---
title: "TArC. Un corpus d'arabish tunisien — Gugliotta 2020"
description: "The paper presents TArC, which the authors describe as the first Tunisian arabish corpus with… 76% accuracy. 26 references."
authors: ["Elisa Gugliotta", "Marco Dinarelli"]
year: 2020
venue: "Actes de la 6e conférence conjointe Journées d'Études sur la Parole (JEP, 33e édition), Traitement Automatique des Langues Naturelles (TALN, 27e édition), Rencontre des Étudiants Chercheurs en Informatique pour le Traitement Automatique des Langues (RÉCITAL, 22e édition). Volume 2 : Traitement Automatique des Langues Naturelles"
source: "https://aclanthology.org/2020.jeptalnrecital-taln.22/"
licence: "CC BY 4.0 (ACL Anthology)"
licence_url: "https://creativecommons.org/licenses/by/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/gugliotta-2020-tarc-un-corpus-d-arabish/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/gugliotta-2020-tarc-un-corpus-d-arabish/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2020](https://md.research.elamur.ai/free/en/papers/2020/page/11/) / TArC. Un corpus d'arabish tunisien

# TArC. Un corpus d'arabish tunisien

**Authors:** [Elisa Gugliotta](https://md.research.elamur.ai/free/en/authors/elisa-gugliotta/) and [Marco Dinarelli](https://md.research.elamur.ai/free/en/authors/marco-dinarelli/) · **Year:** 2020 · **Venue:** [Actes de la 6e conférence conjointe Journées d'Études sur la Parole (JEP, 33e édition), Traitement Automatique des Langues Naturelles (TALN, 27e édition), Rencontre des Étudiants Chercheurs en Informatique pour le Traitement Automatique des Langues (RÉCITAL, 22e édition). Volume 2 : Traitement Automatique des Langues Naturelles](https://md.research.elamur.ai/free/en/venues/jeptalnrecital/) ([JEP-TALN-RECITAL (the French-speaking conferences on speech and natural language processing) 2020](https://md.research.elamur.ai/free/en/venues/jeptalnrecital/2020/)) · **Publisher:** [ATALA](https://md.research.elamur.ai/free/en/publishers/atala/) · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/2020.jeptalnrecital-taln.22/) · [PDF](https://aclanthology.org/2020.jeptalnrecital-taln.22.pdf) · [OpenAlex](https://openalex.org/W3088294933)

**Citations:** 0 citations (OpenAlex, 2026-10-14)

## Summary
The paper presents TArC, which the authors describe as the first Tunisian arabish corpus with morpho-syntactic annotation. Arabish (also called Arabizi) is the informal way Arabic dialects are typed on social media using Latin letters plus digits standing in for Arabic sounds. Like most dialectal varieties, it has very few resources. Existing Tunisian dialect corpora are mostly spoken-language transcripts, and the few arabish transcription systems are not publicly released.

The authors gathered social-media texts from 2010 onward, which lets the corpus track how the writing habit has evolved. They annotate tokens on several layers:
- an Arabic-script transcription following the CODA* orthography guidelines,
- POS tags,
- lemmas,
- an Italian translation.

Annotation is done incrementally in blocks. Automatic models pre-annotate each new block, and humans then correct it. The paper also explains why Tunisian arabish is hard to process: it is quasi-oral, has no fixed spelling, does not map characters one-to-one onto Arabic script, and has no dedicated symbols for emphatic consonants. Early semi-automatic transcription averaged about 76% accuracy. The resulting open resource can support both linguistic study and NLP tool training.

## Key points
- TArC is presented as the first Tunisian arabish corpus annotated with POS and lemmas and also transcribed into Arabic script under the CODA* convention, which unifies orthography for 28 Arabic dialects.
- The texts span 2010 to 2020, so the corpus can be used to study diachronic conventionalisation of arabish spelling.
- Foreign words and para-linguistic elements (emojis, emoticons) are kept, unlike earlier work that filtered them out.
- Prior work compared:
  - a seq2seq transcriber by Younes et al. 2018b, reaching over 95.59% accuracy on cleaned data but not public;
  - a rule-based CODA transcriber by Masmoudi et al. 2015 with roughly 10% error and manual candidate selection;
  - public corpora PADIC, TuDiCoI, STAC, TARIC and TSAC.
- The corpus is built in token blocks with semi-automatic transcription and correction, averaging about 76% transcription accuracy. A multitask model covering all annotation levels is in development.

