Modeling second language learners' interlanguage and its variability: a computer-based dynamic assessment approach to distinguishing between errors and mistakes

Despite a long history, interlanguage variability research is a debatable topic as most paradigms do not distinguish between competence and performance. While interlanguage performance has been proven to be variable, determining whether interlanguage competence is exposed to random and/or systematic variations is complex, given the fact that distinction between competence-dependent errors and performance-related mistakes should be established to best represent the interlanguage competence. This thesis suggests a dynamic assessment model grounded in sociocultural theory to distinguish between errors and mistakes in texts written by learners of French, to then investigate the extent to which interlanguage competence varies across time, text types, and students. The key outcomes include: 1. An expanded model based on dynamic assessment principles to distinguish between errors and mistakes, which also provides the structure to create and observe learners’ zone of proximal development; 2. A method to increase the accuracy of the part-of-speech tagging procedure whose reliability correlates with the number of incorrect words contained in learners’ texts; 3. A sociocultural insight into interlanguage variability research. Results demonstrate that interlanguage competence is as variable as performance. The main finding shows that knowledge over time is subject to not only systematic, but also unsystematic variations.

Name dc:type.qualificationname

phd

Level dc:type.qualificationlevel

doctoral

Grantor dc:publisher.institution

Dublin City University

Year dc:date.issued

2011

Author dc:creator

• Thouësny, Sylvie

• Education
• Computational linguistics
• Educational technology
• Linguistics

Language dc:language

en