Abstract/Summary

Understanding second language (L2) fluency has been an area of central importance for L2 research. While it is generally assumed that fluency improves with increased L2 proficiency, little attention has been paid to the effects of non-L2 specific features of oral performance, such as personal speaking styles in first language (L1). This thesis draws on the findings of two separate but interrelated studies. Study 1 aims to explore the relationship between L1 and L2 fluency behaviours amongst lower proficiency learners (A2, B1 and B2 at CEFR) in an English as a Foreign Language (EFL) context. It also addresses the question of whether the relationships, if any, are mediated by variations in L2 proficiency and task structure. The data were collected from 42 Turkish learners of English who were undergraduate students at a university in Turkey. Language proficiency was measured through two standardized tests of Oxford Placement Test (OPT) and Elicited Imitation Task (EIT), and speech samples were elicited through two structurally different narratives, i.e. tight and loose structure. The data were coded for 17 fluency measures; i.e. eight breakdown, five repair, one speed and three composite measures. The results from the correlational analysis showed that some breakdown and repair measures were positively correlated in L1 and L2, but no correlations were found for speed and composite measures. The relationships were not mediated by variations in L2 proficiency and the task structure. The regression analysis demonstrated a number of models predicting L2 fluency. While L1 fluency contributed significantly to models predicting breakdown and repair fluency, the OPT scores predicted mid-clause pausing and reformulation and the EIT scores predicted speed-related composite measures of speech rate, phonation-time ratio and mean length of run. Study 2 aims to further examine L1-L2 fluency relationships in a study-abroad context amongst learners of higher proficiency levels (B1, B2 and C1 according to CEFR). It also explores the mediating roles of three individual learner variables on these links; i.e. L2 proficiency (measured through the OPT and the EIT), working memory capacity (WMC) (measured through Backward digit span test and serial recognition task) and length of residence (LoR) (represented through the amount of time spent studying abroad). 60 Turkish learners of English who were post-graduate students in the UK participated in Study 2. They performed two oral narrative tasks and their speech were coded and analysed for five fluency measures (two breakdown, one repair, one speed and one composite). The correlational analysis revealed positive correlations between L1 and L2 fluency measures. They also showed that the correlations were overall maintained when the effects of L2 proficiency, WMC and LoR were controlled for. Study 2 was also interested in examining the link between L1 and L2 lexical complexity behaviours and between fluency and lexical complexity aspects in each language separately. Lexical complexity was analysed in terms of lexical diversity (LD) using two measures of type/token ratio (TTR) values and D scores. The findings revealed that LD in L1 and L2 were related to a small extent when measured through TTR values. Weak correlations were observed between fluency and LD in each language, and only between speed fluency and LD measures. When taken together, the results imply that L1 personal speaking styles are carried over to L2 fluency behaviour at least to some extent, and to L2 lexical complexity to a limited extent. The findings have a number of significant implications for L2 research, testing and teaching practices.