Speech Sound Milestones by Age: A Parent Guide

Your child says “wabbit” for rabbit, “fumb” for thumb, and “tat” for cat — and a small worry has been sitting in the back of your mind. Here’s the reassurance to lead with: most of the speech “errors” that worry parents are right on schedule for a young child, not a sign that something is wrong.

The real question isn’t “is my child perfect.” It’s “is my child on track — and when should I actually get help?” This guide answers both, with a chart you can bookmark and come back to.

Speech sounds don’t all arrive at once. They develop in a predictable order over several years, which means a 3-year-old is supposed to still be missing several sounds. That’s not a delay; that’s development working exactly as it should.

The ages here come from the largest modern research review of how children actually learn speech sounds — McLeod & Crowe (2018) together with its U.S.-specific companion review (Crowe & McLeod, 2020), the same norms speech-language pathologists (the professionals who diagnose and treat speech) use day to day. One honest note up front: this chart is for children learning American English as their main language (more on bilingual and dialect differences near the end), and it’s meant to inform you — never to replace a professional evaluation.

How speech sounds develop: the big picture

Here’s the mental model that makes everything else click: sounds made at the front of the mouth with simple movements come early; sounds that need precise, fine tongue control come late.

That’s why p, b, and m are some of the first sounds your baby babbles, while r, s, and “th” can take years to perfect. This is normal motor development — the mouth maturing like any other set of muscles — not a measure of intelligence or effort.

Researchers often group sounds into three buckets, and it helps to hold onto them:

• Early sounds — m, n, w, h, b, p, d
• Middle sounds — ng, t, k, g, f, y
• Later sounds — “th” (both kinds), s, z, sh, ch, j, v, l, r

Think of that last group as “the ones it’s normal to still be working on in kindergarten.”

Now the single most important idea for reading any milestone chart: there are two ages that matter for every sound, and single-number charts confuse parents by hiding one of them.

• The 50% age is when about half of children can say the sound. This is the “just emerging” age — the sound is showing up, but plenty of kids don’t have it yet.
• The 90% age is when 9 in 10 children can say it. This is the “should be solid by now” age — the one speech-language pathologists actually watch.

Throughout this guide you’ll see ages written like 5;0 or 3;6. That’s a shorthand for years and months: 5;0 means 5 years 0 months, and 3;6 means 3 years 6 months.

One last reassurance before the chart. Being on the later side of a range can still be completely typical. Roughly 1 in 10 children master a given sound after the 90% age and are perfectly fine — ASHA is clear that norms are one input, not a verdict.

Speech sound development chart: which sounds develop when

Here’s the centerpiece — every sound with both its 50% “just emerging” age and its 90% “should be mastered” age, plus a plain-English note. Bookmark this one.

The one-line takeaway worth remembering: most consonant sounds are mastered by age 5, and all of them by age 7.

Worked example. Your daughter is 4;6 and says “wabbit,” “thun” for “sun” is occasionally “tun,” and “thumb” comes out “fumb.” Checking the chart: r isn’t due until 6;0, s isn’t fully expected until 5;0, and voiceless “th” not until 7;0. Every one of those is on schedule — there’s nothing here that points to a problem.

Source: McLeod, S., & Crowe, K. (2018). Children’s consonant acquisition in 27 languages: A scoping review. American Journal of Speech-Language Pathology — a large cross-linguistic review of how children acquire consonants; the U.S. English ages reflect its companion review, Crowe & McLeod (2020), Children’s English consonant acquisition in the United States, AJSLP.

Two caveats before you hand this chart to a worried relative:

1. Being a little past an emerging age is normal. It’s the 90% age that flags a sound worth watching.
2. A chart never diagnoses on its own — it sets up the two sections below (intelligibility and when to seek help), it doesn’t replace them.

How much should a stranger understand? Intelligibility by age

Here’s a reframe that takes the pressure off individual sounds: what usually matters more is whether people who don’t know your child can understand them. That’s called intelligibility, and it has its own age expectations, separate from sound mastery.

