How to expand sentence length using an AAC device

Last updated 2026-07-11

If you want your child to string together longer messages on their AAC device, the fastest path is modeling language just above their current level (this is called aided language input), adding one word at a time to what they already do, and building in real reasons to talk throughout the day. Most kids move from single words to two-word combinations over weeks to months of steady modeling, not days, so plan for the long game.

What "expanding sentence length" means on an AAC device

Speech-language pathologists measure this with mean length of utterance, or MLU: the average number of meaningful units a child combines in one message. For a child using an AAC device, that means counting how many symbols or words they string together per turn.

A child at MLU 1.0 hits one button per message: "More." "Eat." "Go." That's genuine communication, and the goal was never to make it look more typical. The goal is giving your child the tools to say more when they have more to say, because you can't argue, explain, or tell a story with a single symbol.

Researchers track MLU in AAC users against the same developmental benchmarks used for spoken language. A 2019 study in the American Journal of Speech-Language Pathology found that AAC-supported interventions reliably increased MLU in children with autism spectrum disorder, with gains averaging 0.5 to 1.5 morphemes over 12 to 20 weeks of structured intervention ^[1]. That's a realistic number to hold onto. Expanding MLU isn't about pushing a child to pack more into every message. It's about making longer combinations available, familiar, and worth reaching for.

Why kids get stuck at one-word messages

Most kids using AAC plateau at single words for a few overlapping reasons, and the device itself is rarely the problem. Single words work, plain and simple. If touching "eat" gets food every time, there's no pull toward adding "I want" in front of it. The environment does the work for the child, a pattern documented in AAC research as communicative efficiency ^[2]. Layout is another common culprit. Many AAC apps default to category-based screens (food, feelings, places), which is fine for requesting but makes subject-verb-object combinations physically clunky to build. You can't put together "I see dog" if those three words live on three different screens with six taps between them. The piece that gets missed most often: the child may simply never have heard longer messages modeled on the device. Spoken language grows through immersion in richer input, and AAC works the same way. If everyone around a child speaks in single symbols, that's what the child learns to do. ASHA's guidance on augmentative and alternative communication names aided language stimulation, modeling on the device itself, as a core strategy for building longer utterances ^[3]. For children with childhood apraxia of speech, motor planning adds another layer entirely: sequencing multiple button presses is its own challenge, separate from language itself. The same goes for children with apraxia of speech, where motor and language goals often need to run side by side. And some children are still learning that the device is for communicating, not a task they perform for therapy. You can't build sentence length on that foundation yet, which is where early intervention services can help lay groundwork first.

Aided language input, and why it works

Aided language input (sometimes called aided language stimulation, or ALgS) means pointing to or activating symbols on the device while you talk, so your child sees and hears language modeled in the same medium they're expected to use. Put simply: you use the device too. The evidence behind this is solid. A 2017 systematic review in Augmentative and Alternative Communication looked at 19 studies of aided language input across a range of ages and diagnoses, and found it consistently increased both vocabulary and utterance length in AAC users ^[4]. The evidence is strongest for children with autism and developmental disabilities, which covers a lot of the kids reading this. In practice, it looks like narrating what's happening through the device. During a snack, you touch "I" + "want" + "banana" while saying the words aloud. You're not asking your child to imitate you. You're just showing, again and again, what the device can do. A few things help this actually land:

• Model one level above your child's current output. If they use one word, model two; if two, model three. Skip ahead to full sentences and you lose them.
• Do it during activities your child is already motivated by, not drills.
• Model imperfectly sometimes ("more shoe"?) so your child sees the device as flexible, not a performance that has to be exact.
• Expect it to feel slow and awkward at first. It speeds up.

ASHA lists aided language stimulation as a recommended strategy in its AAC Practice Portal ^[3], which reflects consensus among practicing SLPs nationally, so this isn't a fringe technique.

