Stimulus generalization and delay of reinforcement during one component of a multiple schedule.

Pigeons pecked a key for food. A light color told them when food would come right away. Another color meant food would arrive ten seconds late.

The birds learned this two-part schedule. Later the researcher tested if the late-food light would now slow pecking when shown in new colors.

Pecking dropped when the late-food light was on. In later tests, the birds pecked least at that exact color and more at colors farther away, making a U-shaped curve.

The delay acted like a stop sign, not just a weak go sign.

Richards (1974) followed up the next year. Bigger delayed reinforcers made the stop signal even stronger, proving the delay itself was the key, not reward size.

Corrigan et al. (1998) later showed that when the delay is not signaled, birds quit pecking and just stare at the feeder hole. Together these papers explain why late rewards hurt responding: the signal turns into a brief conditioned punisher.

HONIBOWER et al. (1964) saw the same U-shaped curve after real punishment. Delayed reinforcement and mild punishment both create stop signals that generalize across similar cues.

If you add even a short signaled delay to reinforcement, the cue can become a mini-punisher. Clients may slow, escape, or avoid that cue. Keep delays short or unsignaled, or fill the delay with another task, so the S-delta effect does not creep into your session.