"Fumb" for thumb. "Dat" for that. "Wif" for with, "bruvver" for brother, "sink" for think. Of all the sound errors parents ask about, this is the one most likely to come with the reassurance that it will sort itself out.
Here is the unusual part: for the "th" sounds, that reassurance is mostly correct. /r/ and /l/ guides exist largely to tell parents when the waiting should stop. "Th" is the one sound where the developmental window genuinely runs later than almost anyone expects — and knowing exactly how late makes the decision much easier.
There Are Two TH Sounds, Not One
English spells two different consonants with the same two letters, and they do not develop at the same time.
Say thumb, then say this. In thumb, think, bath, math, three the sound is voiceless — tongue tip to the teeth, air only, no buzz from the voice box. In this, that, mother, weather, breathe it is voiced, with the vocal folds running. Phonetically they are written /θ/ and /ð/. Put a hand on your throat and the difference is immediately obvious.
This matters for two practical reasons. First, the voiced one is far more common in everyday speech, because it lives in the words we use constantly: the, this, that, them, there, then. A child who struggles with /ð/ affects far more of their talking than a child who struggles with /θ/ in thermometer. Second, as you will see in a moment, the two sounds are a full year apart in typical development. A child can be perfectly on schedule with one and behind on the other.
When Should a Child Say TH?
The best available answer for American English comes from Crowe and McLeod's 2020 review of children's English consonant acquisition in the United States (American Journal of Speech-Language Pathology), which pooled 15 studies covering 18,907 children. Using a 90% criterion — the age by which nine in ten children produce the sound correctly — the picture is this:
- Voiced "th" (/ð/, as in this) arrives at 5;0 to 5;11 (years;months), alongside /r/ and the "zh" of measure.
- Voiceless "th" (/θ/, as in thumb) arrives at 6;0 to 6;11 — and it is the only consonant in that band.
Read that second line again, because it is the whole point of this page: the voiceless "th" is the last consonant English-speaking children master. Everything else is in place before it. The authors note that "most consonants were acquired by 5;0" and that "all fricatives were acquired by 6;11," the "th" sounds being the ones that drag that group deadline out.
So a four-year-old saying "fumb" is on schedule, not behind. A five-year-old is still inside the window for the voiceless one. By 7 or 8, when the sound should have been in place for a year or more, waiting no longer has a rationale — and that is also the age at which it starts to show up in spelling.
The Kinds of TH Errors You Might Be Hearing
- F or V for TH. "Fumb" for thumb, "wif" for with, "bruvver" for brother. The most common substitution in English, and for a good reason: /f/ sits in nearly the same place, uses the lip instead of the tongue, and is acquired around age 3 — three full years before the voiceless "th."
- T or D for TH. "Tink" for think, "dat" for that. ASHA's selected phonological patterns resource files this under stopping — "a fricative and/or an affricate is replaced with a stop sound" — and uses "/dæt/ for that" as one of its own examples. It gives a likely elimination age of 5 for stopping on /ð/.
- S or Z for TH. "Sink" for think, "zis" for this. Less common, and more noticeable to listeners because it collides with real words.
- Omission. "Bruh" for brother, or the sound simply dropping at the end of a word.
- A distorted or dentalized "th." The sound is recognizable but muffled — the tongue is behind the teeth rather than lightly between them. ASHA classifies this kind of error as a distortion rather than a substitution.
One useful note on that second bullet: ASHA's stopping entry puts /ð/ elimination at 5, while Crowe and McLeod put /ð/ acquisition at 5;0 to 5;11. Those are not in conflict. One describes when a system-wide habit of replacing fricatives with stops has typically dropped away; the other describes when the individual sound is reliably produced. A child can stop using "d" for "th" and still produce a slightly muddy "th" for a while afterward.
"Th" for "S" Is a Different Problem Entirely
This is the section most parents arrive needing, because the two situations get searched with the same words.
If your child says sing as "thing" — the tongue coming forward between the teeth where an /s/ belongs — the sound they cannot make is /s/, not "th." ASHA's practice portal uses exactly this example to illustrate a substitution. Clinically that is an interdental or frontal lisp, it has a different developmental timeline, and crucially it is the error most often tied to a forward tongue posture. Our guide to speech therapy for lisps covers the four lisp types, which ones children outgrow, and why a tongue thrust pattern can keep one in place.
This page is about the opposite direction: "th" being the sound that has not arrived yet. If you are hearing both — a lisped /s/ and an "f" for "th" — that is worth saying out loud at an evaluation, because the treatment order is not obvious and the two targets can interfere with each other.
Dialect Comes First
Before any production is called an error, there is a question that has to be settled: what language system is this child actually speaking?
