Core Vocabulary Speech Therapy

The Real Work of Building a Core Vocabulary System

Most people walk into this thinking they just need to pick a set of symbols and laminate them. That is not how it works. Core vocabulary speech therapy is about creating a functional communication system where a small, high-frequency word bank gets reused across every context a client encounters, rather than stacking niche labels that sit unused most of the day. The difference between a working board and a decoration is how aggressively you force generative use.

I spent roughly six months watching a non-speaking teenager with cerebral palsy ignore a 240-symbol board that had every object he could possibly need printed on it. The problem was not the hardware. The problem was that every category on his board was labeled with what the board called the items, not what his therapist called them during modeling. He knew "crackers" under Snack, but never found them because the page said "Food." Once we aligned the icon labels to the vocabulary his AAC device used in his daily routine, he started actually using the system within three sessions. That is a tiny detail that destroys most implementations. The most common misunderstanding is that core vocabulary means removing all nouns. Nouns stay on the board. They shift to fringe pages. Core stays available everywhere, and you teach the student to build sentences using the core words first, then add the fringe word when necessary. It changes the whole trajectory of early AAC training. Next, build a consistent grid layout. I use a four-by-four or five-by-five arrangement depending on the device. The core grid never changes. The fringe pages do. Keeping the core location identical across apps, boards, and communication partners cuts navigation time dramatically. My typical client takes about twelve seconds per cell on a fresh board. With a stable layout, that drops to roughly three seconds after four weeks of consistent use.

Then model. Not prompt. Model. The adult uses the board to speak to the student during natural activities, never demanding the student repeat back. This is called aided language input. It is boring to watch. It is also the single most predictive variable for whether a student ever moves past one-symbol utterances. I track about 45 minutes of modeled input per session across two therapists for a typical beginner case. Anything less than 20 minutes and progress stalls hard. After modeling is routine, add sentence-building drills. Pick a core word, say "want," and create scenarios that force its repeated use. "I want apple." "I want out." "I want help." "I want more." The word changes position, the meaning shifts slightly, but the motor plan stays familiar. You are building syntactic flexibility through repetition, not through explanation.

Edge Case: When Core Word Use Collapses Under Novel Contexts

About two years ago a student I was working with had been doing core vocabulary speech therapy for eleven months. He could generate two-to-three-word combinations on the clinic board at a reliable rate. Then we moved him into the grocery store setting and he went completely silent. Not refusal. Actual loss of access. The environment was overloaded, the board looked different, and his motor planning collapsed.

The workaround was not more therapy. It was reducing the board. I took his full 240-symbol layout and made a stripped six-by-six core-only page with no fringe, no categories, just the top 120 core words arranged in a single screen. We practiced in the store with that reduced page for two weeks before reintroducing the full board. His recovery was fast once the cognitive load dropped. Full boards are overkill for novel environments, and most therapists do not think to simplify before the breakdown happens. Teaching nouns first is the oldest mistake in the book. It delays generative language by months and sometimes years. Every time a student learns "cup" as a label instead of "I want cup," you are reinforcing a single-symbol habit that becomes extremely hard to unlearn. Inconsistent symbol sets between devices and print boards creates confusion. If one app uses Picture Communication Symbols and another uses SymbolSticker, the student has to remap their entire vocabulary each time they switch modalities. Pick one system and stick with it unless there is a compelling reason to change.

Expecting independent use too early. Most students need aided input modeling for 80 to 120 hours before they begin spontaneous core word combinations. That is not a failure of the method. It is the normal trajectory. Pushing for independent use before that window closes usually just produces frustration and board avoidance. Ignoring motor planning. If a symbol location changes between pages or devices, the student loses the spatial memory that makes AAC feasible. I see this constantly with custom-built boards where the clinician rearranges categories for aesthetic reasons. It looks cleaner. It costs the user three to five seconds per symbol search and accumulates into massive fatigue over a school day.

Tools and Resources That Actually Help

For symbol systems, Blissymbols, PCS by Mayer-Johnson, and Stokke Icon Communicator are the standard options. Each has licensing costs. PCS costs roughly $400 to $600 for a school license depending on volume. Stokke is subscription-based at around $200 per year per user. Blissymbols have a free tier with limited export options. For core word lists, the PrAACtical AAC website publishes downloadable PDF sets that are free and regularly updated. The Mayer-Johnson core word frequency lists are also freely available on their site if you register. You do not need to buy anything to get started with the word lists themselves. Speech recognition and voice output devices run from about $500 for basic single-page devices up to $3,000 to $6,000 for full multimodal AAC systems like the Dynavox or Proloquo2Go on an iPad with a mounting arm. Insurance covers a significant portion of high-cost AAC in most US states, but the prior authorization process typically takes six to ten weeks. Start that paperwork before you need the device.

When Core Vocabulary Speech Therapy Falls Flat

This approach does not work well for students with severe visual processing disorders who cannot reliably discriminate symbol arrays, or for clients whose motor impairments prevent consistent touch or gaze selection on any grid layout. In those cases, you pivot to alternative access methods first: switches, scanning protocols, or eye-gaze with larger hit targets. Core vocabulary content still applies, but the delivery mechanism changes entirely. There is also a hard ceiling on how much progress a student can make purely through board-based modeling if their underlying expressive syntax is significantly delayed from neurological factors. Core vocabulary builds the bridge, but it does not replace explicit language intervention for students who need it. Pairing AAC modeling with targeted language therapy usually yields better outcomes than relying on one or the other.

What a Typical Implementation Schedule Looks Like

Weeks one through four focus entirely on adult modeling during familiar routines. No demands on the student. The goal is familiarity with the board and the symbols. Weeks five through eight introduce simple combinations through continuation modeling, where the adult starts a phrase and leaves a symbolic pause for the student to complete if they choose. This is still low pressure. Weeks nine through twelve add explicit teaching of word meanings and spatial locations, paired with continued modeling. By this point, some students begin initiating their own core word combinations without prompting.

Months four through six expand into fringe vocabulary introduction, sentence expansion, and generalization across environments. The core grid stays fixed. Fringe pages multiply. After six months, most students who have had consistent daily input reach a baseline of five to fifteen spontaneous utterances per hour across natural settings. Individual variance is large. Some double that. Some plateau lower. Both are normal given the starting point.