pilot online sample, and open problems. The resource is useful for research on second language acquisition, the creation of teaching materials, automatic grammar checking, and native language identification of a text's author.

Key points

  • The authors plan to collect 4000 samples and over 600 thousand words. Approximately 80% of the samples are written and 20% are spoken (about 800 transcriptions). The corpus is being collected in 2019–2024 from 50–70 students per year with over 20 native languages, at levels from A1 to B2. At least a quarter of the samples belong to level A1. About 4500 written texts and approximately 500 audio files are already available.
  • The corpus has a longitudinal dimension: several dozen learners produced approximately 60 samples each over an academic year. The metadata about learners and tasks follow the core schema of König et al. (2022). The corpus can be searched freely, but it can be downloaded only on request and only for research purposes.
  • Tokens are considered to correspond if their form or lemma matches, including after removal of Romanian diacritics, or if the Levenshtein distance between the forms is less than 2. Each version of the text occupies five columns in the CoNLL-U Plus file, and the 11th column contains the discrepancy type. There are four types: no error, missing word, extra word, spelling error.
  • A dedicated script converts the data into the "vertical" format of NoSketch Engine. The pilot sample is available at lecor.unibuc.ro and supports simple queries and CQL. A known problem: search returns erroneous and corrected forms at the same time, because the student and teacher versions are mixed.
  • In the future, the authors plan multi-level manual error annotation, including word order errors. A proofreading guide already exists for the corrections, and its criteria will be checked through Inter-Annotator Agreement. Separate search by version, possibly via a parallel corpus mode, and subcorpora based on metadata are also planned.

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 · Education

Cite

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

@inproceedings{barbu2023designing,
  title = {{Designing the LECOR Learner Corpus for Romanian}},
  author = {Barbu, Ana Maria and Irimia, Elena and Vasile, Carmen Mîrzea and Păiș, Vasile},
  booktitle = {Proceedings of the 14th International Conference on Recent Advances in Natural Language Processing},
  year = {2023},
  doi = {10.26615/978-954-452-092-2\_016},
  url = {https://doi.org/10.26615/978-954-452-092-2_016}
}

