Versión clásicaVersión móvil
OpenEdition Books

Le français en contextes

Henry Tyne

III - Maîtrise et acquisition : études de données

Do French immersion children sound French when they read aloud?

Lucie Vialettes-Basmoreau y Nathalie Spanghero-Gaillard


Il est admis que les modalités orales et écrites sont liées lors de l’apprentissage de la lecture. Ziegler et Goswami (2005, 2006) ont mis en évidence les difficultés auxquelles les enfants peuvent être confrontés quand ils apprennent à lire. Les enfants qui ont participé à notre étude sont scolarisés dans une école d’immersion française située aux Etats-Unis. Ils sont anglophones mais apprennent à lire en français à l’école. Pour voir dans quelle langue la prononciation est la meilleure quand les enfants lisent à voix haute, nous avons mené une étude longitudinale. Les résultats montrent que les enfants sont meilleurs en anglais qu’en français. Ces résultats nous amèneront à discuter les aspects qui devraient être mis en avant pour améliorer la prononciation des enfants en français.

It is acknowledged that speech and writing are linked as far as learning to read is concerned. Ziegler and Goswami (2005, 2006) have highlighted the difficulties children can be confronted with when learning to read. The children who took part in our study are enrolled in a French immersion school located in the United States. They are Anglophone but learn to read in French at school. To see in which language pronunciation was better when children read aloud, we implemented a longitudinal study. The results show that the children do better in English than in French. These results will lead us to discuss what should be emphasized to improve the children’s pronunciation in French.

Texto completo


1Learning to read is a major step in a child’s education as well as for their cognitive development. Usually, a child learns to read in the language they first learn to talk in. What happens, though, when it is not done with the child’s first language (L1)? This question is raised in a French immersion school located in the United States. In this institution, the children learn how to read in French (as a second language, L2) while English is their L1.

2In which language do the children perform best when they are asked to read aloud? Is it in their L1, since they have a high level of exposure to this language, or is it in French since this language has a shallower orthography (see below) and since they learn how to read in this language at school?

3The study we carried out intends to answer this question. First, the background for the study will be dealt with along with theoretical contributions that link learning to read with the spoken aspect of the language. We will then present our hypothesis. This will lead us to outline the method that was used for the study, looking at the participants, the procedure and the data collected. The results will then be presented and discussed.

Background for the study

4The context in which our study takes place is specific and needs to be explained. It is a French immersion school, Normandale, located in the suburbs of Minneapolis, a city of the Midwest of the United States of America. Minneapolis is close to Canada, a country with two official languages (English and French) since 1969, when French became an official language. In 1972, the Saint-Lambert experiment (Lambert & Tucker 1972) paved the way for French immersion schools throughout Canada. This type of school spread into the USA and Normandale picked French since Canada is close by and it is relatively easy to access teaching materials. French was also privileged because of the historical presence of French back in the 17th century.

5Normandale is an early French immersion school; the children are spoken to in French all day long form kindergarten to 5th grade. The only moments when children can use English are during breaks (lunchtime and recess). Normandale is a strong and free model school (Baker 1995; Gajo 2005) with an educational goal. In other words, this school was not intended to link two communities that would speak two different languages (as is the case in Canada). The parents choose to send their children to Normandale, the general assumption being that if they go to an immersion school (no matter the language), their chances of academic (and also hopefully social) achievement are higher. Children study French so that they can communicate with French speakers if necessary but most of all they study French to get access to other knowledge. So, language is not only an end but it is also a means to have access to new knowledge.

6As far as school learning is concerned, children learn to read in French aged six years old. When they are eight, the school introduces them to reading in English. Parents are asked to provide their children with written material in English so that they are in touch with the written form of their L1. This choice (of differing learning to read in English at school) was made according to Cummins’s Principle of Interdependence (1981: 29) : “to the extent that instruction in Lx is effective in promoting proficiency in Lx, transfer of this proficiency to Ly will occur provided there is adequate exposure to Ly (either in school or environment) and adequate motivation to learn Ly”. In other words, if the child is regularly exposed to their L1, is eager to learn it, and the education they get in L2 is adequate, then a transfer of skills can occur. Cummins emphasizes that the two languages known by the child would be interdependent. Hamers and Blanc (1983) claim that if the children have enough skills in their L1, then they will be able to acquire skills in a L2 without their L1 being affected by exposure to a L2 while learning how to read.

