Most English consonants are in place by age 5, and the order is fairly predictable. Early: p, b, m, n, d. Middle: k, g, f, t. Late: s, sh, ch, l, and last of all r and th (Crowe & McLeod, 2020).
The short version:
- b, n, m, p, h, w, d are usually there by 2;0 to 2;11 (years;months)
- g, k, f, t, ng, y by 3;0 to 3;11
- v, j, s, ch, l, sh, z by 4;0 to 4;11
- r, voiced th and zh by 5;0 to 5;11; voiceless th by 6;0 to 6;11
- A 4-year-old is expected to be at least 50% understandable to a stranger, not 100% (Hustad et al., 2021)
This is articulation, which is a different question from how many words a child has. A 2-year-old with 200 words who says "tup" for "cup" has a sound in progress, not a language delay. Below is the chart, the intelligibility numbers most parents have never seen, and the small set of error patterns that are actually worth an appointment.
Quick Reference: speech sounds by age
| Age | Sounds typically acquired by then | Rough intelligibility |
|---|---|---|
| 2 years | b, n, m, p, h, w, d | Strangers catch roughly half of single words by 31 to 47 months |
| 3 years | g, k, f, t, ng, y | Familiar adults follow most of it; strangers still work for it |
| 4 years | v, j, s, ch, l, sh, z | At least 50% understandable to a stranger is the floor, not the goal |
| 5 years | r, voiced th, zh | 75% understandable falls between 49 and 87 months |
| 6 to 7 years | voiceless th; r still settling for some | Most children are easy to follow, not flawless |
Ages come from a review of 15 studies covering 18,907 US children learning English, at the 90% acquisition criterion (Crowe & McLeod, 2020). Intelligibility ranges come from growth curves built on 538 typically developing children (Hustad et al., 2021).
Which sounds come first, and why
Sounds you can see arrive first. Lips together and apart gives you p, b, m. Tongue tip to the ridge behind the teeth gives you t, d, n. A child can watch your face and copy it.
The late sounds hide. The cross-linguistic picture is consistent: plosives and nasals come earlier than trills, flaps, fricatives and affricates, and sounds made at the lips or the back of the mouth come earlier than sounds needing precise tongue-tip placement (McLeod & Crowe, 2018). Fricatives like s, sh and th need a narrow, steady channel of air that a 3-year-old's tongue cannot hold yet. R needs a tongue shape nobody can demonstrate. Ask an adult where their tongue goes for R. Most genuinely cannot say.
That review also covered 27 languages and 26,007 children, and found children had produced at least 93% of the consonants in their language correctly by age five. Whatever language your child is learning, the tail end of the chart is where the hard sounds live.
Related: R gets its own timeline, because it is the sound that most often brings a school-age child to therapy. See when the R sound develops.
How clear should my child actually sound?
This is the part that surprises people. The old rule of thumb passed around in parenting groups says 50% intelligible at 2, 75% at 3, 100% at 4. The research does not support the top end.
Measured with naive adult listeners transcribing what they heard, single-word intelligibility hits 50% somewhere between 31 and 47 months depending on where the child sits in the distribution, 75% between 49 and 87 months, and 90% between 83 and 120 months or later. The authors' practical benchmark: a child should be at least 50% intelligible by 48 months (Hustad et al., 2021).
So a 4-year-old you understand three quarters of the time is not behind. And a 4-year-old strangers understand less than half the time is worth an appointment, even if you personally follow every word. You are the worst judge of your own child's clarity, because you have the context nobody else has.
Which sound errors are typical, and which are not
Errors are not random. Children simplify in patterns, and the patterns have shapes with names.
| Pattern | What it sounds like | Worth a look after |
|---|---|---|
| Fronting | "tup" for "cup", "doe" for "go" | Age 4 (k and g are typically acquired by 3;11) |
| Stopping | "tun" for "sun", "pish" for "fish" | Age 4 for f and s; age 5 for sh, ch, z |
| Gliding | "wed" for "red", "yeg" for "leg" | Age 5 for l, age 6 to 7 for r |
| Lisp on s and z | "thun" for "sun" | Age 5 for a straight lisp; any age if it sounds wet or sideways |
| Final consonants dropped | "ca" for "cat", "bu" for "bus" | Age 3 and a half (our rule) |
| Clusters collapsing | "poon" for "spoon", "tuck" for "truck" | Age 4 and a half (our rule) |
Two rows say "our rule". That is KidyGrow's threshold for raising it at a check-up, not a published norm, and we would rather label it than dress it up as data. Every other row is derived from the acquisition ages in the review above.
One more distinction that matters more than the chart: a child who says "tar" for "car" every single time is doing something orderly and usually easier to fix than a child for whom the same word comes out three ways in a row. Inconsistency is one important flag, alongside the number and the type of errors.
Wait or get it checked
- Late sound missing, right age band, speech improving โ wait, keep reading and talking.
- Late sound missing but the child is frustrated or teased โ get an evaluation now, regardless of the calendar.
- Early sound (p, b, m, n, d, k, g) missing after 4 โ get an evaluation.
- Under 50% understandable to strangers at 4 โ get an evaluation.
- Many sounds affected at once, or the same word said differently each time โ get an evaluation.
- Sounds that were there are gone โ call your pediatrician the same week.
Hearing sits underneath all of this. Recurrent ear infections and fluid can muffle exactly the high-frequency sounds (s, sh, f, th) that a child needs to hear to copy. A hearing check is part of a proper speech assessment, not an optional add-on (NIDCD).
Related: For the whole arc from first sounds to full sentences, see the toddler speech development guide, and for language rather than articulation, speech delay signs by age.
Common mistakes
Counting sounds instead of listening for patterns. Six errors that are all fronting is one pattern. Two errors scattered across places and manners can be the messier picture.
