This thesis is based upon a longitudinal study of L1 attrition in two bilingual teenage siblings, J and her older brother S, with Korean as their L1 and English as their L2. The two teenagers' initial exposure to English occurred at the age of 9. When the two siblings began to have sustained exposure to and immersion in an English L2 environment, they were at the postpubertal age. They had been attending high school in a Mid-Western city in the U.S for about two years when the study was conducted. The longitudinal study addressed three main questions. The first question concerned the siblings' language (L1 and L2) development, language dominance and cross-lingusitic influence from a bidrectional perspective. The second question examined the evidence for L1 attrition in relation to lexis, morphology and syntax, and their relative levels of vulnerability. The third question examined the role of extra-linguistic factors in the process of L1 attrition. In order to address the above questions, the two siblings were observed in their home over a period of 8 months. The data for the longitudinal study included the siblings' (i) spontaneous speech interactions in the home (ii) oral narratives in the L1 and L2 and (iii) their responses to a language background questionnaire. The results showed that the two siblings' English L2 developed and improved over time, but that they continued to maintain their L1 (Korean) as their dominant language. The evidence from both siblings' L1 and L2 use supported bidirectional cross-linguistic influence (i.e. from the L1 onto the L2 and from the L2 onto the L1). L1 attrition occurred only minimally in relation to morphology (e.g. honorification, case particles, classifiers, and plural marking) and lexical choice, while the siblings' L1 syntax remained relatively stable over time. Overall, the siblings were largely successful in maintaining their L1. In part, this was because their L1 was already firmly established or entrenched, as their immersion in an L2 environment occurred only when they were older (around puberty). Furthermore, their frequent use of their L1 for social networking, along with their positive attitudes toward their heritage language, also played a crucial role in maintaining and stabilizing their L1. Finally, the theoretical and practical implications of the findings of the present study, as well as recommendations for the future research are discussed.

In the field of second language acquisition (SLA), the study abroad context (SA) has gained attention as a site that offers the potential of significant second language (L2) development due to high amounts of input and interaction opportunities compared to at home foreign language (AH) and domestic immersion (IM) contexts (Pérez-Vidal, 2014). In previous research, the SA context has been a country where the L2 is the local language (e.g., English in the United Kingdom). However, with the increase of student mobility programs across Europe, such as ERASMUS, and the status of English as an International Language, another study abroad context is available, one where students can take English-medium classes and use English as a lingua franca in a country where English is not the local language (e.g., Germany, Poland, Spain). In the current study, this new context is operationalized as English as a lingua franca study abroad (ELFSA), the effects of which have received very little attention in SLA to date. By providing an alternative SA context through English medium of instruction on-campus, and English as an international language through off-campus interactions, this new context might bring further insights into the SA phenomenon. Motivated by this gap in the literature, the current longitudinal study aims to investigate the differentiated effects of the SA, ELFSA, and AH contexts on the linguistic development of Turkish undergraduates whose L2 is English. Given the multilingual nature of the ERASMUS context, this study also examines the contextual influences on participants’ perceptions towards multilingualism from a Perceived Positive Language Interaction (PPLI) perspective (Thompson, 2013). The participants of the study were 50 third year Turkish undergraduates, 33 of whom undertook a 16-week ERASMUS exchange semester in Spring 2016. Following a quasi-experimental mixed-methods pretest-posttest design, data were collected via a one-minute spoken and 15-minute written production test to determine linguistic complexity, accuracy, and fluency gains; an Elicited Oral Imitation Test (EIT, Ortega et al., 1999) to measure pre-departure proficiency; a monthly online Language Interaction Questionnaire to