Aversive control versus stimulus control by punishment
Punishment often suppresses behavior through raw aversive strength, not through a learned signal, so check your stimulus control before you claim it.
01Research in Context
What this study did
The team tested whether punishment really works like reinforcement in reverse.
They gave shocks during one signal but not another.
Then they watched if behavior dropped only when the punishment signal was on.
What they found
Only one small group acted as if the signal meant "don’t respond."
All other drops in behavior lined up with the simple pain of the shock.
The authors say most suppression came from direct aversive control, not stimulus control.
How this fits with other research
Fyfe et al. (2007) saw the opposite picture.
Three adults with ID nearly stopped their stereotypy when a red light meant "punishment coming" and bounced back when the light was off.
That clean on-off pattern is classic stimulus control, the very thing Shahan et al. (2023) rarely found.
The clash is only on the surface.
The 2007 study used humans who could talk and had a long history with rules; the 2023 study used shocks in a more basic setup.
Method differences, not a true contradiction, explain the split.
Older pigeon work backs up both views.
HOFFMAN et al. (1964) and Hoffman et al. (1966) showed skewed or two-humped suppression curves after discrimination training, but Shahan et al. (2023) re-frame those curves as side effects of aversive strength, not proof of learned signals.
Why it matters
Before you call a red card or stop sign a "punishment cue," test what is really driving the drop.
Run probe trials with no consequence.
If behavior stays low, the aversive event itself, not the signal, may be doing the work.
Design your interventions to build clear S-deltas, not just pair stimuli with losses.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Probe one "punishment cue" today: present the cue alone and see if behavior still drops.
02At a glance
03Original abstract
Estes (1944) reported that adding electric shock punishment to extinction hastened response suppression but that responding increased when shock was removed. This result contributed to a view that reinforcement and punishment are asymmetrical processes because punishment has only indirect and temporary suppressive effects. Azrin and Holz (1966) suggested the result might be interpreted instead as shock serving as a discriminative stimulus for the absence of reinforcement. Here, to further examine potential stimulus control by punishment in a similar preparation, two groups of rats initially responded for food plus punishment and a third group for food alone. Reinforcement was then removed for all groups for the remaining three phases. With P and N denoting punishment and no punishment, the four phases for the three groups were: P-P-N-N, P-N-P-N, and N-P-N-N. We found some evidence for stimulus control by shock deliveries for group N-P-N-N (as suggested by Azrin and Holz), but all other changes in responding appeared due to introduction or removal of the aversive properties of shock. Although punishment may indeed have temporary effects under many circumstances, we argue that the view that this implies asymmetrical reinforcement and punishment processes was based on the flawed assumption that reinforcement has direct strengthening effects.
Journal of the Experimental Analysis of Behavior, 2023 · doi:10.1002/jeab.805