## Interesting ideas
- Grow the corpus block by block: a model trained on already-corrected blocks pre-annotates the next block, which humans then fix, so annotation cost drops as the model improves.
- Keep noisy material such as emojis, foreign words and code-switched tokens instead of cleaning it away, so the resource reflects real usage.
- Map spontaneous, unstandardised spelling onto an established orthographic convention (CODA*) as a separate annotation layer, so existing tools for the standard script can be reused.
- Choose texts across a decade so the corpus also documents how an informal writing system gradually stabilises.
- Train one multitask model to produce all annotation layers (transcription, POS, lemma) jointly instead of a pipeline of separate tools.

## For Mova

How these ideas could serve Mova, our small open language model — its files, stages and tests:

- **I1** (learn): Use the same loop for silver data: retrain on each corrected batch, then measure how the correction rate falls per block as a progress signal.
- **I2** (tokens): Have the tokenizer and tagger give emojis and foreign tokens explicit categories rather than dropping them, keeping their contribution traceable downstream.
- **I3** (tokens): Add an explicit normalisation layer that keeps both the raw form and a canonical form for chat-style English, letting lemmatiser and parser work on the canonical form.
- **I4** (learn): Date-stratify evaluation sets to check whether accuracy holds on older versus newer informal text, as a drift control.
- **I5** (tokens): A joint tagger-lemmatiser-morphology head could share evidence across layers while still exposing each layer's decision for explanation.

*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

- [Laboratoire d'Informatique de Grenoble](https://md.research.elamur.ai/free/en/labs/laboratoire-d-informatique-de-grenoble/) — Research institute, Grenoble, France · 24 works · 689 citations

📚 Shelves (6): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 18,614 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 17,965 works · [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 17,450 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 17,296 works · [🤝 Social Sciences](https://md.research.elamur.ai/free/en/shelves/social-sciences/) · 1,954 works · [🎓 Education](https://md.research.elamur.ai/free/en/shelves/social-sciences/education/) · 212 works

🧭 Topics (1): [Semitic languages](https://md.research.elamur.ai/free/en/topics/semitic-languages/) · 3,066 works

## Cite

Cite as: Elisa Gugliotta and Marco Dinarelli. 2020. TArC. Un corpus d'arabish tunisien. Actes de la 6e conférence conjointe Journées d'Études sur la Parole (JEP, 33e édition), Traitement Automatique des Langues Naturelles (TALN, 27e édition), Rencontre des Étudiants Chercheurs en Informatique pour le Traitement Automatique des Langues (RÉCITAL, 22e édition). Volume 2 : Traitement Automatique des Langues Naturelles.

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

```bibtex
@inproceedings{gugliotta2020tarc,
  title = {{TArC. Un corpus d'arabish tunisien}},
  author = {Elisa Gugliotta and Marco Dinarelli},
  booktitle = {Actes de la 6e conférence conjointe Journées d'Études sur la Parole (JEP, 33e édition), Traitement Automatique des Langues Naturelles (TALN, 27e édition), Rencontre des Étudiants Chercheurs en Informatique pour le Traitement Automatique des Langues (RÉCITAL, 22e édition). Volume 2 : Traitement Automatique des Langues Naturelles},
  year = {2020},
  url = {https://aclanthology.org/2020.jeptalnrecital-taln.22/}
}
```

- [← previous Studies in Rabbinic Hebrew](https://md.research.elamur.ai/free/en/papers/heijmans-2020-studies-rabbinic-hebrew/)
- [next → TATL at W-NUT 2020 Task 2: A Transformer-based Baseline System for Identification of Informative COVID-19 English Tweets](https://md.research.elamur.ai/free/en/papers/nguyen-2020-tatl-w-nut-2020-task/)
- [↑ back to papers of 2020](https://md.research.elamur.ai/free/en/papers/2020/page/11/)

