Parents are sometimes presented with a choice that sounds urgent:
Should we keep working on speech, or should we introduce AAC?
The question assumes that the two approaches compete.
Current evidence does not support that assumption.
Augmentative and Alternative Communication, or AAC, can include gestures, manual signs, picture boards, communication books, speech-generating devices, communication apps, and other tools that supplement or provide alternatives to spoken communication.
For many young children, AAC is augmentative rather than a replacement for speech. A toddler may point, vocalize, use several spoken word approximations, select symbols on a communication board, and use a speech-generating device within the same interaction.
All of those behaviors can be communication.
ASHA's current early-intervention guidance states that AAC can be introduced to infants and toddlers and should be considered early when spoken communication is not adequately meeting the child's needs. ASHA also states that AAC does not hinder speech development and may support language and spoken-language development.
A particularly useful study for understanding this issue is Walters, Sevcik, and Romski's 2021 analysis of toddlers with developmental delays who participated in parent-implemented language interventions.
Its findings directly address one of families' most persistent concerns:
Will a speech-generating device make my child less likely to talk?
The evidence from this study says no.
But the study also tells a more nuanced story about AAC, spoken vocabulary, speech-sound development, receptive language, caregiver involvement, and what clinicians should—and should not—predict from early performance.
The six steps below explain what that evidence means in clinical practice.
Step 1: Replace “AAC vs. Speech” With Multimodal Communication
The dilemma is usually framed as:
AAC
versus
traditional spoken-language intervention.
That may describe the comparison used in a research study.
It should not define everyday clinical practice.
AAC is not one treatment competing against speech.
It is a broad category of communication strategies and systems that can be used alongside speech.
ASHA defines AAC as communication that supplements or compensates for difficulties in speech-language production or comprehension. AAC may include:
- gestures;
- manual signs;
- pictures;
- communication boards or books;
- speech-generating devices;
- communication applications;
- writing;
- and combinations of these methods.
A toddler might say:
“ba”
while reaching toward a bottle and selecting drink on an AAC system.
The clinician does not need to decide which one counts as the child's “real” communication.
Speech, gesture, symbol selection, and device output can all contribute to the child's message.
The Goal Is Communication Access
The better clinical question is:
Does this child currently have a reliable way to communicate what they want, need, notice, reject, ask, feel, and share?
If spoken language is not yet sufficient, withholding another communication method creates unnecessary limitations.
Young children communicate for many purposes:
- requesting
- rejecting
- protesting
- commenting
- asking
- greeting
- sharing attention
- seeking help
- directing another person
- participating socially
AAC can expand those possibilities while spoken-language intervention continues.
ASHA's current early-intervention guidance explicitly states that AAC can be introduced early and that toddlers do not need to wait until a later developmental stage before receiving multimodal communication support.
What Did the Walters Study Actually Compare?
Walters and colleagues analyzed existing data from two randomized controlled trials of parent-implemented language intervention involving toddlers with developmental delays.
The interventions included AAC conditions using speech-generating devices and a spoken-language condition without AAC.
The 2021 study specifically examined:
- speech-sound development
- spoken target vocabulary
- predictors of spoken vocabulary outcomes
A total of 109 children completed the interventions, but only 45 produced spoken target vocabulary words at the end of intervention that could be analyzed phonetically.
That detail is important.
The paper does not show that every toddler using AAC began speaking.
It examines speech outcomes in children who did produce spoken target vocabulary after intervention.
Earlier Randomized Research Provides Additional Context
The earlier randomized comparison by Romski and colleagues involved toddlers with developmental delays who began intervention with fewer than 10 spoken words.
Sixty-eight toddlers were initially randomized to:
Augmented Communication–Input
Augmented Communication–Output
or
Spoken Communication
conditions.
Sixty-two completed the intervention.
Children in the augmented communication groups used both augmented and spoken target words and developed substantially larger target vocabularies than children in the spoken-communication-only condition.
Together, these studies challenge the assumption that AAC must be delayed until spoken-language intervention has “failed.”
Step 2: Understand What the Evidence Says About AAC and Speech
The strongest conclusion from the Walters study is straightforward:
AAC did not reduce speech development.
Children receiving augmented language intervention produced significantly more spoken target vocabulary words than children receiving the spoken-language intervention without AAC.
The researchers also found no statistically significant difference in speech-sound accuracy between intervention groups.
That means AAC did not produce a tradeoff in which increased communication came at the expense of spoken-sound development.
This finding supports early AAC use without requiring families to choose between:
“communication now”
and
“speech later.”
This Finding Is Consistent With the Larger AAC Literature
The Walters study is not the only evidence addressing this concern.
A classic review of AAC and speech production in people with developmental disabilities identified 23 studies. Among the stronger cases included in the review, none demonstrated a decrease in speech production following AAC intervention, and most demonstrated some increase in speech.
