Last updated 2026-07-10
TL;DR
The expressive language activities with the best evidence for late talkers are self-talk narration, parallel play commentary, sabotage routines, shared book reading with expectant pauses, and choice-offering. Naturalistic modeling inside daily routines beats drill-style practice for kids under 5. Start with 15 minutes of focused interaction a day, then spread the same strategies across meals, baths, and play.
Expressive language is everything your child sends out into the world: words, gestures, pointing, eventually full sentences. It's the output side of things. Receptive language, what they actually understand when you talk to them, is the other half, and it's usually way ahead of what comes out of their mouth. The American Speech-Language-Hearing Association defines a late talker as a child with fewer than 50 words or no two-word combinations by 24 months [1]. Most late talkers understand far more than they say, and that gap is actually good news: it tells you the machinery for language is there. The bottleneck is production, not comprehension.
Causes vary. Some kids have reduced oral motor coordination, some have simply had fewer chances to practice the back-and-forth that language grows from, and some have autism, childhood apraxia of speech, or another condition that changes how the brain plans speech. Somewhere between 70 and 80 percent of late talkers identified at age 2 catch up on their own by age 3, but nobody can reliably predict in advance which kids those will be [2]. The activities below work across most of these causes because they target the conditions language actually grows in: low pressure, high repetition, meaningful context, and a partner who responds the same way every time.
When to start, and why timing matters
Start the moment you notice a concern. You don't need a diagnosis first. The AAP recommends developmental surveillance at every well-child visit and formal screening at 9, 18, and 30 months [3], and if your child is behind at any of those checkpoints, home activities are fair game right away.
Early intervention in the United States runs from birth through age 2 under IDEA Part C, then age 3 to 21 under Part B, and federal law requires eligible children to be evaluated within 45 days of a referral [10]. Eligibility thresholds vary by state, so plenty of kids who don't technically qualify for services still benefit a lot from parent-run strategies at home. The brain's sensitivity to language input peaks in the first three years, which isn't meant to alarm anyone, it just means effort now pays back at a steeper rate than the same effort would later. The activities here suit children roughly 12 months to 5 years, and most scale up as your child grows.
What the research actually supports
A handful of approaches have real evidence behind them, not just clinical opinion. Naturalistic developmental behavioral interventions (NDBIs) are the best-studied category for kids with language delays and autism: Yoder and Stone found that parent-implemented NDBIs produced significant gains in expressive communication in toddlers with autism [4]. The idea is simple, teach language inside natural routines instead of at a table with flashcards.
Milieu teaching is one NDBI strategy: you arrange the environment to create a reason to communicate, then respond to your child's attempt by modeling a slightly harder form. Child points and grunts, you say the word. Child says one word, you expand to two. Researchers call this modeling at the edge of what the child can already do.
Shared book reading with dialogic techniques, open-ended questions, pausing before a predictable word, asking the child to name pictures, has support from multiple randomized trials for vocabulary growth [5]. And responsive interaction is the broadest of these: follow the child's lead, comment on what they're already looking at, and wait. Tamis-LeMonda and colleagues found that maternal responsiveness at 18 months predicted expressive vocabulary at 24 months, independent of maternal education or income [6]. None of this needs special equipment, just your attention and a small shift in how you already talk with your child.
Daily activities worth building into your routine
Self-talk and parallel talk take the least effort and cost nothing. Self-talk means narrating your own actions out loud, slowly, in simple words: "I'm pouring the water. Now I'm stirring. Hot soup." Parallel talk narrates what your child is doing: "You're pushing the truck. It's going fast. Crash!" Both give your child dense language tied to things they can see and touch, so keep sentences short and your voice animated.
Sabotage routines feel wrong at first and work anyway. Set up a moment where your child has to communicate to get what they want: hand over a container they can't open, give them one cracker when they clearly want the whole stack, put their shoes on the wrong feet. Then wait. Don't rescue them right away. Hold eye contact, look expectant, and the second they attempt anything, respond fast and big. You're manufacturing a genuine reason to communicate.
Choice offering works well too. Skip yes/no questions and instead hold up two options where your child can see them: "Banana or apple?" The visual cuts the mental load and raises the odds of a word attempt. Once your child chooses reliably, try pairing one thing they love with one they don't, since that contrast often pulls out a louder, clearer response.
Expansions and extensions take a bit more effort but do a lot of work. When your child says something, repeat it correctly, then add one new piece. Child says "dog," you say "Big dog. He's running." You're not correcting, you're modeling the next rung up with zero pressure to copy you.