investigate the type and amount of language contact; and a dichotomous questionnaire to learn about participants’ perceptions towards multilingualism within PPLI. For triangulation purposes, qualitative data were collected via several open-ended items in the questionnaires and semi-structured interviews. The results of the quantitative and qualitative analysis indicated that the SA and ELFSA contexts were beneficial for English development on most of the measures, yet the former had certain issues for the participants pertaining to the dominant variety and features of spoken English (e.g., weak forms, connected speech, speech rate). The ELFSA was reported to be more multilingual as compared to the SA, also paving the way for seeing further Perceived Positive Language Interaction (Thompson, 2013) among the additional languages that their participants knew. The major difference between the two was described to be the ELFSA participants’ developing an ELF identity towards using English as an L2, as they tended to prioritize fluency over accuracy throughout their exchange semester. The results of the quantitative analysis indicated that the AH group had significantly more gains on written fluency than the two abroad groups. Also, time was found to be a significant factor for lexical development regardless of group differences. In terms of oral gains, main effects of time indicated that the participants as a whole group had significantly different means over time on speech rate and breakdown fluency, the inspection of which showed that the ELFSA had the highest mean differences. As for oral accuracy and lexical complexity, an interaction effect between time and context group was found to approach significance, indicating that both sojourn groups had mean gains on the former, while only the ELFSA had higher mean scores in the posttest for the latter measure. Finally, the data from the semi-structured interviews provided a holistic picture of the interplay between context and development in English. The major finding of the study, thus, indicated that the SA may not be the sole provider for intense L2 input to improve English during a semester abroad; the ELFSA appeared to be equally as beneficial as the SA in terms of linguistic development with a bonus of creating a linguistic identity of an ELF speaker for language learners. The study also provided several empirical and pedagogical implications for those interested in the stay abroad and its influences on L2 development.

Although foreign language instruction in mixed-age (M-A) is gaining popularity (Heizmann and Ries and Wicki 2015; Lau and Juby-Smith and Desbiens, 2017; Shahid Kazi and Moghal and Aziz 2018; Thurn 2011), the research is scarce. Drawing from multiple data sources, this study investigated to what extent do peer interactions among M-A and same-age (S-A) pairs aid L2 development and how students perceive their interactions. In this study, the same learners (N=24) aged between 10 and 12 interacted with the same and different age partners during common classroom lessons in two EFL classrooms. The results suggest that both S-A and M-A peer interactions aided L2 development. Although S-A pairs outperformed M-A pairs on the post-test, the results are not statistically significant. The analysis of students’ perceptions revealed that the majority of students prefer working in S-A to M-A pairs. In addition to age/proficiency differences, factors such as students’ relationships and perceptions of one’s own and partner’s proficiency greatly impact how they interact with one another.

Cette étude quasi-expérimentale a pour but de 1) comparer la prise en compte et les effets de trois conditions rétroactives, à savoir la reformulation, l’incitation et un mélange des deux techniques, 2) déterminer le lien entre la prise en compte et l’apprentissage, et 3) identifier l’effet des perceptions des apprenants quant à la rétroaction corrective sur la prise en compte et l’apprentissage. Quatre groupes d’apprenants d’anglais langue seconde ainsi que leurs enseignants provenant d’un CEGEP francophone de l’île de Montréal ont participé à cette étude. Chaque enseignant a été assigné à une condition rétroactive expérimentale qui correspondait le plus à ses pratiques rétroactives habituelles. La chercheure a assuré l’intervention auprès du groupe contrôle. L’utilisation du passé et de la phrase interrogative était ciblée durant l’intervention expérimentale. Des protocoles de pensée à haute voie ainsi qu’un questionnaire ont été utilisés pour mesurer la prise en compte de la rétroaction corrective. Des tâches de description d’images et d’identification