Moving from one word to two

The two-word stage is the first real sentence milestone, and for most AAC users it's the hardest jump to make. Here's an approach grounded in what the research supports. Start by confirming your child has fast access to core words: "I," "want," "more," "go," "stop," "help," "like," "not," "it," "that." These small words do most of the work in daily language, accounting for roughly 80% of what any person says regardless of topic ^[5]. On a well-designed system, they should live on the home page or one tap away; if they don't, fix the layout before anything else. From there, pick two-word combinations that pay off widely. Carrier phrases like "want + [noun]," "more + [noun]," "stop + [verb]," and "I + [verb]" open up hundreds of sentences using the same two-button sequence. When you model these, slow down. Pause between "want" and "cookie" so your child has time to register that two separate things just happened, something that matters especially for children with autism or apraxia, where processing speed and motor planning can differ from typical development. After they touch one symbol, try expectant pausing: look at them with an open, waiting expression, don't prompt verbally, don't point to the next word, just wait. Five seconds feels longer than it is, and a good number of two-word combinations show up in exactly that pause. If your child produces one word instead of two, treat it as the successful communication it is, then model the fuller version naturally: "Cookie! Yeah, you want cookie," touching both symbols as you say it. And if your child sees an SLP for speech therapy, ask what their specific plan is for MLU expansion. A good clinician can name target combinations and how they're tracking progress.

Vocabulary layouts that support longer sentences

Device setup matters more than most parents realize, and it's an area where you have real influence. Most AAC apps use grid layouts, but how those grids are organized makes all the difference. Category-based layouts ("Food," "People," "Activities") work fine for requesting single items but slow sentence building to a crawl, since every phrase means hopping between screens. Core-first layouts flip that: the most grammatically useful words sit on the home screen, with fringe vocabulary built out from there. Proloquo2Go (using the LAMP Words for Life layout in its updated version), TouchChat with WordPower, and Snap Core First are popular systems with strong core-first setups. LAMP (Language Acquisition through Motor Planning) organizes vocabulary so consistent motor patterns build automaticity, which matters a lot for children with apraxia ^[5]. None of this comes cheap: costs range from about $200 for a tablet with a free or low-cost app up to $8,000 or more for a dedicated speech-generating device with a funded core-word system, though Medicaid and private insurance often cover dedicated devices when an SLP documents medical necessity ^[6]. A few features worth looking for: a core home page that puts pronouns, verbs, and modifiers within reach without navigating away; consistent button placement, since the same word in the same spot every time builds motor memory faster; and a sentence strip or display bar across the top of the screen that shows what's been touched so far, giving your child visible proof that two symbols just happened. Predictive vocabulary, where the system guesses the next likely word, can help too, but use it cautiously: it can support expansion or it can breed dependence on the prediction instead of independent building. Worth asking your SLP before turning it on broadly. If your child's current layout isn't supporting combinations, that's worth raising with an SLP who specializes in AAC. They can assess the setup and recommend changes that fit your child specifically.

What are the best daily activities for practicing longer AAC sentences?

Structured practice works better when it doesn't feel like practice. The goal is to set up moments where the child actually wants to say something, and where a longer message gets a better result than a short one.

Snack time is a natural fit for two-word combinations like "want cracker." Reading books with repetitive lines invites three-word strings such as "bear go home." Bubbles and other cause-effect toys pull for "more blow." Video or show breaks give a reason to combine words into "I want Bluey." Playdates and sibling play tend to produce phrases like "stop that please," while narrating the morning routine, brushing teeth, getting dressed, can stretch toward four-word sentences like "I brush my teeth."

What ties these together is that the child genuinely wants to communicate in each one. The book-reading example is especially well studied: shared reading with aided language input modeling has produced MLU gains across several controlled studies with preschool-aged AAC users ^[4].

One more thing worth saying plainly: don't hold back the cracker or the bubbles until your child produces a longer utterance. That turns communication into a compliance game. Model the longer form, accept and reward the shorter attempt, and trust that length builds over time with enough input.

How long does it take to see progress?

This depends on enough variables that a single number would mislead you, but the research gives some rough footing. The 2019 AJSLP analysis mentioned earlier found MLU gains averaging 0.5 to 1.5 morphemes over 12 to 20 weeks of structured intervention ^[1], which works out to roughly 3 to 5 months of steady practice, not days or weeks. Some kids move faster, some slower. Pace depends on how long the child has used AAC, how consistently people model language across settings, whether the vocabulary layout fits the child's motor and cognitive access, and whether something like autism spectrum disorder is affecting how skills generalize.

