Some factors involved in the stimulus control of operant behavior.

MORSE et al. (1958) wrote a theory paper. They asked how reinforced and extinguished stimuli shape what an animal learns. They used pigeons as the model.

The paper is not an experiment. It is a map of ideas that later labs tested.

The authors said reinforcement at one stimulus value and extinction at nearby values make a steep generalization curve. The curve gets sharp peaks and valleys.

They also said more training makes the curve even steeper.

Snapper et al. (1969) ran the test. Pigeons got either discrimination training (S+ vs S-) or plain extinction. Only the discrimination group made a steep inhibitory curve. This matches the 1958 claim.

Powell et al. (1968) added detail. They gave pigeons 30, 90, or 180 discrimination sessions. More sessions made steeper gradients. Again, the 1958 idea held.

HEARSKELLEHER et al. (1964) looked at reinforcement schedules. DRL and long VI schedules made flat curves. This seems to clash with the steep-curve claim, but the methods differ. E et al. held stimulus values the same and varied timing. H et al. kept timing steady and varied stimulus values. Both can be true: stimulus contrast sharpens control, while loose timing blurs it.

When you want tight stimulus control, use differential reinforcement, not extinction alone. Pick a clear S+ and S-. Give many contrast trials. Keep your schedule brisk. These moves steepen the gradient and cut stimulus over-selectivity.