--- title: "Bilingual resources for Moroccan Sign Language Generation and" description: "The authors argue that bilingual teaching through Moroccan Sign Language (MSL) and Standard Arabic (SA) works best, and… 1 reference." authors: ["Abdelhadi Soudi", "Corinne Vinopol", "Kristof Van Laerhoven"] year: 2025 venue: "Proceedings of the 18th Workshop on Building and Using Comparable Corpora (BUCC)" source: "https://aclanthology.org/2025.bucc-1.1/" 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/soudi-2025-bilingual-resources-moroccan-sign-language/" --- [Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/soudi-2025-bilingual-resources-moroccan-sign-language/) [home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2025](https://md.research.elamur.ai/free/en/papers/2025/page/17/) / Bilingual resources for Moroccan Sign Language Generation and Standard Arabic Skills Improvement of Deaf Children # Bilingual resources for Moroccan Sign Language Generation and Standard Arabic Skills Improvement of Deaf Children **Authors:** [Abdelhadi Soudi](https://md.research.elamur.ai/free/en/authors/abdelhadi-soudi/), Corinne Vinopol and [Kristof Van Laerhoven](https://md.research.elamur.ai/free/en/authors/kristof-van-laerhoven/) · **Year:** 2025 · **Venue:** [Proceedings of the 18th Workshop on Building and Using Comparable Corpora (BUCC)](https://md.research.elamur.ai/free/en/venues/bucc/) ([BUCC 2025](https://md.research.elamur.ai/free/en/venues/bucc/2025/)) · **Publisher:** [Association for Computational Linguistics](https://md.research.elamur.ai/free/en/publishers/association-for-computational-linguistics/) · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/) **Links:** [Publisher's page](https://aclanthology.org/2025.bucc-1.1/) · [PDF](https://aclanthology.org/2025.bucc-1.1.pdf) ## Summary The paper addresses the poor access to education for Deaf children in Morocco. Most of them do not attend school, instruction is given only in spoken or written Arabic or French, and almost no interpreters or trained teachers are available. The authors argue that bilingual teaching through Moroccan Sign Language (MSL) and Standard Arabic (SA) works best, and they present several tools that support it. The first is the MSL Generator, a rule-based machine translation system. It analyses Arabic input with BAMA and the Stanford parser, then looks each word up in a sign dictionary and outputs sign graphics or videos. When a word has several senses, the user picks the intended one. The second is MSL Clip and Create, which offers: - a bilingual database of 3,000 entries, - a Publisher for adding signs to reading passages, - six game and worksheet templates, - a story-builder. The third is Madrasati-Signs, a vetted crowdsourcing platform for collecting signs, including regional variants. The authors chose rules over data-driven methods because MSL has almost no parallel corpora. The paper reports no quantitative evaluation. Its main result is adoption: the Moroccan Ministry of Education endorsed the tools and helped distribute them free to Deaf associations. ## Key points - The Generator pipeline runs in this order: - tokenization, - Buckwalter Arabic morphological Analyzer (BAMA; 40,648 lemmas plus three affix–stem compatibility tables), - Stanford parser, - dictionary lookup, - Named Entity dictionary fallback, - "unknown word" or a context-based suggestion. - For polysemous words, every sense is displayed with its own sign graphic and the user chooses one. One worked example is a word that can mean "dissertation," "letter," or "mission." - An avatar output mode is in development. It converts HamNoSys descriptions to SiGML and also draws on GessyCA. A pictorial, easier-to-read HamNoSys keyboard supports the work, and animation of the 3,000 dictionary words is underway. - MSL Clip and Create includes: - a 3,000-sign SA–MSL database covering 21 domains, - a Publisher, - six templates (crosswords, word searches, Bingo, matching games, flashcards, fingerspelling scrambles), - a story-builder with three Hispanic folktales and three Moroccan textbook stories, with one-click Arabic diacritization. - Madrasati-Signs is modelled on AfricaSign. It accepts webcam recordings or video uploads and labels regional variants as sign 1, sign 2, and so on. Contributions are limited to signers vetted by regional Deaf associations, who record their region and Deafness affiliation (e.g., CODA). ## Interesting ideas - When a word