A Systematic Review of Studies on Pronunciation Instruction and Practice in L2 (2005-2025)

Abstract

This systematic review (SR) synthesizes sixteen research articles published by the author between 2005 and 2025 on pronunciation pedagogy and practice in L2 contexts, with a particular focus on Arabic–English learners. The studies covers multiple instructional settings and modalities and were categorized into five thematic clusters: (i) AI generated speech and pronunciation errors, (ii) segmental errors in specialized domains, (iii) prosody, discourse, and spoken Arabic phenomena, (iv) technology enhanced pronunciation and listening practice, and (v) cognitive foundations of listening, decoding, and pronunciation. Across these clusters, the findings converge on a central issue: pronunciation is not an isolated phonetic skill but a cognitive–linguistic process shaped by phonological transfer, lexical access, decoding efficiency, background knowledge, and input quality. When any of these components is weak, learners rely on predictable compensatory strategies, such as nonsense forms, rhyming analogies, literal translation, or overgeneralization, revealing the cognitive pressures involved in real time speech production. The SR also shows that effective pronunciation pedagogy extends beyond segmental practice. Instruction that integrates decoding, grapheme–phoneme correspondences, sustained exposure to accurate models, and guided oral reading results in more consistent improvement in accuracy and fluency. Technology as text to speech software, YouTube videos, mobile apps and AI generated speech, emerges as a promising support for pronunciation practice when learners engage with it regularly and receive structured opportunities to connect written and spoken forms. Nevertheless, several gaps still exist in the field: suprasegmental features are underrepresented, longitudinal studies are limited, and many technology based interventions lack rigorous phonological evaluation. Addressing these gaps is essential for advancing pronunciation pedagogy. Overall, this study contributes a clearer, more integrated understanding of pronunciation pedagogy by positioning pronunciation as a cognitive linguistic skill shaped by perception, decoding, vocabulary development and repeated exposure to accurate models. The review offers a foundation for future research and provides educators with evidence based insights for designing pronunciation instruction that is both pedagogically sound and responsive to learners’ needs.