An evaluation of resurgence of appropriate communication in individuals with autism who exhibit severe problem behavior.

Katherine and her team worked with nine autistic children who used words or pictures to ask for things. Each child had one mand that earned treats and another mand that did not.

The kids first got praise and snacks for using Mand A. Then the team stopped all rewards for Mand A. Next they taught and rewarded Mand B. Finally they stopped rewards for Mand B too. They watched to see if the old Mand A would pop back up.

In eight of nine tests the first mand came roaring back as soon as the new mand lost its payoff. The effect was large and steady across kids.

Even though the children now had a brand-new way to ask, the old way returned the moment reinforcement stopped.

Greer et al. (2024) saw the same bounce-back with destructive behavior. They showed that big, sudden cuts in reinforcement cause the strongest resurgence. Katherine’s data match this pattern: when Mand B went to zero reinforcement, Mand A returned.

Fisher et al. (2019) used functional communication training and found that richer early reinforcement makes later resurgence worse. Their finding extends Katherine’s warning: the more you reinforce at first, the bigger the relapse risk.

Podlesnik et al. (2023) reviewed 200 basic lab studies and note that almost all use pigeons or rats. Katherine’s 2014 study is one of the few to prove the effect with real children and real words, closing the gap between lab and clinic.

When you fade reinforcement for a new mand, expect the old one to sneak back. Plan for this by thinning gradually, keeping brief practice of the old mand under extinction, and teaching caregivers what a relapse looks like. A five-minute booster session now can save ten hours of retraining later.