Conditioned reinforcement and discrimination in second-order schedules.

Hamilton et al. (1978) worked with pigeons in a small lab chamber.

The birds pecked a key for food, but only after a brief red light flashed.

Some birds saw the red light paired with every food delivery. Other birds saw the same light, but it never matched up with food.

The team counted how fast the birds pecked during the early part of each trial.

Pigeons pecked almost twice as fast when the red light was paired with food.

When the light was unpaired, pecking stayed slow—no better than a plain tandem schedule with no light at all.

The paired light had become a tiny reward on its own: a conditioned reinforcer.

Berler et al. (1982) saw the same speed-up in rats. Their rats pressed a lever during a tone that came right before free food. The tone wasn’t food, yet it jolted the rats back into action on a DRL schedule that usually slows them down.

Horner (1971) looks like the opposite story. He gave pigeons free food no matter what they did. Response rates crashed. E’s study shows the flip side: when a brief cue is tied to food, rates soar. Same lab species, opposite results—because contingency, not the food itself, drives the change.

Crossman et al. (1973) showed that the way you end a session changes how long responding lasts in extinction. E’s work adds a second layer: the little cues inside the schedule can also keep behavior alive by becoming rewards of their own.

If you add a brief praise, click, or light before the real reinforcer, that cue can start pulling its own weight. Use this to boost early responding in long tasks or chain schedules. Test it next session: pair a 2-s chime with every token, then watch if your client works faster before the first token even drops.