Parent question

The articulation hierarchy: how speech therapy builds a sound step by step

Published 21 October 2026 · Updated 21 October 2026

The short answer

Many speech therapists teach a new sound in steps: first the sound on its own, then in syllables, words, phrases, sentences, and finally everyday conversation. A child moves up when they say the sound correctly most of the time at the current step. Research programs often use a target of around 80 to 85%, but there’s no single standard. The goal is for the sound to become automatic, so your child uses it without thinking.

The steps

This step-by-step approach is usually credited to Van Riper and is often called traditional articulation therapy (Herreras Mercado et al., 2019). In a US survey, more speech-language pathologists (SLPs) said they used it with 3 to 6-year-olds than any other approach (Brumbaugh & Smit, 2013).

The levels of the articulation hierarchy, with S as an example
Level	Example for S	What it builds
The sound alone	“sss”	Finding the right tongue position.
Syllables	“see”, “so”, “oss”	Joining the sound to vowels.
Words	“sun”, “messy”, “bus”	The sound at the start, middle and end of words.
Phrases	“my bus”, “sunny day”	Keeping it right with other words around it.
Sentences	“The sun is hot.”	Longer, faster speech.
Conversation	Talking about their day	Using it without thinking.

ASHA describes this as carrying the sound over to harder and harder levels, from syllables and words to phrases, sentences and conversation.

Where therapy starts

What your child can already copy. SLPs check whether your child can copy the sound when they hear a model, called stimulability. Sounds a child can already copy tend to improve more easily (Powell, Elbert & Dinnsen, 1991).
Easier spots. Some sounds are easier next to certain vowels or in certain places in a word, so an SLP may start wherever your child already gets it right (ASHA).

When a child moves up a level

Traditional programs move up once a child reaches a set accuracy in a session, such as 85% at the syllable level (Preston, Leece & Storto, 2019). One structured method moves up after 5 correct out of 6 tries, and one study lowered it to 3 out of 4 so practice moved up faster (Benway & Preston, 2024). So research programs often use around 80 to 85%, but there’s no single standard and SLPs adjust it for each child.

Getting it right in the session isn’t the same as learning it. Some children are accurate during practice but don’t keep it afterwards, which is why SLPs check again later, not just at the end of a good session (Preston, Leece & Storto, 2019).

It isn’t always a strict ladder

Mixing levels. In a small study of 4 children with S errors, mixing words, sentences and storytelling in the same session worked without moving strictly from easy to hard (Skelton, 2004).
Mixing practice. Motor learning research suggests repeating the same word many times helps a child get it right in the moment, while mixing up words helps them keep it. Less frequent feedback may also help long-term learning (Preston, Leece & Storto, 2019; Maas et al., 2008).

That’s why your child’s SLP may not follow the table above exactly.

The goal: automatic speech

The last stage is making the sound automatic and teaching your child to notice and fix their own errors (ASHA). Once sentences are mostly right, therapy can shift to conversation and self-monitoring (Preston, Leece & Storto, 2019). This is the hardest step for many children, and practice outside therapy matters. See articulation carryover.

How to help at home

Ask your child’s SLP which level your child is working on, and practise at that level, not above it.
Short, frequent practice adds up. See practice at home and speech practice games.
Keep a simple practice log so the SLP can see how home practice is going.

Questions parents ask

What are the levels of the articulation hierarchy?

The sound on its own, then syllables, words, phrases, sentences and conversation. Some programs also include storytelling.

What percentage does my child need to move up a level?

There’s no single standard. Research programs often use around 80 to 85% correct at one level before moving up, and SLPs adjust it for each child.

Why can my child say the sound in words but not in conversation?

Conversation is usually the last and hardest level. Making the sound automatic takes lots of practice. See articulation carryover.