Resurgence Under Delayed Reinforcement

Abstract

Four pigeons were studied under a resurgence procedure in which, in successive phases, 1) pecking the left response key (the target response) was reinforced while responding on the right response key (the alternative response) was not reinforced; 2) both responses were extinguished; 3) the alternative response was reinforced while the target response continued to be extinguished; and 4) the target response continued to be extinguished while unsignaled delays of reinforcement that progressively increased across several sessions were added to the alternative response key. As the delays progressively reduced alternative response rates and reinforcement rates, rates of pecking on the originally trained, but now extinguished, key-peck response recurred (resurged) in three of four subjects. The present findings suggest that resurgence is not restricted to conventional extinction where reinforcement is completely eliminated. They also suggest a means of controlling resurgence in the absence of access to the behavioral history that is the wellspring of that resurgence.