When parents practice daily at home alongside weekly SLP sessions, progress tends to outpace clinic-only work. A 2020 study in the Journal of Speech, Language, and Hearing Research found that parent-implemented AAC interventions produced gains comparable to clinician-implemented ones, as long as parents received adequate training first ^[7]. That training part matters: watching a few videos online is a start, not a substitute for guided coaching.

If there's been no movement in utterance length after three months of consistent, well-run modeling, that's worth bringing back to your SLP. It might mean the vocabulary layout needs adjusting, the targets need to shift, or there's a motor or access barrier nobody has caught yet. And progress rarely moves in a straight line anyway. Most children show spurts followed by plateaus, and a plateau usually just means they're consolidating what they've learned before jumping ahead.

Do visual supports help build longer sentences?

Yes, with some nuance. Visual sentence strips, picture-supported scripts, and printed carrier phrase cards can all support longer AAC messages because they take the load off working memory: instead of holding a whole sentence in mind while operating the device, the child glances at the strip and knows what's next.

They work best under a few conditions. They should use the same symbols as the AAC device, not a different set of images or vocabulary. They should be faded out gradually as the child internalizes the pattern. And they work better tied to real communication than to drill: an "I want + ___" strip taped by the snack cabinet does something useful, while the same strip used only at a table drill does less.

For children with echolalia, visual and scripted supports interact in interesting ways. Some children can learn to map scripted phrases they already say aloud onto AAC sentence frames, essentially translating spoken scripts into device sequences. It's a legitimate approach and probably underused. The piece on echolalia meaning goes into more detail on how that connects to functional language.

Social stories built with AAC symbol sequences are another variation worth trying: a story showing a character who "wants" something, using the same symbols your child already uses, can prime sentence combinations in a low-stakes way. Underneath all of this is a simple research principle: hearing a word, seeing its symbol, and watching a gesture together supports language learning better than any single channel alone ^[4]. Visual supports are one piece of that, not the whole intervention.

What does an SLP do here, and how do you find one?

An SLP is the clinical anchor for this work. They assess baseline MLU, set targets, evaluate the device layout, train you on modeling, and track progress. But you spend more waking hours with your child than any therapist does, so your daily follow-through carries as much weight as the weekly session, maybe more.

Look for an SLP with real AAC experience, since not all of them have it. ASHA's directory at asha.org lets you filter by specialty area, including AAC ^[3]. Find someone who has worked with your child's diagnosis and who spends session time training you as much as working with the child.

If in-person access is limited, online speech therapy has expanded a great deal since 2020, and the evidence for telepractice-delivered AAC services is strong enough that ASHA recognizes it as an appropriate service delivery model ^[8].

On the school and insurance side: children who qualify under IDEA (Individuals with Disabilities Education Act) are entitled to speech-language services through their IEP when a communication impairment affects educational performance ^[9], and that includes AAC support. But school-based goals tie to educational access rather than broad language development, so many families end up needing both a school SLP and a private one to cover everything.

One app worth mentioning: Little Words (littlewords.ai) offers an AI speech companion built to support families working on these same language goals between therapy sessions. It isn't a replacement for clinical care, but it's built to give parents more guided practice time, and the research suggests that time is what moves the needle.

Common mistakes parents make

Most of these come from good intentions. One is prompting instead of modeling: "Say 'I want juice.' Touch 'I.' Now touch 'want.'" This turns AAC into a compliance task and can increase prompt dependence without increasing spontaneous language. It works better to model it yourself, touch "I want juice," hand over the juice, and move on.

Another is expecting carryover to happen on its own. A child might produce three-word combinations in therapy and none at home. Generalization doesn't happen automatically; you have to practice the same combinations across settings and people. That's not a failure on the child's part, it's how language learning works for everyone, autistic children included ^[1].

Skipping steps is another common one. If a child is at MLU 1.0, jumping straight to four-word sentences because they're "supposed to" be there by some age doesn't work. Build in order, one word to two, two to three, and let each level settle before moving up.

Reshuffling the device layout too often is a quieter mistake. Every reorganization erases motor memory, since children build speed through repeating the same physical sequence. Move a word and the child has to relearn where it lives. A stable layout beats a constantly "improved" one.

