Since 2020, the use of music in English language teaching (ELT) has moved beyond sporadic enrichment activities toward a more systematic, design-conscious pedagogy grounded in clear mechanisms of learning. This review combines broader framing and mechanistic research with post-pandemic classroom practices to explain how music can influence L2 outcomes and why its use has increased since 2022. The argument traces a transition from incidental songs to music as a medium of instruction (MMI) embedded in multimodal, translanguaging, and care-centred designs; it then consolidates evidence for three main pathways, affective regulation, rhythmic–prosodic entrainment, and memory/repetition, while highlighting social-relational and reward-based processes that strengthen these mechanisms (García et al., 2022; Meighan, 2023; Madkur et al., 2022; Renihan et al., 2024).

Emergency remote teaching normalised audio and video workflows, subtitles, and learner media, reducing logistical barriers for music lessons. Teachers adopted webinars, LMS, YouTube, and recording apps; students made short artefacts for repeated practice (Adisti, 2022; Devarajoo, 2021; Hidayati et al., 2021; Khadka, 2023; Öztürk, 2023; Poonpon, 2022; Rose et al., 2020). However, access inequalities and digital literacy limit music-rich designs, especially in low-resource areas (Devarajoo, 2021; Khadka, 2023; Öztürk, 2023). By 2022–2025, short, repeatable “micro-sequences” combining lyric exposure, guided practice, and quick feedback in blended formats supported by lightweight tools emerge (Kim and Namkung, 2024; Metruk, 2024; Sulindra et al., 2024).

While earlier work viewed songs or nursery rhymes mainly as stimuli for vocabulary or rhythm (Ahire, 2021), post-2020 scholarship sees music as an instructional channel within structured frameworks. García et al. (2022) define MMI as using music to deliver language input and focus, such as active listening tasks, while multimodal and translanguaging approaches incorporate local and heritage knowledge into ELT, viewing music as a resource for identity and participation (Madkur et al., 2022; Meighan, 2023). This shift connects with movement-based designs (Barrio and Arús, 2024), participatory models using Gen-Z media (Bahaudin et al., 2021), and a care-focused approach in higher education that sees music-infused pedagogy as relational work (Renihan et al., 2024). Policies promoting holistic, interdisciplinary learning, like India’s NEP initiatives, also legitimise music-centred ELT in curricula (Anandvardhan, 2025; Sharma and Sharma, 2025).

Umbrella reviews across clinical and educational settings show that music can reduce arousal and subjective anxiety and improve mood, effects that plausibly lower foreign-language anxiety and increase willingness to communicate (WTC) in classroom interactions (Harney et al., 2022; Chen et al., 2021; Weijden et al., 2021). Classroom studies on affect-oriented pedagogy similarly link enjoyment and reduced anxiety to greater participation, with song-based routines offering a low-stakes pathway when framed as collaborative, optional, and care-attentive (Hickey et al., 2022; Toyama and Yamazaki, 2021; Weerakoon et al., 2023). The care-centred literature suggests that attention to classroom relationships is not incidental but constitutive: radical care pedagogies situate musical activity as a space for safety, voice, and mutual recognition—conditions that enable speaking and risk-taking (Renihan et al., 2024).

English intelligibility largely relies on stress timing, reduction, and intonation. Music’s metre and beat accentuate these temporal and pitch patterns; combining rhythm with gesture or stepping (sensorimotor entrainment) improves temporal understanding and phrase-level organisation (Barrio and Arús, 2024). Evidence from intervention and rehabilitation contexts shows that singing and rhythm-based tasks can recalibrate prosodic control, supporting the idea that classroom designs that emphasise stress and contour before transitioning to speech are plausible (Ong et al., 2024). Classroom technology—such as slow playback, visual feedback, and mobile recording—has made such prosody-focused micro-sequences easier to scaffold and assess (Kim and Namkung, 2024; Metruk, 2024). The main point is not that music offers a general cognitive benefit but that it makes specific suprasegmental cues more accessible for perception and production.

Lyrics have repetitive structures like choruses, aiding exposure without boredom. Research shows retrieval practice and spacing are key for long-term vocabulary learning, achievable through lyric activities like gap-fill and re-listening (Koval, 2021; Nakata, 2015; Rice and Tokowicz, 2019). Neuroscience adds two insights: first, reward from enjoyable music activates dopaminergic systems, so learner choice matters (Ferreri et al., 2013, 2021; Ferreri and Rodríguez-Fornells, 2022; Murty et al., 2018). Second, involuntary musical imagery (“earworms”) can support verbal memory through spontaneous replay, complementing study (Kubit and Janata, 2022, 2023; Williamson et al., 2011). Effectiveness varies based on item features and context; sometimes rest works better than music, especially for abstract words (Martini et al., 2022). Reviews show active musical engagement surpasses passive listening for recall, and movement enhances memory via multisensory coupling (Chin and Rickard, 2010; Gkintoni et al., 2025; Stekić, 2024; Vongpaisal et al., 2016).

