Effects of stimulus control and deprivation upon discriminative responding.

The team used pigeons in a lab. Birds pecked a key under two cues. A tone meant food came on a variable-interval schedule. A white keylight meant no food.

After training, they ran extinction in both cues. They also tested the birds when food had been withheld for many hours. They wanted to see if hunger would bring back pecking in the white-light extinction component.

Strong stimulus control acted like a lock. When the birds clearly knew tone = food and white light = no food, extinction responding stayed low even when the birds were hungry.

Weak control was like an open gate. When the cues were less clear, hunger quadrupled pecking during the white-light extinction periods. Resurgence slipped through.

Jones (1969) saw the same lock effect. Pigeons given separate stimuli for each link of a chain kept responding longer in extinction. Strong cue-to-link control slowed the drop.

Crossman et al. (1973) found a second way to keep the lock closed. Ending sessions with one large reinforcer after a long pause also protected stimulus control during later extinction probes.

Badia et al. (1972) showed how to build the lock in the first place. Differential reinforcement created >90% correct discrimination in just one or two sessions. Non-differential reinforcement never built control, matching W’s weak-control condition.

When you fade reinforcement, check how well the learner can already tell the difference between reinforcement and extinction cues. If the discrimination is shaky, resurgence is more likely when motivation rises—think end-of-day hunger or a skipped snack. Tighten the stimulus control first: use clear, separate cues for reinforcement and extinction periods, and reinforce only when those cues are present. A sharper discrimination now means fewer surprise spikes later.