Speech sound development chart: when children learn each English sound

Most English consonants are produced correctly by 90 percent of children by age 5, and the last few (r, zh and the two th sounds) are usually in place between 5 and a half and 6 and a half. Early sounds such as b, p, m, n, h, w and d are typically accurate by around age 3. The chart below lists all 24 English consonants with the age at which 90 percent of children in a large review of studies were producing each one correctly.

The chart: 24 English consonants by age

The ages come from Crowe and McLeod (2020), who pooled published studies of English-speaking children in the United States and calculated the mean age at which 90 percent of children produced each consonant accurately. Ages are given in months, with the years and months equivalent beside them. “4;6” means four years and six months.

Crowe and McLeod grouped the consonants into early sounds (acquired between 2;0 and 3;11), middle sounds (4;0 to 4;11) and late sounds (5;0 to 6;11). The grouping is an easy way to remember the chart. Lip, nose and simple tongue-tip sounds come first. The longer “hissing” sounds, l, and the ch and j sounds arrive in the preschool years. The r and th sounds come last.

What the 50, 75 and 90 percent criteria mean

Speech sound studies report the age at which a certain share of children produce a sound correctly. Three cut-offs are common.

• 50 percent. Half the children in the sample say the sound correctly. This is sometimes called the age of customary production. It tells you a sound is emerging in the population, not that every child should have it.
• 75 percent. Three in four children have the sound. Many clinicians treat this as a sensible point to start paying attention.
• 90 percent. Nine in ten children have the sound. This is the conservative mastery benchmark used in the chart above.

Crowe and McLeod (2020) reported all three criteria, and the 50 percent ages are earlier than the 90 percent ages for every sound. If you find a chart online with much younger ages than the one above, it is probably showing a 50 or 75 percent criterion, or using older data.

Because the 90 percent figure is the latest of the three, a child who has not reached it is in the last 10 percent of their age group for that sound. That is not the same as having a speech sound disorder, but it is a reasonable point to ask a registered speech-language pathologist. Reading the chart as a set of deadlines is the most common mistake parents make with it. Reading it as a guide to what is usual, and what is unusual, is more helpful.

What “mastered” means: all word positions

Speech sound norms usually count a sound as acquired when the child says it correctly at the beginning, middle and end of words. A child may say “sun” clearly but leave the s off “bus”, or say “cat” but say “tootie” for “cookie”. This is a normal part of the learning stage, and it is one reason a quick listen at home can give a different picture from a formal assessment, where an SLP samples each sound in every position.

Position matters for another reason. Consonant clusters, where two or more consonants sit together as in “spoon”, “tree” or “milk”, usually come after the single sounds. A 3-year-old who says “poon” for “spoon” may already say s on its own perfectly well. These simplification patterns are covered in our guide to phonological processes.

Vowels

Vowels are learned earlier than consonants. In the cross-linguistic review by McLeod and Crowe (2018), vowels were generally accurate by around age 3. Persistent vowel errors after 3 are unusual in English. They are worth mentioning to an SLP, because the ASHA Practice Portal lists vowel errors among the features clinicians consider when they look at motor speech difficulties such as childhood apraxia of speech. On their own, vowel errors do not mean a child has apraxia. They are simply one of several things a clinician will listen for.

How understandable should my child be?

Alongside individual sounds, SLPs look at intelligibility, which is how much of a child’s speech a listener can understand. A widely used rule of thumb is that unfamiliar adults understand about 50 percent of a 2-year-old’s speech, about 75 percent at 3, and nearly all of it by 4. Parents usually understand more than strangers do at every age.

A related measure is percentage of consonants correct (PCC). In the McLeod and Crowe (2018) review, average PCC was roughly 64 percent at age 2, 86 percent at 3, 92 percent at 4 and 94 percent at 5. If you want a rough sense of your own child’s number, our PCC calculator walks through it, and the intelligibility estimator gives a quick age comparison. Both are general information tools, not screening tests.

What the chart does not tell you

A chart is a group average. It cannot say whether a particular child has a speech sound disorder, and it does not account for several things an assessment would.

• Error type. Substituting one sound for another (t for k) is a different pattern from distorting a sound (a slushy s). Distortions are less tied to age, as our guide to lisps explains.
• Consistency. Saying the same word three different ways in a row can matter more than which sounds are missing.
• Hearing. Repeated ear infections and fluid behind the eardrum can muffle the very sounds a child is trying to learn.
• Other languages. Children learning two languages may acquire some sounds on a different timeline, and some sounds do not exist in both languages. The chart is based on English-speaking children in the United States. See our guide on bilingual children and speech delay.
• Frustration and participation. A child who avoids talking because they are not understood deserves support regardless of where they sit on the chart.

When to ask for an assessment

The chart suggests some reasonable prompts. Consider contacting a registered speech-language pathologist if:

• your child is 3 and people outside the family understand less than about three quarters of what they say,
• your child is 4 and is still missing several early or middle sounds, or is often not understood by strangers,
• your child is 5 or older and sounds like r, s, l or th are still absent or distorted,
• any sound is produced with air escaping over the sides of the tongue (a lateral lisp), at any age,
• you have any concern about hearing, or
• your child is frustrated, teased or avoiding talking.

In Ontario, parents of children who have not yet started school can self-refer to the free Preschool Speech and Language Program. No doctor’s referral or diagnosis is needed. Our guides on when to see a speech-language pathologist and how to get speech therapy in Ontario cover the options in more detail. Earlier support is generally easier, and an assessment gives clarity either way.

What you can do at home

• Use the free speech sound checker to compare the sounds your child uses with the chart above. It is general information, not a screening tool.
• Model rather than correct. If your child says “wabbit”, reply naturally with “yes, a rabbit” and carry on. Hearing the word said correctly in conversation gives them a model without pressure.
• Talk face to face so your child can see your mouth, and cut background noise when you can.
• Have hearing checked if your child has had repeated ear infections or often asks you to repeat yourself.
• If an SLP has given you a target sound, the articulation practice tool has word lists organized by sound and word position for short, frequent practice. Drilling a sound the child cannot yet produce tends to build frustration, so wait for guidance before practising at home.

A note for clinicians

Crowe and McLeod (2020) reported mean ages of acquisition at the 50, 75 and 90 percent criteria across studies, which is why the figures above can differ from a single normative sample such as a test manual. When sharing the chart with families, it is worth naming the criterion and noting that the data describe single-word productions by monolingual US English speakers. The early, middle and late grouping is a convenient frame for discussing target selection alongside stimulability, error consistency and the child’s phonetic inventory, and it sits comfortably beside the cross-linguistic findings in McLeod and Crowe (2018) for children learning English as an additional language.