Preference for signalled reinforcement.

Shimp et al. (1974) worked with pigeons in a small chamber.

Two keys sat side by side. One key gave grain on a variable-interval schedule and a red light always came on two seconds before each grain delivery.

The other key gave grain on the same schedule but with no warning light.

The bird could hop to either key at any time.

Every pigeon spent more than ninety percent of the session on the key with the red light.

When the researchers swapped the lights, the birds swapped with them.

The birds were not just choosing grain; they were choosing the signal that grain was coming.

Liberman et al. (1973) saw the same flip one year earlier, but with shock. Rats picked a lever that turned on a tone before shock instead of a lever that gave no warning, even when the tone meant eight times more shocks.

Together the two papers show that signals are wanted for both good and bad events.

Smith et al. (2010) moved the idea into the clinic. They gave non-contingent reinforcement to reduce problem behavior. When each delivery was preceded by a brief beep, the behavior dropped twice as fast.

The pigeon lab result now has a real-world speed-up tool.

You already deliver reinforcers. Add a two-second cue—a click, a word, a flash—right before each one. The cue itself becomes a mini-reinforcer and the learner stays with you longer. It costs nothing and may cut problem behavior faster, as E et al. showed. Try it in your next DRL or NCR session.