10-Minute Speech Practice That Doesn't Require Sitting Still

Last updated 2026-07-10

Why do late talkers struggle to ask questions in the first place?

Asking a question is one of the harder things a young child can do with language. It requires knowing you don't have some piece of information, believing another person does, and having enough grammar to invert a sentence or attach a question word. For a child whose language system is still under construction, that's a lot of demands stacked on top of each other.

The American Speech-Language-Hearing Association notes that question forms emerge in a rough developmental order: single words and rising intonation come first, then 'what' and 'where', then 'who', then 'when' and 'why' and 'how' ^[1]. Late talkers often have the intent to ask something but lack the verbal scaffolding to form the request. They pull your hand toward a cabinet instead of saying 'open' or 'what's in there?' The drive is there. The form isn't yet.

For children on the autism spectrum, there's an extra layer. Research on asking for information (sometimes called 'manding for information') found it to be one of the later-emerging verbal behaviors in autistic children, and it often needs explicit teaching rather than showing up on its own the way it does for neurotypical peers ^[2]. That doesn't mean it can't be learned. It means you have to be more deliberate about setting up the conditions for it.

Echolalia shapes how some children approach questions too. A late talker may echo the question back to you rather than answering or generating their own. That's not a failure. It's evidence of phonological memory, and it gives you something to build on. Our article on echolalia covers how to work with that behavior rather than against it.

What is the right order to teach question words to a late talker?

Order matters here. You don't teach 'why' before 'what', and working against the sequence makes everything harder. The research on acquisition order is pretty clear.

Here is the approximate developmental order, with typical ages of emergence in neurotypical children for reference ^[1][3]:

For most late talkers, you work on 'what' and 'where' first. Full stop. Jumping to 'why' because it seems more interesting will frustrate a child who hasn't locked in the earlier forms.

Two notes on the table. These are medians from typically developing children, and late talkers may show a compressed or reordered version. Treat it as a guide, not a checklist. If your child's speech-language pathologist has a different target based on their assessment, follow that. The SLP has information you don't.

What does 'modeling questions' actually look like at home?

Modeling means you say the question out loud, in the moment, as if you were the child wondering it. You're not asking your child to repeat after you. You're narrating the internal question a curious person would have in that situation.

You open the fridge and say, out loud, 'What's in here? I wonder what's in here.' Then you answer it yourself: 'Oh, apples!' No demand that the child says anything. You've just handed them one more exposure to the form.

Speech therapists call this 'aided language stimulation' or 'self-talk', and the evidence base is solid. A systematic review in the American Journal of Speech-Language Pathology found that naturalistic language intervention, including modeling and responsive interaction, produced significant gains in expressive vocabulary and early grammar for children with language delays ^[4]. Naturalistic is the key word. The gains come from embedding the input into real activities, not from sitting at a table with flashcards.

For modeling to do its job, you repeat the same question form dozens of times across many different situations. Not dozens of times in one sitting, which turns into pressure. Dozens of times across a week, spread through bath time, snack, the car, getting dressed. ASHA recommends at least 20-30 meaningful exposures to a target form per week for children in speech therapy ^[1]. At home, aim for the same density.

One honest note: modeling feels awkward at first. You will feel like you're talking to yourself. That's normal. Most parents adjust within a week.

Which everyday routines work best for question-asking practice?

The best routines already have built-in uncertainty. Uncertainty is the natural trigger for questions. If your child already knows what's coming, there's no reason to ask.

Here are routines that create real information gaps:

Snack or meal prep. Keep some snacks in opaque containers or paper bags. Before opening, model 'What's in the bag? I wonder what's in here.' Over time, children who are ready start initiating that question form themselves because the uncertainty is genuine.

Books with hidden elements. Lift-the-flap books, or simple picture books where you cover part of a page with your hand, work well for 'what' and 'where'. Point to the covered area and model 'Where is it? Where did it go?' Then reveal. The reveal is the reward.

Toy play with missing pieces. Take one piece from a familiar puzzle or toy before play starts. When your child notices it's gone, that's your moment. Model 'Where's the piece? Where did it go?' You're not forcing anything. You're meeting a need they already have with the words that fit.

Getting dressed. Hide one sock. Ask 'Where's the other sock?' Sounds trivial. It isn't. You've created a real question moment inside a routine your child does every single day.

Surprise bags from the park or yard. Put an interesting object (a rock, a leaf, a small toy) in a paper bag and let your child feel it without looking. Model 'What is it? What's inside?' This one works well for children who respond to sensory input.