## References

26 references, 9 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Al-Badrashiny M., Eskander R., Habash N. & Rambow O. (2014). [Automatic transliteration of romanized dialectal arabic](https://md.research.elamur.ai/free/en/papers/badrashiny-2014-automatic-transliteration-romanized-dialectal-arabic/). In Proceedings of the eighteenth conference on computational natural language learning, p. 30–38.
2. Buckwalter T. & Parkinson D. (2014). A frequency dictionary of Arabic : Core vocabulary for learners. Routledge. [↗](https://doi.org/10.4324/9780203883280)
3. Coulmas F. (2003). Writing systems : An introduction to their linguistic analysis. Cambridge University Press. [find](https://scholar.google.com/scholar?q=Coulmas+F%2E+%282003%29%2E+Writing+systems+%3A+An+introduction+to+their+linguistic+analysis%2E+Cambridge+University+Press%2E)
4. Darwish K. (2014). [Arabizi detection and conversion to Arabic](https://md.research.elamur.ai/free/en/papers/darwish-2014-arabizi-detection-conversion-arabic/). Proceedings of the EMNLP 2014 Workshop on Arabic Natural Language Processing (ANLP), p. 217–224. doi : 10.3115/v1/W14-3629. [↗](https://doi.org/10.3115/v1/w14-3629)
5. Dinarelli M. & Grobol L. (2019a). Hybrid neural models for sequence modelling : The best of three worlds. CoRR. arXiv preprint 1909.07102. [↗](https://arxiv.org/abs/1909.07102)
6. Dinarelli M. & Grobol L. (2019b). Seq2biseq : Bidirectional output-wise recurrent neural networks for sequence modelling. CoRR. arXiv preprint 1904.04733. [↗](https://arxiv.org/abs/1904.04733)
7. Durand O. (2009). Dialettologia araba. ’Sapienza’ University of Rome, ’Studi Orientali’ Faculty. 17 Au moment de la soumission finale, 2 bloques de tokens additionnels ont été téléchargés et ajoutés au corpus. Ces bloques sont donc à transcrire et à corriger avec le bloque 5. Nous sommes en train de développer un système d’apprentissage multitâche pour effectuer tous les niveaux d’annotations avec un seul modèle. [find](https://scholar.google.com/scholar?q=Durand+O%2E+%282009%29%2E+Dialettologia+araba%2E+%E2%80%99Sapienza%E2%80%99+University+of+Rome%2C+%E2%80%99Studi+Orientali%E2%80%99+Faculty%2E+17+Au+moment+de+la+soumission+finale%2C+2+bloques+de+tokens+additionnels+ont+%C3%A9t%C3%A9+t%C3%A9l%C3%A9charg%C3%A9s+et+ajout%C3%A9s+a)
8. Graja M., Jaoua M. & Hadrich-Belguith L. (2010). Tunisian dialect corpus interlocutor (tudicoi). In Arabic Natural Language Processing Research Group (ANLP) : MIRACL Laboratory, Sfax (Tunis). [find](https://scholar.google.com/scholar?q=Graja+M%2E%2C+Jaoua+M%2E+%26+Hadrich%2DBelguith+L%2E+%282010%29%2E+Tunisian+dialect+corpus+interlocutor+%28tudicoi%29%2E+In+Arabic+Natural+Language+Processing+Research+Group+%28ANLP%29+%3A+MIRACL+Laboratory%2C+Sfax+%28Tunis%29%2E)
9. Graja M., Jaoua M. & Hadrich-Belguith L. (2013). Discriminative framework for spoken tunisian dialect understanding. In International Conference on Statistical Language and Speech Processing, p. 102–110: Springer. [find](https://scholar.google.com/scholar?q=Graja+M%2E%2C+Jaoua+M%2E+%26+Hadrich%2DBelguith+L%2E+%282013%29%2E+Discriminative+framework+for+spoken+tunisian+dialect+understanding%2E+In+International+Conference+on+Statistical+Language+and+Speech+Processing%2C+p%2E+102%E2%80%931)