Drilling a sound the mouth cannot make yet. Asking a 3-year-old to say "rabbit" correctly fifty times teaches frustration. Nothing else.
Correcting mid-sentence. Say the word back correctly and move on. "You want the tar? Here's the car." No pause, no lesson.
Comparing to an older sibling. Sound acquisition ranges are wide, and there is real variability between typically developing children in both studies above.
Assuming a bilingual child gets extra time. Different languages have different sound inventories and different timelines, but learning two does not buy a grace period on either. For a bilingual child, an assessment needs to take both languages into account, rather than only comparing the child against the norms of one language.
Treating the chart as a test. Those studies used single words, elicited one at a time in a quiet room. Your child talks in sentences, at speed, with a mouth full of cracker. A sound can be present in "sun" and gone in "spaghetti sauce" and that is normal for months.
When to seek professional help
Start with your pediatrician, who can rule out hearing loss and refer you on. A speech-language pathologist does the actual articulation assessment: which sounds are present in isolation, in words, in sentences, and how much of it a stranger can follow. The American Speech-Language-Hearing Association publishes a public chart of sound development by age (ASHA), and the AAP's parent guidance covers what to expect from a language check (HealthyChildren.org).
Ask for an assessment rather than waiting for a referral to feel urgent. An assessment commits you to nothing. It is information.
Related: When to seek help for speech delay covers the referral path, and if the wobble is fluency rather than sounds, read why toddlers stutter when excited.
Frequently asked questions
At what age should a child say all their sounds correctly?
Most consonants are acquired by age 5, with r, th and zh landing between 5 and 7 for many children (Crowe & McLeod, 2020). "All sounds, all words, all the time" is closer to 7 than to 5. A 6-year-old who still says "wed" for "red" is inside the typical range, though it is worth mentioning at a check-up.
Is "tup" for "cup" at age 3 something to worry about?
Usually not. That is fronting, and k and g are typically acquired somewhere in the fourth year. If it is still there at 4, or if it is happening alongside several other patterns, get an articulation assessment.
How many words should my child have before I worry about sounds?
Those are separate questions. Word count is language, sound production is articulation, and a child can be ahead on one and behind on the other. If word count is the worry, start with how many words a 2-year-old should say.
My child is understandable to me but not to their teacher. Who is right?
It is worth taking seriously what the teacher hears. You have a lot of context, so you often understand more than someone outside the family does. An outside listener is the better measure of how clear the speech is without that context, which is exactly how the intelligibility studies were designed.
Does a dummy or thumb-sucking cause sound errors?
Prolonged sucking habits are associated with dental and bite changes that can affect s and z, and most guidance suggests winding them down through the preschool years. It is rarely the whole explanation for a broad pattern of errors, so do not stop at the dummy if several sounds are affected.
Do all languages follow the same order?
Roughly, at the level of sound classes. Across 27 languages, plosives and nasals came before fricatives, affricates and trills, and most consonants were acquired by five (McLeod & Crowe, 2018). The specific late sounds differ, because a rolled R and an English R are not the same motor task.
How KidyGrow helps you track sounds over months
Sound progress is invisible week to week and obvious across a season, which is exactly the wrong shape for a tired parent's memory.
KidyGrow learns your child instead of an average child. Note down what a word actually sounded like when it catches your ear, and by the second or third week the Daily Brief stops offering generic milestone lines and starts sounding like something that remembers: "Two weeks logged. K appears in 'car' and 'cookie' but not in 'cracker', so it is coming in short words first. R and TH still absent, which is expected at this age."
You usually do not notice fronting on the day it starts. You notice it when a grandparent asks what "tat" means. The app is there to hold the thread you cannot hold at 7pm.
Sometimes it will not find a pattern. Four weeks of a cold and a house move produce noise, not signal, and the honest output then is "not enough clear data yet". That is still better than a guess.
About this guide: KidyGrow is an app built by a mother that learns your child's development and changes as they do. This article draws on the Crowe & McLeod consonant acquisition review, the Hustad intelligibility growth curves, and parent guidance from ASHA, the AAP and NIDCD. It is educational and is not a diagnostic tool. If you are worried, ask your pediatrician. That is always the right next step. Last updated September 2026.
Charts describe groups. Your child is one child, and the useful comparison is your child in October against your child in June. Write down five words this week. Do it again in six weeks. The answer is usually in the difference.
Sources
- Crowe, K. & McLeod, S. (2020). Children's English Consonant Acquisition in the United States: A Review. American Journal of Speech-Language Pathology, 29(4), 2155-2169. https://pubmed.ncbi.nlm.nih.gov/33181047/
- McLeod, S. & Crowe, K. (2018). Children's Consonant Acquisition in 27 Languages: A Cross-Linguistic Review. American Journal of Speech-Language Pathology, 27(4), 1546-1571. https://pubmed.ncbi.nlm.nih.gov/30177993/
- Hustad, K. C., Mahr, T. J., Natzke, P. & Rathouz, P. J. (2021). Speech Development Between 30 and 119 Months in Typical Children I: Intelligibility Growth Curves. Journal of Speech, Language, and Hearing Research, 64(10), 3707-3719. https://pubmed.ncbi.nlm.nih.gov/34491793/
- National Institute on Deafness and Other Communication Disorders. Speech and Language Developmental Milestones. https://www.nidcd.nih.gov/health/speech-and-language
- American Speech-Language-Hearing Association. Speech Sound Development Chart. https://www.asha.org/public/speech/development/chart/
- American Academy of Pediatrics, HealthyChildren.org. Language Development in 2-Year-Olds. https://www.healthychildren.org/English/ages-stages/toddler/Pages/Language-Development-2-Year-Olds.aspx