The classic, easy-to-remember benchmark is the “rule of fourths” (Coplan & Gleason, 1988):

Now the honest part that a lot of milestone articles leave out: real children vary enormously in the preschool years (Hustad et al., 2021 found a wide spread in how understandable typically-developing kids are at the same age — many typical 4-year-olds still measure nearer 75–80% in conversation). So treat those percentages as a guide, not a stopwatch. A 2-year-old who is hard for strangers to understand is common and not automatically a concern.

There is one number worth keeping in your back pocket as a practical signal: by age 4, an unfamiliar listener should understand at least about two-thirds of what your child says (Gordon-Brannan & Hodson, 2000). A 4-year-old well below that is worth getting checked — we’ll come back to this in the evaluation section.

Notice how this fits with the sound chart: a 3-year-old who is still missing the later sounds can also be around 75% intelligible. Both things are normal at the same time. Missing some sounds and being mostly understandable aren’t a contradiction — they’re exactly what a typical 3-year-old looks like.

What’s typical at each age (2, 3, 4, 5, 6–7)

Scan to your child’s age. Each one covers which sounds are usually in place, which are still fair game to be missing, the expected intelligibility, and one “this is normal” line.

Age 2 (around 24 months)

Early sounds are emerging — p, b, m, n, h, w, and d. Lots of sounds are still missing, and that’s expected. Strangers understand roughly half of what your toddler says, and word attempts matter far more than precision right now. Normal: dropping the ends of words (“ca” for cat) and simplifying longer words. A 2-year-old is a communicator first and a pronouncer second.

Age 3

By now p, b, d, m, n, h, and w are usually mastered, and t, k, g, and f are coming online. Strangers understand about 75% of speech. Around this age, “tat” for “cat” (called fronting) and dropping final consonants are on their way out. Normal: no r yet, no s-blends (like “spoon” or “star”), and no “th.” Three-year-olds still have years of fine-tuning ahead.

Age 4

Most early and middle sounds are solid now — k, g, f, y, and ng. Your child should be mostly understandable to strangers — often around 75–80% and climbing — even if some sounds are still imperfect. Meanwhile s and l are emerging. Normal: “wabbit” for rabbit, a mild lisp on s, and still no “th.” A 4-year-old can be easy to understand and still have a handful of sounds in progress.

Age 5

Most consonants are mastered by kindergarten — s, l, sh, ch, j, v, and z are all expected to be solid. Normal: r and “th” are still developing, and a mild lisp may be fading out. If the only sounds your child is still working on are r and “th,” they are right where the research expects them to be.

Ages 6–7

This is when the last pieces fall into place. The r sound and voiced “th” (as in “this”) master around age 6, and voiceless “th” (as in “thumb”) — the very last English consonant — arrives around age 7. By 7, the full consonant system should be in place. This is also the window where lingering r or “th” errors start to be worth a closer look.

If your 5- or 6-year-old is down to just a sound or two, the specific sounds they’re still working on are a natural fit for short, playful practice at home — the kind of low-pressure, few-minutes-a-day reinforcement that complements (not replaces) any guidance from an SLP and fits easily into a bedtime routine. Sound Safari is built around exactly that idea, with a library of practice words organized by sound and position.

Normal vs. red-flag error patterns

Speech-language pathologists don’t just track individual sounds — they look at patterns (called phonological processes). Here’s a parent-friendly map of which patterns are normal and scheduled to disappear, and which ones are worth flagging at any age.

The presence of a normal pattern on schedule is good news — it means development is unfolding the way it should. What matters is the timeline (still doing it well past the resolve-by age) or an atypical pattern. The three red-flag patterns above aren’t a diagnosis; they’re simply patterns worth mentioning to a professional, because they aren’t part of the usual path. (For a deeper walkthrough of these patterns, see our guide to phonological processes.)

A few other at-any-age signals belong on the watch list too: a child losing speech skills they previously had, or being far harder to understand than other children the same age.