7Furthermore, research on literacy indicates that speech and writing are linked when the child learns how to read. The phonological loop (Baddeley 1993, 2000), located in the working memory, and phonological awareness (Gombert 1990) are two tools that can be taken into account.

8The phonological loop consists of two parts: a phonological store containing auditory information that decays rapidly, and an articulatory loop, or articulatory rehearsal component, that is in charge of reviving the auditory information from the phonological store. Thus, the phonological loop is relevant to perception (through the role of the phonological store) and also to production (via the articulatory loop).

9Phonological awareness can be defined as “the ability to identify the components of the linguistic units and to manipulate these components deliberately” (Gombert 1990, in Demont 2001: 274). This concept is based on the linguistic analysis of alphabetical languages’ writing systems. In these languages, a word can be divided into syllables, onset and rimes, that are infra-syllabic units, and phonemes, the smallest linguistic units.

10It appears that phonological awareness is at the same time a prerequisite to learning to read and a consequence of this learning. On the one hand, the first two levels of phonological awareness, syllable awareness and onset-rime awareness, seem to be a prerequisite to learning how to read (Goswami & Mead 1992; Demont & Gombert 1996; Stuart & Coltheart 1988; Sprenger-Charolles et al. 2003). Indeed, an early training in phonological awareness improves reading proficiency (Bradley & Bryant 1983; Lundberg et al. 1988; Hatcher et al. 2004). On the other hand, the third level of phonological awareness, i.e. phonemic awareness, can be considered as a consequence of the exposure to written stimuli (Perfetti 1985; Ehri 1998).

11Frost et al. (1987) insist on the fact that alphabetical languages can be divided into two groups, as far as spelling-sound correspondences are concerned: languages that have a shallow orthography and languages that have a deep orthography. In the first group, the spelling-sound correspondences are transparent, that is to say that one phoneme matches one grapheme and vice versa. For the second group, the correspondences are not univocal; one phoneme can be transcribed in different ways and one grapheme can be pronounced in different ways.

12Interactive models of learning to read insist on the link between speech and writing, and take into account the orthographical depth of the languages. The model designed by Ziegler and Goswami (2006) features three types of difficulties children who learn to read can be confronted with:

  • For the young reader (5-6 years old), all the phonological units are not directly attainable. Phonological awareness of the beginner reader does not account for the full range of the units offered by the language: it is an availability problem.

  • In a language that has a deep orthography (which is the case for French and English), on the one hand, the same phoneme can be transcribed in different ways, on the other hand, the same grapheme can be pronounced in various ways: it is a consistency problem.

  • The third element derives from the second: when spelling-sound relations are not consistent, it is not enough to identify spelling-sound correspondences; it is also necessary to retain units that are larger than the phoneme. The novice reader is therefore compelled to remember a larger number of units when learning to read in a language that has a deep orthography. Learning is more costly. The reader requires a more fine-grained tool (granularity) in order to identify precisely spelling-sound correspondences.

13An experiment was devised to investigate the Normandale situation, i.e. involving English-speaking children going to a French immersion school and learning to read in French, looking at the link between speech and writing, and the effects it can have on learning to read according to orthographical depth.


14French is a language that is considered to have a more ‘transparent’ orthography than English (Jaffré & Fayol 1997; Sprenger-Charolles 2004) even though its writing system is opaque. The children at Normandale learn how to read in French in order to improve their pronunciation when reading aloud. We are interested in finding out whether the fact that children first learn to read in French, a shallower language than their L1, implies that their pronunciation in French, when reading aloud, is better than when they read in English, the language to which they are more exposed. A year-long study was carried out from January 2009 to January 2010.



15The participants are four children enrolled in first grade at Normandale: two girls, Sophie and Gillian, and two boys, Gabriel-Pierre and Nathan. Their mean age in January 2009 was 7:1 years old. They had already been learning French for one year (in kindergarten).

16A questionnaire was completed by their parents to find out more about their background and about their contact with English written material at home. The answers to the questionnaire helped to establish distinct profiles for the children (see Fig. 1).

