Abstract
Vocabulary learning is fundamental in language learning, and form-meaning links of vocabulary knowledge are pivotal for successful comprehension (Devine, 1988, p. 49; Laufer, 2003). Second language (L2) reading has long been acknowledged as the primary source of this learning (Boers, 2022; Huckin & Coady, 1999; Krashen, 1989; Pigada & Schmitt, 2006; Waring & Takaki, 2003; Webb & Chang, 2015). Constantly evolving e-learning and artificial intelligence (AI) tools present new prospects and challenges for effective L2 vocabulary learning through digital reading. Particularly, personalized text recommendation (PTR) systems, which adapt reading texts to each learner’s learning interest, need, and L2 ability, have been found to significantly enhance L2 vocabulary learning (Hsieh et al., 2012; Hsu et al., 2013; Wang, 2016). However, these systems overlooked learners’ differences in knowledge of in-text words. As a result, learners may not be able to notice or obtain the correct meaning of unknown words within the recommended reading texts, rendering L2 vocabulary learning less effective. To address this issue, personalized gloss (i.e., glosses that select each learner’s unknown words as target words to be glossed with explanatory information of word meanings) holds promise for enhancing the effectiveness of L2 vocabulary learning through digital reading. Nevertheless, personalized gloss is yet to be developed and investigated.To fill this gap, the present system development project started from scratch by justifying the facilitative potential of personalized gloss with theoretical considerations and synthesized empirical evidence. After reviewing how existing tools address the influential factors of incidental vocabulary learning through reading, a meta-analysis was conducted to explore the overall effect of digital reading on incidental L2 vocabulary learning as well as the moderating effects of L2 proficiency, test formats, and tools that aid incidental L2 vocabulary learning through reading. Results showed that digital reading had a large overall effect on all aspects of incidental L2 vocabulary learning, particularly for intermediate to advanced learners using computerized glosses or PTR systems. However, existing glosses and PTR systems suffered from two problems: the identification of unknown words and the evaluation of text readability. To solve these problems, it is necessary to calibrate the likelihood of words being known for specific learner groups. In addition to the meta-analysis highlighting the efficacy of L2 reading for intermediate to advanced learners’ incidental vocabulary learning, this PhD research project administered a large-scale vocabulary test to over 500 Chinese EFL learners at top-tier universities. Results indicated that the lemma is the optimal lexical unit and Nation’s (2017) headword list is the optimal approach to word frequency for estimating the likelihood of words being known among this learner cohort. To expand the findings to a wider learner group while enhancing their validity, two innovative calibration methods were devised to integrate learner-based word knowledge from past, current, and future studies with word frequency to calibrate the likelihood of words being known. Based on these foundation studies, a system architecture of personalized gloss was proposed, and a case study was conducted to support its effectiveness in accurately identifying individual learners’ unknown words and supporting their vocabulary learning.
| Date of Award | 16 Apr 2025 |
|---|---|
| Original language | English |
| Awarding Institution |
|
| Supervisor | Derek Irwin (Supervisor) & Yanhui Zhang (Supervisor) |