Mediating Corpus Data for Novice Users: Data-Driven Learning as a Digital Interpretive Practice

Abstract

Corpus-based Data-Driven Learning (DDL) is often linked to autonomous, expert language learners, but its use with young A1 learners requires careful attention to how digital language data is displayed, explained, and made pedagogically usable. Corpus examples are treated not as fixed or self-evident knowledge, but as linguistic evidence whose value depends on interpretation, task design, and representation. Using a digital epistemological perspective and the 4Is framework of Illustration, Interaction, Induction and Intervention, the study investigated the selection, adaptation and mediation of corpus-based resources for young learners, and the constraints and opportunities for engagement of learners and teachers in contexts of low expertise. The study drew on an eight-week sequence of 16 DDL lessons in three A1-level Movers classes in Vietnam. It qualitatively analyzed eight grammar-focused observed lessons because their pattern-noticing tasks allowed comparison of how corpus examples were selected, adapted, and mediated across the 4Is stages. Eight vocabulary-focused lessons formed part of the wider implementation context but were not analyzed in detail. The findings show that teachers acted as epistemic mediators by organizing, simplifying, and sequencing corpus examples, guiding learners’ interpretation of linguistic patterns, and sustaining clarity through explicit instruction and task design. Constraints such as limited access to data, smaller datasets and close supervision of teachers did not only constrain DDL; they provided tractable conditions for guided pattern noticing. The study further suggests teacher-mediated work with selected corpus examples can serve as a developmental bridge toward more learner-centered interpretation of less manipulated and more authentic corpus data over time.