Thanks to the help of the French National Agency for Research (ANR), manufacturers, researchers on spoken interactive systems and researchers in French language didactics are developing a software designed to help young French learners (students aged 6 to 7) and students learning French as a foreign language (FFL - level A1 to B1) to produce written texts. The innovative feature of the research lies in the development and the use of vocal synthesis allowing students to re-read their texts as they are writing them in order to help them evaluate the difference between what is written and what they want to write (Crinon, 2000). The tool seeks to develop students’ phonological and alphabetical awareness through writing and proofreading (Mangenot, 96 ; Mangenot 2000) and, then, to develop creativity by supporting the students’ efforts through out their proofreading process.

Didactic activities covering different phases of language acquisition are proposed; they range from activities targeting phonological acquisition to more traditional exercises such as word or phrase dictations, exercises on punctuation, word order or coherent liaisons and activities on speech styles. Most activities are based upon texts coming from literature heritage: aesop's fables, nursery rhymes, fairy tales; they are starting points for creative activities, writing games or text transformations.

Five research directions are explored:

- Offering high-quality synthetic voices as understanding the activities and proofreading efficiently in order to correct encoding mistakes largely depend on the quality of voices.

- Defining the types of activities allowing working on specific linguistic skills together with respecting students’ mixed abilities.

- Offering a tool able to vocalize exercises available on a Web 2.0 platform and to collect the students’ answers.

- Setting up a device allowing keeping track of students’ profile

- Offering a tool used to create new exercises adapted to students’ profiles automatically.

This ANR project seeks to provide a prototype that will be tested both with young primary students and A2-B1 level learners of French as a foreign language within two years.

The scientific added value of this ANR project lies in the control of the prosodic relevance in didactic context. Its aim is to allow students to master more efficiently the skills required to create free texts. The tool innovation mainly lies in its modularity and adaptability likely to foster larger autonomy.

Current platforms dedicated to learning languages are often limited to prefabricated exercises associated to audio files recorded beforehand. They don’t offer the possibility to have access to new exercises if specific skills need further work. This adaptability to the learner’s needs is nevertheless crucial in learning (less-able students starting writing, having dysgraphia problems, FFL, etc.) as the acquisition rhythms differ from one individual to another. Two things are essential make this modularity possible: exercises are classified according to the skills they enable to develop; learners’ productions are assessed according to a grid reference related to the type of exercises.

To our knowledge, there is no solution of this type for the acquisition of French on the market.

- Crinon J., "L'aide à l'écriture", Repères n° 11 Ecriture et traitement de texte, INRP, p. 139-157, 1995. - Crinon J., "Ecrire mieux avec le traitement de texte ?", in Anis J. et Marty N. Lecture-écriture et nouvelles technologies, Paris, CNDP, p. 47-58, 2000. - Eskenazi M., "An overview of spoken language technology for education", Speech Communication, vol.51, n°10, pp.832-844, 2009. - Gelan A., "Language and Text-to-Speech Technologies for Highly Accessible Language & Culture Learning", Int. Journal of Emerging Technol. in Learning (iJET) Vol 6, No 2, pp. 11-14, 2011. - Guyomard M., Siroux J., "A general public application of pedagogic and linguistic vocations of speech synthesis : ordictée", ACL Anthology Proc. of the Research on Computational Linguistics Int. Conf., pp.316-324, 1997. - Handley Z., "Is Text-to-speech synthesis ready for use in computedassisted language learning ?", Speech Communication, vol.51, n°10, pp. 906-919, 2009. - Handley Z., Hamel M.-J., "Establishing a methodology for benchmarking speech synthesis for computer-assisted language learning (CALL)", Language Learning Technol., vol.9, n°3, pp.99–119, 2005. - Mangenot F., "Les aides logicielles à l'écriture", Paris, CNDP, 1996. - Mangenot F., "Aide à l'écriture ou environnements d'écriture ?", in Anis J. et Marty N. Lecture-écriture et nouvelles technologies, CNDP, p. 59-67, 2000. - Martin P., "WinPitch LTL, un logiciel multimédia d'enseignement de la prosodie", Alsic (Apprentissage des Langues et systèmes d'information et de communication), vol 8/2, pp.95-108, 2005. - Marty N., Informatique et nouvelles pratiques d'écriture, Nathan, 253 p., 2005. - Mercier G., Guyomard M., Siroux J., Bramoullé A., Gourmelon H., Guillou A., Lavannant P., "Courseware for breton spelling pronunciation and intonation learning", Proc. of InSTIL, pp.145-148, 2000. - Santiago-Oriola C., "Vocal synthesis in a computerized dictation exercise", Proc. Of Eurospeech, pp.191-194, 1999. - Saussure (de) F., "Cours de linguistique générale", 1916, réédition 1979 : Payot, Paris, 1916.