What Core Words Speech Therapy Actually Looks Like in a Clinic
Most people hear "core words" and immediately picture a laminated card with ten pictures taped to a board. That is not wrong, but it is also not the whole thing. Core Words Speech Therapy is a vocabulary strategy where you teach a small, reusable set of high-frequency words first — things like go, stop, more, open, close, help, want, yes, no — and then build communication around them instead of trying to master hundreds of low-frequency vocabulary items upfront. I have spent years watching this approach work and watching it fail for the same reasons every time, which is that the method itself is fine and the implementation is usually sloppy. The core idea comes from AAC research, originally developed around 15 years ago by therapists who noticed that clients using augmentative and alternative communication devices were being taught too many specific phrases instead of learning how to combine a few powerful words into new sentences on demand. The core vocabulary size tends to sit between 50 and 100 words depending on the population, and the ratio of core to fringe words in a communication system is usually kept somewhere around 80/20 or 90/10 in the beginning.
How Core Words Speech Therapy Works in Practice
You pick a set of core words. You make sure they are available everywhere — on a low-tech board, on a speech device, in a binder, on a phone app, wherever the person actually communicates. Then you model them constantly throughout the day without forcing the user to produce them correctly first. That modeling piece is what most people mess up. Here is the practical part. If a child or adult uses a device, you tap your core word set while you are talking to them. Not at them. With them. You say "Go stop go" while you are pushing the swing. You say "Open close open" while you are handing them a snack container. You do this dozens of times a day across different routines. The user picks up the pattern eventually. It is not instant. In my experience it usually takes about six to twelve weeks of consistent modeling before you see the first real two-word combinations emerging from someone who was previously only pressing single buttons or making sounds. The fringe words come later. Fringe words are the specific nouns and labels — dog, cookie, bathroom, soccer ball, grandma. Those still matter, obviously. But teaching them one by one before the person has a way to combine them into actual sentences is backwards. A person who only knows "cookie" can ask for a cookie but cannot say "I want cookie now" or "More cookie please" or "Stop cookie no good." The core words give them the grammar skeleton. The fringe words just hang on it afterward.
The Implementation Details Most People Skip
There is a difference between having a core word set and actually using core words in therapy. I ran into this exact problem last year with a client who had a gorgeous 150-word vocab pack on her tablet device, fully laid out with core and fringe organized by category. Beautiful setup. Completely useless for two months because nobody had ever modeled the core words in a functional sentence structure during activities. She was tapping "mommy" and "want" and "drink" separately like a word list, not building actual messages. The workaround was brutal but simple. I deleted most of the fringe words from her active pages. Pulled everything down to roughly 40 core words plus maybe six or seven high-demand fringe nouns. We worked exclusively with those for three weeks. Only core modeling during meals, transitions, play, and breaks. No pressure to produce. Just me talking and tapping the core words in real time. By week four she started combining them without prompting. By week six she was saying things like "I want go park now please" using seven core words in a row. That was the first time she had produced a multi-word sentence in the entire therapy history. This usually cuts the process down from maybe six months of slow vocabulary building to about eight or ten weeks of real communicative output, depending on the person and how consistent the team is.
Counter-Intuitive Things Nobody Tells You
First, core words are not easy words. They are high-frequency words, which is a different thing entirely. Words like "more" and "stop" and "help" sound simple but they require abstract thinking. A child with developmental delays often struggles with them more than concrete nouns because concrete nouns map directly to objects they can see and touch. Abstract words need a level of cognitive processing that takes time to develop. That is why some clients plateau on core words for months even after they have mastered dozens of fringe nouns. It is not resistance. It is a real cognitive gap. Second, you need to model core words in grammatically correct sentences even if the user only produces single words. This sounds obvious but therapists constantly dumb it down. They say "Go go go" or "More more more" instead of "I want more juice now." The user hears simplified speech and ends up producing simplified output. Keep your modeled sentences full and natural. The core words are the content, not the grammar level. If you say "I want more" while looking at the child's empty cup, you are modeling both the core word and the sentence frame at the same time. That is the whole point. Third, core word boards should change location. Put them on the table during meals. Put them in the car. Put them in the bathtub if that is where communication happens. A board that lives in a clinic drawer is not a communication system. It is a prop. I have seen boards with hundreds of core words sitting unused in a corner of a bedroom for two years because nobody ever carried them into the actual routines where the person communicates.
When Core Words Speech Therapy Fails Completely
It fails when the team treats it as a vocabulary list instead of a communication framework. It fails when the user has motor planning issues like apraxia and the core word locations keep changing from session to session. It fails when the device battery dies and there is no low-tech backup. And it fails when people expect results in three weeks. Motor planning is a real bottleneck. If someone uses a device, the physical location of each word matters. Every time you rearrange the core word layout, you force that person to rebuild their motor map from scratch. I had a client whose therapist changed the core word page arrangement every Monday because "the data showed better organization." That client went backwards three steps every single week. We kept one fixed layout for six months and saw massive progress. The data was better because the layout was stable, not because it was rearranged. Another failure mode is overloading the core set. Some programs push 100 or more words into the core category immediately. That defeats the purpose. A smaller core set used consistently across multiple partners and settings beats a larger one used sporadically. I usually recommend starting with 50 core words or fewer and adding from there only after the person is reliably combining them into sentences without prompting for at least two weeks straight.
If core words are not working after three months of consistent modeling, the issue is rarely the method itself. It is usually the fidelity of implementation. Check whether the people around the user are actually modeling core words during their daily routines or just tapping the board occasionally during structured therapy time. Those are two very different things.
Where Core Words Speech Therapy Connects to Other Approaches
Core words work best when combined with visual support strategies and naturalistic teaching. They pair well with PECS if the person is transitioning from picture exchange to a more flexible system. They also integrate smoothly with Hanen program principles because both approaches emphasize parent modeling and embedding language into everyday activities. The downside is that core words alone will not address motor speech disorders or severe articulation deficits. If apraxia is present, you need the motor planning consistency I mentioned earlier plus potentially a dedicated speech motor intervention alongside the core word work. The real takeaway is that core words are a framework, not a cure. They give the person a functional vocabulary engine. Everything else — semantics, syntax, pragmatics, motor output — has to be built around that engine separately. When teams remember that, it works. When teams treat it like a silver bullet, it disappoints everyone involved.
