Speaker
Abstract section 3: Content/method
Using a within-subjects design, the study involved 20 target nouns and 69 B1-level participants from a university in Vietnam. The pretraining phase involved self-study of nouns and their Vietnamese translations. In the retrieval phase, three blocks of uninformative contexts were practised, with each block containing ten AI-generated sentences for half of the nouns and ten human-written for the other half. Form and meaning recall tests were administered immediately and one week later to test retention.
Abstract section 1: Relevance
Vocabulary is instrumental in language learning, yet memorising words is challenging. Retrieval, the process of actively recalling learnt knowledge, can facilitate memory (Barcroft, 2007; Karpicke & Roediger, 2008). One type of retrieval practice is using contexts to trigger memory and enhance retention (van den Broek et al., 2022; Zahar et al., 2001). Operationalised on the basis of contextual richness (Mondria & Boer, 1991), these contexts are manipulated to promote vocabulary retrieval rather than meaning comprehension. In addition, AI tools are increasingly utilised to generate content that supplements vocabulary learning (Jeon, 2023). However, investigations into context-driven retrieval and AI remain scarce.
Abstract section 4: Outcomes/results
Analysed using lme4 linear mixed-effects modelling in R (Bates et al., 2015), the results showed that human-written contexts were as effective as AI-generated ones in activating retrieval. Also, although AI displayed an advantage over humans in tailoring contextual information to the point of uninformativeness, human-written elements made the contextual information more relatable and memorable. These findings suggest that contextualised word retrieval might be best implemented with AI when familiar elements are reduced, maximising the effectiveness of uninformative contexts in retrieving vocabulary. This could inform classroom instruction, as AI can generate contextual sentences that help students actively recall learnt lexical items.
Abstract section 5: References
Barcroft, J. (2007). Effects of opportunities for word retrieval during second language vocabulary learning. Language Learning, 57(1), 35–56. https://doi.org/10.1111/j.1467-9922.2007.00398.x
Bates, D., Mächler, M., Bolker, B., & Walker, S. (2015). Fitting linear mixed-effects models using lme4. Journal of Statistical Software, 67(1). https://doi.org/10.18637/jss.v067.i01
Jeon, J. (2023). Chatbot-assisted dynamic assessment (CA-DA) for L2 vocabulary learning and diagnosis. Computer Assisted Language Learning, 36(7), 1338–1364. https://doi.org/10.1080/09588221.2021.1987272
Karpicke, J. D., & Roediger, H. L. (2008). The critical importance of retrieval for learning. Science, 319(5865), 966–968. https://doi.org/10.1126/science.1152408
Mondria, J.-A., & Boer, M. W.-D. (1991). The Effects of contextual richness on the guessability and the retention of words in a foreign language. Applied Linguistics, 12(3), 249–267. https://doi.org/10.1093/applin/12.3.249
Abstract section 2: Contribution/research questions
The present study aims to advance the line of research on contextual richness by testing the effectiveness of AI-generated and human-written contexts in supporting word retrieval. Research questions are as follows:
1. What is the difference between AI-generated and human-written contexts?
2. How effective are AI-generated and human-written contexts in triggering vocabulary retrieval?
| Title | Vocabulary Retrieval: AI-generated vs Human-written contexts |
|---|---|
| Teaching Context | College and university education |