Post-2020 practice divides into three overlapping modes. Movement-embedded pedagogy combines rhythm with clapping, stepping, or gestures to externalise stress feet and phrase boundaries, improving timing and inclusion in mixed-ability groups (Barrio and Arús, 2024; Vongpaisal et al., 2016). Performative and participatory models leverage learner identity and social reinforcement through short, low-stakes performances or recorded artefacts to promote effortful retrieval in motivating contexts (Bahaudin et al., 2021). Care-centred approaches use music to build safety and reciprocity, reducing performance anxiety and increasing participation, especially in tertiary settings (Renihan et al., 2024). In blended courses, platform-mediated tasks such as subtitled YouTube videos introduce controllable tempo, captions, and multimodal cues. However, they depend on curation, digital skills, and equitable access (Adisti, 2022; Devarajoo, 2021; Sulindra et al., 2024).

Progress relies on the alignment of mechanisms and measurement. In prosody research, acoustic features such as F0 range, duration, and rhythm metrics like %V and nPVI, along with intelligibility ratings, complement teacher assessments. Existing pipelines require adaptation for singing and fast speech (Algethami and Hellmuth, 2023; Ding and Xu, 2016; Kim and Namkung, 2024; Metruk, 2024). Forced-alignment tools, like the Montreal Forced Aligner, can accelerate labelling but should be used cautiously with singing; methods that incorporate singing awareness, joint pitch-lyrics alignment, and manual checks are recommended (Huang et al., 2022; Liu, 2024; MacKenzie and Turton, 2020; Mahr et al., 2021; McAuliffe et al., 2017; Wu et al., 2023). For connected speech, quantifying weak forms- such as durations and centralised formants- and analysing linking features like boundary continuity and formant transitions is achievable via semi-automatic extraction and spot-checks (Adi et al., 2015; McAuliffe et al., 2017; Wu et al., 2023). Algorithms can flag deletion or assimilation, but these should be validated against human transcriptions (Lipani, 2020; Yuan and Liberman, 2011). Delayed outcome assessments are essential to evaluate long-term effects across various domains (Koval, 2021; Rice and Tokowicz, 2019). Literature on affect and WTC emphasises triangulating self-reports with behavioural and light physiological measures, while also recording teacher immediacy and classroom climate as moderating factors (Derakhshan et al., 2022; Dewaele et al., 2019; Jin, 2023; Reinders and Wattana, 2014; Toyama and Yamazaki, 2021; Weijden et al., 2021).

Four forces help explain sustained adoption. Digital infrastructures and habitualised media workflows lowered the activation energy for integrating music (Adisti, 2022; Poonpon, 2022; Rose et al., 2020). Methodological innovation, MMI and prosody-facing micro-sequences, reframed music as a vehicle for explicit linguistic targets (García et al., 2022; Metruk, 2024). Policy legitimation encouraged multimodal, interdisciplinary practice in formal programmes (Anandvardhan, 2025; Sharma and Sharma, 2025). Epistemic shifts towards translanguaging and local knowledge validated culturally situated repertoires (Madkur et al., 2022; Meighan, 2023). Yet uptake ultimately depends on teacher beliefs, i.e., value, self-efficacy, and identity, and enabling conditions (PD, aligned materials, workload, leadership), which determine whether pilots become routine practice; programme studies consistently call for scaffolded professional development and syllabus-mapped resources to sustain change (Pitychoutis et al., 2025; Kim and Namkung, 2024; Sulindra et al., 2024).

Two issues recur. First, causal precision: many reports are brief, involve small samples, and lack comparison conditions, making it difficult to distinguish effect sizes and mechanisms. Utilising cluster RCTs or within-class cross-over designs that manipulate chant exaggeration, retrieval/feedback schedules, and spacing would strengthen inference (Koval, 2021; Nakata, 2015; Rice and Tokowicz, 2019). Second, context and inclusion: adult EFL/EMI, non-music vocational settings, and under-researched regions remain underrepresented; considerations of equity (devices, bandwidth) and cultural acceptability must be integrated from the start (Devarajoo, 2021; Khadka, 2023; Öztürk, 2023; Novianti et al., 2025). Methodologically, singing-aware alignment, mixed acoustic–perceptual batteries, and routine delayed tests are accessible improvements. Substantively, the most promising focus areas are prosodic and phrasal phenomena (weak forms, linking, stress/intonation), where musical scaffolds closely align with the linguistic constructs; segmental reductions are more conditionally trainable and likely require paired spoken-mode practice and intensive perception work (Algethami and Hellmuth, 2023; Ding and Xu, 2016; Ong et al., 2024).

The post-2020 literature supports a cautious but practical claim: music works best in ELT when used as pedagogy, not merely as ambience, i.e., when tasks make prosodic cues audible and practised, or provide repetition-rich lexical encounters with retrieval and spacing. Affective, reward, and social-bonding processes foster engagement, while entrainment and mnemonic structure carry the instructional weight. To strengthen gains, future work should combine principled designs (MMI, translanguaging, movement-embedded micro-sequences) with more robust measurement (acoustic-perceptual batteries, delayed tests) and supportive conditions (PD, syllabus mapping, equitable access). Done well, music shifts from occasional novelty to scaffolded micro-pedagogy that is culturally responsive, ethically attuned, and empirically testable (García et al., 2022; Renihan et al., 2024; Rice and Tokowicz, 2019).