ASHA is unambiguous about this. Its portal opens the topic with "we all speak with an accent and use one or more dialects," describes dialects as "rule-governed language systems that reflect the regional and social background of its speakers," and warns that a dialect's own rules can be "misattributed as a phonological error pattern instead of a dialectical difference." Its phonological patterns resource goes further: "dialectal varieties of a language that do not produce certain sounds would not be considered an error pattern," and a speech-language pathologist "would not use the listed terms to describe those dialectal rules."
The "th" sounds are where this matters most, and there is specific research on it. Pearson, Velleman, Bryant, and Charko's 2009 study in Language, Speech, and Hearing Services in Schools tested all the consonants of Mainstream American English in a nationwide sample of 537 children who spoke African American English as a first dialect and 317 who spoke Mainstream American English as a first dialect, ages 4 through 12. Comparing the age at which each group reached the same 90% criterion, they found that the voiced "th" was the only sound that differed between the two dialects in both word-initial and word-final position. Their conclusion for clinicians is direct: initial diagnosis of impairment "should focus only on mastery of noncontrastive segments and structures that share a similar developmental profile for the 2 dialect groups."
In plain terms: "dat" for that in a child whose home dialect produces it that way is a dialect feature, not a delay — and treating it as a delay is a documented way to over-diagnose. It is worth adding that even the intelligibility benchmarks clinicians lean on carry this caveat; ASHA notes that in the study behind its own figures, all participants were monolingual American English speakers and "most of the participants were White."
None of this means a child who speaks a dialect cannot have a speech sound disorder. It means the evaluation establishes the child's language system first, and looks for errors inside that system. If a family's goal is specifically to add Mainstream American English pronunciation as a second dialect, that is a legitimate and different goal from treating a disorder, and it should be named as such rather than folded into a therapy plan.
The One Sound Where Tongue-Between-Teeth Is Correct
Here is where "th" breaks the pattern of every other sound guide on this site.
The "th" sounds are interdental by definition. The tongue tip belongs lightly between the front teeth — that is not a mistake to be corrected, it is the target. Every other sound guide spends its time getting the tongue off the teeth.
Which is why "th" is conspicuously absent from one particular list. ASHA's practice portal on orofacial myofunctional disorders names the speech signs typical of those disorders as "distorted productions of /s, z/, often with an interdental lisp, or abnormal lingual dental articulatory placement for /t, d, l, n, ʧ, ʤ, ʃ, ʓ/." The "th" sounds appear nowhere in it. The "sh," "ch," and "j" sounds, by contrast, are on that list — which is why SH sound speech therapy treats tongue resting posture as a live question in a way this page does not.
The practical consequence is one of the more counterintuitive things in this work. A child whose tongue habitually rests low and forward, or pushes between the teeth during a swallow, will often have a perfectly good "th" — and a distorted /s/ sitting right next to it. For that child the therapy goal is not keeping the tongue back; it is teaching the contrast between two sounds that live millimeters apart, so the tongue learns to come forward for think and stay back for sink.
That distinction is the reason Laura Friedman's dual credentials — CCC-SLP and Qualified Orofacial Myologist — change the evaluation rather than just the brochure: the same oral posture can be an asset for one sound and the obstacle for the next. Our comparison of myofunctional therapy vs. speech therapy explains where that line sits, and if the pattern turns out to be a whole class of sounds being simplified rather than one sound missing, our comparison of phonological vs. articulation disorder covers why that calls for a different plan entirely.
What TH Therapy Actually Involves
- Settling the dialect question, as above, before anything is labeled an error.
- Checking stimulability. ASHA defines this as "the child's ability to accurately imitate a model of the target speech sound that the child currently misarticulates," and says testing it helps determine "whether the sound is likely to be acquired without intervention, and which targets are appropriate for therapy." For a six-year-old with an "f" for "th," a child who can produce a clean "th" on request is often in a very short course of therapy.
- Using the visibility. "Th" is the most visible consonant in English. There is no hunting for a hidden tongue shape the way there is with /r/; the instruction is close to "put your tongue between your teeth and blow." Mirrors do a great deal of the work, and so does simply having the child watch your mouth.
- Pairing voiced and voiceless. Because /θ/ and /ð/ differ only in voicing, a child who finds one can usually be walked across to the other in the same session by adding or removing the buzz.
- Prioritizing by payoff. The voiced "th" sits in the, this, that, them — so for a child who is hard to understand, it often earns more intelligibility per session than the voiceless one, even though it is the earlier-developing sound.
- Establishment, generalization, maintenance. ASHA sets out the sequence plainly: establishment is "eliciting target sounds and stabilizing production on a voluntary level"; generalization is "facilitating carryover of sound productions at increasingly challenging levels"; maintenance is "stabilizing target sound production and making it more automatic; encouraging self-monitoring of speech and self-correction of errors."
