O desenvolvimento de repertórios verbais é reconhecido como sendo fundamental para a inserção do indivíduo na cultura. Pesquisas têm indicado que existem características de interação adulto-criança, que, quando presentes desde muito cedo, favorecem esse desenvolvimento. A escola é um contexto privilegiado para que a criança desenvolva repertórios verbais. No Brasil, a educação infantil objetiva promover o desenvolvimento integral (inclusive da linguagem) de crianças de 0 a 5 anos. Por isso, um dos objetivos deste estudo foi averiguar a presença de condições estruturais e de interação (adulto-criança), que a literatura destaca como sendo importantes para a aprendizagem da linguagem por crianças, em escolas municipais de educação infantil. Além disso, objetivou-se comparar essas mesmas condições em turmas de educação infantil com faixas etárias diferentes (de aproximadamente 2 a 3 anos Maternal I; e 4 anos de idade Etapa I). Para tanto, foi realizada uma pesquisa de campo exploratória, com método observacional. Participaram do estudo duas turmas, uma com crianças que tinham aproximadamente 3 anos (Maternal I M1) e a outra 4 anos (Etapa I E1) quando a pesquisa foi iniciada (n total = 35 crianças e duas professoras). O procedimento envolveu cinco etapas: i) ingresso nas salas de aula e avaliação das crianças; ii) observações e filmagens na sala de aula; iii) análise do ambiente escolar, entrevista com as professoras e preenchimento de escalas padronizadas; iv) devolutivas para as escolas; v) transcrição e análise categorizada das filmagens. A análise de dados foi feita pelo registro das frequências (absoluta e relativa) de cada categoria, para cada turma. Também foi calculado o coeficiente de correlação de Pearson para verificar se os desempenhos dos alunos no teste de linguagem tinham correlação com o quanto eles falavam nas aulas. Os resultados indicam que as crianças mais bem avaliadas quanto à linguagem também foram aquelas que falaram mais durante as aulas. Em ambas as turmas, os alunos emitiram mais respostas verbais estritamente relacionadas àquilo que as professoras haviam perguntado. Mas, os alunos de E1 iniciaram mais interações com a professora, emitiram mais perguntas e responderam fazendo mais expansões àquilo que a professora havia perguntado. As professoras, por sua vez, arranjaram contingências muito semelhantes para as duas turmas: emitiram mais respostas verbais que não propiciam o desenvolvimento de linguagem das crianças (e.g., perguntas fechadas e instruções relacionadas à atividade); apresentaram a mesma variabilidade de palavras (types e tokens); e organizaram atividades que demandavam desempenhos com o mesmo nível de complexidade para as duas faixas etárias. As condições da escola foram bem avaliadas apenas no que diz respeito à qualidade das interações e aos cuidados pessoais dispensados às crianças, em detrimento de condições estruturais que efetivamente favorecessem a realização de atividades acadêmicas de qualidade. Esses resultados indicam que a escola, como uma instituição na qual devem existir agentes culturais, que ensinem repertórios verbais ao indivíduo, tem cumprido pouco seu papel / The development of verbal repertoires is recognized as being fundamental to insert the individual into the culture. Researches have indicated that there are characteristics of adult-child interaction, which, when present from a very early age, favor this development. The school is a privileged context for the child to develop verbal repertoires. In Brazil, early childhood education aims to promote the integral development (including language) of children from 0 to 5 years old. Therefore, one of the objectives of this study was to verify the presence of structural and interaction conditions (adult-child), which the literature highlights as being important for children\'s language learning in municipal schools for early childhood education. In addition, it aimed to compare the same conditions in preschool classes of children with different ages (approximately 2 years old - Maternal I and 4 years old - Etapa I). For that, an exploratory field research was carried out, with an observational method. Two groups participated in the study, one with children who were approximately 3 years old (Maternal I - M1) and the other 4 years (Etapa I - E1) when the research was started. In total, there were 35 children and their respective teachers. The procedure involved five steps: i) entrance into the classrooms and evaluation of the children; ii) observations and filming in the classroom; iii) analysis of the school environment, interview with teachers and filling of standardized scales; iv) devolutive for schools; v) transcription and categorized analysis of filming. Data analysis was done by recording the frequencies (absolute and relative) of each category, for each class. Pearson\'s correlation coefficient was calculated to verify if students\' performances in the language test correlated with how much they spoke in class. The results indicate that the children who were better evaluated for language were also the ones who spoke the most during the lessons. In both groups, students spoke more verbal responses strictly related to what teachers had asked. But the students at E1 started more interactions with the teacher, asked more questions and answered by making more expansions to what the teacher had asked. The teachers, on the other hand, arranged very similar contingencies for the two classes: they issued more verbal responses that did not support children\'s language development (e.g., closed-ended questions and instructions related to the activity); they spoke the same word variability; and organized activities that demanded performances with the same level of complexity for the two age groups. School conditions were well evaluated only in terms of the quality of interactions and personal care provided to children, but not about structural conditions that effectively favored quality academic activities. These results indicate that the school, as an institution in which there must be cultural agents who teach verbal repertoires to the children, has been accomplishing little its role

