Punishment of an alternative behavior generates resurgence of a previously extinguished target behavior

Fontes and team worked with lab rats in three phases. First, the rats pressed one lever for food. This was the target behavior.

Next, the lever stopped paying off. The rats learned to poke their nose in a slot for food instead. This was the alternative behavior.

Last, the nose slot still paid, but now it also gave a mild foot shock. The scientists watched to see if the rats went back to the first lever.

When the new choice was shocked, the old lever pressing came back. Stronger shocks made the comeback bigger.

Even a tiny shock brought back some pressing. The rats did not just sit still; they returned to the behavior that had been put out.

Sullivan et al. (2020) saw the same bounce-back in kids with autism. They also showed that untreated problem behaviors in the same class popped up again. The animal lab result lines up with the clinic.

Diaz-Salvat et al. (2020) looks like it disagrees. They found that giving three ways to get food kept resurgence low, even without punishment. The key gap is method: they used extra choices, not aversives. More options cuts relapse; adding shock brings it back.

Ghaemmaghami et al. (2018) taught complex FCRs without any resurgence at all. They moved goals slowly and kept reinforcement rich. Their work shows the flip side: plan the teaching pace and you may skip the relapse that punishment invites.

If you add a reprimand, response cost, or other aversive to one behavior, watch for the old problem to return. Track the original target during maintenance probes. Before you punish, ask if you can add more acceptable ways to get the same reinforcer instead. Planning for resurgence saves you from surprise comebacks.