10. Guellil I., Saâdane H., Azouaou F., Gueni B. & Nouvel D. (2019). [Arabic natural language processing : an overview](https://md.research.elamur.ai/free/en/papers/guellil-2019-arabic-natural-language-processing-overview/). Journal of King Saud University-Computer and Information Sciences. doi : 10.1016/j.jksuci.2019.02.006. [↗](https://doi.org/10.1016/j.jksuci.2019.02.006)
11. Gugliotta E. & Dinarelli M. (2020). [Tarc : Incrementally and semi-automatically collecting a tunisian arabish corpus](https://md.research.elamur.ai/free/en/papers/gugliotta-2020-tarc-incrementally-semi-automatically-collecting/). cs.CL. arXiv preprint 2003.09520. [↗](https://arxiv.org/abs/2003.09520)
12. Habash N., Eryani F., Khalifa S., Rambow O., Abdulrahim D., Erdmann A., Faraj R., Zaghouani W., Bouamor H., Zalmout N., Hassan S., Al-Shargi F., Alkhereyf S., Abdulkareem B., Eskander R., Salameh M. & Saddiki H. (2018). [Unified guidelines and resources for Arabic dialect orthography](https://md.research.elamur.ai/free/en/papers/habash-2018-unified-guidelines-resources-arabic-dialect/). In Proceedings of the Eleventh International Conference on Language Resources and Evaluation (LREC 2018), Miyazaki, Japan : European Language Resources Association (ELRA).
13. Haspelmath M. & Tadmor U. (2009). Loanwords in the world’s languages : a comparative handbook. Walter de Gruyter. [find](https://scholar.google.com/scholar?q=Haspelmath+M%2E+%26+Tadmor+U%2E+%282009%29%2E+Loanwords+in+the+world%E2%80%99s+languages+%3A+a+comparative+handbook%2E+Walter+de+Gruyter%2E)
14. Hert P. (1999). Quasi-oralité de l’écriture électronique et sentiment de communauté dans les débats scientifiques en ligne. Réseaux, 17(97). [↗](https://doi.org/10.3406/reso.1999.2171)
15. Masmoudi A., Ellouze Khemakhem M., Estève Y. & Hadrich-Belguith L. (2014a). Tunisian arabic railway interaction corpus. In Arabic Natural Language Processing Research Group (ANLP) : MIRACL Laboratory, Sfax (Tunis). [find](https://scholar.google.com/scholar?q=Masmoudi+A%2E%2C+Ellouze+Khemakhem+M%2E%2C+Est%C3%A8ve+Y%2E+%26+Hadrich%2DBelguith+L%2E+%282014a%29%2E+Tunisian+arabic+railway+interaction+corpus%2E+In+Arabic+Natural+Language+Processing+Research+Group+%28ANLP%29+%3A+MIRACL+Laboratory%2C)
16. Masmoudi A., Habash N., Ellouze M., Estève Y. & Hadrich-Belguith L. (2015). Arabic transliteration of romanized tunisian dialect text : A preliminary investigation. In Computational Linguistics and Intelligent Text Processing - 16th International Conference, CICLing 2015, Proceedings, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), p. 608–619: Springer-Verlag. doi : 10.1007/978-3-319-18111-0_46. [↗](https://doi.org/10.1007/978-3-319-18111-0_46)
17. Masmoudi A., Khmekhem M. E., Estève Y., Belguith L. H. & Habash N. (2014b). [A corpus and phonetic dictionary for Tunisian Arabic speech recognition](https://md.research.elamur.ai/free/en/papers/masmoudi-2014-corpus-phonetic-dictionary-tunisian-arabic/). In Proceedings of the Ninth International Conference on Language Resources and Evaluation (LREC’14), p. 306–310, Reykjavik, Iceland : European Language Resources Association (ELRA).
18. Medhaffar S., Bougares F., Estève Y. & Hadrich-Belguith L. (2017). [Sentiment analysis of Tunisian Dialects : Linguistic ressources and experiments](https://md.research.elamur.ai/free/en/papers/medhaffar-2017-sentiment-analysis-tunisian-dialects-linguistic/). In Proceedings of the third Arabic Natural Language Processing Workshop, p. 55–61: Association for Computational Linguistics. doi : 10.18653/v1/W17-1307. [↗](https://doi.org/10.18653/v1/w17-1307)