The pattern across all of these is the same. Create a real information gap, model the question form that fits, then answer it or let the child answer it. No pressure, no demand, just supply.

What are the best question-asking games for late talkers at home?

Games work because they lower the stakes. A child who is playing isn't thinking about whether they're saying words correctly. They're thinking about the game. That drop in self-monitoring can open a window for new language.

Here are games that target question forms, organized by stage:

For children working on 'what' (roughly 12-24 months developmental level): Peekaboo variations are underrated. Instead of 'peekaboo', say 'Where's mama? There she is!' You're modeling 'where' every single time. 'What is it?' guessing bags (described above) fit here too.

For children working on 'where' (roughly 18-30 months developmental level): Hide-and-seek with toys rather than people. 'Where's the bunny? Let's find the bunny!' Model the question, do the search together, find the item. Repeat with different objects. Simple. It works.

For children working on 'who' (roughly 24-36 months developmental level): Family photo games. Hold up a photo and wait. If your child doesn't initiate, model 'Who's that? Who is it?' Then answer. Later you can stretch to 'What's grandma doing in this picture?'

For children working on 'what doing' (roughly 24-36 months developmental level): Action figure or doll play with a simple script. Move a figure and say nothing. Wait a beat. Then model 'What's he doing?' and answer yourself: 'He's jumping!' This one pairs well with cause-and-effect toys.

For children working on 'why' (roughly 36-48 months developmental level): Storybooks are your best tool. After a key event, close the book and model 'Why did he do that? Why?' Then think aloud with a simple answer. Don't expect the child to answer 'why' for a while. Modeling 'why' in context is the first step.

None of these require buying anything, and that's the point. The evidence does not support the idea that specialized toys or apps beat well-run naturalistic interaction with common household objects ^[4].

How does wait time help a late talker ask more questions?

Wait time is one of the most powerful and least-used tools parents have, and it costs nothing. When you create a question moment, then wait silently for 5 to 10 seconds instead of filling the silence, you give the child processing time and a clear turn to communicate.

The concept comes from classroom research but applies straight to home practice. Most adults wait about 1 to 2 seconds before jumping in. For a child with a language delay, that's nowhere near enough.

Hart and Risley (1995) documented that children with larger vocabularies had caregivers who were more responsive and less directive in conversation ^[5]. Responsiveness includes waiting. When you wait, you signal that you expect a contribution. That expectation, held gently without pressure, is an invitation to talk.

Here's how to do it. Create the information gap, model the question once, then look at the object or situation expectantly rather than at the child. Staring at the child can feel like pressure. Looking at the thing you're both wondering about is more like a shared puzzle. Wait a full count of five. If nothing comes, model the question again and answer it yourself. No correction, no 'say it', no 'use your words'. Just supply and move on.

Over weeks, many children start to fill that wait with something, even if it's a vocalization or a point. That's progress. The full question form often comes later.

Should I use AAC tools or visual supports to help with question asking?

Yes, and this is where a lot of parents miss a real opportunity. Augmentative and alternative communication supports, including picture symbol boards, core word displays, and speech-generating devices, are not a last resort for children who can't speak at all.

ASHA's position is that AAC supports language development across all children, including those who already have some speech ^[6]. Giving a child a visual symbol for 'what' or 'where' hands them a way to ask before the verbal form is available.

The simplest version is a small laminated card with two or three question word symbols that you keep in your snack area or reading corner. When a question moment shows up, you touch the 'what' symbol as you say 'what's in here?' Now you've given both the spoken and the visual model at once. That dual-channel input helps children who process visual information more reliably than sound.

Families deeper into AAC will find that most full-featured aac devices have core word pages that include question words. If your child uses a device, make sure question words are programmed and sitting on the main page, not buried three screens deep.

One clarification worth making: using visual supports or AAC does not reduce a child's motivation to talk. A 2006 meta-analysis in the American Journal of Speech-Language Pathology found no evidence that AAC intervention suppresses speech development, and found positive effects on both communication and speech ^[7]. If someone tells you a picture board will make your child 'lazy' about talking, that claim has no support behind it.

How do I know if my child is making progress with question asking?

Progress with question asking is nonlinear, and early on it usually looks like something other than actual questions. Here's what to watch for, roughly in order of emergence:

1. The child pays attention when you model a question (even just eye contact or turning toward the object). 2. The child echoes the question word back, even without intonation ('what', said flatly). 3. The child uses a gesture (pointing, reaching) paired with a vocalization in a question-like context. 4. The child uses a question word in isolation ('what?') to request information. 5. The child produces a two-word question form ('where ball?', 'what that?'). 6. The child produces a grammatically complete question with correct inversion.

