AUSTRALIAN EXPECTATIONS FOR WORD MAPPING

A national Teacher Test of Phonological-Orthographic Mapping (POM) requires a nationally agreed standard for calculating the English alphabetic code.
 

Before teachers can be assessed against that standard, we need to establish exactly what Australian teachers are expected to know, the research and theoretical basis for those expectations, and how consistently the underlying phoneme–grapheme principles are reflected in curriculum guidance and the phonics programmes teachers are using.

The issue is not whether all phonics programmes teach the same grapheme–phoneme correspondences (GPCs), use the same sequence or adopt the same instructional approach. They do not need to.

The issue is whether teachers are given a consistent way to map the phonemes and graphemes in any word, regardless of the programme they use.

Australian teachers are expected to teach children to connect the phonemes in spoken words with the graphemes that represent them in written words, for both reading and spelling.

The Australian Curriculum: English states:

“Students develop knowledge about the sounds of English (phonemes) and learn to identify the sounds in spoken words.”

It goes on to state that students:

“apply their knowledge of phonemes and graphemes to decode words.”

The National Literacy Learning Progression defines the relationship explicitly:

“A phoneme is a spoken sound and a grapheme is the letter or group of letters that represent each phoneme.”

This establishes an expectation that phonemes and graphemes are connected at word level. For reading, children use the graphemes to identify the phonemes and decode the word. For spelling, they need to identify the phonemes and select the graphemes that represent them.

In other words, teachers need to understand bidirectional word mapping, not simply the list of GPCs explicitly taught within a phonics programme.

THE RESEARCH BASIS FOR AUSTRALIAN PHONICS AND WORD RECOGNITION

These Australian expectations have a published evidence and theoretical basis.

ACARA states that the National Literacy Learning Progression was developed from:

“peer-reviewed, internationally recognised research”

across fields including cognitive psychology, linguistics and sociolinguistics.

ACARA identifies Paris's (2005) distinction between constrained and unconstrained skills as an important theoretical underpinning of the progression. Phonological awareness and phonic knowledge are identified as foundational constrained skills that are explicitly taught.

ACARA also identifies Freebody and Luke's Four Resources Model as informing the progression, including the role of the reader as code breaker.

The current Australian Curriculum continues to identify Phonological awareness and Phonic knowledge and word recognition as constrained skills and states that the number of progression levels is determined by the available research and evidence.

At the word level, the National Literacy Learning Progression describes children becoming increasingly proficient in using letter–sound relationships as code-breaking skills. Children are expected to apply phoneme–grapheme knowledge to increasingly complex and unfamiliar words.

This establishes the Australian curriculum expectation. It does not, however, answer the question at the centre of this work:

What is the consistent phonological-orthographic analysis teachers are expected to apply when mapping those words?

That question becomes particularly important because the explicitly taught code cannot include every phoneme–grapheme correspondence children will encounter.

EXPLICIT PHONICS KICK-STARTS SELF-TEACHING

A phonics programme cannot explicitly teach every phoneme–grapheme correspondence children will encounter in English.

Nor does it need to.

Share's self-teaching hypothesis proposes that phonological recoding provides a mechanism through which children independently acquire word-specific orthographic knowledge. Successful decoding of unfamiliar written words therefore provides opportunities for further orthographic learning (Share, 1995).

Experimental research has provided further evidence for orthographic learning through self-teaching (Cunningham et al., 2002).

Ehri's research on orthographic mapping provides a complementary account of how phonemic awareness and grapheme–phoneme knowledge enable connections to be formed between spellings and pronunciations so that written words become stored in memory (Ehri, 2005; 2014).

Explicit phonics therefore provides children with the alphabetic knowledge needed to kick-start this process. The explicitly taught code is not the whole English alphabetic code.

For example, a programme might explicitly teach:

‹sh› → /ʃ/ in ship

and:

‹ch› → /ʃ/ in machine

but perhaps not:

‹s› → /ʃ/ in sugar

or:

‹ti› → /ʃ/ in fraction.

Children do not need a separate explicit phonics lesson for every possible grapheme representing /ʃ/ before encountering words containing those correspondences. They can use their existing alphabetic knowledge to decode unfamiliar words and, through successful encounters, extend their orthographic knowledge (Share, 1995; Cunningham et al., 2002).

Teach them what they need. Show them the rest.

But children differ in how successfully they make this transition to independent word learning. When a child cannot yet figure out an unfamiliar correspondence, the adult needs to be able to show the mapping.

This creates an important distinction between what children need to be explicitly taught and what teachers need to understand.

Teachers need sufficient knowledge of phonological-orthographic mapping to support words containing correspondences outside the explicitly taught code.

To show the rest, the adult needs to understand how the rest maps.

WORD-MAPPING TECHNOLOGY ALSO REQUIRES AN AGREED STANDARD

Technology can give teachers immediate access to the bidirectional mapping of words at the point of need.

We currently provide teachers with word-mapping technology and training using Ortho-GraphiX®, including mappings of words containing correspondences outside those explicitly taught within a phonics programme.

