Writing in a non-alphabetic language using a keyboard: Behaviours, cognitive activities and text quality

Lu, Xiaojun (2020) Writing in a non-alphabetic language using a keyboard: Behaviours, cognitive activities and text quality. University College London, Doctoral Thesis.

Abstract

Cognitive processes in second language (L2) writing have received growing attention for the past three decades. Yet, few studies have examined the processes involved in non-alphabetic language writing. Adopting Rijlaarsdam and Van Den Bergh's (1996) writing model as the theoretical framework, this thesis investigates the separate and joint effects of writing stages, genre and proficiency on the cognitive processes in which L2 writers of a non-alphabetic language engage, and how writing behaviours predict L2 text quality.

Thirty-two L2 users of Chinese wrote two argumentative and two narrative essays using the Pinyin input method. Their keystrokes during writing were captured. All participants described their thoughts about pauses and revisions prompted by the recording of their last writing performance. A cloze test was used to establish participants' level of L2 proficiency. Data was also obtained from 32 first language (L1) users for a fuller picture of non-alphabetic language writing processes. Writing behaviours were analysed in terms of speed fluency, pausing and revision, and the stimulated recall comments were classified as planning, translation or monitoring. Text quality was determined via holistic rating.

Mixed-effects regressions and qualitative analyses yielded the following results. First, a small stage effect emerged for the cognitive processes in non-alphabetic language writing, with a slight difference between L1 and L2. Second, L2 writers engaged in different behaviours and cognitive activities when writing in different genres, while L1 writers only behaved differently. Third, L2 proficiency affected speed fluency and certain pausing and revision behaviours, and this effect tended to be mediated by stages. Finally, L2 text quality was predicted by speed fluency and pauses between smaller textual units, and these relationships were modulated by stages, genre and proficiency. L1 text quality was only predicted by pauses between larger textual units in certain stages; the strength of the relationships was genre-dependent.

This record has no associated files available for download.

Contributors

Author: Xiaojun Lu

Download statistics