As Speech Pathologists, Have We Forgotten How to Teach Sounds?
An opinion piece
Over the years, I have heard several ideas repeated in professional development courses, university programmes and clinical discussions:
Van Riper's articulation therapy is outdated.
If an entire class of sounds is affected, the child needs phonological therapy.
Articulation therapy is primarily for distortions rather than teaching new sounds.
If a child can produce a sound in isolation, the sound has been established and we can move on.
I have spent a lot of time thinking about these statements.
Not because I am opposed to phonological approaches. Minimal pairs, contrastive approaches and other phonological techniques can be extremely useful when carefully matched to a child.
But when I look at what I actually do when a child cannot produce a sound, or can produce it only in limited contexts, I keep coming back to articulation-based teaching and the articulatory hierarchy.
Whether we call it sound elicitation, phonetic placement, shaping, successive approximation or speech motor learning, we are helping a child learn, refine and stabilise a speech movement.
And when that movement does not yet exist, teaching it is not optional.
Stimulable does not mean established
One of the most important distinctions I make in therapy is the difference between a sound being possible and a sound being established.
When a child produces a sound once, I do not automatically conclude that they have learnt that sound.
Instead, I ask:
How established is it?
Can they use it:
in different word positions?
with different vowels?
in different syllable shapes?
in longer words?
in phrases or sentences? (if these are accessible to them)
without immediate imitation?
with reasonable consistency?
A child may be stimulable for a sound without having acquired it.
For example, a child may produce /k/ accurately in isolation or only in word-final position, yet continue to front it everywhere else.
It is easy to say:
"The child can say /k/, therefore the remaining difficulty must be phonological."
But I am not convinced that conclusion automatically follows.
Producing duck is not the same motor task as producing key. Producing key is not the same as producing cookie. Producing cookie is not the same as maintaining /k/ accurately within a sentence.
The movement may exist in one context without being sufficiently stable or flexible across others.
My first question is therefore:
Have we actually finished teaching the sound?
Three clinical situations
I find it helpful to think about speech sounds in three broad categories.
1. The sound is absent
The child cannot yet produce the sound, even with support.
My job is to help them access the movement.
2. The sound is emerging
The child can produce the sound in isolation, under imitation or in a limited context.
My job is to strengthen it and expand it into other contexts.
3. The sound is established
The child can use the sound across contexts but is not consistently generalising it.
At this point I may explore broader issues such as self-monitoring, phonological representation or contrastive use.
The problem is that clinicians sometimes jump directly from Situation 1 to Situation 3.
A child produces /k/ once in isolation and suddenly we behave as though the production problem has been solved.
It hasn't.
Stimulability is valuable information, but stimulability is not acquisition.
What does this mean for CAS?
The issue becomes particularly important when working with children with Childhood Apraxia of Speech.
In severe CAS, therapy often involves two simultaneous goals:
Sounds we are building | Syllable shapes we are building |
/p/, /b/, /m/ | CV |
/t/, /d/ | VC |
Early consonants | CVC |
Later fricatives and affricates | CVCV |
Vowels | More complex shapes |
Most of my CAS goals sit within one of these two columns.
More accurately, they usually sit in both.
CAS therapy is not sounds versus syllable shapes. We build both side by side.
Once I establish a new sound, I do not keep a child practising it in isolation for weeks.
I quickly move into achievable syllable shapes and functional words.
But I have not abandoned the articulatory hierarchy.
I am still teaching the sound while simultaneously teaching the child how to combine that sound into increasingly successful speech sequences.
Sometimes I become more traditional, not less
One of the misconceptions about articulation therapy is that it applies only to early sound elicitation.
My experience has often been the opposite.
As a child's speech becomes longer and more complex, I sometimes become more traditional, not less.
A child may use /t/ accurately in simple words but lose it in longer words. Another may produce /k/ only in one word position. Another may distort a sound whenever linguistic complexity increases.
In these situations, I often return to systematic sound teaching.
