The Elephant in the Phonics Room: Rethinking Phonics in Diverse Classrooms
By: Dr Hyejeong Ahn
Why this matters?
Phonics is once again at the centre of literacy teaching. Across many English‑speaking systems, teachers are now working with synthetic phonics programmes that emphasise explicit, sequential grapheme–phoneme instruction, often accompanied by diagnostic tools such as phonics screening checks.x
Yet an uncomfortable question lingers—one we often sense but rarely name:
Is it possible that, in our current approach to phonics instruction and screening checks, our efforts to support decoding unintentionally end up privileging certain ways of pronouncing words or specific accents?
A quick reminder: What phonics is for
Phonics is based on the alphabetic principle, linking letters to spoken sounds for effective reading comprehension. It is intended to support decoding skills, not to assess accent or enforce speech norms.
Systematic phonics includes methods like synthetic and analytic phonics. In places like Australia and England, synthetic phonics (SP) is often mandated with phonics screening checks. SP explicitly teaches grapheme–phoneme correspondences in a set order, and screening typically involves students blending and segmenting words orally.
So where is the problem?
The concern is not with explicit phonics itself. Rather, the concern is that some commonly used sequences and screening practices rest on assumptions that do not hold in linguistically diverse classrooms.
1.Problematic Assumption
Assumption 1: There is a single optimal order for teaching sounds.
Phonics progressions typically prioritise “simplicity” and “utility” for early decodables, often based on monolingual speakers of dominant dialects, like the SAPIN set, selected from a 1960 American English Accent corpus. In multilingual classrooms, however, the best instructional sequence may vary, as sound contrasts that are easy for one group may not be clear for another.
Assumption 2: Oral responses in decontextualised checks are treated as neutral indicators of decoding.
In practice, scoring is shaped by listener expectations about how sounds should be realised. Decoding is therefore often judged through how closely a student’s pronunciation aligns with expected norms. A child may demonstrate accurate grapheme–phoneme knowledge yet be marked incorrect if their pronunciation differs. In decontextualised tasks, accent influences not only how students produce responses, but how those responses are interpreted, often without educators recognising this influence.
For example, a student might:
pronounce goat with a monophthong rather than a diphthong (closer to got), reflecting a feature of Philippine English
produce this as dis (/d/ for /ð/)
produce three as tree (/t/ for /θ/)
reduce vowel contrasts in sheep and ship, making distinctions less salient to listeners attuned to Australian English
These are not random mistakes. They are:
features of many English varieties
shaped by learners’ linguistic histories
replicated in students’ homes and communitiesBeryl Exley is a Professor of English Curriculum and Literacies Education at Griffith University, Australia. She advocates for child-centred practices in the early years, primary and middle years of schooling. Beryl was awarded Life Member of the Australian Literacy Educators’ Association (ALEA) in 2019.
2. Assessment equity: Whose pronunciation counts?
When assessment relies on decontextualised oral judgements, it may privilege familiarity with dominant pronunciation norms.
Research in the fields of linguistics and languages such as World Englishes, and phonology complicates these assumptions:
English is pluricentric (multiple versions) and changing consistently and rapidly.
English phonemes vary across contexts and speakers
There is no single, universally agreed “Standard English” accent.
While there is limited research specifically on phonics screening checks and accent variation, broader work on language attitudes shows that less prestigious accents are more likely to be judged negatively or deficit.
The issue is not difference itself, but which differences are treated as errors. Everyone has an accent; what counts as “accented” depends on context. Australian English may sound accented in the United States, just as American or Indian English may be perceived as accented in Australia. However, not all differences are treated equally. For instance, rhotic /r/ (car with a strong /r/) in General American English is rarely penalised in Australian classrooms, whereas features associated with less prestigious varieties are more likely to be marked as problematic.
This raises questions:
If an Australian English speaker in the United States does not produce postvocalic /r/, would their decoding be marked incorrect?
If an Indian English speaker produces west closer to vest, is this treated as a decoding error?
3. Linguistically under-informed practices
Many phonics approaches remain insufficiently attuned to how phonological systems vary across languages.
When sounds don’t exist: Phoneme absence in EAL/D learners
For many multilingual learners, some English phonemes do not exist in their home languages. This affects how sounds are perceived, produced, and categorised.
A few common features include:
Phoneme substitution (contrast not in the home language): learners may substitute one English sound for another when that sound contrast is not used in their home language (e.g., /p/–/f/, /r/–/l/, /v/–/w/). For example, fly may be produced as ply; rice as lice; raw as law; vine may sound closer to wine.
Reduced sensitivity to vowel length/quality contrasts: distinctions that are salient in Australian English may be less obvious for some learners (e.g., long–short contrasts such as ship vs sheep, bit vs beat).