This was a focused mini review that synthesises post-COVID, Scopus-indexed empirical research on the use of music, song, chant, or rap in English language teaching and learning (ELT/EFL/ESL). We followed PRISMA 2020 guidance to structure the search, screening, and reporting processes appropriate for a concise, narrative mini-review format (Figure 1).

The primary information source was a Scopus-indexed corpus, containing 89 numbered records relevant to the topic. The eligibility window was 1 January 2020 to 31 August 2025, to reflect post-COVID teaching ecologies and recent classroom practice. The reason Scopus was chosen was its broad coverage across Education and Humanities (Martín-Martín et al., 2018).

Inclusion criteria:

Exclusion criteria:

Two reviewers independently screened the 89 records at title/abstract and full-text stages against the criteria above. Two reviewers independently screened titles/abstracts and then full texts against the eligibility criteria. Disagreements were resolved by discussion; a third reviewer was available but not required.

From the 89 Scopus-indexed records provided, 11 empirical, peer-reviewed journal articles (2020–2025) met the inclusion criteria (English as the target language; music/song/chant/rap central to instruction or exposure; measurable L2 or validated affect outcomes; no conference papers, chapters, or reviews). We also acknowledged an adjacent observational study that was not used to support causal claims, bringing the total to 12. The remaining records were excluded mainly because they were not about music-mediated ELT instruction (n = 45) or not ELT/EFL/ESL (n = 42); further exclusions included reviews (n = 12), non-empirical/conceptual pieces (n = 9), items published before 2020 (n = 6), and conference proceedings (n = 1).

For each included article, we extracted details such as author(s), year, title, outlet, target population and setting, as well as intervention or task descriptions. This includes whether singing, chanting, rapping, lyrics, or YouTube music videos were used, along with the approximate dose, and the outcome domains and measures, such as acoustic or production metrics, lexical tests, listening or speaking performance indicators, and validated affect scales. Extraction was performed by one reviewer and verified by a second reviewer against the article text and the Scopus entry.

Due to heterogeneity, pooled effects were not calculated. Instead, we summarised effect directions and constructed alignment by outcomes close to the target. We also counted studies by year and main theme using a hybrid deductive–inductive coding scheme (with skills prioritised first and modality adjustments afterwards).

Given the scope of the mini-review and the heterogeneous designs, we did not apply a formal risk-of-bias tool. Instead, we recorded the design type (e.g., classroom quasi-experiment, case study, survey), the dose/duration when reported, the alignment of measures to targets, and the presence/absence of delayed post-tests or acoustic/production metrics. This information informed our narrative weighting in the Results and Discussion (e.g., stronger claims where designs and outcomes closely matched the stated mechanism, more cautious interpretation where measures were distal or designs were brief) (Table 1).

Five studies explore lexical improvements through lyric-based exposure and practice: Chen (2020) used mnemonic melody-pairing; Tilwani et al. (2022) employed a Solomon design with five 90-min sessions and a three-week delay; Lu and Murao (2023) manipulated musical input during vocabulary learning; Zhang et al. (2023) found singing more effective than recitation for pronunciation and vocabulary among adolescent ESL learners; and Mannarelli and Serrano (2024) observed that explicit vocabulary teaching via songs in EFL classrooms yields more durable gains than implicit methods. A common finding is that structured repetition combined with retrieval, often through chorus or refrain cycles and prompted recall, consistently leads to improvements. However, when music is used without a structured practice, the effects depend on learner characteristics like working memory and proficiency.

Three studies train suprasegmentals and oral performance through sing/chant-then-speak or rhythmic priming: Kitjaroonchai and Sukman (2025) (15-week song-based curriculum; gains in CAF with delayed retention), Wang and Liu (2025) (song-lyrics/rhyme tasks; phoneme categorisation improvements among underachievers), and Sugiura and Hori (2025) (lab-based rhythm priming; acoustic gains in rhythm and vowel-length contrast, moderated by proficiency/phonological STM). The addition of Sugiura and Hori (2025) strengthens claims that explicit rhythmic scaffolding can shift measurable acoustic indices, not just performance ratings.

Busse et al. (2021) conducted a classroom RCT in primary EFL demonstrating that singing outperforms speaking/control in improving grammar and spelling. This expands the benefits of music-mediated instruction beyond vocabulary and prosody, strengthening the case for form-focused lyric activities in young learners.

AlSmadi (2020) reports increased motivation with song-based routines compared to other methods, based on validated self-report data. Feng and Guo (2025) conducted a ten-week case study using YouTube music videos, documenting behavioural, emotional, social, and cognitive engagement in real-time. Their focus is on feasibility conditions rather than establishing causal effects.

Zhang (2025) connects music perception to ESL reading comprehension using MANOVA and PLS-SEM without providing musical instructional treatment. We consider this as a mechanistic background and do not interpret it as causal evidence for teaching methods; all the effect claims mentioned are based on the eleven intervention and evaluation studies and the one observational study.