I want to know:
Can they produce the sound initially?
Medially?
Finally?
Across different vowels?
In different word shapes?
In phrases and sentences they can otherwise manage?
If the answer is no, then the sound may still be emerging.
The child may not need an entirely different therapy approach.
Sometimes they simply need more teaching.
This is why I become cautious when somebody says:
"They can already produce the sound, so the remaining difficulty must be phonological."
My next question is always:
Can they produce it everywhere?
What if an entire class of sounds is missing?
Another common statement is:
"If an entire class of sounds is affected, the child requires phonological therapy."
I continue to question that assumption.
Consider a child with no established fricatives. Why do we automatically assume that this pattern requires a phonological explanation?
Young children typically acquire simpler speech sounds before more complex ones. Early developing sounds such as /m/, /n/, /p/, /b/, /t/, /d/ and /w/ are often acquired before fricatives such as /f/, /s/, /ʃ/, /v/ and /z/.
Fricatives place different demands on the speech system. They require the child to create and maintain a precise constriction while controlling continuous airflow.
If a child has not yet learnt that skill, is it really surprising that multiple fricatives may be absent?
For me, the absence of an entire class of sounds does not automatically indicate that the child needs a different therapy approach. It may simply indicate that the child has not yet developed the articulatory skill shared by those sounds.
My first question is not:
"Which phonological approach should I use?"
My first question is:
"How am I going to teach this child fricatives?"
One criticism sometimes levelled at articulation therapy is that it is inefficient because sounds are taught one at a time. That has not been my experience.
When I establish one fricative, I am often teaching a skill that supports the acquisition of others.
The child is learning how to control airflow. They are learning how to create a constriction. They are learning new tongue and lip placements. They are learning how different speech movements feel.
Those skills frequently transfer.
Once a child learns to maintain airflow for one fricative, that experience can help support the acquisition of another. Once a voiceless sound is established, its voiced cognate may require far less teaching because much of the articulatory groundwork is already present.
In that sense, articulation therapy is not simply teaching isolated sounds. It is gradually building a broader speech system.
Learning one sound often prepares the way for the next.
What about phoneme collapse?
One area where my thinking has evolved considerably is the concept of phoneme collapse.
The longer I do this work, the more concerned I become that phoneme collapse is often interpreted as evidence of a phonological disorder without sufficient consideration of CAS.
Children with CAS commonly present with a severely reduced phonemic repertoire and a heavy reliance on a small number of preferred sounds. In fact, a restricted phonemic inventory forms part of the well-known Mayo Clinic characteristics of CAS.
As a result, a child may use one or two familiar sounds in place of many different targets. Large portions of the speech sound system appear to collapse into a very small number of productions.
In phonological therapy literature, this pattern is often labelled phoneme collapse.
The longer I work with children with CAS, however, the more I question whether many examples of so-called phoneme collapse are actually reflecting an underlying motor speech difficulty rather than a primary phonological difficulty.
After all, most children acquire a reasonably broad range of early-developing sounds such as /m/, /n/, /p/, /b/, /t/, /d/ and /w/ relatively early in development. If a child is reducing much of their speech sound system to one or two preferred sounds, my first question is not:
"Which phonological approach should I use?"
My first question is:
"Why has this child not acquired those early speech movements?"
For me, that is a motor speech question.
In fact, recent work by Freese, Masso, Fricke and Fox-Boyer, presented at the Apraxia Kids Conference, identified the absence of early-developing sounds as an important marker in the differential diagnosis of CAS.
So when I see a child with a severely restricted phonemic repertoire who is relying heavily on one or two preferred sounds, I become concerned about an underlying motor speech difficulty.
For that reason, phoneme collapse does not make me less suspicious of CAS.
If anything, it makes me more suspicious.
Particularly when it occurs alongside other red flags such as:
severely reduced intelligibility compared with age peers;
receptive language that is substantially stronger than expressive language;
early sounds missing from the inventory;
unusual or atypical error patterns;
a markedly reduced phonemic repertoire;
heavy reliance on a small number of preferred sounds.