Technology does not replace teacher knowledge. Teachers need to understand the underlying phonemes, graphemes and mapping principles and be able to apply them independently.

But clear guidance matters just as much to us as developers of word-mapping technology.

If we provide an Australian teacher with the mapping of a word, we need to know that the analysis displayed is the analysis that teacher is expected to use.

We cannot simply decide what constitutes the Australian mapping.

The same applies to training and assessment. We cannot train teachers to apply a particular analysis, assess them against it in the Teacher Test, or programme Ortho-GraphiX® to display it as the Australian mapping unless there is a sufficiently clear basis for determining that it is correct.

This becomes particularly important where pronunciation differs. Part 2 of the Teacher Test therefore requires teachers to select the reference variety being assessed, for example Australian English or British English, and map the supplied pronunciation.

Our technology needs the same clarity.

For Australian English, we need to know which reference pronunciation and phoneme inventory should be used, how grapheme boundaries should be determined, how all letters within the word should be accounted for, and what underlying mapping principles Australian teachers are expected to apply.

The issue is therefore not simply whether teachers know the code.

Teachers, training, assessment and technology all need to be working from the same agreed standard.

The formula must come first.

PROGRAMMES CAN VARY. THE UNDERLYING WORD-MAPPING FORMULA SHOULD NOT.

Phonics programmes can teach different GPCs, use different sequences and adopt different instructional approaches.

That variation does not require different underlying phonological-orthographic analyses.

If two teachers using different programmes are asked to map the same pronunciation of the same word, there should be an agreed basis for identifying the phonemes and the graphemes representing them.

At present, this cannot necessarily be assumed.

Differences can occur between programmes. There can also be inconsistencies within individual programmes.

A programme may teach children to segment spoken words into individual phonemes and connect those phonemes with graphemes in one context, while elsewhere using instructional categories such as consonant blends or silent letters without making clear how those categories should be reconciled with phoneme-by-phoneme and grapheme-by-grapheme word mapping.

This may not become obvious while children are working with the words explicitly provided and mapped within the programme. It becomes much more important when a child encounters a word outside that content and the teacher has to apply the underlying principles independently.

What formula are they expected to use?

CONSONANT BLENDS AND CLUSTERS

Australian guidance makes clear that consonant blends and clusters contain separate phonemes.

The Australian Curriculum requires Year 1 students to:

“Segment words into separate phonemes (sounds) including consonant blends or clusters at the beginnings and ends of words.”

NSW Department of Education phonological-awareness guidance provides a particularly useful example. It segments frog into the individual phonemes:

/f/ /r/ /ɒ/ /ɡ/

However, Australian curriculum materials also use consonant blend as an instructional description for written letter sequences.

The National Literacy Learning Progression, for example, refers to students writing:

“two-letter consonant blends in words correctly (sl in slip)”

and curriculum materials have used written examples such as ‹bl› and ‹st›.

This creates an important word-mapping question.

If frog contains four individual phonemes:

/f/ /r/ /ɒ/ /ɡ/

and a grapheme is the letter or group of letters representing a phoneme, should the corresponding graphemes be:

‹f› ‹r› ‹o› ‹g›?

If so, ‹fr› is not one grapheme within that mapping. It contains ‹f› and ‹r›, corresponding with two separate phonemes.

The term consonant blend may have an instructional purpose, but it does not itself establish the grapheme boundaries required for phonological-orthographic mapping.

The phonological guidance tells teachers that the sounds are separate.

The corresponding grapheme analysis needs to be equally explicit.

WHAT ABOUT ‘SILENT LETTERS’?

A similar issue arises with so-called silent letters.

The National Literacy Learning Progression states:

“A phoneme is a spoken sound and a grapheme is the letter or group of letters that represent each phoneme.”

Yet it also refers to:

“words with silent letters in digraphs (kn, mb)”.

What mapping is intended?

In knee, is it:

‹kn› → /n/

or:

‹k› = silent + ‹n› → /n/?

In thumb, is it:

‹mb› → /m/

or:

‹m› → /m/ + ‹b› = silent?

These are not simply different descriptions of the spelling. They produce different phonological-orthographic mappings.

NSW guidance raises a related question.

Its glossary defines a grapheme as:

“The smallest unit of writing used to represent one phoneme. A letter or combination of letters corresponding to or representing a single phoneme.”

Yet it defines a consonant digraph as:

“Two consonant graphemes used to represent one phoneme”

and includes ‹kn› in knee as an example.

If a grapheme is the smallest written unit representing one phoneme, is ‹kn› one grapheme consisting of two letters, or two graphemes representing one phoneme?

This may appear to be a small terminological distinction.

It is not small if teachers are being asked to identify graphemes when mapping words.

It also matters when developing technology expected to show those grapheme boundaries to teachers.

Before a Teacher Test can mark one analysis correct and another incorrect, or word-mapping technology can present one analysis as the appropriate Australian mapping, the standard itself needs to be clear.

WHOSE PHONEMES ARE BEING MAPPED?

There is a further fundamental question:

Whose phonemes are being mapped?

Pronunciation varies.