Epenthetic vowel insertion (adding a vowel to break up consonant clusters): in some syllable‑based languages, consonant clusters and word‑final consonants are restricted, so a short vowel may be inserted. For example, learners from Japanese or Korean backgrounds may pronounce student as su‑tu‑den‑teu. Similarly, cat → ca‑teu; bat → ba‑teu; boat → bo‑a‑teu.
To speakers of Australian English, these pronunciations might seem like mistakes, but they are actually systematic and predictable. If the main objective is reading comprehension, it's important to clarify what is being assessed: decoding or pronunciation. For example, when a student says wine instead of vine, this instead of dis, or ship in place of sheep, as long as the meaning is understood and the students consistently apply these sound patterns, should this be considered weak grapheme-phoneme knowledge and require correction?
Why this matters for multilingual learners
For EAL/D learners and varieties of English speakers the implications are serious. Their speech may reflect:
phonological patterns from home languages
features of their own English variety
developing control of English sound systems
1. Check for Validity and Reliability Questions
Validity asks: what are we actually assessing?
Are we evaluating students’ ability to decode written words, or whether their speech aligns with a particular accent or pronunciation norm? If “correctness” is judged by how closely a student’s pronunciation matches the assessor’s expectations, the focus shifts away from decoding and towards accent conformity. In this case, the assessment risks becoming a form of accent policing, where students are evaluated not for their reading ability, but for how they speak. Ensuring validity therefore requires that phonics assessment remains firmly focused on decoding rather than enforcing specific pronunciation norms.
Reliability asks: would the same response be scored the same way by different assessors?
If scoring depends on the listener’s accent exposure and expectations, the same response may be judged differently across teachers, or even by the same teacher in different contexts. This variability undermines consistency and fairness, as outcomes become shaped by the assessor’s linguistic background rather than the student’s decoding ability. Ensuring reliability therefore requires that scoring is not contingent on familiarity with particular accents, especially in linguistically diverse classrooms.
2. Audit the programme sequence as a set of assumptions, not a law of nature. Identify which correspondences and contrasts are introduced early, what “mastery” looks like, and whose speech norms are implicitly treated as the benchmark. Use local classroom evidence (student language profiles) alongside research evidence to justify any sequence used.
3. Redesign how ‘correct’ is determined in oral checks
Listen for patterns, not isolated “errors”
Is the variation consistent across words?
Consistency often signals systematic knowledge, not confusion.
Keep asking yourself: Am I assessing decoding? Or Am I unintentionally devaluing students’ own Englishes and home language practices?
4. Make variation visible in teaching
Explicitly tell students:
“In English, sounds can be said in different ways.”
“This is one way, you may hear others.”
This builds flexibility and reduces the idea of a single “correct” sound.
5. Be explicit about what you are teaching (decoding vs accent norms)
In teaching, modelling a socially recognised variety can be pedagogically inevitable and sometimes strategically helpful for students’ participation and future opportunities. But it is still a choice. Name it clearly for yourself and for students: are you teaching decoding for reading, or also (implicitly) encouraging accent conformity for speaking?
Key takeaways and what we don’t know yet
Phonics is designed to support reading, not accent conformity
English sounds are realised in multiple ways.
Decontextualised oral assessment increases the risk of conflating decoding with accent.
Multilingual learners may be disproportionately affected by current assessment practices
Teachers can respond by attending to patterns of phonological knowledge, rather than differences in pronunciation
Professional development is essential to ensure phonics assessment does not reproduce accent‑based inequities
There is an urgent need:
for research to revisit phonics for multilingual classrooms.
to test the validity and reliability of phonics screening checks for accent‑related bias in scoring.
for research on teachers' knowledge and attitudes about English varieties to develop research informed professional training
The elephant in the phonics room is that we know phonics is meant to assess decoding, yet in practice it often evaluates how “correctly” students’ sound.While language fundamentally facilitates communication, in educational settings “good English” operates as a marker of social status. Certain ways of speaking are legitimised, while others are marginalised.
Although teaching a socially recognised variety may support access to future opportunities, it also positions phonics within systems of social mobility that can reproduce gatekeeping practices, rather than focusing on reading development. When phonics assessments prioritise sounding “correct,” they function less as measures of decoding and more as mechanisms of gatekeeping, diverging from their intended purpose.
If you’d like to explore further
This blog draws on:
Ahn, H., & Sangster, C. (2026). The elephant in the phonics room: Rethinking pronunciation norms in early reading instruction. The Australian Journal of Language and Literacy. https://doi.org/10.1007/s44020-026-00106-3