Phonics instruction has been re-emphasised in contemporary literacy policy as the foundation of early reading, reinforced through major reports and curriculum reforms. Grounded in the alphabetic principle, phonics teaches grapheme-phoneme correspondences to support decoding. Yet despite its prominence, limited attention has been paid to assumptions about spoken English that shape how phonics is taught and assessed. Many phonics practices, particularly synthetic phonics and oral phonics screening checks, rely on a single, standardised pronunciation model, overlooking the pluricentric and evolving nature of English. In classrooms where learners and teachers speak diverse varieties of English, such assumptions may risk narrowing what counts as accurate decoding. The paper outlines the historical development of phonics and the debates that shaped its implementation, before proceeding to draw on the Oxford English Dictionary’s coverage of world Englishes to show that English phonemes are realised in multiple systematic ways. Due to systematic variation in spoken language across accents and linguistic backgrounds, oral decoding responses may be evaluated not only in relation to grapheme-phoneme knowledge but also according to how closely they align with assessors’ expectations of a socially familiar pronunciation norm. In such contexts, judgements about decoding accuracy may therefore be shaped not only by phonological knowledge underlying decoding but also by social expectations about which pronunciations are recognised as legitimate forms of English. The paper argues for a linguistically responsive approach to phonics instruction and assessment that recognises phonetic variation across Englishes and avoids the risk of treating accent-related pronunciation variation as evidence of decoding difficulty. It concludes by identifying implications for multilingual learners, speakers of diverse English varieties, and teachers, along with key gaps in research on teacher knowledge and judgement in linguistically diverse classrooms.
Ahn et al. (Mon,) studied this question.