A four-year-old pushes the worksheet off the table and the adult, reasonably, gives him a minute. Two weeks later he pushes it off faster, and the minute has become five.
Nobody taught him that. The arrangement did: pushing the worksheet reliably produces a break, and a behavior that reliably produces something valuable is going to happen more often. Functional communication training starts from that observation and does something with it — instead of trying to make the pushing stop, it gives the break a second, easier door.
Step one, and everything depends on it
What is the behavior achieving? In practice there are four common answers:
- Escape — the task ends, gets shorter, or pauses.
- Attention — an adult comes over, talks, reacts.
- A thing — the tablet, the snack, the toy that got put away.
- How it feels — the behavior produces something in the body, and no other person is involved.
You cannot read this off the behavior itself. Two children can scream identically and be achieving opposite things, which is why the same “strategy” works brilliantly with one and makes the other worse. Working out the answer is its own procedure, and if the behavior is dangerous or you cannot tell, this is the point where a supervising BCBA gets involved rather than the point where you guess.
The fourth answer is the one to be careful with. When the behavior produces its own consequence directly — nobody else delivers it — there is usually no request that reproduces the same thing, because the adult does not control that consequence. Functional communication training is not the right tool there; that is a question for a supervising BCBA.
Step two: choose the replacement
The new response has to produce the same outcome. If the screaming buys escape, the request has to buy escape — teaching “please” and then continuing the task teaches nothing except that words are useless.
Figure — Asking for a break and asking for a toy
- Activity → break request → break
- Interest in the toy → toy request → access to the toy
Three properties decide whether the child adopts it:
It costs less than the old behavior. Screaming is fast and reliable. A five-word sentence is slower and less reliable, so a child who can choose will keep screaming, sensibly. Start with the cheapest form that works: a card handed over, a single sign, one word, a button. Elaborate it later, once the easy version is established.
The child can already do it, or nearly. Do not pick a form that itself needs weeks of teaching. You are replacing a behavior that works today.
It is recognizable to everyone. A gesture only you understand will fail with the grandparent, the teacher and the swimming instructor — and the child will go back to the version that works with all of them.
Step three: make it work, every time
At the beginning the request gets honored immediately and always. Every time, straight away, even when it is inconvenient, even the fifteenth time in an hour. That fifteenth request matters as much as the first — a single unanswered one is what teaches the old behavior it still has a chance.
And the old behavior stops producing the outcome. This is the hard half. Those two rules only work together — a request that sometimes works, alongside a scream that also sometimes works, teaches the child to try both, and usually to escalate.
The old behavior may get louder before it fades, though it does not always — a review of extinction procedures found that pattern in roughly a quarter of the cases studied, not most of them. Where it happens, it is not evidence that the plan is failing, and it is exactly the moment when adults give in and accidentally teach that louder is what works. If the escalation becomes intense, unsafe, or hard to manage alone, that is a reason to bring in a supervising BCBA rather than to push through it by yourself.
Step four: make it survive real life
Fifteen requests an hour is not a place to stay. But the way out is slow and the schedule is written in advance, not decided in the moment:
- Wait a little. “Yes, one minute” — and then actually one minute, and then the break.
- Ask for a little. Two more items, then the break, then three.
- Move it around. New rooms, new people, new tasks — a request that only works with you is a request that works nowhere.
Go too fast and the old behavior comes back, which is worth knowing in advance so it reads as information rather than as failure: when it returns, step back to the last level that was working and move more slowly from there.
What it does not do
Functional communication training gives the child a better way to get something they were already getting. That is a large change and a limited one.
It does not teach language beyond that request. It does not settle whether the demand was reasonable in the first place — a child escaping a task that is genuinely too hard is telling you something about the task, and what that calls for is fixing the task as well as teaching the request. And it does not survive on its own: a replacement that everyone honors for three weeks and nobody honors in the fourth is a replacement the child will drop, correctly.
Further reading
- Task analysis and chaining — for when the task really is too long, and the request was telling you so.
- ABA therapy activities — what a session looks like once the child has a way to say no.