Inhibition and the stimulus control of operant behavior.

Hearst et al. (1970) wrote a think-piece, not an experiment. They looked at fifty years of animal lab work and said, “We still don’t know what ‘inhibitory stimulus’ means.”

The authors asked researchers to pick one of two meanings before they write. Does the cue directly stop behavior, or is it just the weakest cue on a generalization gradient?

The paper found confusion, not numbers. Papers used the same word for different processes. The team said clearer labels would speed up research.

Harrison et al. (1975) answered the call. They ran a single-case study with pigeons and showed the S- from errorless training really does suppress pecks. That gave the field a clean way to measure inhibition.

Tiger et al. (2017) took the same S- idea into a clinic. Two children with autism learned to stop hand-flapping when a red card appeared, even after staff stopped blocking the behavior. The concept now helps people, not just pigeons.

Blanchard et al. (1979) seemed to disagree. They found weak S- elements did not suppress responding. The gap is method: B used successive discriminations with very faint cues, while M used clear errorless steps. Weak cues may not reach the strength needed to show inhibition.

When you write a program, say what you mean by “inhibitory stimulus.” If you want a cue that actively stops behavior, test it by removing all extra contingencies and watch the rate drop. Pick clear S- signals, not weak shades, and check that the learner really contacts the cue. Your data will be cleaner and your graphs will make sense to the next BCBA.