References

33 references, 1 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Margarita Alonso-Ramos. 2016. Spanish Learner Corpus Research: Current Trends and Future Perspectives. Benjamins, Amsterdam. find
  2. Verginica Barbu Mititelu, Radu Ion, Radu Simionescu, Elena Irimia, and Cenel-Augusto Perez. 2016. The romanian treebank annotated according to universal dependencies. In Proceedings of The Tenth International Conference on Natural Language Processing (HrTAL2016). find
  3. Guy Brook-Hart. 2009. Learning from common mistakes. Cambridge University Press, Cambridge. find
  4. Marcus Callies and Sandra Götz. 2015. Learner Corpora in Language Testing and Assessment. John Benjamins Publishing Company, Amsterdam. find
  5. Mihaela-Viorica Constantinescu and Gabriela Stoica. 2020. Româna ca limbă străină: Corpus. Editura Universităt , ii din Bucures , ti, Bucures , ti. find
  6. Elisa Corino and Carla Marello. 2017. Italiano di stranieri: I corpora VALICO e VINCA. Guerra, Perugia. find
  7. ‘ Roberts Dar gis, Ilze Auzin , a, Inga Kaija, Kristı̄ne Levāne-Petrova, and Kristı̄ne Pokratniece. 2022. LaVA – Latvian language learner corpus. In Proceedings of the Thirteenth Language Resources and Evaluation Conference, pages 727–731, Marseille, France. European Language Resources Association. find
  8. Ana Díaz-Negrillo and Paul Thompson. 2013. Ana Díaz-Negrillo, Nicolas Ballier, and Paul Thompson (eds). Automatic Treatment and Analysis of Learner Corpus Data. John Benjamins Publishing Company, Amsterdam / Philadelphia. find
  9. Amália Mendes, Sandra Antunes, Maarten Janssen, and Anabela Gonçalves. 2016. The COPLE2 corpus: a learner corpus for Portuguese. In Proceedings of the Tenth International Conference on Language Resources and Evaluation (LREC’16), pages 3207– 3214, Portorož, Slovenia. European Language Resources Asso… find
  10. Francesca Gallina. 2017. Anna Gudmundson, Laura Alvarez Lopez, Camilla Bardel (eds.), Romance languages. Multilingualism and Language Acquisition. Peter Lang, Frankfurt am Mein. find
  11. Rosamond Mitchell. 2021. The Routledge Handbook of Second Language Acquisition and Corpora, chapter Corpora and Instructed Second Language Acquisition. Routledge, Taylor & Francis, London/New York. find
  12. Sylviane Granger, Estelle Dagneaux, Fanny Meunier, Magali Paquot, et al. 2009. International corpus of learner English, volume 2. Presses universitaires de Louvain Louvain-la-Neuve. find
  13. Vasile Păis , . 2020. Multiple annotation pipelines inside the relate platform. In The 15th International Conference on Linguistic Resources and Tools for Natural Language Processing, pages 65–75. find
  14. Sylviane Granger, Gaëtanelle Gilquin, and Fanny Meunier. 2015. The Cambridge Handbook of Learner Corpus Research. Cambridge University Press, Cambridge. find
  15. Vasile Păis , , Radu Ion, Andrei-Marius Avram, Maria Mitrofan, and Dan Tufis , . 2021. In-depth evaluation of Romanian natural language processing pipelines. Romanian Journal of Information Science and Technology (ROMJIST), 24(4):384–401. find
  16. Sylviane Granger, Helen Swallow, and Jennifer Thewis-sen. 2022. The louvain error tagging manual. version 2.0. find
  17. Heather Hilton. 2009. Annotation and analyses of temporal aspects of spoken fluency. Calico Journal, 26(3):644–661. find
  18. Scott Jarvis and Magali Paquot. 2015. Native language identification. Cambridge handbook of learner corpus research, pages 605–628. find
  19. Adam Kilgarriff, Vít Baisa, Jan Bušta, Miloš Jakubíček, Vojtěch Kovář, Jan Michelfeit, Pavel Rychlỳ, and Vít Suchomel. 2014. The sketch engine: ten years on. Lexicography, 1(1):7–36. find
  20. Emma Marsden, Florence Myles, Sarah Rule, and Rosamond Mitchell. 2002. Oral french interlanguage corpora: Tools for data management and analysis. occasional paper. find
  21. Tony McEnery, Richard Xiao, and Yukio Tono. 2006. Corpus-based language studies: An advanced resource book. Routledge, London/New York. find
  22. Tony McEnery, Richard Xiao, and Yukio Tono. 2015. The Cambridge Handbook of Learner Corpus Research, chapter Learner corpora and natural language processing. Cambridge University Press, Cambridge. find
  23. Beáta Megyesi, Lena Granstedt, Sofia Johansson, Julia Prentice, Dan Rosén, Carl-Johan Schenström, Gunlög Sundberg, Mats Wirén, and Elena Volodina. 2018. Learner corpus anonymization in the age of gdpr: Insights from the creation of a learner corpus of swedish. In 7th Workshop on NLP for Computer Ass… find
  24. Vasile Păis , , Radu Ion, and Dan Tufis , . 2020. A processing platform relating data and tools for Romanian language. In Proceedings of the 1st International Workshop on Language Technology Platforms, pages 81–88, Marseille, France. European Language Resources Association. find
  25. Vasile Păis , , Dan Tufis , , and Radu Ion. 2019. Integration of romanian nlp tools into the relate platform. In International Conference on Linguistic Resources and Tools for Natural Language Processing. find
  26. Nives M. Preradović, Monika Berać, and Damir Boras. 2015. Cergol Kovačević and Sanda Lucia Udier (eds.) Multidisciplinary Approaches to Multilingualism, chapter Learner Corpus of Croatian as a Second and Foreign Language. Peter Lang. find
  27. Alexandr Rosen, Jiří Hana, Barbora Hladká, Tomáš Jelínek, Svatava Škodová, and Barbora Štindlová. 2020. Compiling and annotating a learner corpus for a morphologically rich language. CzeSL, a corpus of non-native Czech. Karolinum Press, Charles University. find
  28. Michael Rundell. 2007. Macmillan English dictionary advanced learner. MacMillan. find
  29. Pavel Rychlỳ. 2007. Manatee/bonito-a modular corpus manager. In RASLAN, pages 65–70. find
  30. Diane Schmitt and Norbert Schmitt. 2011. Focus on Vocabulary 2: Mastering the Academic Word List. Pearson Education, White Plains, NY. find
  31. Stefania Spina, Irene Fioravanti, Luciana Forti, Valentino Santucci, Angela Scerra, and Fabio Zanda. 2022. Il corpus celi: una nuova risorsa per studiare l’acquisizione dell’italiano l2. Italiano LinguaDue, 14(1):116–138. find
  32. Milan Straka, Jan Hajic, and Jana Straková. 2016. Udpipe: trainable pipeline for processing conll-u files performing tokenization, morphological analysis, pos tagging and parsing. In Proceedings of the Tenth International Conference on Language Resources and Evaluation (LREC’16), pages 4290–4297. find our card
  33. Iunia-Lavinia Vasiu. 2020. Achizit , ia limbii române ca L2. Interlimba la nivelul A1. Presa Universitară Clu-jeană, Cluj-Napoca. find

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/papers/barbu-2023-designing-lecor-learner-corpus-romanian/