des différences entre les images ont été administrées avant l’intervention (pré-test), immédiatement après l’intervention (post-test immédiat) et 8 semaines plus tard (post-test différé) afin d’évaluer les effets des différentes conditions rétroactives sur l’apprentissage des formes cibles. Un questionnaire a été administré pour identifier les perceptions des apprenants quant à la rétroaction corrective. En termes de prise en compte, les résultats indiquent que les participants sont en mesure de remarquer la rétroaction dépendamment de la forme cible (les erreurs dans l’utilisation du passé sont détectées plus que les erreurs d’utilisation de la phrase interrogative) et de la technique rétroactive utilisée (l’incitation et le mélange d’incitation et de reformulations sont plus détectés plus que la reformulation). En ce qui a trait à l’apprentissage, l’utilisation du passé en général est marquée par plus de développement que celle de la phrase interrogative, mais il n'y avait aucune différence entre les groupes. Le lien direct entre la prise en compte et l’apprentissage ne pouvait pas être explicitement établi. Pendant que la statistique inférentielle a suggéré une relation minimale entre la prise en compte du passé et son apprentissage, mais aucune relation entre la prise en compte de la phrase interrogative et son apprentissage, les analyses qualitatives ont montrés à une association entre la prise en compte et l’apprentissage (sur les deux cibles) pour certains étudiants et augmentations sans prise en compte pour d'autres. Finalement, l’analyse factorielle du questionnaire indique la présence de quatre facteurs principaux, à savoir l’importance de la rétroaction corrective, la reformulation, l’incitation et les effets affectifs de la rétroaction. Deux de ces facteurs ont un effet modérateur sur la prise en compte de la rétroaction sans, toutefois, avoir d’impact sur l’apprentissage. / This quasi-experimental study sought to investigate the often assumed yet little investigated relationship between noticing of corrective feedback (CF) and L2 development in relation to learner beliefs about error correction. Specifically, it aimed to (1) uncover the noticeability and effectiveness of three CF techniques (namely, recasts, prompts, a combination of the two) (2) to determine a relationship between noticing of CF and learning of the past tense and questions in the past, and (3) to determine whether learner beliefs about CF mediate what is noticed and learned in the language classroom. The participants were four groups of high-beginner college level francophone ESL learners (n = 99) and their teachers. Each teacher was assigned to a treatment condition that fit his CF style, but the researcher taught the controls. CF was provided to learners in response to their production problems with the simple past and questions in the past. While noticing of CF was assessed through immediate recall and questionnaire responses, learning outcomes were measured by way of picture description and spot the differences tasks administered through a pre-test, post-test, and delayed post-test design. Learner beliefs about CF were probed by means of a 40-item questionnaire. To elicit the learner and teacher perspectives on the study, semi-structured interviews were held with the three teachers and 20 learners, drawn randomly from the participating classes. The results indicated that the noticeability of CF is dependent on the grammatical target it addresses (i.e., feedback on past tense errors was noticed more) and that the feedback techniques that push learners to self-correct alone or in combination with target exemplars are more effective in bringing out the corrective intent of a feedback move. In relation to the learning outcomes, the overall past tense accuracy increased more than that for questions, but there were no differences between the groups. The direct link between noticing and learning could not be unequivocally established. While the inferential statistics suggested a minimal relationship between noticing and past tense scores, especially if the CF was provided with recasts, but no relationship between noticing and questions scores, the qualitative analyses pointed to an association between noticing and test scores (on both targets) for some learners and gains without noticing for others. Finally, in relation to the beliefs about CF, the participants’ responses centered on four common themes (the importance of oral CF, recasts as CF technique, prompts as CF technique, and affective consequences of CF), two of which mediated the noticeability of the supplied CF, but none impacted the learning outcomes.

