Toward construction of a corpus of english learners' utterances annotated with speaker proficiency profiles: Data collection and sample annotation

0Citations
Citations of this article
6Readers
Mendeley users who have this article in their library.
Get full text

Abstract

The first author has designed and implemented college English classes emphasizing face-to-face oral interactions within small groups of students in class, presupposing and expecting further cultivation of learners' ability to learn for themselves, by themselves and among themselves. Previous experiences confirm such expectations, and the authors are currently working on collecting learners' spoken interactions with high-quality digital audio and video recording devices along with written materials, scores of language proficiency tests and questionnaire responses of those students. In this presentation, we describe the scope and objective of our project, summarize class activities and recording procedures, and then touch on expected transcription procedures and possible tools for annotation. It may be interesting to notice, in passing, how introduction of various recording devices positively affect students' motivations and performances in their language learning activities. © 2008 Springer-Verlag Berlin Heidelberg.

Cite

CITATION STYLE

APA

Harada, Y., Maebo, K., Kawamura, M., Suzuki, M., Suzuki, Y., Kusumoto, N., & Maeno, J. (2008). Toward construction of a corpus of english learners’ utterances annotated with speaker proficiency profiles: Data collection and sample annotation. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 4938 LNAI, pp. 171–178). https://doi.org/10.1007/978-3-540-78159-2_16

Register to see more suggestions

Mendeley helps you to discover research relevant for your work.

Already have an account?

Save time finding and organizing research with Mendeley

Sign up for free