Most late talkers spend weeks at steps 1 through 3 before moving to step 4. That's not stalling. That's the process.

Keep a simple log in your phone's notes app: date and what you observed. 'Said where? pointing at cabinet' with a date is useful information for your SLP. Patterns across weeks tell you things individual days can't.

If your child has had consistent modeling for 8 to 12 weeks with no movement through these stages at all, raise it with an SLP. It may mean question asking isn't the right target right now, or that something else needs assessment. Early intervention services can evaluate children under 3 at no cost to families in the United States under IDEA Part C ^[8].

What should I avoid when teaching question asking at home?

A few patterns show up again and again in research on why home practice stalls, and they're worth naming directly.

Asking your child to repeat questions on demand. 'Say: where is the ball. Say it.' This puts the child in a performance role, and performance anxiety suppresses the very language you're trying to build. The research on demand in communication is consistent: child-led or low-demand contexts produce more spontaneous language than adult-directed drill ^[4].

Answering too fast. If you model 'what's in the bag?' and answer it yourself with no wait, you've given no communicative opportunity. You've done a monologue.

Correcting the form. If your child says 'where ball go?', resist the urge to say 'say it the right way'. Instead, expand: 'Oh, where did the ball go? I wonder where the ball went!' You've handed them a natural model of the correct form without signaling theirs was wrong.

Using questions as tests. 'What color is that? Tell me.' You already know the answer. The child knows you know. These don't feel like real questions because they aren't. Real questions have genuine information gaps. Test questions are drills in disguise, and children figure that out fast.

Doing all of this in one long session. Twenty minutes of concentrated question practice is harder for both of you than five minutes spread across four natural moments in the day. Distributed practice beats massed practice for language learning ^[3].

If your child is also working on speech sound production, our article on apraxia of speech covers how motor speech factors might change the way you structure modeling.

When should I get a speech-language pathologist involved instead of just doing home activities?

Home activities supplement professional support. They don't replace it. That's not a hedge, it's just accurate.

The American Academy of Pediatrics recommends developmental surveillance at every well-child visit and formal screening at 18 and 24 months ^[9]. If a child is flagged at either visit, or if a parent raises a concern, a referral for a speech-language pathology evaluation is the right next step. Trust your instincts here. Parent concern is one of the most reliable early indicators of language delay.

Reach out to an SLP if:

• Your child has no words by 12 months, fewer than 50 words by 24 months, or no two-word combinations by 24 months ^[9][12].
• Your child's question asking isn't progressing despite consistent modeling for 10 to 12 weeks.
• Your child shows significant frustration or distress around communication.
• You suspect something beyond a language delay, such as autism, hearing loss, or motor speech difficulties.

For families in the U.S., children under 3 qualify for free evaluation and services through Early Intervention (IDEA Part C) ^[8]. Children 3 and older can be evaluated through the public school district at no cost. You don't need a doctor's referral to request either evaluation, though a referral can help.

Our speech therapy and speech therapist overview walks through what to expect from a formal evaluation and how to find a qualified provider.

If you want structured support between therapy sessions, the Little Words app has an AI-guided speech companion that helps parents model language during everyday routines, with targets you can align to what your SLP is working on. It's not therapy. It's practice scaffolding.

Do these activities work differently for autistic children compared to other late talkers?

The honest answer: mostly the same principles apply, with a few specific adaptations.

Autistic children often have interests that are narrower or more intense than their peers. That's an advantage for this work. A child who is deeply into trains will tolerate far more repetitions of a question model if the question is about trains. The content interest supplies the motivation that keeps engagement up. Find the interest and build your question activities there first.

Many autistic children also have stronger visual processing relative to auditory processing ^[10]. That makes AAC supports and visual symbols especially useful, not as a fallback but as a primary channel. Pairing your spoken question with a visual symbol or with pointing to a picture is often more efficient than spoken input alone.

Sensory factors matter too. A question-asking activity in a loud, chaotic environment won't produce the same results as the same activity in a quiet setting. Dysregulation competes directly with language learning. A child who is overwhelmed is not in a state to process or produce new language, no matter how good your model is.