Music and song routines with predictable gaps are reliable triggers: "Old MacDonald had a farm, E-I-E-I..." and then you stop. The melody and rhythm lower the motor planning demand of speech, and plenty of kids who produce almost no words will fill a song gap before they'll label an object out of nowhere. Use that.
For play-based vocabulary building, repeat one target word across three to five play settings in a single day. Targeting "up"? Use it lifting a toy, when the ball flies into the air, when you scoop up your child, when the block tower falls. Same word, different contexts every time. Word-learning research in toddlers consistently shows that varied context speeds up generalization [7].
And for book reading, pick titles with repetitive text and clear picture-word matches ("Brown Bear, Brown Bear" is the classic). Read it again and again, and once your child knows it, pause before the predictable word, look at them, and wait three to five seconds. If nothing comes, just say it yourself, no drama. It's the repetition across days that builds the attempt.
Play, correction, and what actually helps
Play is the original language classroom. Kids pick up word meanings fastest when the word is attached to something they're actively doing, not when they're sitting still being taught. The type of play matters too: symbolic play, using one object to stand for another (a banana as a phone, a block as a car), tracks closely with language, and children who do more of it at 18 months tend to have larger expressive vocabularies at 24 months [8]. If your child does little pretending, building pretend play alongside language work is worth doing.
Joint attention, the moment you and your child are both locked onto the same thing at the same time, is what makes any of this work. You can't teach a word your child isn't looking at, so instead of pulling them toward what you want to talk about, get on the floor and follow their eyes. Keep toys plain, too: ones that do the work for the child (lights, sounds, buttons) shrink how much your child needs to say or imagine. Blocks, a few small figures, and some cars will pull more language out of your child than almost any electronic toy on the shelf.
As for correcting errors: don't, at least not the blunt "that's wrong, say it like this" way. Direct correction tends to shut down communication attempts, and a child who's already shaky about talking doesn't need proof their attempt fell short. Expansion gets you the same result without the sting. If your child says "boo" for "blue," you say "Blue! Yes, the blue ball." You've modeled the target naturally without making them repeat it, and with enough exposures the accurate form tends to crowd out the approximation on its own.
For children with apraxia of speech, error patterns are especially tangled and specific, and correction helps even less. Those kids need an SLP's structured plan rather than a parent drilling them at home. The one move that does help across the board is waiting after your child communicates, even when it comes out wrong: pause, look interested, stay close. Sometimes the second try comes out clearer.
How is expressive language different for autistic late talkers?
Autistic late talkers often communicate differently than neurotypical late talkers, so the activities that help need a few tweaks.
Many autistic kids rely on echolalia, repeating phrases from TV, books, or earlier conversations, as a main way to express themselves. For years people dismissed this as meaningless. It isn't. Prizant and Duchan showed that echolalia carries communicative intent even when the phrase doesn't sound right to the listener [9]. Build on it rather than trying to stamp it out.
Visual supports pull a lot of weight here too. Pairing words with pictures, objects, or actions gives an autistic child something stable to hang language on. That's a big reason many AAC devices work so well: they offer a consistent visual and motor route to expression that doesn't depend on the day-to-day ups and downs of spontaneous speech.
Motivation to communicate varies enormously from kid to kid. Some autistic children will talk endlessly about what they love. Others mostly communicate to get a need met rather than to connect socially. Meet them where they are. If your child only talks about trains, talk about trains with them. The language they build in that one channel spreads out over time.
For approaches built specifically for this population, see autism spectrum speech therapy.
How many words should my late talker have at each age?
Here's a reference table built from ASHA and CDC developmental norms. These are averages with a lot of spread around them, and a child at the 10th percentile is not the same as a child with a disorder.
| Age | Typical expressive vocabulary | Typical phrase/sentence use |
|---|---|---|
| 12 months | 1 to 3 words (beyond mama/dada) | Gestures + vocalizations |
| 18 months | 10 to 25 words | Single words, some two-word attempts |
| 24 months | 50+ words | Two-word combinations ("more milk", "daddy go") |
| 36 months | 200 to 1,000 words | Three-word sentences, strangers understand ~75% |
| 48 months | 1,000+ words | 4 to 6 word sentences, asks and answers simple questions |
The 50-word mark with two-word combinations at 24 months is the clinical threshold most often used for "late talker" status [1]. If your child sits below these ranges by more than a month or two, it's worth a screening conversation with the pediatrician, and possibly a referral to speech therapy. CDC milestone norms line up with these figures [12].
What to do during everyday routines
The 20 minutes you carve out for "speech time" matters less than the 8 to 10 hours of routine your child already moves through every day. Routines are where the real repetition lives.