If you’d like a low-stakes way to see which sounds your child is actually producing, Sound Safari includes a built-in quick sound check (24 sounds across 44 items) that walks through each sound in a word. It’s purely informational — a snapshot to share or to help you decide whether to check in with a professional — never a diagnosis.

When to consider a speech-language evaluation

Most of this guide is reassurance, because most worries turn out to be on-schedule development. But here’s the practical part: concrete, age-anchored reasons to check in. Read these as “reasons to get a professional read,” not alarm bells.

Consider an evaluation if:

• A stranger understands less than about half of a 3-year-old, or under about two-thirds of a 4-year-old.
• A sound is still missing or substituted past its 90% age — for example, no k or g at 4, or no s or l at 5.
• You notice any of the red-flag patterns at any age: dropping the first sound of words, backing, or replacing sounds with a throaty catch.
• Your child loses speech skills they used to have.
• Your child is frustrated, avoids talking, or is being teased about how they sound.

Just as important — the reassuring flip side, the part competitors bury:

These are usually fine to simply monitor:

• “Wabbit” for rabbit before age 6.
• Missing “th” before age 7.
• A mild lisp at age 4.
• Being on the later-but-typical edge of a range.

Therapy for r is rarely indicated before age 6, and therapy for “th” rarely before age 7 — that’s directly in line with McLeod & Crowe and ASHA’s developmental-norms guidance.

So what is an evaluation? It’s a no-pressure way to get a professional’s read on whether your child is on track. Early support is easier than waiting, and in the United States, evaluations are available free through Early Intervention (for children under 3) or your public school district (for children 3 and up). Many private speech-language pathologists also offer screenings. When in doubt, an evaluation gives you an answer — it doesn’t commit you to anything.

One important boundary: this article, the chart above, and any screening tool — including Sound Safari’s quick sound check — are informational, not diagnostic. Sound Safari is a clinical tool, not a medical device; it does not provide diagnoses or treatment recommendations. Those decisions belong to a qualified speech-language pathologist who can evaluate your child directly.

Speech delay vs. speech disorder, and other common questions

A few of the most common parent confusions, cleared up quickly:

Delay vs. disorder. A delay means your child is on the right track, just behind — the patterns look normal, they’re simply arriving late. A disorder means the pattern itself is atypical, like an error not seen in typical development (dropping the first sound of words is a classic example). One is a question of timing; the other is a question of pattern — and a speech-language pathologist is the one who can tell them apart.

Do kids outgrow lisps and r errors on their own? Often yes for very young children, since these are late sounds anyway. But past the 90% age (s and l by 5, r by 6, “th” by 7), a persisting error is worth checking rather than waiting indefinitely.

Will practicing a sound “wrong” at home make it worse? Don’t worry — gentle modeling helps. If your child says “wabbit,” simply saying “rabbit” back warmly is supportive. What backfires is constant correction or drilling a sound your child can’t yet make — that just creates frustration. The rule of thumb: model, don’t drill. Light, positive, playful practice is always fine.

A note on bilingual children and dialects. This chart is for children whose main language is mainstream American English. Bilingual children, and speakers of dialects such as African American English, follow different, equally-normal timelines (Crowe & McLeod, 2020). Don’t apply this chart to them without a speech-language pathologist who knows the child’s full language background. Children with hearing differences, cleft palate, or apraxia also follow different paths.

Frequently Asked Questions

What speech sounds should my child have at each age?

By age 3, most children have p, b, d, m, n, h, and w. By 4, the middle sounds (t, k, g, f, y, ng) are usually solid. By 5, most consonants are mastered, including s, l, sh, ch, j, v, and z. The latest sounds — r, voiced “th”, and voiceless “th” — round out the system by ages 6 to 7. These are 90% mastery ages from McLeod & Crowe (2018) and Crowe & McLeod (2020), the research norms speech-language pathologists use.

How much of my child’s speech should a stranger understand at age 2, 3, and 4?