For autistic children who use scripted or echoed language, question forms sometimes show up first as scripts ('I wonder where it went') before they generalize to flexible, spontaneous questions. That scripted route is a real pathway, not a dead end. Our article on autism spectrum speech therapy goes deeper on this.

One more thing. 'Manding for information', the technical term for asking questions in applied behavior analysis, has a solid evidence base for autistic learners specifically. Published research found that question asking could be systematically taught to minimally verbal autistic children through naturalistic behavioral intervention ^[2]. The techniques overlap a lot with the naturalistic modeling approach described here.

What questions should I ask my child's SLP about question-asking targets?

Walking into sessions with specific questions makes the partnership more productive. Here are the ones worth asking.

'What question form are you targeting right now, and what's next after that?' This tells you where you are in the sequence and what to model at home.

'What level of response are we expecting, and what counts as progress?' The SLP might count any vocalization as a response, or they might be looking for a two-word form. Knowing this lets you track the same thing at home.

'What's the current frequency target?' If the SLP is aiming for 20 opportunities per session, you need to know how many sessions per week that is and how to match that density at home.

'Are there activities you've tried in session that seemed to work well?' Therapists develop instincts about individual children. Ask for those observations.

'Should I correct errors or just expand?' Most SLPs say expand rather than correct. Confirm it anyway, because approaches vary and you want to stay consistent with the sessions.

'What should I stop doing?' Sometimes parents are running patterns that work against the target without knowing it. Asking directly gives the SLP room to tell you.

If your child gets services through a school or early intervention program, you have the right to a copy of their Individualized Education Program or Individualized Family Service Plan, which spells out language goals including communication targets. Those documents are yours, and they're written in language you can act on at home ^[8].

Frequently asked questions

At what age should my child be asking questions?

Most children start using rising intonation to ask simple questions around 12-18 months, produce 'what' and 'where' by 18-24 months, and ask 'why' and 'how' by age 3-4. These are medians for typically developing children. Late talkers follow the same order on a shifted timeline. If your child is using no question forms by 30 months, an evaluation by a speech-language pathologist is a reasonable next step.

What if my late talker just echoes my question back instead of asking their own?

Echoing a question is a sign of phonological memory and communicative awareness, not a problem. It's often a stepping stone toward spontaneous question use. Respond as if the echo were a real question: if you said 'where's the ball?' and your child echoes 'where's the ball?', say 'yes! Where IS the ball? Let's look!' You're validating the attempt and extending the interaction. Our article on echolalia has more on working with this pattern.

How many times do I need to model a question word before my child will use it?

Nobody has precise data on this for late talkers specifically. Research estimates suggest neurotypical children need 10-15 exposures for a new word, but late talkers often need many more, with some clinician guidelines pointing to 50-100 meaningful exposures before spontaneous use. Meaningful is the key word: exposures embedded in real situations where the child actually wants the information count far more than drill repetitions.

Can I use YouTube videos or TV to help teach question asking?

Screen-based language exposure has weak evidence as a standalone strategy, especially for children under 2. The AAP recommends limiting screen time and prioritizing live interaction for language development. That said, watching a video together and pausing to model questions ('what's that? Where is he going?') turns passive screen time into interactive input. The interaction is what matters. A screen without an adult engaging alongside it is much less effective.

What's the difference between a late talker and a child with a language disorder?

A late talker is typically a child under 3 with fewer words than expected but otherwise typical development in play, social interaction, and comprehension. Many catch up by age 3-4 without intervention, though a meaningful percentage do not and go on to have lasting language difficulties. A language disorder is a longer-term profile that persists beyond the preschool years and affects comprehension or expression or both. Only a formal evaluation can tell the two apart.

Should I teach question asking in my home language or in English?

Teach question asking in the language your child hears most and the language you speak most fluently. Bilingual and multilingual children may acquire question forms at slightly different rates in each language, but the developmental sequence within each language is the same. Your strongest language produces better-quality models. Research on bilingual development does not support the idea that speaking one language at home delays a second.

Do question-asking activities also help with answering questions?

Yes, and the relationship goes both ways. Children who understand question forms well enough to ask them also understand what's expected when they receive one. Modeling questions and answers together, which you do naturally when you ask a question aloud and then answer it yourself, builds both sides of the conversational exchange at once. Most SLPs work on asking and answering in parallel for this reason.

What if my child gets frustrated during question-asking activities?

Stop the activity and move to something your child enjoys. Frustration signals that the demand level is too high or the situation is too pressured. Regroup and go back to pure modeling with no expectation that the child will produce anything. Language learning needs a regulated nervous system. If your child is consistently frustrated during communication activities, raise it with your SLP; it may mean the targets need adjusting or that other factors need assessment.