A systematic review focusing on autistic children similarly concluded that AAC did not impede speech production. Many studies reported speech increases, although the gains were generally modest.
An updated 2021 review of AAC and speech production in autistic children again found overall improvements in speech production, while emphasizing that spoken gains generally did not exceed the child's AAC communication.
That final point deserves attention.
AAC should not be offered only because adults hope it will eventually produce speech.
It has value because it provides communication.
AAC Success Should Not Be Defined by Whether the Device Becomes Unnecessary
Suppose a toddler begins using an AAC system and later develops more spoken words.
That is an excellent outcome.
Suppose another toddler develops only a small amount of speech but becomes able to:
- ask for help
- reject
- choose activities
- comment
- tell a parent what hurts
- participate in preschool
- communicate through AAC
That is also an excellent communication outcome.
AAC should not be considered successful only when:
“The child graduates from the device.”
Some individuals use AAC temporarily.
Some use it alongside speech.
Some rely on AAC as their primary expressive communication system.
Some use AAC only when:
- speech is difficult
- communication demands increase
- they are fatigued
- the communication partner cannot understand their speech
AAC needs can change throughout development.
Step 3: Use Baseline Language to Understand the Child—Not to Restrict AAC Access
Walters and colleagues found that children's baseline receptive and expressive language abilities predicted the number of spoken target vocabulary words produced after intervention.
Children with stronger language skills at the beginning tended to produce more spoken target vocabulary after treatment.
This is clinically useful information.
It helps researchers understand variation in outcome.
It may also help clinicians develop realistic expectations.
But it should not be interpreted as:
“Children need strong receptive language before they can benefit from AAC.”
or:
“A child has to demonstrate certain cognitive or language prerequisites before receiving a device.”
Current AAC guidance rejects that type of restrictive eligibility rule.
ASHA's National Joint Committee states that access to communication supports should be based on the individual's communication needs rather than predetermined criteria such as:
chronological age;
diagnosis;
cognitive level;
or the absence of presumed prerequisite skills.
Prediction Is Not the Same as Eligibility
This distinction is easy to miss.
A factor can be associated with a particular outcome without being a prerequisite for treatment.
For example:
baseline expressive language may help predict how much spoken vocabulary develops.
That does not determine whether the child deserves communication access.
The two questions are different.
Research question
Which baseline characteristics predict subsequent spoken vocabulary?
Clinical access question
Does this child currently need additional ways to communicate?
AAC decisions should be driven primarily by the second question.
AAC Assessment Should Examine the Whole Communication System
For toddlers, AAC assessment is often:
- play based
- dynamic
- collaborative
- routines based
ASHA recommends considering:
- current communication attempts
- receptive and expressive language
- motor abilities
- vision and hearing
- sensory access
- interests
- communication environments
- family routines
- communication partners
- potential AAC access methods
A young child may be communicating already through:
- eye gaze
- reaching
- pointing
- vocalizations
- body movements
- facial expressions
- word approximations
- taking an adult to an object
AAC assessment builds from those existing strengths.
Feature Matching Is More Useful Than Choosing a Device by Diagnosis
There is no single “autism AAC device,” “developmental-delay device,” or “toddler AAC app.”
A system should be matched to the individual child.
ASHA's feature-matching framework considers factors such as:
- symbol type
- display organization
- access method
- voice output
- physical positioning
- portability
- ability to support different communication functions
- use across settings
- communication-partner needs
- the child's preferences
The process can involve trials of:
- low-tech boards
- pictures
- recordable buttons
- communication books
- tablet-based AAC
- dedicated speech-generating devices
The goal is not to find the most technologically advanced option.
It is to find a communication system the child can access and continue to develop.
Step 4: Interpret Speech-Sound Errors Correctly
A common interpretation is:
“As the number of words produced increased, so did speech-sound errors.”
That statement needs careful interpretation.
Walters and colleagues found that children producing more spoken target words also produced more speech-sound errors.
But this should not be interpreted as evidence that AAC causes more speech errors.
Nor should errors simply be celebrated as proof of progress.
The more likely clinical explanation is that children producing more words also create more opportunities for speech-sound errors to occur.
Most importantly, speech-sound accuracy did not differ significantly between the AAC and spoken-language intervention groups.
AAC therefore was not associated with poorer speech-sound accuracy.
Vocabulary Growth and Speech Accuracy Are Different Outcomes
A toddler may begin saying more words before those words are produced clearly.
That is common in early development.
Consider a child who begins using:
“ba” for ball
“do” for go
“nana” for banana
Those productions may represent substantial progress in expressive vocabulary even though speech-sound production remains immature.
The clinician should evaluate separately:
How many messages and words is the child expressing?
and
How accurately are those words being produced?
Those are related but different developmental outcomes.
