Blocking, unblocking, and overexpectation in autoshaping with pigeons.
Classic conditioning effects—blocking, unblocking, overexpectation—appear in single-subject pigeon data, giving BCBAs a clear model for teaching stimulus control concepts.
01Research in Context
What this study did
The team ran three small experiments with individual pigeons in an operant chamber.
Each bird learned to peck an illuminated key that sometimes delivered grain.
The researchers then added or removed extra light cues to test three classic learning effects: blocking, unblocking, and overexpectation.
What they found
All three effects showed up cleanly in the single-subject graphs.
When a second cue predicted the same food, pecking to that cue was weak—classic blocking.
Boosting or cutting the food amount after adding a cue changed pecking speed—unblocking.
Pairing two previously trained cues together raised pecking above baseline—overexpectation.
How this fits with other research
Shimp (1966), Nevin (1969), and Yuwiler et al. (1992) all used pigeons to study choice, but they focused on how birds split responses between two keys.
Those papers found matching or momentary-maximizing rules.
The current study does not contradict them; it simply asks a different question about how cues gain or lose strength, not how responses are allocated.
Together they show that the same pigeon preparation can reveal both choice rules and cue-learning rules without conflict.
Why it matters
You can replicate these effects with clients to show how stimulus control forms.
Try pairing a new cue with an already mastered one; if the new cue barely gains control, you have just demonstrated blocking.
Use that visual in staff training or parent meetings to explain why some prompts fail.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Run a brief blocking demo: pair a new colored card with an already effective instruction; graph the child’s response to the new card to show weak control.
02At a glance
03Original abstract
Three experiments used pigeons in an autoshaping procedure and a single-subject design to examine compound stimulus control in classical conditioning. Experiment 1 examined the blocking effect, and Experiment 2 examined the unblocking effect. In both experiments, response-independent food was first delivered intermittently in the presence of one distinctively colored houselight but not another. Then, conventional autoshaping trials were carried out in the presence of each houselight. In Experiment 1, the keylight readily elicited responding in the presence of the houselight that had been negatively correlated with food, but not in the presence of the houselight that had been positively correlated with food. In Experiment 2, the keylight readily elicited responding in the presence of the houselight positively correlated with food, but only when the amount of food used on the autoshaping trials was either greater or less than that previously delivered in the presence of the houselight. Experiment 3 examined the overexpectation effect. Conventional autoshaping trials were first carried out by presenting each of two keylights individually. Then, additional autoshaping trials were carried out by presenting the two keylights as a compound, with either the same amount of food or a greater amount of food per trial. Finally, the keylights were retested by again presenting them individually. The number of responses per trial elicited by the keylights decreased when the amount of food used in compound trials was the same as that used in individual trials. However, the number of responses per trial remained approximately the same when the amount of food used in compound trials was greater than that used in individual trials. Taken together, the results of the three experiments demonstrate (a) the generality of the blocking, unblocking, and overexpectation effects by virtue of their extension to appetitive unconditioned stimuli; (b) the suitability of pigeons as subjects and autoshaping as a procedure for studying classical conditioning; and (c) the appropriateness of single-subject designs.
Journal of the experimental analysis of behavior, 1996 · doi:10.1901/jeab.1996.65-575