Dans ce travail de thèse, j’ai entrepris d’éclairer le mode de structuration du répertoire langagier, concept sociolinguistique introduit par l’ethnographe de la communication John Gumperz dans les années soixante, afin de pouvoir proposer des applications pédagogiques en vue de la formation des enseignants de langues à cette notion. Le concept de répertoire langagier constitue, en effet, un concept phare de la didactique du plurilinguisme, didactique qui a donné lieu à de nombreux projets éducatifs en Europe depuis une vingtaine d’année, et qui vise au développement d’une compétence plurilingue et pluriculturelle. Aussi, après avoir essayé de cerner au mieux le mode de structuration du répertoire langagier, qui implique que l’on considère un niveau intermédiaire de genres de discours permettant de passer des formes linguistiques aux formes de la communication, j’ai proposé une recherche-action dont l’objectif était l’élaboration d’un modèle de répertoire langagier. Cette recherche exploratoire a été menée grâce au concours de douze étudiants de licence 3 de science du langage de l’université de Rouen, déjà sensibilisés à la notion de répertoire langagier, et dont la majorité se destinait à l’enseignement du français langue étrangère. Le dispositif comprenait deux parties. La première partie, à visée réflexive, consistait à demander aux étudiants de représenter sous la forme d’une représentation graphique l’idée qu’ils se faisaient de leur répertoire langagier. La seconde partie plus analytique visait à sensibiliser les étudiants à la notion de genre discursif, grâce à l’étude d’extraits issus de l’article de John Gumperz (1964) dans lequel il explicite le mode de structuration du répertoire langagier. Chacune des parties comportait une succession d’opérations intellectuelles susceptibles de faire surgir les représentations des étudiants et de les interroger. En effet, réfléchir au mode de structuration du répertoire langagier implique nécessairement de conscientiser les comportements langagiers que l’on adopte au cours des différentes activités langagières auxquelles on participe dans la vie de tous les jours. Ainsi, les étudiants étaient amenés à prendre du recul sur la notion de langue, « invention des linguistes » (Calvet, 2004), pour s’interroger davantage sur les pratiques langagières qui, en vérité, constituent les langues. Ce changement de paradigme n’a pas été facile à opérer car les représentations se construisent « à travers des formules primitives stéréotypées » (Py, 2000 :6), qui circulent dans les milieux sociaux, et non sur la réalité des pratiques langagières et des comportements. Par ailleurs, les noms des langues qui constituent très souvent des archilexèmes qui recouvrent et masquent la pluralité des pratiques langagières, et par là-même l’expression de la pluralité des individus d’une part, et d’autre part les outils linguistiques véhiculés notamment par l’école : dictionnaires, manuels de langues, grammaires, qui adoptent rarement une approche sociolinguistique des langues, accréditent la représentation d’entités homogènes et bien délimitées. Ainsi, les discours des étudiants, qu’ils portent sur leurs dessins réflexifs ou sur le cours, montrent qu’ils sont réceptifs à la diversité langagière qui les entoure, qu’ils ont conscience d’accommoder leur discours dans certaines circonstances, de faire des choix langagiers en fonction du degré de formalité d’une situation, et qu’ils distinguent différentes variétés linguistiques qui relèvent notamment de la variation diatopique et de la variation diastratique. / This research aims at highlighting the structuration of the verbal repertoire, which is a sociolinguistic concept introduced by ethnographer of communication John Gumperz in the sixties, with the view to raising language teachers’awareness on this notion. Indeed, verbal repertoire has become a core concept in didactic of plurilingualism which has been developed in Europe over the past two decades, and whose goal is the development of a plurilingual and pluricultural competence. Thus, in the first part of this work, I set out to better understand the structuration of verbal repertoire, assuming that an intermediate level consisting of registers lies between linguistic forms and communicative forms. I then proposed an action-research project with the aim of shaping a verbal repertoire model. This exploratory research involved 12 students carrying out graduate studies in sociolinguistics at the University of Rouen, already acquainted with verbal repertoire notion, and whose goal is to teach French as a foreign language. This educational setting comprised two parts. The first part required a reflective approach by the students who had to propose a graphic representation of their own verbal repertoire; the second part was more analytical, and should raise their awareness about the concept of registers, by analysing extracts of John Gumperz’s article (1964) in which he clarifies the structuration of the verbal repertoire. Each part of the action-research consisted of operating intellectual sequences aiming at bringing out students’representations and questioning them. Indeed, examining ways of structuring the verbal repertoire necessarily means developing a self-awareness of language behaviour about everyday communication. Thus, students were encouraged to step back from an abstract conception of language to adopt a more concrete point of view exploring language practices. This paradigm shift had been very challenching for the students, as social representations are built on stereotypical primitive ideas (Py, 2000:6) circulating in the social circles, not on the observation of language practices. Moreover, both the use of generic names regarding languages on the one hand, which covers and conceals the plurality of language practices and at the same time the plural identity of individuals; and on the other hand, the linguistic approach taken by most of the educative tools, such as dictionaries, method books, grammars, confirm the representation of well bounded and clearly defined languages. Thus, speeches provided by student regarding their graphic representation of their verbal repertoires and the course content, show that they are receptive to the linguistic diversity around them, they are aware of accommodating to others while interacting in certain circumstances and making language choices to adjust to the formality of the situation, and make distinctions between different linguistic varieties especially diatopic and diastratic variations