The role of the response-reinforcer contingency in negative automaintenance.
Animals quickly spot when a response turns reinforcement off and will shift away, proving they read response-reinforcer contingencies.
01Research in Context
What this study did
Duncan et al. (1972) worked with eight pigeons in a lab chamber.
Two keys lit up. One gave grain only if the bird did NOT peck it. The other gave grain no matter what.
After sessions, the team flipped the rules to see if the birds would notice and switch keys.
What they found
Every bird pecked most on the non-contingent key — the one that paid no matter what.
When the rules reversed, every bird shifted to the new non-contingent key within one session.
The birds acted as if they could read the fine print: "Pecking here turns food off."
How this fits with other research
Byrd (1980) later showed the same birds peck faster when food truly depends on the peck, building on the 1972 finding that birds already sense the difference.
Julià (1982) moved the lens from single pecks to runs of pecks, proving contingencies can shape whole response patterns, not just one hit at a time.
Macht (1971) had already shown that non-contingent grain alone can pull birds to one side of a chamber; the 1972 study adds the twist that birds will also stop pecking if that peck cancels the grain.
Why it matters
Your client may keep doing a problem behavior even when it loses reinforcement — the contingency feels negative. This study says, "Yes, they notice," so your next move is to make the new, desired response pay off more clearly and immediately. Flip the key, then highlight the new key.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Arrange a brief reversal: let the problem behavior cancel a reinforcer for one trial, then immediately deliver that same reinforcer for the replacement response so the learner feels the switch.
02At a glance
03Original abstract
When a response key is briefly illuminated before a grain reinforcer is presented, key pecking is reliably developed and maintained in pigeons, even if pecking prevents reinforcement (negative automaintenance). This experiment demonstrated that pigeons are sensitive to a negative response-reinforcer contingency, even though it does not eliminate responding. Within individual pigeons, two kinds of trials were compared: red key trials, in which reinforcement was negatively contingent on responding, and white key trials, in which reinforcement was unrelated to responding. Reinforcement frequency in non-contingent trials was yoked to the obtained reinforcement frequency in negatively contingent trials. All eight pigeons pecked substantially more on the non-contingent key than on the negative key, and preferred the non-contingent key to the negative key on occasional "choice" trials where both were presented together. When the stimuli correlated with the two conditions were reversed, the pigeons' behavior also shifted. These response differences are taken as evidence that pigeons are sensitive to the negative response-reinforcer contingency.
Journal of the experimental analysis of behavior, 1972 · doi:10.1901/jeab.1972.17-351