FRENCH-YMCA: A FRENCH Corpus meeting the language needs of Youth, froM Children to Adolescents

Cherifa Ben Khelil; Jean-Yves Antoine; Anaïs Halftermeyer; Frédéric Rayar; Mathieu Thebaud

cs.CLApr 7, 2026

FRENCH-YMCA: A FRENCH Corpus meeting the language needs of Youth, froM Children to Adolescents

Authors: Cherifa Ben Khelil, Jean-Yves Antoine, Anaïs Halftermeyer, Frédéric Rayar, Mathieu Thebaud

Organizations: EFREI Research Lab - University of Paris Panthéon Assas, France · LIFAT - University of Tours, France · LIFO - University of Orleans, France · CMRRF, Mutuality of Morbihan, France

Abstract

In this paper, we introduce the French-YMCA corpus, a new linguistic resource specifically tailored for children and adolescents. The motivation for building this corpus is clear: children have unique language requirements, as their language skills are in constant evolution and differ from those of adults. With an extensive collection of 39,200 text files, the French-YMCA corpus encompasses a total of 22,471,898 words. It distinguishes itself through its diverse sources, consistent grammar and spelling, and the commitment to providing open online accessibility for all. Such corpus can serve as the foundation for training language models that understand and anticipate youth's language, thereby enhancing the quality of digital interactions and ensuring that responses and suggestions are age-appropriate and adapted to the comprehension level of users of this age.

Figures & tables

Corpus name	Text source	Text genre	Number of files	Age range
CHILDES	Transcriptions from audio and video recordings of children’s conversations	Conversation	2,275	6 months to 11 years
E-calm	Student Essays and writings	Formal writing	4,500	6 to 18 years
Philosophèmes	Transcriptions of philosophical discussion	Philosophical conversation	22	5 to 18 years
Grenouille	Student Essays	Formal writing	400	8 to 14 years
C’est pas moi	Transcriptions of discussions on ”Lying” and writings	Conversation and formal writing	53	6 to 8 years
PréCPhi	Philosophical Writings	Formal writing	1,000	8 to 15 years

Table 1: Overview of varied corpora for children and adolescent.

Figure 1: Example of a text file and its corresponding JSON metadata file from the French-YMCA Corpus.

Text source	Age	Language register	Number of files	Number of words	Avg. sentence length	Data availability
Books and stories for children	Children	Literature	2 211	2 935 108	12	¿80% open; rest by agreement
Children e-journals	Children	Formal newspaper	16 026	5 103 070	17	By agreement only
Adolescents e-journals	Adolescents	Formal newspaper	2 079	949 753	16	By agreement only
Vikidia	Adolescents	Formal informative	17 237	7 380 229	20	Public
Animated movie subtitles	Children	Informal	1 647	6 103 738	7	By agreement only
Total			39 200	22 471 898	12

Table 2: Distribution of the collected texts based on their source type.

AraYoungVoices: A Diverse L1/L2 Corpus of Arabic Child and Adolescent Speech
Oct 4, 2026Shammur Absar Chowdhury, Zien Sheikh Ali, Houssam Eddine-Othman Lachemat +1Child Speech Recognition Automatic Speech Recognition
Qatar Computing Research Institute, Qatar
Measuring Form and Function in Language Models
May 27, 2026Héctor Javier Vázquez Martínez, Charles YangLLM Evaluation Child Language Acquisition
University of Pennsylvania Department of Linguistics and Computer and Information Science
Learning from Child-Directed Speech in Two-Language Scenarios: A French-English Case Study
Mar 13, 2026Liel Binyamin, Elior SulemLanguage Model Pretraining Language Modeling
Faculty of Computer and Information Science, Institute for Applied AI Research Data Science Research Center Ben-Gurion University of the Negev

Oct 4, 2026J/K move · Enter open · S save

AraYoungVoices: A Diverse L1/L2 Corpus of Arabic Child and Adolescent Speech

Shammur Absar Chowdhury, Zien Sheikh Ali, Houssam Eddine-Othman Lachemat, Hamdy Mubarak

Qatar Computing Research Institute, Qatar

Child Speech Recognition Automatic Speech Recognition Speech Corpus Construction

Open paper AskPDF arXiv

State-of-the-art ASR systems primarily target native adult speech, leading to substantial performance gaps for children, adolescents, and L2 speakers. We introduce AraYoungVoices, a 151.72-hour Arabic read-speech corpus from 286 speakers aged 7--18, comprising AraKids (7--12) and AraTeens (13--18). The corpus includes 146 native Arabic (L1) and 140 second-language (L2) speakers, with native speakers spanning Egyptian, Gulf, Levantine, and North African dialectal backgrounds and L2 speakers representing diverse linguistic backgrounds across the Americas, Asia, Africa, and Europe. We benchmark four pretrained ASR models under zero-shot and fine-tuned settings using unseen-speaker-

\&

-unseen-prompt (USUP) and unseen-speaker-

\&

-seen-prompt (USSP) evaluations. Results show that L2 speech remains substantially more challenging than L1 speech, with the largest errors observed mainly for younger L2 speakers. Age-specific fine-tuning improves the matched age group, while joint fine-tuning provides a stronger balance across populations. ASR hypotheses are also consistently closer to the standard reading prompt than to the verbatim transcription, particularly for L2 speech, suggesting partial normalization of reading deviations.

Abstract

Explore similar work

AraYoungVoices: A Diverse L1/L2 Corpus of Arabic Child and Adolescent Speech

Measuring Form and Function in Language Models

Learning from Child-Directed Speech in Two-Language Scenarios: A French-English Case Study