Last, some families treat AAC as a last resort, worried it will replace speech. The evidence doesn't support that fear. A 2006 review and several later studies found that AAC use does not suppress speech development and often supports it ^[10]. Starting earlier just means more time to build language. If you're still weighing that decision, the early intervention article has useful background.

How do I track whether my child's AAC sentence length is actually improving?

You need some kind of measurement, otherwise you're just guessing. Here's a practical way to do it that doesn't require clinical training.

Once a month, record 10 to 15 minutes of your child using the device during a natural activity: snack time, play, a favorite book. Count the number of symbols they touch per turn, then average them. That's your MLU estimate for the month. A simple spreadsheet is enough to track it. This is essentially what SLPs do at a more formal level: ASHA's practice guidelines recommend language sampling as a primary measure of AAC communication progress ^[3].

Each separate symbol press counts as one unit. If the child touches "I" + "want" + "juice," that's MLU 3. If they touch the same symbol three times in a row ("more more more"), that only counts once, since it's the same word repeated. Sentence-ending symbols like "please" or "thank you" count if the child chooses to select them on purpose.

Real progress looks like this: the child starts messages on their own instead of just answering questions, combines symbols nobody explicitly taught them, produces longer messages in more than one setting, and repairs a breakdown in communication by adding more information rather than giving up.

Pay attention if MLU hasn't moved after 8 to 12 weeks of consistent work, if longer messages only show up in therapy, or if the child avoids the device or gets upset when it comes out. Any of those, and it's worth bringing your recordings to your SLP. Actual samples beat descriptions every time; a 30-second clip tells an SLP more than any parent summary can.

Toward the end of a treatment cycle, apps like Little Words (littlewords.ai/start) can give you structured activities matched to your child's current level, along with progress data you can hand to your SLP, which helps close the gap between clinic sessions.

Frequently asked questions

At what age should a child start combining two words on AAC?

In typical spoken language development, two-word combinations show up between 18 and 24 months. For AAC users, the timeline shifts depending on when they started using the device and how consistently aided language input has been provided. Rather than chasing a specific age, aim for one MLU level above wherever the child currently is.

Does modeling on the AAC device really help, or is it just something therapists say?

The evidence backs it up. A 2017 systematic review in Augmentative and Alternative Communication found that aided language stimulation consistently increased both vocabulary and utterance length across 19 studies. It's the most well-supported single strategy for AAC language development, and the key is using it consistently across settings, not just in therapy sessions.

My child only uses AAC to request things. How do I get them to use it for other purposes?

Requesting is the easiest function to teach because it gets reinforced immediately. To build other uses, like commenting, protesting, or asking questions, set up situations where requesting isn't the point. Narrate play with the device, use it to label things that surprise you, model phrases like "I see that" or "that's funny" during shared activities. Commenting shows up once the environment around the child starts rewarding it.

Can a child use AAC and still develop spoken language?

Yes. The worry that AAC suppresses speech isn't supported by research. A widely cited 2006 review found no evidence that AAC use reduces speech development, and several studies found it actually supports verbal output. ASHA and the American Academy of Pediatrics both endorse introducing AAC early. The device gives language a route out while speech skills keep developing alongside it.

What is LAMP and does it help with sentence length?

LAMP stands for Language Acquisition through Motor Planning. It organizes AAC vocabulary so each word has its own consistent motor pattern, building automaticity over time. For children with motor-based speech challenges like apraxia, this reduces the planning load per word, which in theory frees up cognitive resources for combining words. It's most widely studied in young children with autism and apraxia.

How many words should be on my child's AAC device home page?

There's no universal right number, but research on core vocabulary suggests 20 to 40 high-frequency words cover roughly 80% of everyday communication needs. The home page should prioritize pronouns, verbs, and modifiers over nouns, since those are the grammatical glue that makes longer sentences possible. Nouns can sit one tap away.

Should I correct my child when they produce an incomplete sentence on AAC?

No. Correcting turns AAC from a communication tool into a performance test, which suppresses spontaneous use. Instead, respond to what they meant, then model the fuller version. If they touch "juice," say "juice! You want juice," while touching "want" + "juice" on the device yourself. This is called expansion in speech-language research, and it works better than correction.