Do Not Withhold AAC Until Speech Becomes Intelligible
If a toddler speaks but is difficult to understand, AAC may still be useful.
For example, a child with:
childhood apraxia of speech;
a severe speech-sound disorder;
motor impairment;
or another developmental condition
may have more language to express than their speech system can currently convey.
AAC can reduce the mismatch between:
what the child knows or wants to communicate
and
what unfamiliar listeners can understand.
The communication system can support the child while speech intervention addresses the speech-production difficulty.
Step 5: Expand Vocabulary—and Expand Communication Functions
The Walters study focused specifically on target vocabulary.
AAC intervention supported larger expressive target vocabularies and, in the 2021 analysis, more spoken target words.
Vocabulary is important.
But modern AAC intervention should not stop at:
“more words.”
The larger objective is communicative competence.
A toddler should eventually have ways to communicate more than:
“I want ___.”
Requesting Is Important, but It Should Not Dominate AAC
Requesting is often one of the earliest AAC intervention targets because it is:
- easy to arrange
- highly motivating
- straightforward to measure
But children communicate for many reasons.
They need vocabulary for:
requesting
more, go, open
rejecting
no, stop, don't
commenting
funny, big, wow
seeking help
help, stuck
directing
come, look, put
social interaction
hi, bye, me, you
asking
where, what
sharing states
hurt, tired, scared, happy
participating in routines
again, finished, turn
A 2025 systematic review of speech-generating-device interventions for autistic preschoolers found that most studies still focused primarily on requesting. The authors specifically called for greater attention to communication across a broader range of functions.
That lesson extends beyond autism.
A communication system should allow a toddler to say more than:
“Give me something.”
Core and Personalized Vocabulary Both Matter
Early AAC systems often include highly reusable vocabulary such as:
- more
- go
- stop
- help
- want
- open
- turn
- you
- me
These words can be used across many situations.
But highly personal vocabulary matters too.
A toddler may need:
- family members' names
- favorite foods
- favorite characters
- specific toys
- places
- songs
- routines
- phrases relevant to their family
A functional system should reflect the child's actual life.
Language Input Matters Too
Children learning spoken language hear thousands of examples before adults expect sophisticated speech from them.
AAC learners also benefit from seeing their communication system used by competent partners.
One common strategy is aided AAC modeling, sometimes called augmented input or aided language input.
The adult speaks while also selecting relevant symbols on the child's AAC system.
For example:
Parent says:
“The car is going!”
while selecting go.
Or:
“You need help.”
while selecting help.
A systematic review of 48 studies involving 267 children and youth found that aided AAC modeling interventions were generally effective in improving expressive communication outcomes.
Another review found that augmented input could improve single-word vocabulary and multi-symbol expression, although the evidence for more complex receptive language remained limited.
Modeling AAC therefore has evidence.
It should not be treated as magical exposure.
Children still need individualized teaching, communication opportunities, and responsive partners.
Step 6: Treat Parents as Communication Partners, Not Assistant Therapists
The Walters study used parent-implemented language intervention.
That matters.
Toddlers communicate throughout:
- meals
- play
- bath time
- getting dressed
- car rides
- books
- family visits
- everyday problem-solving
Families therefore create communication opportunities that no clinic session can reproduce completely.
Research increasingly supports caregiver involvement in AAC intervention.
A 2025 systematic review of caregiver-implemented AAC interventions for children with intellectual and developmental disabilities found caregiver implementation to be a promising and effective way to support communication outcomes, while also identifying important gaps in whose families have been represented in the research and how caregiver outcomes are measured.
Parent Involvement Does Not Mean Turning the Home Into a Therapy Clinic
Caregiver coaching should not create the expectation that parents conduct therapy every waking hour.
The clinician's role is to help caregivers recognize opportunities that already occur naturally.
For example:
Snack
Parent models:
open
when opening the container.
Bubbles
Parent models:
go
before blowing.
Toy Stops Working
Parent models:
help
Child Pushes Something Away
Parent acknowledges:
stop
or
no
Favorite Activity Ends
Parent models:
again
or
finished
These interactions can take seconds.
The goal is not to quiz the toddler continuously.
It is to make communication available within ordinary life.
Communication Partners Need Training Too
A sophisticated AAC system will not be useful if everyone around the child:
- asks questions constantly
- moves the device out of reach
- speaks for the child
- waits too little for responses
- does not know how to model the system
ASHA identifies communication-partner involvement as integral to AAC assessment and intervention and recommends ongoing training for parents, teachers, caregivers, and other communication partners.
Partner strategies may include:
- modeling AAC
- pausing
- responding to all communication modes
- creating communication opportunities
- avoiding excessive testing
- expanding messages
- keeping the system available
AAC intervention changes the communication environment, not just the child.
.png)
.png)