Abstract
The provision of corrective feedback (CF) on errors in essay writing is a controversial topic. In second language acquisition (SLA), the controversy focuses whether CF contributes to the development of language systems, and if so, what type of feedback is most effective. However, individual second language learners differ in their first languages, goals for second language learning, and experience with the second language, as well as many other personal characteristics, so it should be no surprise that a single type of CF on errors does not suit all learners. Specific second language learning situations also differ; both teachers and students may be affected by a variety of changeable factors, from their moods and well-being to external conditions and unexpected distractions, so it should also be no surprise that a single type of feedback on errors does not suit the same learner on all occasions. Sociocultural theory (SCT) provides a framework in which this idea can be operationalized. Through Vygotsky’s concept of zone of proximal development (ZPD), one can draw on appropriate mediation for individuals on specific occasions. Furthermore, Vygotsky also thought that working in a learner’s ZPD would help the learner move towards self-regulation and independent functioning. However, he also held that such development was ‘revolutionary’: a nonlinear, nonmonotonic process characterized by forward and backward movement and varying quality.
This dissertation contributed to existing scholarship by investigating: (1) Does the type of feedback affect learner’s performance and uptake of correct form? (2) Is the provision of feedback based on individual learner’s ZPD more effective than random CF in which the techniques are not scaled? (3) How does ZPD graduated mediation lead to learners’ development? These questions were investigated by comparing ZPD graduated mediation and random CF on two specific types of errors: English articles and prepositions. Nine ESL learners were placed into three language groups. The individuals in each group were matched on English language proficiency level and L1. Each group had a ZPD graduated pair and one random CF treatment. Nine (lower and higher) intermediate undergraduate ESL learners were asked to independently write essays based on the given prompts. Data were collected during six one-on-one tutorial sessions between participants and the researcher. The learners created four written essays, and feedback sessions were audio-recorded. The goal of the tutorial sessions was to address the targeted errors that were identified in essay writings. Following enrichment, the participants’ development was explored through two delayed post-tests.
Findings confirmed that the type of feedback plays a role in learners’ performance; mediated corrections were more effective than random CF. Learners development in the ZPD is a multidimensional process that involves independent accurate performance, the quality and quantity of mediations required to resolve an error, and learner reciprocity in the interactions through which mediation occurred. The study also contributed to the field by developing a two-dimensional reciprocity scale for identifying learner levels of responsiveness and types of responses.