A handy rule of thumb (the “rule of fourths”) puts strangers at roughly 50% understanding at age 2, 75% at age 3, and approaching fully understandable by age 4 — though newer growth-curve research (Hustad et al., 2021) finds many typical 4-year-olds still nearer 75–80% in conversation with strangers. Treat these as a guide, not a stopwatch. A common research-backed flag: by age 4, an unfamiliar listener should follow at least about two-thirds of what your child says.

Is it normal for a 4-year-old to say “wabbit” instead of “rabbit”?

Yes. Replacing r with w (“wabbit”) is typical well into the preschool years. The r sound is one of the latest to develop, reaching 90% mastery around age 6. Therapy for r is rarely recommended before then, so a 4- or 5-year-old saying “wabbit” is usually right on schedule. For how r is actually taught when the time comes, see our r sound therapy guide.

At what age should a child say the “th” sound?

The “th” sounds are the last English consonants to arrive. Voiced “th” (as in “this”) reaches 90% mastery around age 6, and voiceless “th” (as in “thumb”) around age 7. Saying “fumb” for “thumb” at age 5 or 6 is within the typical range, so “th” is rarely targeted before age 7.

Is saying “tat” for “cat” normal for a 3-year-old?

Yes. Saying “tat” for “cat” is a pattern called fronting — making a back sound (k, g) at the front of the mouth (t, d). It’s typical and usually resolves on its own by around age 3 and a half. If it’s still happening well past then, it’s worth mentioning to a speech-language pathologist.

What is the latest speech sound to develop?

Voiceless “th” (as in “thumb”) is the latest English consonant to develop, reaching 90% mastery around age 7. The r sound and voiced “th” (as in “this”) are close behind, around age 6. That’s why lingering r and “th” errors in a 5- or 6-year-old are usually nothing to rush about.

When should I worry about my child’s speech and see a speech therapist?

Consider an evaluation if a stranger understands less than about half of a 3-year-old or under about two-thirds of a 4-year-old, if a sound is still missing past its 90% mastery age (for example, no k or g at 4, or no s or l at 5), if your child drops the first sound of words at any age, or if they’re losing skills they used to have. Frustration, avoiding talking, or teasing are also good reasons to check in. An evaluation is informational and low-stakes — it gives you an answer without committing you to anything.

What’s the difference between a speech delay and a speech disorder?

A delay means your child is on the typical path, just behind schedule — the patterns look normal but are arriving late. A disorder means the errors themselves are atypical — patterns not usually seen in typical development, like dropping the first sound of words. A speech-language pathologist is the one who can tell the two apart.

Do children outgrow lisps on their own?

Often, yes — for very young children, since s and z are still developing. But past the 90% mastery age (s by 5, r by 6, “th” by 7), a persisting error is worth checking rather than waiting indefinitely. A mild lisp at 4 is usually fine to monitor; a clear, steady lisp at 6 or 7 is a reasonable reason to get a professional read. Our s sound therapy guide and l sound therapy guide walk through how those sounds are shaped.

Does being bilingual delay a child’s speech sounds?

No — being bilingual does not cause a speech delay. Bilingual children follow their own typical timelines that differ from a single-language chart. The ages in this guide are for children whose main language is mainstream American English, so they shouldn’t be applied to a bilingual child without a speech-language pathologist who knows the child’s full language background.

Here’s the recap to carry with you: most of the “errors” that worry parents — “wabbit,” “fumb,” “tat” — are on schedule, not a problem. Use the chart as your guide, and let the 90% mastery ages and the intelligibility signals be your cue to check in rather than guess. And if you ever do have a concern, an evaluation is low-stakes and gives you a real answer. If you’d like to see exactly which sounds your child is producing and turn the ones they’re still working on into a few minutes of playful daily practice, Sound Safari has a built-in quick sound check (24 sounds across 44 items) and a parent-friendly practice library — informational, never diagnostic, with a 14-day free trial and no credit card required.