I am not suggesting that every child described as having phoneme collapse has CAS.
However, I am increasingly concerned that clinicians sometimes stop at the phoneme-collapse label and move directly towards a phonological explanation without adequately investigating whether an underlying motor speech disorder may be contributing to the pattern.
My concern is not simply the terminology.
My concern is that once the pattern is labelled phoneme collapse and assumed to be phonological in nature, the child may never receive a sufficiently thorough motor speech assessment.
This places children with CAS at risk of delayed identification, delayed diagnosis and delayed access to intervention that more directly addresses the underlying difficulty.
For me, phoneme collapse is not a reason to stop considering CAS.
It is one of the reasons I start looking more carefully for it.
CAS Is Not a Three-Tick Checklist
Another reason I sometimes see children moved towards a severe phonological disorder label is that they do not appear to demonstrate all three of the commonly cited ASHA characteristics of CAS at the time of assessment.
In practice, however, children do not always present with every characteristic equally clearly at every stage of development.
A child who has received years of therapy may become more consistent over time, even while continuing to rely on inaccurate or restricted speech movements.
Another child may be so severely affected that they have not yet developed the connected speech needed for prosodic difficulties to be easily observed.
The absence of one characteristic does not automatically exclude CAS.
For me, a severely unintelligible child with a markedly restricted phonemic repertoire and missing early-developing sounds deserves careful consideration of CAS, regardless of whether every item on a checklist is immediately obvious.
Where do minimal pairs fit?
I may occasionally use minimal pairs within an articulation-based session to highlight the contrast between the child’s production and the target. However, I do not see the use of contrasting words as evidence that the child now requires a separate phonological treatment approach.
Even in Van Riper's traditional articulation therapy, discrimination and comparison of correct versus incorrect productions were central components of treatment. The use of contrast is therefore not unique to modern phonological approaches, although contemporary minimal pair therapy developed these ideas in a different way.
For me, the contrast does not replace articulation therapy.
The child still needs to be able to produce the target sound. If that movement is absent, unstable or restricted to one context, presenting contrasting words does not remove the need to teach it.
This is why I do not generally feel the need to replace articulation therapy with approaches such as maximal oppositions, cycles or core vocabulary therapy when my treatment goal is to establish and generalise sounds.
I teach the sounds.
Where contrast helps the child recognise meaning differences or monitor production, I may use it. But contrastive practice sits within the sound-teaching process. It is not a substitute for that process.
A pattern tells us what, not always why
I continue to question how confidently we infer an underlying linguistic or phonological deficit from a speech-production pattern alone.
Speech sound patterns are descriptions.
They tell us what the child is doing.
They do not necessarily tell us why the child is doing it.
When a child is missing several sounds, we can describe that pattern.
When multiple sounds are substituted by the same sound, we can describe that pattern.
When an entire class of sounds is absent, we can describe that pattern. But,
the pattern itself does not establish the underlying mechanism.
For me, the crucial question is whether the child is actually demonstrating a difficulty understanding or representing the contrast, or whether they simply cannot yet produce the sounds required to express that contrast.
When I look at many children on my CAS caseload, I see children who can readily demonstrate auditory discrimination of target sounds and words. They recognise when productions are different. They can often identify the correct word when asked to point to it. Their primary difficulty is producing those contrasts accurately, consistently and across contexts.
If a child can hear the difference between cap and tap, recognise the target when it is presented and understand the distinction being communicated, I become cautious about assuming that the primary problem lies within the child's phonological system.
The difficulty may instead lie in the ability to access, coordinate and stabilise the speech movements needed to produce the contrast.
Importantly, I am not concerned about the child who says wabbit for rabbit or tun for sun while remaining otherwise highly intelligible. Many children with mild or isolated speech sound errors make excellent progress over time, and some will master those sounds with maturation alone.