Before a teacher can connect phonemes with graphemes, they first need to identify the phonemes present in the pronunciation being mapped.

Different pronunciations of the same written word may therefore legitimately result in different phonological analyses. This needs to be distinguished from incorrectly identifying the phonemes within a specified pronunciation.

Programmes therefore need to make clear whether the mapping reflects the child's pronunciation, the teacher's pronunciation, or a specified reference pronunciation.

There is already Australian precedent for making this explicit.

NSW Department of Education GPC resources use International Phonetic Alphabet (IPA) symbols to guide the pronunciation of phonemes in Standard Australian English.

IPA provides a way to specify the intended phoneme independently of the orthography being analysed.

This principle is already built into the Teacher Test of Phonological-Orthographic Mapping.

In Part 1, teachers say the word in their own accent, identify the phonemes they use and map those phonemes to the graphemes in the written word.

In Part 2, teachers select a specified reference pronunciation, for example Australian English or British English, hear that pronunciation, identify its phonemes and map those phonemes to the graphemes.

This allows pronunciation variation to be separated from knowledge of the mapping process.

The same clarity is necessary for word-mapping technology.

If Ortho-GraphiX® is providing a mapping for Australian teachers, we need to know which reference pronunciation, phoneme inventory and mapping principles the technology is expected to represent.

The purpose is not to declare one accent correct.

It is to distinguish legitimate differences in pronunciation from differences in the analysis of the phonemes and graphemes representing a specified pronunciation.

WE NEED TO EXAMINE THE PROGRAMMES TEACHERS ARE USING

The next stage is therefore to examine phonics programmes being used in Australian schools.

For each programme, we need to establish how it defines phonemes and graphemes; how it maps words for reading and spelling; whether the same underlying analysis is used bidirectionally; and which GPCs are explicitly taught.

We also need to examine what happens beyond that explicitly taught set. How are consonant blends and clusters treated? How are so-called silent letters treated? Are spellings such as ‹kn›, ‹gn›, ‹wr› and ‹mb› treated as graphemes representing phonemes, or are individual letters treated as silent? Are these principles applied consistently throughout the programme?

And we need to establish whose pronunciation is represented, whether phonemes are explicitly identified using IPA or another unambiguous reference, and what guidance is provided when a child's pronunciation differs from the programme's reference pronunciation.

The purpose is not to require programmes to teach identical content.

Programmes can vary.

What needs to remain consistent is the underlying formula used to determine the phonological-orthographic mapping of the word.

TOWARDS A NATIONAL STANDARD

Before teachers can be assessed nationally, there needs to be an agreed basis for identifying the phonemes in the pronunciation being mapped, identifying the graphemes representing those phonemes, determining grapheme boundaries, accounting for the letters within the written word, mapping bidirectionally for reading and spelling, and dealing transparently with differences in pronunciation.

This matters to everyone expected to apply that standard.

Teachers need it to map unfamiliar words consistently.

Training providers need it to know what they should teach.

The Teacher Test needs it to determine what constitutes an accurate mapping.

And as developers of word-mapping technology, we need it to ensure that the mappings we provide to Australian teachers are actually right for the context in which those teachers are working.

Technology can apply an agreed formula consistently and at scale. It should not determine what that formula is.

The formula must first be agreed.

Once that standard is established, programmes can continue to vary in how much of the code they explicitly teach and how they teach it, while teachers, training, assessment and technology share a common understanding of how the English alphabetic code is mapped.

Children who need more support can then be shown the code in the words they actually encounter, rather than having to wait for every correspondence to be explicitly taught.

Programmes can vary. The underlying word-mapping formula should not.

Teach them what they need. Show them the rest.

REFERENCES

Australian Curriculum, Assessment and Reporting Authority (ACARA). Australian Curriculum: English, Version 9.0.

Australian Curriculum, Assessment and Reporting Authority (ACARA). Literacy General Capability, Version 9.0.

Australian Curriculum, Assessment and Reporting Authority (ACARA). National Literacy Learning Progression.

Australian Curriculum, Assessment and Reporting Authority (ACARA). National Literacy Learning Progression: Appendix 1 – The evidence base for development.

Cunningham, A. E., Perry, K. E., Stanovich, K. E., & Share, D. L. (2002). Orthographic learning during reading: Examining the role of self-teaching. Journal of Experimental Child Psychology, 82(3), 185–199.

Ehri, L. C. (2005). Learning to read words: Theory, findings, and issues. Scientific Studies of Reading, 9(2), 167–188.

Ehri, L. C. (2014). Orthographic mapping in the acquisition of sight word reading, spelling memory, and vocabulary learning. Scientific Studies of Reading, 18(1), 5–21.

Freebody, P., & Luke, A. (1990). “Literacies” programs: Debates and demands in cultural context. Prospect: Australian Journal of TESOL, 5(3), 7–16.

Paris, S. G. (2005). Reinterpreting the development of reading skills. Reading Research Quarterly, 40(2), 184–202.

Share, D. L. (1995). Phonological recoding and self-teaching: Sine qua non of reading acquisition. Cognition, 55(2), 151–218.