19. Meftouh K., Harrat S., Jamoussi S., Abbas M. & Smaili K. (2015). [Machine Translation Experiments on PADIC : A Parallel Arabic DIalect Corpus](https://md.research.elamur.ai/free/en/papers/meftouh-2015-machine-translation-experiments-padic-parallel/). In The 29th Pacific Asia Conference on Language, Information and Computation, shanghai, China. hal : hal-01261587.
20. Meftouh K., Harrat S. & Smaïli K. (2018). PADIC : extension and new experiments. In 7th International Conference on Advanced Technologies ICAT, Antalya, Turkey. hal : hal-01718858. [find](https://scholar.google.com/scholar?q=Meftouh+K%2E%2C+Harrat+S%2E+%26+Sma%C3%AFli+K%2E+%282018%29%2E+PADIC+%3A+extension+and+new+experiments%2E+In+7th+International+Conference+on+Advanced+Technologies+ICAT%2C+Antalya%2C+Turkey%2E+hal+%3A+hal%2D01718858%2E)
21. Tobaili T. (2016). [Arabizi identification in twitter data](https://md.research.elamur.ai/free/en/papers/tobaili-2016-arabizi-identification-twitter-data/). In Proceedings of the ACL 2016 Student Research Workshop, p. 51–57.
22. Younes J., Achour H. & Souissi E. (2015). Constructing linguistic resources for the tunisian dialect using textual user-generated contents on the social web. In International Conference on Web Engineering, p. 3–14: Springer. [find](https://scholar.google.com/scholar?q=Younes+J%2E%2C+Achour+H%2E+%26+Souissi+E%2E+%282015%29%2E+Constructing+linguistic+resources+for+the+tunisian+dialect+using+textual+user%2Dgenerated+contents+on+the+social+web%2E+In+International+Conference+on+Web+Enginee)
23. Younes J., Achour H., Souissi E. & Ferchichi A. (2018a). Survey on corpora availability for the tunisian dialect automatic processing. In 2018 JCCO Joint International Conference on ICT in Education and Training, International Conference on Computing in Arabic, and International Conference on Geocomputing (JCCO : TICET-ICCA-GECO), p. 1–7: IEEE. [find](https://scholar.google.com/scholar?q=Younes+J%2E%2C+Achour+H%2E%2C+Souissi+E%2E+%26+Ferchichi+A%2E+%282018a%29%2E+Survey+on+corpora+availability+for+the+tunisian+dialect+automatic+processing%2E+In+2018+JCCO+Joint+International+Conference+on+ICT+in+Education+a)
24. Younes J., Achour H., Souissi E. & Ferchichi A. (2020). Romanized tunisian dialect transliteration using sequence labelling techniques. Journal of King Saud University-Computer and Information Sciences. [↗](https://doi.org/10.1016/j.jksuci.2020.03.008)
25. Younes J., Souissi E., Achour H. & Ferchichi A. (2018b). A sequence-to-sequence based approach for the double transliteration of tunisian dialect. Procedia computer science, 142, 238–245. [find](https://scholar.google.com/scholar?q=Younes+J%2E%2C+Souissi+E%2E%2C+Achour+H%2E+%26+Ferchichi+A%2E+%282018b%29%2E+A+sequence%2Dto%2Dsequence+based+approach+for+the+double+transliteration+of+tunisian+dialect%2E+Procedia+computer+science%2C+142%2C+238%E2%80%93245%2E)
26. Zribi I., Ellouze M., Hadrich-Belguith L. & Blache P. (2015). Spoken tunisian arabic corpus ”stac” : Transcription and annotation. Research in Computing Science, 90, 123–135. [↗](https://doi.org/10.13053/rcs-90-1-9)

*Updated 2 November 2026*