Cette étude quasi-expérimentale a pour but de 1) comparer la prise en compte et les effets de trois conditions rétroactives, à savoir la reformulation, l’incitation et un mélange des deux techniques, 2) déterminer le lien entre la prise en compte et l’apprentissage, et 3) identifier l’effet des perceptions des apprenants quant à la rétroaction corrective sur la prise en compte et l’apprentissage. Quatre groupes d’apprenants d’anglais langue seconde ainsi que leurs enseignants provenant d’un CEGEP francophone de l’île de Montréal ont participé à cette étude. Chaque enseignant a été assigné à une condition rétroactive expérimentale qui correspondait le plus à ses pratiques rétroactives habituelles. La chercheure a assuré l’intervention auprès du groupe contrôle. L’utilisation du passé et de la phrase interrogative était ciblée durant l’intervention expérimentale. Des protocoles de pensée à haute voie ainsi qu’un questionnaire ont été utilisés pour mesurer la prise en compte de la rétroaction corrective. Des tâches de description d’images et d’identification des différences entre les images ont été administrées avant l’intervention (pré-test), immédiatement après l’intervention (post-test immédiat) et 8 semaines plus tard (post-test différé) afin d’évaluer les effets des différentes conditions rétroactives sur l’apprentissage des formes cibles. Un questionnaire a été administré pour identifier les perceptions des apprenants quant à la rétroaction corrective. En termes de prise en compte, les résultats indiquent que les participants sont en mesure de remarquer la rétroaction dépendamment de la forme cible (les erreurs dans l’utilisation du passé sont détectées plus que les erreurs d’utilisation de la phrase interrogative) et de la technique rétroactive utilisée (l’incitation et le mélange d’incitation et de reformulations sont plus détectés plus que la reformulation). En ce qui a trait à l’apprentissage, l’utilisation du passé en général est marquée par plus de développement que celle de la phrase interrogative, mais il n'y avait aucune différence entre les groupes. Le lien direct entre la prise en compte et l’apprentissage ne pouvait pas être explicitement établi. Pendant que la statistique inférentielle a suggéré une relation minimale entre la prise en compte du passé et son apprentissage, mais aucune relation entre la prise en compte de la phrase interrogative et son apprentissage, les analyses qualitatives ont montrés à une association entre la prise en compte et l’apprentissage (sur les deux cibles) pour certains étudiants et augmentations sans prise en compte pour d'autres. Finalement, l’analyse factorielle du questionnaire indique la présence de quatre facteurs principaux, à savoir l’importance de la rétroaction corrective, la reformulation, l’incitation et les effets affectifs de la rétroaction. Deux de ces facteurs ont un effet modérateur sur la prise en compte de la rétroaction sans, toutefois, avoir d’impact sur l’apprentissage. / This quasi-experimental study sought to investigate the often assumed yet little investigated relationship between noticing of corrective feedback (CF) and L2 development in relation to learner beliefs about error correction. Specifically, it aimed to (1) uncover the noticeability and effectiveness of three CF techniques (namely, recasts, prompts, a combination of the two) (2) to determine a relationship between noticing of CF and learning of the past tense and questions in the past, and (3) to determine whether learner beliefs about CF mediate what is noticed and learned in the language classroom. The participants were four groups of high-beginner college level francophone ESL learners (n = 99) and their teachers. Each teacher was assigned to a treatment condition that fit his CF style, but the researcher taught the controls. CF was provided to learners in response to their production problems with the simple past and questions in the past. While noticing of CF was assessed through immediate recall and questionnaire responses, learning outcomes were measured by way of picture description and spot the differences tasks administered through a pre-test, post-test, and delayed post-test design. Learner beliefs about CF were probed by means of a 40-item questionnaire. To elicit the learner and teacher perspectives on the study, semi-structured interviews were held with the three teachers and 20 learners, drawn randomly from the participating classes. The results indicated that the noticeability of CF is dependent on the grammatical target it addresses (i.e., feedback on past tense errors was noticed more) and that the feedback techniques that push learners to self-correct alone or in combination with target exemplars are more effective in bringing out the corrective intent of a feedback move. In relation to the learning outcomes, the overall past tense accuracy increased more than that for questions, but there were no differences between the groups. The direct link between noticing and learning could not be unequivocally established. While the inferential statistics suggested a minimal relationship between noticing and past tense scores, especially if the CF was provided with recasts, but no relationship between noticing and questions scores, the qualitative analyses pointed to an association between noticing and test scores (on both targets) for some learners and gains without noticing for others. Finally, in relation to the beliefs about CF, the participants’ responses centered on four common themes (the importance of oral CF, recasts as CF technique, prompts as CF technique, and affective consequences of CF), two of which mediated the noticeability of the supplied CF, but none impacted the learning outcomes.