My child's school SLP says AAC sentence building isn't an IEP goal. Can I push for it?

Yes. Under IDEA, IEP goals have to address the child's needs in the educational environment. If short communication is limiting the child's ability to participate in instruction, talk with peers, or advocate for themselves at school, sentence expansion counts as educationally relevant. You can request a specific MLU goal be added, and if the team resists, bring language sample data and ask them to explain why it isn't included.

What's the difference between core vocabulary and fringe vocabulary on AAC?

Core vocabulary is the set of words used most often across contexts by most speakers: pronouns, common verbs, prepositions, basic modifiers. Studies consistently find that roughly 200 to 400 core words make up about 80% of what people say day-to-day. Fringe vocabulary is specific to a topic or person, like the name of a favorite character or a food they eat often. Both matter, but core words are what drive sentence building.

How do I get other caregivers and teachers to model on the AAC device too?

Consistency across the people around a child is one of the biggest predictors of AAC generalization, and one of the hardest things to pull off. A few things that actually work: short in-person demonstrations (watching someone model for two minutes beats any handout), laminated quick-reference cards with three or four target combinations for the week, and asking the SLP to train teachers directly instead of relaying the information yourself.

Are there specific apps or AAC systems that are better for building longer sentences?

Systems with strong core vocabulary home pages and consistent button placement tend to support sentence growth best. Proloquo2Go with the LAMP Words for Life layout, TouchChat with WordPower, and Snap Core First are commonly used options. The right system depends on the child's motor access, cognitive level, and what their SLP recommends: no app is universally best, fit to the child matters more than brand.

How do I know if my child's AAC device layout is holding back their sentence development?

A few warning signs: the child needs more than two or three taps to combine any subject and verb, core words like "I," "want," "more," and "stop" are buried in category folders, or the child consistently avoids certain types of messages. An SLP with AAC expertise can run a device audit and recommend layout changes. This is a legitimate clinical service, not just a tech preference.

What's a realistic MLU goal for the end of a school year for a five-year-old AAC user?

It depends entirely on where the child starts. A reasonable growth target based on published research is 0.5 to 1.0 morphemes over a school year with consistent intervention. A child starting at MLU 1.0 might realistically reach 1.5 to 2.0. A child already at 2.0 might reach 3.0 with good implementation. Ask your SLP to set a specific, measurable MLU goal instead of a vague "will use longer sentences" objective.

Sources

1. American Journal of Speech-Language Pathology (2019), AAC interventions and MLU in autism: AAC-supported interventions increased MLU in children with ASD by an average of 0.5 to 1.5 morphemes over 12 to 20 weeks of structured intervention
2. Augmentative and Alternative Communication, communicative efficiency and AAC: Single-word requests are reinforced efficiently, reducing motivation to produce longer messages without explicit intervention
3. ASHA Practice Portal, Augmentative and Alternative Communication: ASHA lists aided language stimulation as a recommended strategy for building utterance length in AAC users and endorses language sampling as a primary progress measure
4. Augmentative and Alternative Communication (2017), systematic review of aided language input: A systematic review of 19 studies found that aided language stimulation consistently increased vocabulary and utterance length in AAC users
5. Beukelman & Mirenda, Augmentative and Alternative Communication (4th ed.), core vocabulary research: Approximately 200 to 400 core words account for roughly 80% of what any person communicates in daily interaction, regardless of topic
6. Centers for Medicare & Medicaid Services, speech-generating devices coverage: Medicaid and private insurance often cover dedicated speech-generating devices when an SLP documents medical necessity
7. Journal of Speech, Language, and Hearing Research (2020), parent-implemented AAC intervention outcomes: Parent-implemented AAC interventions produced comparable gains to clinician-implemented ones when parents received adequate training first
8. ASHA, Telepractice in Speech-Language Pathology: ASHA recognizes telepractice as an appropriate service delivery model for AAC services
9. U.S. Department of Education, Individuals with Disabilities Education Act (IDEA): Children who qualify under IDEA are entitled to speech-language services in their IEP if communication impairment affects educational performance, including AAC services
10. American Journal of Speech-Language Pathology (2006), AAC and speech development review: A review found no evidence that AAC use reduces speech development; multiple studies found AAC supports verbal output rather than suppressing it