L'autisme est un handicap caractérisé par une altération qualitative de la communication et de la socialisation, une intolérance au changement et des comportements stéréotypés. Les interactions de tutelle sont au coeur du dispositif de prise en charge, leurs caractéristiques, autant linguistiques que pragmatiques, sont mal connues. Les déterminants de la compréhension entre tuteur et enfant autiste ont fait l'objet de peu de recherches naturalistes. Une méthodologie d'étude a été élaborée, reposant sur l'utilisation croisée du paradigme pragmatique et du programme informatisé CHILDES (Child Language Data Exchange System). Elle compare les interactions duelles de tutelle entre un tuteur et un enfant autiste (10), déficient intellectuel (10) ou témoin (10). Les résultats montrent qu'aucune des stratégies directives utilisées par le tuteur n'est plus efficace qu'une autre pour les enfants autistes modérément déficients. En revanche, les stratégies à forte directivité sont plus efficaces lorsque la déficience est sévère. La capacité de compréhension apparaît supérieure chez les enfants témoins mais elle est comparable entre les enfants autistes et déficients. Par ailleurs, le taux d'incompréhension est supérieur pour le groupe autiste, rendant ce phénomène déterminant. Enfin, si les enfants autistes comprennent moins bien le tuteur que les témoins à l'échelle des actes de communication, cette différence n'apparaît pas à celle de la thématique, manifestant l'adaptation des exigences du tuteur aux difficultés de l'enfant. D'autres résultats plus traditionnels confortent la validité de notre méthode d'analyse, les perspectives offertes sont discutées. / Autism consists in qualitative impairment in social interaction and communication, with restricted repetitive and stereotyped patterns of behavior. Tutorial interactions are widely used, but their characteristics, linguistic as well as pragmatic, remain partly unknown. Moreover, only few naturalistic works are dedicated to the understanding of determinants between tutor and autistic child. A methodology has been elaborated which involves the use of pragmatic paradigm crossed with the computerized program CHILDES (Child Language Data Exchange System). This method compares tutorial interactions in dual situation between a tutor and a child with autism (10), with intellectual disability (10), and control (10). Results show that none of the strategies used by the tutor is more efficient than the others for children with autism and moderate intellectual disability. However, the strategies with strongest directivity are more efficient with children with severe intellectual disability. The understanding ability appears to be superior in control children but comparable in children with autism or intellectual disability. Besides, the misunderstanding rate is superior for autistic children, showing the importance of this phenomenon. Finally, if autistic children, at the communication act level, have a lower understanding of the tutor productions than control children, this difference does not appear at the thematic level. This result shows the adaptation of the tutor demand to child difficulties. Other results, more traditional, reinforce the validity of our analysis method, possible perspectives are discussed.

This thesis presents the findings of a study of writing center tutorial practices in a Middle Eastern university where the language of instruction is English. Data from stimulated recall activities, written observations, and interviews were analyzed to answer the following research questions: 1. How do tutees perceive the effectiveness of writing center tutorials? 2. How do tutors perceive the effectiveness of writing center tutorials? 3. Which type of tutoring approach do tutees find more effective? 4. Which type of tutoring approach do tutors find more effective? The data revealed that tutees noticed an improvement in their assignments, believed that their concerns had been addressed, and that they had acquired transferable skills. Most tutees assessed their tutors positively, valuing tutors who inspired confidence and were able to explain concepts clearly. Although tutees appreciated knowledgeable tutors, they valued egalitarian peer-tutoring relationships. Tutors reported that tutorial sessions improved their tutees' assignments and that tutees had acquired transferable skills. Nevertheless, tutors were critical of their own performance. Some tutors admitted to lacking the knowledge necessary to explain certain writing concepts, including grammatical concepts; some felt they dominated the tutorials; and others felt their approach was too directive. The data revealed that both tutors and tutees preferred the directive approach for lower order concerns and a non-directive approach for higher order concerns. This study shows that diverse tutoring models that accommodate the background and experiences of Middle Eastern students, and their particular strengths and weaknesses, should be considered. It recommends tutorial training that emphasizes flexibility and recognizes the distinctive nature of each tutorial situation and the opportunity it presents to address the needs and expectations of individual students. These findings could signal a direction for the development of writing center pedagogy that focuses on the linguistically and culturally diverse students in the Middle East.