This study investigates the efficacy of integrating corpus-based tools and bilingual scaffolding to explicitly instruct emotional vocabulary (EV) among secondary school ESL learners. Despite its criticality for interpersonal communication and identity construction, EV remains marginalized in ELT curricula, a research gap driven by the inherent abstractness of affective lexis and a lack of systematic, data-driven frameworks for its instruction. To address this, a quasi-experimental intervention was conducted with 90 Class VIII students at a government high school in Tadepalli, Andhra Pradesh. The study utilized a mixed-methods approach, leveraging Sketch Engine and AntConc to generate contextualized concordance lines, which were further supported by L1 (Telugu) bilingual glosses to mitigate semantic ambiguity and cognitive load. The analysis employed a robust quantitative comparison of three exposure thresholds, EXP 7, EXP 14, and EXP 20, complemented by a qualitative “Trace & Talk” protocol to capture learners’ cognitive processing of emotional nuances. Findings revealed statistically significant gains across all groups (p < .001), demonstrating a strong linear correlation between frequency and mastery. Although the EXP 20 group exhibited the highest immediate growth in productive knowledge of meaning (PKM), increasing by 105.3%, delayed post-tests indicated that the EXP 14 group achieved superior long-term retention stability. This suggests that in resource-constrained settings, a moderate but cognitively engaging repetition threshold may be more effective for longitudinal consolidation than intensive exposure alone. The broader implications of this study extend to the global advancement of data-driven learning (DDL), demonstrating how high-tech corpus methodologies can be democratized for low-resource educational contexts through “paper-based” adaptations. By bridging the gap between digital tools and L1 resources, the study offers a replicable model for fostering affective literacy and communicative agency. Future research should build on this by applying similar corpus-informed frameworks to larger, more diverse learner populations to further investigate the intersection of digital mediation and cultural identity in language acquisition.
Muthyala et al. (Sun,) studied this question.