Abstract

The aim of this dissertation is to explore the use of automatic speech recognition (ASR) for writing purposes and to assess its pedagogical effectiveness as a writing tool. Considering that ASR is useful for enhancing second language (L2) oral skills and pronunciation (Liakin et al., 2015), it is worth examining ASR’s pedagogical features and
limitations; whether ASR can be accepted as a writing tool; and how ASR-assisted writing can contribute to the development of writing skills. This dissertation begins with an overview of the ASR technology and its architecture, as well as the technology’s advancements throughout time, followed by an analysis of the components of writing skill, transcription
modes, and related challenges. The remainder comprises three interconnected manuscript-based chapters (Chapters 2 to 4, respectively), each of which addresses a distinct aspect of three overarching research questions: (a) What are the pedagogical benefits and limitations of ASR use for L2 pedagogy and for writing purposes? (Manuscript A); (b) What are learners’ perceptions of ASR use as a writing tool? (Manuscript B); and (c) What behaviors do learners employ when engaged in ASR-assisted writing? (Manuscript C).
The first manuscript systematically reviews the pedagogical potential and limitations of using the ASR technology in L2 learning and, more particularly, in L2 English writing. The second manuscript explores writers’ perceptions of employing ASR for writing to reveal whether writers accept ASR as a tool for composing texts. This manuscript also compares
native and nonnative English writers’ perceptions in terms of using ASR for writing. Finally, the third manuscript examines the behaviors that writers employ when composing texts with ASR and whether there is a relationship between the choice of behavior utilized and the lexical diversity of an ASR-written text.
This dissertation will further our understanding of users’ perceptions and acceptance of ASR as a writing tool by identifying both the potentials and limitations of the technology. Given that very little research has been conducted to investigate the use of ASR systems for L2 writing by adult writers (as far as the researcher is aware, there is only one study in this
field, and the handful of studies available focuses on writing by native speakers or on writers with learning disabilities), this dissertation will open a new line of inquiry in the field of L2 writing development.