ABA Fundamentals

STIMULUS GENERALIZATION OF THE EFFECTS OF PUNISHMENT.

HONIG et al. (1964) · Journal of the experimental analysis of behavior 1964
★ The Verdict

Punishment locks onto the punished cue, so shift cues slowly or suppression may vanish.

✓ Read this if BCBAs who use punishment or response-cost in clinics or schools.
✗ Skip if Practitioners who work only with reinforcement-based plans.

01Research in Context

01

What this study did

HONIBOWER et al. (1964) punished pigeons for pecking a key lit with a 550 mμ color.

They then tested how much the birds slowed down when the color was slightly changed.

The design mapped a generalization gradient after punishment.

02

What they found

Shock cut responding to the exact color that was punished.

Responding rose again as the color moved away, forming a steep hill-shaped gradient.

When punishment stopped, the gradient flattened and responding returned.

03

How this fits with other research

Rose et al. (2000) repeated the idea with adult humans and point-loss instead of shock.

People also slowed most at the punished stimulus, showing the same gradient.

SIDMAELLIOTT et al. (1962) had already shown that punishing one key can make an unpunished key busier.

K’s gradient work adds the detail: suppression is sharpest right at the punished cue.

04

Why it matters

The study reminds you that punishment suppresses the exact signal, not just the behavior.

If you thin or move the punishment cue, responding can pop back.

Check for stimulus control before you fade prompts or change rooms.

Free CEUs

Want CEUs on This Topic?

The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.

Join Free →
→ Action — try this Monday

Probe the punished response in two new settings to see if suppression travels.

02At a glance

Intervention
other
Design
single case other
Sample size
3
Finding
negative

03Original abstract

Three pigeons were trained to respond to seven spectral stimulus values ranging from 490 to 610 mmu and displayed in random order on a response key. After response rates had equalized to these values, a brief electric shock was administered when the subject (S) responded to the central value (550 mmu) while positive reinforcement for all values was maintained. Initially, there was broad generalization of the resulting depression in response rate, but the gradients grew steeper in the course of testing. When punishment was discontinued, the rates to all values recovered, and equal responding to all stimuli was reattained by two of the Ss. Stimulus control over the effects of punishment was clearly demonstrated in the form of a generalization gradient; this probably resulted from the combined effects of generalization of the depression associated with punishment and discrimination between the punished value and neutral stimuli.

Journal of the experimental analysis of behavior, 1964 · doi:10.1901/jeab.1964.7-21