Are there specific toys or materials I should buy to help with question asking?

Honestly, no. The research does not show that specialized toys produce better language outcomes than naturalistic interaction with common household objects. Lift-the-flap books, opaque bags, familiar toys with missing pieces, and family photos work as well as anything sold for speech development. Save your money and put the time into consistent modeling during routines you already have.

How do question-asking activities connect to early literacy skills?

Question forms are tied closely to narrative comprehension and reading readiness. Children who ask 'why' and 'what happens next' during shared book reading show better reading comprehension in early elementary school. Modeling question words during book reading, especially 'what', 'where', and 'why' about story events, builds both the language targets and the inferential thinking skills that support later reading.

What if my child only communicates through pointing or gestures right now?

That's a real and important communication foundation to build on. When your child points, narrate the question their point implies: if they point at the cabinet, say 'what's in there? You want to know what's in there!' You're mapping their intent onto question language. Gestures and points that get responded to consistently tend to turn into words over time. Never discourage pointing or gesturing in favor of speech; both channels support development.

My child asks the same question repeatedly. Is that normal for late talkers?

Repetitive questioning is common in both late talkers and autistic children and can serve different functions. For some children it regulates them, for others it's a scripted routine that provides comfort, and for others it's genuine uncertainty about an answer that's hard to hold onto. Respond consistently and briefly each time. If the repetition is intense or distressing, bring it up with your SLP or pediatrician, because the underlying function affects how you respond most helpfully.

Sources

1. ASHA (American Speech-Language-Hearing Association) - Late Language Emergence Practice Portal: Question forms typically emerge in a developmental order, with 'what' and 'where' preceding 'why' and 'how'; ASHA recommends 20-30 meaningful exposures to a target form per week during intervention.
2. Journal of Applied Behavior Analysis - manding-for-information research (Sundberg et al.): Manding for information (question asking) is a later-emerging verbal behavior in autistic children and often requires explicit instruction; it can be taught through naturalistic behavioral intervention.
3. Paul, R. (2007) Language Disorders from Infancy Through Adolescence - typical question development sequence: Developmental sequence for question words: 'what' and 'where' emerge by 18-24 months, 'who' by 24-30 months, 'when' and 'why' by 36-48 months in typically developing children.
4. American Journal of Speech-Language Pathology - systematic review of naturalistic language intervention: Naturalistic language intervention approaches including modeling and responsive interaction produced significant gains in expressive vocabulary and early grammar for children with language delays; specialized toys showed no advantage over common objects.
5. Hart & Risley (1995) Meaningful Differences in the Everyday Experience of Young American Children - caregiver responsiveness and vocabulary: Children with larger vocabularies had caregivers who were more responsive and less directive in conversation; wait time and responsiveness are associated with better language outcomes.
6. ASHA - Augmentative and Alternative Communication (AAC) Practice Portal: ASHA's position is that AAC supports language development across all children including those with some speech; it is not a last resort.
7. American Journal of Speech-Language Pathology - Millar, Light & Schlosser (2006) meta-analysis on AAC and speech development: A meta-analysis found no evidence that AAC intervention suppresses speech development, and found positive effects on both communication and speech production.
8. U.S. Department of Education - IDEA Part C Early Intervention Program: Under IDEA Part C, children under 3 in the U.S. qualify for free evaluation and early intervention services; children 3 and older can be evaluated through public school districts at no cost; parents do not need a doctor's referral.
9. American Academy of Pediatrics - Developmental Surveillance and Screening: The AAP recommends formal developmental screening at 18 and 24 months; a child with fewer than 50 words by 24 months or no two-word combinations by 24 months warrants referral for speech-language evaluation.
10. Journal of Autism and Developmental Disorders - visual processing in autism review: Autistic children often show stronger visual processing relative to auditory processing, supporting the use of visual symbols and AAC as primary rather than supplementary channels for language input.
11. ASHA - Late Language Emergence Practice Portal: Late talkers are children under 3 with fewer words than expected but otherwise typical development; a meaningful percentage do not catch up without intervention and go on to have persistent language difficulties.
12. National Institute on Deafness and Other Communication Disorders (NIDCD) - Speech and Language: Children typically produce single words by 12 months, 50 or more words by 24 months, and two-word phrases by 24 months; delays beyond these milestones warrant professional evaluation.