Prosodic encoding and the development of durable lexical representations in L2 English: a randomised controlled longitudinal study

Robust L2 lexical representations require not only semantic and orthographic knowledge but also precise phonological specification, yet the role of lexical stress remains underexplored. Intermediate Arab EFL learners (n = 132) were cluster-randomized to either a Prosodic Encoding Group, receiving a multi-cue encoding package with explicit stress-perception and stress-production activities, or an Active Control Group, receiving meaning-focused vocabulary instruction without systematic stress emphasis. Participants learned forty mid- to low-frequency academic words over six weeks. Outcome measures included immediate and delayed retention tests, timed lexical decision tasks, controlled written and oral production tasks, stress perception assessments and acoustic production analysis. Linear mixed-effects models and multilevel mediation analyses were employed. The Prosodic Encoding Group demonstrated superior delayed retention with medium-to-large effects (ds ≈ 0.8–1.1), faster lexical retrieval and higher productive accuracy. Stress perception gains were statistically associated with delayed retention, consistent with a role for suprasegmental specification in lexical durability. These findings support integrating prosodically oriented, multi-cue training within L2 vocabulary instruction, while acknowledging that stress specification’s unique contribution was not isolated from other prosodic and multimodal supports.