Fig. 1. Children’s profiles

Fig. 1. Children’s profiles

17The participants have different profiles. First, the L1 of the parents is not always English. Gillian’s parents are both native English speakers so are Sophie and Gabriel-Pierre’s mothers. The fathers of these two children are French native speakers. While Sophie’s father hardly speaks French with her, Gabriel-Pierre’s father, on the contrary speaks French with his son most of the time. As for Nathan’s parents, their L1 is Tagalog (a language from the Philippines) but his mother told us that they only speak English to their child.

18It is also important to note that the involvement of the parents in English reading is high. It can be seen from the results in the following categories: monthly reading duration (amount of time when they read with their child for a month); time of reading (they read to their children at an appropriate time, not immediately after school, Fortune & Tedick 2003); monthly reading frequency (they read every day to / with their child). The parents also enjoy reading with their child. As far as the assessment of their child is concerned, differences appear. Sophie’s parents notice that she reads using smaller units (letter by letter) than the other children at the same time of the year.


19The material for the study consists of five short narratives (with humans or animals as protagonists) in French and English. The narrative form was chosen in an effort to make the test as ecological as possible. The texts contain 36 words in French and 34 words in English. These texts were not translated from one language to another. The texts were created in relation with the first grade curriculum: children work according to topics that were included in the texts so as to build on existing knowledge and vocabulary. The texts were not illustrated.


  • 1 Access to this platform was granted by Benoît Jeunier, Research Director at IUFM (teacher training (...)

20As mentioned before, we carried out a longitudinal study that lasted one year. Meetings with the children were carried out via the Internet platform, Centra 71. This tool allowed us to talk to the children, to share documents with them and most importantly to record them while they read. The order in which the children read the texts (French and English) was random from one month to another but it was the same for all the children for a given month. The spoken data was fully transcribed orthographically, and the moments when the children read were transcribed using the International Phonetic Alphabet.

21In order to address the question we asked (i.e. is pronunciation better in French or in English when reading aloud?), we identified phoneme errors in French and English in the transcript. Judges (native French speaking and native English speaking) also listened to children’s reading; they were asked to rate the children’s pronunciation using a scale ranging from 0 (“the pronunciation is very bad”) to 5 (“the pronunciation is very good”), according to what they considered to be a ‘good’ French accent.


22Figures 2 and 3 show that the number phoneme errors is higher in French than in English, for all the children and also that the judges gave them better grades for pronunciation in English.

Fig. 2. Number of phoneme errors in French and in English (all children)

Fig. 2. Number of phoneme errors in French and in English (all children)

Fig. 3. Pronunciation score for all the children in French and in English

Fig. 3. Pronunciation score for all the children in French and in English

23A U Mann-Withney non-parametric test (see Fig. 4) showed that the number of errors in French was significantly higher than the number of errors in English (U=138.0, z=3 .10, p < .001). The difference between the children’s scores in the two languages is not significant.

Fig. 4. Phoneme errors in French and in English (all children)

Fig. 4. Phoneme errors in French and in English (all children)

24Given that the number of vowel errors was higher than the number of consonant errors, it was decided to refine our results and to focus on vowels when the children read in French (see Fig. 5).

Fig. 5. French vowel errors (all children)

Fig. 5. French vowel errors (all children)

25The lighter sections (Fig. 5) indicate errors due to English, for example diphthongization, absence of nasalization, or over-lengthening of vowels. The section in dark grey (intralanguage substitutions, 25 %) accounts for errors which involve the substitution of one French phoneme by another. Finally, it was decided to study the progression of these types of errors from January 2009 to January 2010. The results show (Figs. 6 and 7) that there are fewer errors in January 2010 and some completely disappear (weakened vowels, unexpected interlanguage substitution).

Fig. 6. Mistakes for French vowels by all the children, January 2009

Fig. 6. Mistakes for French vowels by all the children, January 2009

Fig. 7. Mistakes for French vowels by all the children, January 2010

Fig. 7. Mistakes for French vowels by all the children, January 2010


26The results show that children have better pronunciation in English than in French when they read aloud. Despite the fact that French has a shallower orthography than English and that the children learn how to read in this language, they still have better scores in their L1 (English). Thus, as far as learning to read in an immersion situation is concerned, exposure to the first language (English) appears to be more important than the fact that spelling-sound correspondences are transparent or the fact that the language is taught (L2 French). Even if the children are immersed in French language every day of the week for five to six hours, and even if this language has a shallower orthography, they still have more difficulties reading aloud than they have in English. The only child who has results that are close in French and in English is Gabriel-Pierre whose father is a French native speaker and talks to him in French.