is ambiguous, the system shows every candidate sense to a human and asks them to choose, rather than guessing silently. Unresolved ambiguity becomes an explicit, inspectable decision point. - Lookup follows a fixed fallback chain: main lexicon, then named-entity list, then an explicit "unknown" result or a context-based suggestion. Each failure is therefore traceable to a specific stage. - Regional variants of the same concept are stored as numbered alternatives instead of being merged into one canonical form. This keeps variation visible without splitting the vocabulary. - Crowdsourced data comes only from contributors vetted by the community, and each contributor attaches demographic metadata. This lets data quality and variation be filtered after collection. - Signs are described in a notation (HamNoSys) and compiled into a separate rendering format (SiGML) before animation. This separates linguistic description from presentation. ## For Mova How these ideas could serve Mova, our small open language model — its files, stages and tests: - **I1** (world): When a lemma maps to several WordNet senses that context cannot separate, Mova could record all of them and flag the choice for a teacher tier, instead of committing to one sense unseen. - **I2** (tokens): Lemma and lexicon lookup could carry a provenance tag (lexicon / NE gazetteer / guessed / unknown), so later steps know how much to trust each token. - **I3** (back): The generator could keep numbered surface variants per concept (e.g., British vs. American wording) and choose among them through an explicit style parameter. - **I4** (learn): Silver annotations could carry source metadata (which witness parser, which teacher tier), so agreement can be weighted and checked against negative controls. - **I5** (back): Mova could compile the meaning graph into an intermediate surface specification before realizing text, so generation errors can be located either in content planning or in rendering. *Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.* 📚 Shelves (2): [🖥️ 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 🧭 Topics (2): [Semitic languages](https://md.research.elamur.ai/free/en/topics/semitic-languages/) · 3,066 works · [✍️ Text generation](https://md.research.elamur.ai/free/en/topics/text-generation/) · 214 works ## Cite Cite as: Abdelhadi Soudi, Corinne Vinopol and Kristof Van Laerhoven. 2025. Bilingual resources for Moroccan Sign Language Generation and Standard Arabic Skills Improvement of Deaf Children. Proceedings of the 18th Workshop on Building and Using Comparable Corpora (BUCC). The paper itself, not this page. BibTeX from the library's record of the work. ```bibtex @inproceedings{soudi2025bilingual, title = {{Bilingual resources for Moroccan Sign Language Generation and Standard Arabic Skills Improvement of Deaf Children}}, author = {Abdelhadi Soudi, Corinne Vinopol and Kristof {Van Laerhoven}}, booktitle = {Proceedings of the 18th Workshop on Building and Using Comparable Corpora (BUCC)}, year = {2025}, url = {https://aclanthology.org/2025.bucc-1.1/} } ``` - [← previous Bibliography of Ethiopian Semitic, Cushitic, and Omotic Linguistics XXVII: 2023](https://md.research.elamur.ai/free/en/papers/krzyzanowska-2025-bibliography-ethiopian-semitic-cushitic-omotic/) - [next → Binary_Bunch at AraHealthQA Track 1: Arabic Mental Health Q&A Classification Using Data Augmentation and Transformer Models](https://md.research.elamur.ai/free/en/papers/bhattacharjee-2025-binary-bunch-arahealthqa-track-1/) - [↑ back to papers of 2025](https://md.research.elamur.ai/free/en/papers/2025/page/17/) ## References 1 references, 0 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet. 1. Britta Bauer, Sonja Nießen, and Hermann Hienz. 1999. Towards an automatic sign language translation system. In Proceedings of the international workshop on physicality and tangibility in interaction: towards new paradigms for interaction beyond the desktop, Siena, Italy. *Figure 8: Screenshots of bingo and fingerspelling templates* Robert Bayley, Clayton Valli, and Ceil Lucas. 2001. Sociolinguistic variation in American sign language. Gallaudet University Press. [find](https://scholar.google.com/scholar?q=Britta+Bauer%2C+Sonja+Nie%C3%9Fen%2C+and+Hermann+Hienz%2E+1999%2E+Towards+an+automatic+sign+language+translation+system%2E+In+Proceedings+of+the+international+workshop+on+physicality+and+tangibility+in+interaction%3A+) *Updated 2 November 2026*