Bath time is ideal: the vocabulary stays the same (water, splash, wash, hair, dry), the context is shared, and your child is usually calm. Label everything. Pause at the predictable moments. Say "Ready, set..." and wait.
Mealtimes are built for choices and requests. Serve small portions so your child has to ask for more. Hold the cup just out of reach and wait for an attempt before you hand it over.
Car time is underrated. You're a captive audience, the eye contact pressure is off (which suits some kids), and you can narrate whatever rolls by: "Big truck! It's red. Beep beep."
Bedtime pulls together predictability, connection, and the shared book reading the research backs. Same book, same routine, same pauses. Boring is productive here.
One principle runs through all of it: fewer questions, more comments. Questions put kids on the spot, while comments model language and leave room for a child to jump in. Aim for roughly one question for every five comments.
When does a late talker need a speech-language pathologist?
Home activities don't replace an evaluation when certain markers show up. Get a referral to a speech-language pathologist if your child has fewer than 10 words by 18 months, has no words at all by 16 months, loses words or skills they used to have at any age, has fewer than 50 words and no two-word combinations by 24 months, is hard for familiar adults to understand at age 3 (intelligibility below roughly 75 percent), shows frustration or shuts down around communication attempts, or has a family history of language delay or learning differences.
ASHA lists "early identification and treatment of speech, language, and hearing disorders" as a core part of the SLP scope of practice, and names parent coaching as an evidence-based service delivery model for toddlers [1]. In other words, a good SLP isn't only working with your child in a room. They're teaching you how to do this at home.
If access is the problem, online speech therapy has expanded a lot and works especially well for parent coaching.
If you want a starting point, Little Words has a short quiz at /start that maps your child's communication profile to specific activities. It isn't a clinical assessment, just a structured place to begin.
How can I tell if these activities are actually helping?
Parents often feel like they're putting in the work with no way to tell if it's landing. Here's a tracking system that needs no clinical training.
Keep a word list. Every time your child says a new word or clearly attempts one, write it down with the date. Don't count words they used to say, only spontaneous, unprompted productions. At 24 months, 50 unique words is the threshold; at 18 months, 10 to 25 is typical.
Note where each word shows up. If your child only says "more" at the table and never during play, that tells you something. Generalization across settings is the real goal.
Film a short clip of natural play once a week. On video you'll catch changes in how much your child vocalizes, how they take turns, and how many communicative jobs they're doing (requesting, labeling, protesting, greeting) that slip past you in the moment.
Bring this to your pediatrician and SLP. It's genuinely useful clinical information: a child who added 12 new words in 6 weeks is on a different track than one who added 2.
Nobody has great data on exactly how fast home activities should show results; the research mostly measures outcomes at 6 to 12 month follow-ups. A rough rule of thumb: if you've been consistent for 8 to 12 weeks with no visible change in word count or communication attempts, treat that as your signal to get a professional evaluation, not a reason to grind harder alone.
If your child has fewer than 10 words by 18 months, or fewer than 50 words and no two-word phrases by 24 months, that's a real threshold for concern. Any loss of words a child used to say, at any age, deserves an immediate call to the pediatrician rather than a wait-and-see approach. The AAP recommends formal developmental screening at 9, 18, and 30 months, so those visits are the natural moment to raise concerns instead of letting them sit. It helps to know what you're actually dealing with. A late talker usually struggles only with expressive language: comprehension and social skills are mostly fine, and many catch up by age 3 without any intervention. A language disorder is different because it doesn't resolve on its own and often affects comprehension, grammar, and social use of language too. The distinction tends to sharpen between ages 3 and 5, which is exactly why ongoing monitoring and early evaluation matter so much. Screen time plays a bigger role than many parents expect. The AAP recommends none for children under 18 months (other than video chat) and firm limits from 2 to 5. Background TV in particular cuts into both the amount and quality of parent-child talk, which is the main input late talkers need to progress. Swapping even 30 minutes of screen time for narrating your child's play counts as a direct language intervention. Simple signing can help too. Pairing basic signs with spoken words doesn't delay speech, and it often speeds things up by giving a child an easy way to communicate before words come. When a child can signal what they want, frustration drops and the drive to communicate tends to rise. Start with 5 to 10 core signs (more, all done, eat, drink, up, help, open, please) and use them every time you say the word aloud. You don't need to switch into baby talk. Child-directed speech (slower, higher-pitched, repetitive) has real research behind it, but that's not the same as trading real words for nonsense sounds. Say real words clearly, keep sentences short, repeat the key ones, and pause often. Skip long, complicated sentences. The goal is clear and engaging, not babyish. Late