27Furthermore, it must be underlined that when children speak in English they interact with their families and friends, they are a part of the society they live in. French is used at school in order to acquire new skills in mathematics, science, history, etc. The relation they have with the two languages is different by nature; it is more affective with English than it is with French.

28As far as vowel sounds are concerned, when reading in French, children appear to be inclined to refer to the phonological system of English. Regarding the production of vowels, the phonological recoding stage (the moment the child converts the spelling seen in the corresponding sounds) is under the influence of English. This is an example of a negative transfer (Lecocq et al. 2004), that is to say that the L1 is creating an interference with the immersion language. For instance, Gillian pronounces the words et, des, maison with a diphthong, which gives us [eɪ], [deɪ] [meɪzõ].

29Exposure to oral input seems to be a major element to take into account when the children begin to read. A greater exposure to French as well as phonetic practice might allow the English-speaking children to pronounce better in French from kindergarten. It would be easier for them to match spelling to sound when they begin to read texts. It would be interesting for schools to develop pronunciation activities in kindergarten. For instance, it could be an activity in which the children would have to focus on the sounds of the sentence and to mimic the feeling conveyed by the intonation. These exercises combined with phonetic correction (we especially refer to the verbo-tonal method as advocated by Renard 2002) might improve pronunciation from the first year of school on.

30Recent studies (Zedda 2006; Cornaz et al. 2010) have shown the advantages of exercises involving singing to improve the learners’ pronunciation. Indeed, various studies have demonstrated the positive impact of having a musical activity on the language acquisition process (Bancroft 1985; Lowe 1998; Magne et al. 2006)


31The small number of participants for this study does not allow us to generalize the results. However, it gives an accurate idea of what happens for these children and it also emphasizes the fact that in immersion school children have different profiles: some come from English-speaking families, others from families in which the target language is spoken by one of the parents and others in which the L1 of the parents is neither the language of the society nor the language of the school. More research (both quantitative and qualitative) on immersion children is needed. Quantitative research enables us to generalize the results but qualitative research reveals what cannot be taken into account when there are a large numbers of participants: the differences between the participants, the social and family background of the children.


Baddeley, A. 1993, La mémoire humaine: théorie et pratique, Grenoble: PUG.

Baddeley, A. 2000, “The episodic buffer: a new component of working memory?”, Trends in cognitive sciences 4(11), p. 417-423.

Baker, C. 1995, A parents’ and teachers’ guide to bilingualism, Clevedon: Multilingual Matters.

Bancroft, W. 1985, “Music therapy and education”, Journal of the society for accelerative learning and teaching 10(1), p. 3-16.

Bradley, L., & P. Bryant 1983, Categorizing sounds and learning to read: a causal connection, Nature 301, p. 419-421.

Cornaz, S., N. Henrich & N. Vallée 2010, “L’apport d’exercices en voix chantée pour la correction phonétique en langue étrangère : le cas du français langue étrangère appliqué à des apprenants italiens d’âge adulte”, Cahiers de l’APLIUT 24(2), p. 103-119.

Cummins, J. 1981, “The role of primary language development in promoting educational success for language minority students”, in California State Department of Education (ed.), Schooling and language minority students: A theoretical framework, Los Angeles: National Dissemination and Assessment Center, p. 3-49.

Demont, E. 2001, “Contribution de l’apprentissage précoce d’une deuxième langue au développement de la conscience linguistique et à l’apprentissage de la lecture”, International journal of psychology 36(4), p. 274-285.

Demont, E., & J. Gombert 1996, “Phonological awareness as a predictor of recoding skills and syntactic awareness as a predictor of comprehension skills”, British journal of educational psychology 66(3), p. 315-332.

Ehri, L. 1998, “Grapheme-phoneme knowledge is essential for learning to read words in English”, in J. Metsala & L. Ehri (eds.), Word recognition in beginning literacy, Mahwah, NJ: Erlbaum, p. 3-40.

Fortune, T. & D. Tedick 2003, What parents want to know about foreign language immersion programs, Center for Applied Linguistics (CAL),

Frost, R., L. Katz & S. Bentin 1987, “Strategies for visual word recognition and orthographical depth: a multilingual comparison”, Journal of experimental psychology: human perception and performance 13(1), p. 104-115.

