The Development of Receptive and Productive Knowledge of Idiomatic Phrasal Expressions in Saudi EFL Learners: A Multi-Phase, Multi-Modal Corpus Study

Abstract

While corpus linguistics has revolutionized vocabulary teaching, idiomatic phrasal expressions (IPEs) remain underrepresented in Saudi mainstream syllabi. Existing studies often lack longitudinal scope, multimodal analysis, or robust corpus design, leaving critical questions unanswered about how idiomatic competence develops across time, tasks, and instructional conditions. This dissertation investigates how Saudi EFL learners acquire, use, and retain idiomatic phrasal expressions (IPEs) across three phases: baseline, intervention, and retention, while examining modality effects (writing versus speaking). The study adopts a longitudinal, mixed-methods design, and draws on the Saudi Female English Corpus (SFEC), a purpose-built learner corpus of approximately 100,000 words, comprising 300 spoken and written texts produced by 60 Saudi female university learners (experimental = 30; control = 30). Productive idiomatic use was analyzed using Martinez and Schmitt’s Phrasal Expressions List (2012), organized into five frequency bands (B1–B5), alongside a receptive idiomatic expressions test. Quantitative analyses examined idiomatic density, lexical variety, frequency-band distribution, and receptive–productive alignment, while qualitative analyses explored pragmatic functions, syntactic patterning, and cross-linguistic influence. Baseline findings revealed a pronounced receptive–productive gap, with learners recognizing high-frequency IPEs (Bands 1–2) but rarely producing them accurately. Following a corpus-informed, explicit instructional intervention, the experimental group demonstrated substantial gains in productive idiomatic use, particularly in writing, where idiomatic density increased by 129% and expanded into Bands 3–5. Speaking gains were more modest (≈19.5%) but stylistically meaningful, involving discourse markers and stance expressions. Retention analysis indicated partial attrition two months post-intervention, with Band 3 IPEs showing the strongest durability. Importantly, non-canonical and L1-influenced forms produced by the experimental group are interpreted as developmental risk-taking, reflecting deeper conceptual engagement and emerging figurative awareness rather than simple error, whereas the control group’s largely accurate output reflects more conservative, low-risk strategies with limited stylistic expansion. Pedagogically, the findings support corpus-informed, frequency-sensitive IPEs instruction grounded in Data-Driven Learning (DDL), highlighting a reciprocal relationship in which learner corpora function as positive pedagogical exemplars while simultaneously informing corpus-based instructional design.