talking can be one feature of autism, but the two aren't the same thing: not every late talker is autistic, and not every autistic child is a late talker. Other signs to watch for include reduced eye contact, not responding to their name by 12 months, few gestures like pointing, limited social smiling, and any loss of skills. A developmental pediatrician or psychologist handles autism evaluations, since speech delay alone doesn't diagnose it. For toys, simple and open-ended beats electronic. Small figures, toy vehicles, blocks, a play kitchen with pretend food, and basic puzzles all create natural chances to label, request, and comment. Bubbles and balloons work especially well because kids want them badly, have to ask for them, and the short cycles create repeated back-and-forth turns. Most parent-implemented language studies use sessions of 15 to 20 minutes of focused play, but the real gains come from weaving strategies like self-talk, parallel talk, and expansion into everyday routines: meals, bath time, free play. Consistency across the whole day matters more than any single session length. One technique worth knowing is wait time: pause 3 to 10 seconds after setting up a chance to communicate, with an expectant look, before you jump in and fill the silence. It works because most parents instinctively rescue kids from that silence, which removes their reason to speak up. Milieu teaching studies build in this kind of expectant waiting as a core piece, and it reliably increases how often children attempt to communicate. Reading daily helps too, especially with dialogic reading: pausing before predictable words, asking open-ended questions about the pictures, and following your child's comments rather than pushing through the text. Reviews of shared reading show real gains in expressive vocabulary in children under 5. Rereading the same books again and again actually works better than always reaching for something new, since the anticipation pulls out more word attempts. If your child talks freely at home but says little or nothing at preschool, and this has gone on for more than a month, that pattern is called selective mutism. It's a separate condition from late talking, rooted more in anxiety than language, and needs its own treatment approach. Mention it to your pediatrician and ask for a referral to an SLP or child psychologist who has experience with it. The strategies for helping a late talker stay largely the same across ages (modeling, expansion, following the child's lead), but what you're working on shifts. At 2, you're focused on single words and early combinations. At 4, a late talker is more likely to need help with sentence grammar, telling stories about what happened, and building out vocabulary. A 4-year-old with a persistent, unevaluated delay really should see an SLP, since catching up on their own is far less likely at that age than at 2. Bilingual exposure doesn't make delays worse. Bilingual children often split their words across two languages, which can make the count in any single language look low, but their combined vocabulary matches monolingual peers. Bilingualism itself doesn't cause delays. If a child is behind in both languages, that's worth evaluating, but telling bilingual families to drop one language is outdated advice that ASHA no longer supports.Sources
- American Speech-Language-Hearing Association (ASHA), Late Language Emergence: Late talkers defined as children with fewer than 50 words or no two-word combinations by 24 months; parent coaching is an evidence-based service delivery model for toddlers
- Rescorla LA, Journal of Speech Language and Hearing Research, 2011: Approximately 70 to 80 percent of late talkers identified at age 2 catch up by age 3, though predicting which children will catch up remains unreliable
- American Academy of Pediatrics, Developmental Surveillance and Screening: AAP recommends developmental surveillance at every well-child visit and formal screening at 9, 18, and 30 months
- Yoder PJ & Stone WL, Journal of Speech Language and Hearing Research, 2006: Parent-implemented naturalistic developmental behavioral interventions produced significant gains in expressive communication in toddlers with autism
- Whitehurst GJ et al., Developmental Psychology, 1988 (dialogic reading original trial): Dialogic reading techniques including open-ended questions and pausing before predictable words produced significant gains in expressive vocabulary in children under 5
- Tamis-LeMonda CS et al., Journal of Child Language, 2001: Maternal responsiveness at 18 months predicted expressive vocabulary at 24 months independent of maternal education or socioeconomic status
- Storkel HL & Maekawa J, Journal of Speech Language and Hearing Research, 2005: Variability of context across word-learning exposures accelerates generalization of new vocabulary in toddlers
- McCune L, Developmental Psychology, 1995: Children who show more symbolic play at 18 months tend to have larger expressive vocabularies at 24 months
- Prizant BM & Duchan JF, Journal of Speech and Hearing Disorders, 1981: Echolalia in autistic children carries communicative intent even when phrases do not appear contextually appropriate to the listener
- US Department of Education, IDEA Part C regulations, 34 CFR Part 303: IDEA Part C guarantees early intervention services from birth through age 2; federal law requires evaluation within 45 days of referral
- AAP Council on Communications and Media, Pediatrics, 2016: AAP recommends no screen time for children under 18 months other than video chat; background TV reduces quantity and quality of parent-child verbal interaction
- CDC Developmental Milestones, Learn the Signs Act Early: CDC developmental milestone norms for expressive language at 12, 18, 24, 36, and 48 months