Children need to map words in both directions:
 

from print to speech (decoding)
from speech to print (encoding)

These are not separate skills. They are two directions of the same process. When both are secure, children can move towards self-teaching with ease.

If systematic synthetic phonics is positioned mainly as the decoding of grapheme–phoneme correspondences, moving from print to speech, while speech-to-print is treated primarily as spelling and taught separately, teachers can miss opportunities to support this integrated development.

Direction is built into the system

In England, the Department for Education mandates systematic synthetic phonics as the prime approach to early reading. The direction is explicit: print to speech.

Programmes must present SSP as the prime approach to decoding print. Children are shown graphemes, taught the phonemes associated with them, and expected to blend these from left to right through the word in order to read.

Segmenting spoken words into phonemes for spelling is included, but it is defined as the reverse of blending. One direction therefore takes precedence. Print to speech is foregrounded. Speech to print is included, but positioned as the reverse process rather than as an equal starting point.

This matters because what is made explicit in national guidance affects what teachers understand to be central. If decoding is consistently foregrounded, while encoding is framed as the reverse, teachers can come to see print-to-speech mapping as the core of phonics and speech-to-print as something additional.

The Department for Education reinforces this in its public guidance, explaining phonics as children learning to:

recognise the sounds that each individual letter makes
identify the sounds that combinations of letters make
blend these sounds from left to right to read a word

This is a print-to-speech model.

Why this creates a problem

This wording is not an accurate description of English orthography.

Individual letters do not simply “make” a sound. The letter a, for example, can represent multiple different phonemes in English. Nor do phonics programmes teach all the sounds that each individual letter, or each combination of letters, can represent.

What they teach is a limited, programme-defined subset of grapheme–phoneme correspondences in an incremental sequence. In practice, SSP programmes cover around 100 correspondences, not the full code. English orthography contains far more correspondences than this.

So the claim that children are taught “the sounds that each individual letter makes” and “the sounds that combinations of letters make” is misleading in two ways. It oversimplifies the structure of English, and it overstates the scope of what children are actually taught.

This can encourage an incomplete model of how the writing system works, as if:

• letters have fixed sounds

• children are learning the sounds of letters

• the system is largely consistent and transparent

That interpretation would only make sense in a transparent orthography. English is not.

In English:

• graphemes represent phonemes within spoken words

• a grapheme can represent more than one phoneme

• a phoneme can be represented in multiple ways

If this is not made explicit, children can form an incomplete model of the code, and teachers may not be encouraged to make that representational relationship clear.

The starting point matters

The Department for Education guidance for SSP programme developers begins teaching at Phase 2 and does not require phonemic awareness to be fully developed beforehand.

Children are introduced to graphemes early, while phoneme awareness develops alongside.

A speech-to-print approach would start differently.

Children would first:

• listen for phonemes in spoken words

• segment those words into sounds

• blend sounds back together

This establishes the key principle that speech is represented in print, rather than suggesting that print is the starting point.

It is also more schema-driven. It builds on what children already have, spoken language. Instead of asking children to infer the relationship between speech and print from graphemes on a page, it starts with the sounds in words they already know and shows how those sounds are represented in writing. That makes the logic of the alphabetic system clearer from the outset.

It also helps identify children who begin school with weak phonemic awareness and may be at higher risk of difficulty.

The claim that “almost all children” will succeed

The Department for Education states that “almost all children” who receive high-quality phonics teaching will learn the skills needed to tackle new words.

However, national data does not support this claim.
 

 

• 1 in 5 children do not meet the expected standard in the phonics screening check after Year 1

• 11% still do not meet this standard after Year 2

• 1 in 4 children do not reach the expected standard in reading by age 11

These are not small numbers. They represent a significant group of children for whom the current model is not sufficient.

Decoding is not the same as knowing

Children who decode can still reach the correct word without securely mapping it.

They may rely on:

• context

• memory

• partial decoding

This can give the impression of success without secure underlying knowledge.

Encoding is different.

To spell a word, a child must:
 

• identify the full sequence of phonemes

• assign graphemes to represent each phoneme

• This requires precision and cannot be bypassed.

Spelling (encoding) is the stronger test

While decoding (print to speech mapping) may allow a child to recognise a word, spelling (speech to print mapping) shows more clearly whether they have actually mapped it.

If a child can spell a word, it indicates that they have:
 

• identified the phonemes

• linked them to graphemes

• stored the structure

For this reason, spelling provides a clearer reflection of orthographic learning and memory.

What is not being evaluated

The Department for Education stated that, although programmes may include wider aspects of reading, writing and spelling, these elements were not included in the validation of SSP programmes.

This matters because it affects what is measured, what is valued, and what teachers are held accountable for.

As a result:
 

 

• print-to-speech is clearly defined and prioritised

• speech-to-print is less explicitly specified

• decoding is assessed more directly than encoding

This creates an imbalance in how children are supported to learn words, and in what teachers may come to see as the main goal of phonics teaching. It also means that phonics programmes aren't giving guidance relating to phoneme to grapheme mapping. Accent variation will make this much harder than print to speech mapping. 

While teachers in England are expected to support children to apply their knowledge of the alphabetic code across the school day, including in both reading and writing, there is little guidance on how they actually make phoneme–grapheme mapping decisions in practice, particularly beyond the correspondences explicitly taught in phonics programmes. What happens when the sounds used by individuals (their Personal Pronunication Code) don't align with the Phonics Pronunication Code?
If Adele's son says /f r i:/ for 3, what does the teacher tell him as the phonics programme doesn't teach <th> for that sound.   

The reform

Bidirectional word mapping means:
 

• recognising that reading and spelling are one connected process

• ensuring children understand that print represents speech

• supporting mapping in both directions from the start

Children need print to speech for decoding. They also need speech to print for encoding. But if the goal is secure orthographic learning, speech to print is the more demanding test and the clearer indication that the word has actually been mapped.

This is what is currently missing for too many children, and another reason why Phonics Reform England was launched.