Amanda Huensch, University of South Florida
This talk highlights the benefits and challenges of using multiple tasks to investigate the development of second language (L2) fluency across multiple languages.
Data come from 49 English L1 learners of French (n=25) or Spanish (n=24) (university students on their year abroad, http://langsnap.soton.ac.uk). Data were collected in the form of picture-based narratives and semi-structured interviews before and during a 9-month residence abroad. Comparable data were collected in the participants’ first language (L1), English, and from ten native speakers (NS) of French and Spanish.
Given the relatively robust finding that mid-clause pausing differentiates native from non-native speech, an investigation of mid-clause pausing is underway that compares L2 data to both English-L1 data and NS data. Preliminary results indicate that task (narrative vs. interview) affects the distribution of silent pauses for NSs (in all three languages), such that NSs pause less within clauses in the narratives. The same trend does not appear to be true for learners. In addition, a comparison of learners to NSs in both tasks appears to demonstrate that learners are more ‘native-like’ (regarding mid-clause pausing) in the interview task. These results provide important implications for the selection of task when investigating L2 fluency.
Methodological challenges of conducting the analyses will be discussed, such as coding fluency across multiple languages (e.g., choosing/marking clause and ASU boundaries) and tasks (e.g., removal of prompted text from narratives, segment selection from interviews).