Gajo, L. 2005, “Le français langue seconde d’enseignement. Choix de modèles, de langues et de disciplines”, Le français dans le monde: recherches et applications (spécial issue) “Français langue d’enseignement, vers une didactique comparative”), p. 47-57.

Gombert, J.-E. 1990, Le développement métalinguistique, Paris: PUF.

Goswami, U., & F. Mead 1992, “Onset and rime awareness and analogies in reading”, Reading research quarterly 27(2), p. 153-162.

Hamers J. & M. Blanc 1983, Bilingualité et bilinguisme, Bruxelles: Mardaga.

Hatcher, P., C. Hulme & M. Snowling 2004, “Explicit phoneme training combined with phonic reading instruction helps young children at risk of reading failure”, Journal of child psychology and psychiatry 45(2), p. 338-358.

Jaffré, J.-P., & M. Fayol 1997, Orthographes, des systèmes aux usages: un exposé pour comprendre, un essai pour réfléchir, Evreux: Flammarion.

Lambert, W., & G. Tucker 1972, Bilingual education of children: the St. Lambert experiment, Rowley, Massachusetts: Newbury House.

Lecocq, K., P. Mousty, R. Kolinsky, V. Goetry, J. Morais & J. Alegria 2004, Evaluation des programmes d’immersion en communauté française: une étude longitudinale comparative du développement des compétences linguistiques d’enfants francophones immergés en néerlandais, ULB: Ministère de la Communauté française,

Lowe, A.-S. 1998, “L’intégration de la musique et du français au programme d’immersion française: avantages pour l’apprentissage de ces deux matières”, Revue des sciences de l’éducation 24(3), p. 621-646.

Lundberg, I., J. Frost & O. Petersen 1988, “Effects of an extensive program for stimulating phonological awareness in preschool children”, Reading research quarterly 23(3), p. 263-284.

Magne, C., D. Schön & M. Besson 2006, “Musician children detect pitch violations in both music and language better than non-musician children”, Journal of cognitive neuroscience 18, p. 199-211.

Perfetti, C. 1985, Reading ability, Oxford: OUP.

Renard, R. (ed.) 2002, Apprentissage d’une langue étrangère seconde. La phonétique verbo-tonale, Bruxelles: De Boeck Université.

Sprenger-Charolles, L. 2004, “Linguistic processes in reading and spelling. The case of alphabetic writing systems : English, French, German and Spanish”, in T. Nunes & P. Bryant (eds.), Handbook of children’s literacy, Dordrecht: Kluwer Academic Publishers, p. 43-66.

Sprenger-Charolles, L., L. Siegel, D. Béchennec & W. Serniclaes 2003, “Development of phonological and orthographic processing in reading aloud, in silent reading and in spelling: a four year longitudinal study”, Journal of experimental child psychology 84, p. 194-217.

Stuart, M., & M. Coltheart 1988, “Does reading develop in a sequence of stages?”, Cognition 30(2), p. 139-181.

Zedda, P. 2006, “La langue chantée: un outil efficace pour l’apprentissage et la correction phonétique”, Les cahiers de l’Acedle 2, p. 257-282.

Ziegler, J. & U. Goswami 2005, “Reading acquisition, developmental dyslexia, and skilled reading across languages: a psycholinguistic grain size theory”, Psychological bulletin 131(1), p. 3-29.

Ziegler, J. & U. Goswami 2006, “Becoming literate in different languages: similar problems, different solutions”, Developmental science 9, p. 429-436.


1 Access to this platform was granted by Benoît Jeunier, Research Director at IUFM (teacher training college), Midi-Pyrénées, France.

Índice de ilustraciones

Título Fig. 1. Children’s profiles
Archivo image/png, 49k
Título Fig. 2. Number of phoneme errors in French and in English (all children)
Archivo image/jpeg, 48k
Título Fig. 3. Pronunciation score for all the children in French and in English
Archivo image/jpeg, 52k
Título Fig. 4. Phoneme errors in French and in English (all children)
Archivo image/jpeg, 48k
Título Fig. 5. French vowel errors (all children)
Archivo image/png, 40k
Título Fig. 6. Mistakes for French vowels by all the children, January 2009
Archivo image/png, 38k
Título Fig. 7. Mistakes for French vowels by all the children, January 2010
Archivo image/png, 32k

© Presses universitaires de Perpignan, 2017

Condiciones de uso: