%20(1)%20-%20Copy.jpg)
Phonics Reform England: Not reading reform. Phonics reform. Identifying the parts that are failing too many children.
English Pronunciation Code (EPC)
The Pronunciation Code describes how individuals produce and organise speech sounds, which vary by accent and experience, and understanding this is essential for accurately mapping speech to print and achieving Word Mapping Mastery®.
What is the Pronunciation Code?
The Pronunciation Code refers to how speech sounds are organised and produced when mapping spoken words to print.
In English, this is not fixed. It varies according to accent, speech development, and individual differences.
This means that even when children are learning the same words, they may not be working from the same underlying sound structure.
Children do not all start with the same pronunciation code, even though phonics programmes are designed as if they do.
This matters because learning to read and spell depends on mapping speech to print. If the speech sound structure is not secure or not shared, this mapping becomes more difficult. Without a bridge between the child’s pronunciation code and the programmes used, some children will experience considerable difficulty aligning spoken and written English.
Phonics introduces how print represents speech, but it assumes that children can already access and work with the sounds in words, and that they all use the same sound system. For some children, this is not yet in place. For others, they do not use that system. Some children face both challenges.
The International Phonetic Alphabet (IPA) provides a consistent way to represent speech sounds. However, it is complex and primarily used in specialist fields. It is not designed for everyday classroom use or for young learners.
Attempts have been made to make the English code more transparent, most notably through the Initial Teaching Alphabet (ITA) develped by James Pitman . This was developed to create a simpler system in which each sound is represented by one symbol, so that children could see a direct and consistent link between speech and print. The aim was to make reading and spelling effectively “phonetic”.
Children were expected to learn this simplified system first and then transfer that knowledge to standard written English once reading was established.
However, this does not align with how reading develops. Children learn to read and spell through self-teaching, building knowledge of the written code by mapping speech to print across many words. If the code they are learning is not the one they will continue to encounter, this process is disrupted.
English is not a fully phonetic system. When children move from a simplified one sound–one symbol code to standard spelling, the correspondences they have learned no longer apply. They cannot simply transfer what they know. They have to learn a different code.
As discussed by Stanilas Dehaene (2009, pp. 31–37), the complexity of English spelling reflects the need to encode language beyond a simple transcription of speech. Writing does not aim to reproduce speech exactly, but to encode it at a level that allows efficient access to meaning.
However, while this explains why a fully transparent system is not possible, it does not address how individual learners navigate this complexity when their own pronunciation code differs.
It is a problem no-one has found a solution for, until now.
AI cannot do this.
This ground-breaking work from Emma Hartnell-Baker, who defines the logic of the pronunciation and spelling code, working in partnership with a tech team who make this logic operational through coding, focuses on that gap.
An algorithm has been developed to convert words into defined phoneme arrays and grapheme arrays, enabling bidirectional mapping between speech and print.
This makes the relationship between each learner’s pronunciation code and the written code visible and usable.
The Pronunciation Code also has wider applications. It is being used to develop speech-first technology, including augmentative and alternative communication (AAC) approaches that support children who are not yet able to speak. It also supports learners of English as an additional language by making the sound structure of words visible, helping them understand and produce English pronunciation.
Understanding the pronunciation code used within phonics, as seen in the English Code Overlay and the Phonics Pronunciation Code, and then comparing it to the child’s own pronunciation code, helps educators understand not only the need for a bridge, but how that bridge can be created.
Word Mapping Mastery® (WMM) is the goal for all learners before the age of 7. However, many learners require additional visual and structural support. This is why we build word mapping technology.
The Pronunciation Code underpins this work. While the output is consistently structured, it can be adapted so that users can align it with their own pronunciation code, ensuring access for all learners.
Dehaene, S. (2009). Reading in the brain: The new science of how we read. Viking.
These earlier attempts reflect an important scientific insight: that reading depends on understanding how speech is represented in print. The drive to make this relationship clearer was well founded.
However, changing the written code itself proved to be the wrong solution, as it disrupted access to the language children ultimately need to read. Learning depends on self-teaching through interaction with the actual written code. When that code is replaced, the opportunity to build this process is disrupted.
The issue is not whether some children can transfer between systems, but why others cannot. The focus must be on those at risk, often around one in five learners, who do not develop this mapping easily. This has not changed whether 'three cueing' was used, or synthetic phonics.
What this work shows is that the original problem was correctly identified, but not fully resolved. The challenge is not to replace the code, but to reveal it. By making the structure of standard written English visible, while allowing it to be aligned with each learner’s pronunciation, it becomes possible to support self-teaching in the way the system requires. This is where the science, the code, and the technology now come together.