Where there is a facilitating context, therapy starts there — ASHA's guidance is to begin instruction "in the syllable context(s) where the sound can be produced correctly." For "th" that is frequently a long, exaggerated, slightly silly production in isolation, which children tend to find easy to copy.
The TH Sound in Adults
ASHA reports that 1% to 2% of young adults have residual or persistent speech errors. "Th" also arrives in adult sessions by a second route: many of the world's languages do not contain these sounds at all, so they are a standard target in accent modification, which Bloom offers for adults alongside articulation, voice, and fluency work.
Adults generally move quickly here. The sound is visible, the instruction is concrete, and the hard part is the same as it is for every sound — carrying a new production into conversation that is moving too fast to think about the tongue.
Getting a TH Evaluation
An evaluation for the "th" sounds answers five questions: which of the two sounds is affected, which positions in a word, whether the error is a substitution, an omission, or a distortion, whether your child is stimulable for the target, and — first — what the child's home language system expects. It also checks what else is traveling alongside it; ASHA reports that roughly 40% of children with speech sound disorders have a co-occurring language impairment, which changes the plan.
For "th" more than for most sounds, the honest answer at four or five is often "this is on schedule, let us check again in six months," and you will hear that if it is the answer. Our guide to the signs your child may need speech therapy covers what else an evaluation weighs, our speech milestones by age guide places "th" in context with everything else developing at the same time, and our articulation disorder treatment page describes the full range of sound errors we work on as part of speech therapy at Bloom. If the sound in question is actually /l/ rather than "th," our guide to L sound speech therapy is the nearest companion to this one.
Sessions are available in person at our Dallas office and through online speech therapy nationwide.
Frequently Asked Questions
At what age should a child be able to say the TH sound?
Later than almost everyone expects. Crowe and McLeod's 2020 review of 15 studies covering 18,907 children in the United States found that 90% of children produce the voiced "th" of this between 5;0 and 5;11 (years;months), and the voiceless "th" of thumb between 6;0 and 6;11 — the only consonant in English that late. A four-year-old saying "fumb" is on schedule. By 7 or 8 it is overdue and worth evaluating.
Why does my child say "f" instead of "th"?
Because /f/ is the nearest easy neighbor: nearly the same place in the mouth, made with the lip instead of the tongue, and acquired around age 3 while the voiceless "th" is not acquired until 6;0 to 6;11. "Fumb" and "wif" are the most common substitutions in English and are usually a timing issue rather than a disorder. The same logic explains "dat" for that, which ASHA files under stopping — a fricative replaced with a stop sound.
Is "dat" for "that" a speech problem or an accent?
It can be either, and the order of those questions matters. ASHA states that "we all speak with an accent and use one or more dialects," and that a clinician "would not use the listed terms to describe those dialectal rules." Pearson and colleagues' 2009 study of 537 African American English speakers and 317 Mainstream American English speakers aged 4 to 12 found the voiced "th" was the only sound that differed between the dialects in both word-initial and word-final position — making it the single sound where establishing a child's home language system before judging the production matters most.
My child says "th" instead of "s" — is that the same thing?
No, it is the opposite problem. When "th" replaces /s/ — ASHA's example is sing produced as "thing" — the target sound is /s/ and the pattern is an interdental or frontal lisp. This page is about the reverse: "th" being the sound the child cannot yet make. Different targets, different plans, so it is worth being clear about which direction you are hearing.
Can tongue thrust cause a TH sound problem?
Rarely, and this is the one sound where it works in reverse. The "th" sounds are interdental by definition — the tongue tip belongs lightly between the teeth. ASHA's list of speech signs under orofacial myofunctional disorders names distorted /s/ and /z/ productions and placement errors on /t, d, l, n/ and the "ch," "j," and "sh" sounds; "th" is not on it. A child with a forward resting posture often produces a good "th" and a distorted /s/, so the goal becomes the contrast between the two sounds rather than keeping the tongue back.
Can TH sound therapy be done online?
Yes, and "th" is arguably the best sound in English for telehealth. It is the most visible consonant we have — the tongue tip is literally outside the teeth — so a child can watch the therapist's model on screen and check their own mouth in a mirror at the same time. Bloom provides online speech therapy nationwide alongside in-person sessions in Dallas.
Not Sure Whether to Wait?
With "th," the answer genuinely is often "wait" — but it should be a decision rather than a guess. Schedule a free consultation with Laura Friedman, MS, CCC-SLP, QOM — in person in Dallas or online nationwide — to find out which of the two sounds is affected, whether what you are hearing is an error or a dialect feature, and whether now is the right time to start.