The children I worry about are different.
I worry about the child who remains substantially unintelligible beyond the age at which most peers can be understood. The child with a markedly restricted phonemic repertoire. The child who is missing early-developing sounds. The child whose speech system appears stalled despite ongoing development in other areas.
My concern is that these children are sometimes described simply as having a severe phonological disorder, and that the label itself can unintentionally narrow the clinical investigation.
Once the pattern is assumed to be phonological in origin, we may stop asking important motor speech questions.
Why has this child not acquired early speech movements that most children master relatively easily?
Why is their phonemic repertoire so restricted?
Why does intelligibility remain so profoundly affected?
Why are new sounds so difficult to establish and stabilise?
For me, these are not simply phonological questions. They are often the very questions that lead me to consider whether an underlying motor speech disorder, including CAS, may be contributing to the child's presentation.
The risk is not merely one of terminology. The risk is that children with significant motor speech disorders may be overlooked because the observable pattern has already been explained by a phonological label.
So, is traditional articulation therapy outdated?
No.
I do not use the articulatory hierarchy rigidly or identically with every child.
With a young child with severe CAS, I may establish a rough approximation and quickly place it within a simple functional syllable shape.
With a child in the middle stages of CAS therapy, I may systematically teach a consonant or vowel across word positions, vowel contexts, phrases and sentences.
With an older child who has a residual speech sound error, I may follow the hierarchy almost by the book.
The hierarchy has not disappeared from my therapy.
It has become flexible.
When a child is struggling with a sound, I ask:
Is the sound absent, emerging or sufficiently established?
Is it available only under direct imitation?
Is it restricted to one word position or vowel context?
Have the other contexts actually been taught?
Is there a facilitating phonetic context?
Can an existing sound be shaped towards the target?
Can I introduce it within an achievable syllable shape?
Am I expecting generalisation from a movement that is not yet stable?
Am I calling the problem phonological before fully teaching the sound?
Those questions guide my intervention more directly than a pattern label alone.
Why I still use the articulatory hierarchy to teach consonants and vowels
Articulation-based principles are present whenever I elicit a first approximation, shape one sound towards another, establish a vowel, teach airflow, introduce voicing, extend a sound across word positions, practise clusters or carry a residual sound into conversation.
The therapy may look different at each stage. The level of accuracy and complexity may change.
But the child is still learning, refining and stabilising speech movements.
For me, the clinical reality is clear:
A sound is not established simply because a child can produce it once.
A child cannot use a sound consistently until they can access it across the contexts in which it is needed.
Perhaps, in our enthusiasm for classifying patterns and matching them to named programmes, we have overlooked something fundamental. Children still need us to teach sounds.
That is why I continue to use the articulatory hierarchy. Not because it is old or traditional, but because, when thoughtfully adapted, it remains one of the clearest and most useful roadmaps for helping children establish sounds and carry them into everyday speech.
Olga xo
Want to learn more about how I make these clinical decisions?
Inside the OKAT CAS Academy, I teach speech pathologists how I assess children with suspected CAS, select meaningful targets, adapt cueing and help children progress from early approximations towards increasingly complex, functional and intelligible speech.
The next OKAT CAS Academy cohort is planned for early 2027.
Join the OKAT CAS Academy waitlist to be the first to hear when enrolments open.
References
To, C. K. S., McLeod, S., Sam, K. L., & Law, T. (2022). Predicting which children will normalize without intervention for speech sound disorders. Journal of Speech, Language, and Hearing Research, 65(5), 1724-1741. https://doi.org/10.1044/2022_JSLHR-21-00444
Freese, W., Masso, S., Fricke, S., & Fox-Boyer, A. (submitted). Tools for differential diagnosis of speech sound disorders in German-speaking children. Findings presented at Apraxia Kids Conference.
Van Riper, C. (1978). Speech correction: Principles and methods (6th ed.). Englewood Cliffs, NJ: Prentice-Hall.
Comments