Let's Show the Code! The Orthographic Mapping Tool
The Code Mapping® display, using the English Code Mapping Tool, makes the grapheme structure of words fully visible, complementing the representation of speech sounds shown through the IPA. Every word is segmented into graphemes, and each grapheme is shown as a complete unit in sequence across the word. The alternating colours are not decoration, they simply mark the boundaries between graphemes so the structure can be seen at a glance. When you track each grapheme across the word, the structure of the code is immediately visible. This is not additional information layered onto the word, it is the word, with its full grapheme structure made visible. Have a free play here!
Picture-embedded mnemonics, including systems such as Letterland, are often used to introduce grapheme–phoneme correspondences. A character such as Clever Cat may help a child recall that the grapheme <c> can represent a /k/ sound. As Linnea Ehri notes, this kind of support can be useful in the earliest phase of learning, where children form partial connections and benefit from salient visual cues.
However, this support is transitional, and its limitations become clear very quickly in English. English is an opaque orthography, where graphemes do not map consistently to a single phoneme. The same grapheme <c> appears in cat and cent, representing different sounds. A mnemonic tied to one sound does not prepare the learner for this shift. The child may remember the character, but not understand the structure of the word. The association must then be overridden or abandoned, creating instability at the point where precision is needed.
This is not a minor issue. Orthographic mapping depends on forming direct, bonded connections between phonemes and graphemes within words. When attention is directed towards characters, stories, or visual prompts rather than the internal structure of the word, this bonding is weakened. The learner may recall the mnemonic but fail to establish a stable representation of the word itself.
The problem lies in where the mnemonic is anchored. Picture-based systems attach meaning to the grapheme, yet in English the instability lies at the grapheme level. A grapheme does not carry a fixed sound value across words, so any system that binds it to a single cue will break down as soon as variability is encountered.
A different approach is to anchor representation at the level of the phoneme. The phoneme remains stable even when its spelling changes. When learners can reliably identify and represent each phoneme in a word, they are better able to map those sounds to the available graphemes without relying on fixed cues or memorised associations. This provides a stable foundation for both reading and spelling in an opaque orthography.
What about mnemonics?
For children like Alf who have not succeeded with print-first phonics this shift is significant. It provides a way into the code that does not depend on prior success with letters. It allows them to build from what they can already do, hear and say sounds, and develop accurate representations that can then be mapped to print.
Phonemies® are the starting point. Each Phonemie® represents a speech sound, allowing even very young children to identify, segment, and work with the sounds in words before needing to engage with spelling. For toddlers, this provides an immediate and accessible way into language, grounded in what they can hear and say.
This matters because English is an opaque orthography. Graphemes do not consistently represent the same sound across words, so relying on letters as the entry point introduces variability from the outset. By contrast, phonemes are stable. When learners begin with clearly identified phonemes, they are working with a consistent system.
Phonemies® make these phonemes visible and usable. Each symbol represents a sound that can be said, heard, and tracked in sequence. This allows learners to build a secure understanding of the sound structure of words without needing to rely on memorisation, rules, or external prompts.
Once the phoneme sequence is secure, graphemes can be introduced as the written representations of those sounds. At this point, the task is not to work out what the word says, but to select how each phoneme is represented in print. This shifts the process from uncertainty to precision.
This aligns with the process described by Linnea Ehri, where orthographic learning depends on forming connections between phonemes and graphemes. Phonemies® support this by ensuring that the phoneme, the stable element in the system, is clearly identified from the start.
Phonemies® therefore provide a foundation that begins before print and continues as written language is introduced. They allow learners to work with the structure of spoken words first, so that mapping to graphemes becomes clear, consistent, and meaningful.
For children at risk of reading and spelling difficulties, starting with print introduces complexity too early. English orthography is variable and unpredictable at the grapheme level. When learners are asked to work from letters to sounds before they have secured the phoneme structure of words, they are more likely to guess, memorise, or rely on partial cues.
A speech-first approach removes this instability. Phonemes can be heard, produced, and segmented. They provide a consistent and accessible starting point, particularly for children with speech, language, or processing differences. By using Phonemies® as the reference point, learners can work with the structure of spoken words in a way that is clear and repeatable.
This is especially important for children who struggle with attention, memory, or phonological processing. When the task is anchored in sound, they can track each unit in sequence and build a complete representation of the word. This reduces cognitive load and supports accuracy. Instead of trying to interpret print while simultaneously working out the sounds, they are able to focus on one stable system first.
Once the phoneme structure is secure, mapping to graphemes becomes a process of selection rather than inference. This supports the formation of precise phoneme–grapheme connections, as described by Linnea Ehri, and reduces the likelihood of errors becoming embedded.
👉 Phonemies® start with the sound, so children can build the code from what is stable, before meeting the variability of print.
